Climate Models And Their Evaluation

8
Climate Models
and Their Evaluation
Coordinating Lead Authors:
David A. Randall (USA), Richard A. Wood (UK)
Lead Authors:
Sandrine Bony (France), Robert Colman (Australia), Thierry Fichefet (Belgium), John Fyfe (Canada), Vladimir Kattsov (Russian Federation),
Andrew Pitman (Australia), Jagadish Shukla (USA), Jayaraman Srinivasan (India), Ronald J. Stouffer (USA), Akimasa Sumi (Japan),
Karl E. Taylor (USA)
Contributing Authors:
K. AchutaRao (USA), R. Allan (UK), A. Berger (Belgium), H. Blatter (Switzerland), C. Bonﬁ ls (USA, France), A. Boone (France, USA),
C. Bretherton (USA), A. Broccoli (USA), V. Brovkin (Germany, Russian Federation), W. Cai (Australia), M. Claussen (Germany),
P. Dirmeyer (USA), C. Doutriaux (USA, France), H. Drange (Norway), J.-L. Dufresne (France), S. Emori (Japan), P. Forster (UK),
A. Frei (USA), A. Ganopolski (Germany), P. Gent (USA), P. Gleckler (USA), H. Goosse (Belgium), R. Graham (UK), J.M. Gregory (UK),
R. Gudgel (USA), A. Hall (USA), S. Hallegatte (USA, France), H. Hasumi (Japan), A. Henderson-Sellers (Switzerland), H. Hendon (Australia),
K. Hodges (UK), M. Holland (USA), A.A.M. Holtslag (Netherlands), E. Hunke (USA), P. Huybrechts (Belgium),
W. Ingram (UK), F. Joos (Switzerland), B. Kirtman (USA), S. Klein (USA), R. Koster (USA), P. Kushner (Canada), J. Lanzante (USA),
M. Latif (Germany), N.-C. Lau (USA), M. Meinshausen (Germany), A. Monahan (Canada), J.M. Murphy (UK), T. Osborn (UK),
T. Pavlova (Russian Federationi), V. Petoukhov (Germany), T. Phillips (USA), S. Power (Australia), S. Rahmstorf (Germany),
S.C.B. Raper (UK), H. Renssen (Netherlands), D. Rind (USA), M. Roberts (UK), A. Rosati (USA), C. Schär (Switzerland),
A. Schmittner (USA, Germany), J. Scinocca (Canada), D. Seidov (USA), A.G. Slater (USA, Australia), J. Slingo (UK), D. Smith (UK),
B. Soden (USA), W. Stern (USA), D.A. Stone (UK), K.Sudo (Japan), T. Takemura (Japan), G. Tselioudis (USA, Greece), M. Webb (UK),
M. Wild (Switzerland)
Review Editors:
Elisa Manzini (Italy), Taroh Matsuno (Japan), Bryant McAvaney (Australia)
This chapter should be cited as:
Randall, D.A., R.A. Wood, S. Bony, R. Colman, T. Fichefet, J. Fyfe, V. Kattsov, A. Pitman, J. Shukla, J. Srinivasan, R.J. Stouffer, A. Sumi
and K.E. Taylor, 2007: Cilmate Models and Their Evaluation. In: Climate Change 2007: The Physical Science Basis. Contribution of
Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning,
Z. Chen, M. Marquis, K.B. Averyt, M.Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New
York, NY, USA.

Climate Models and Their Evaluation
Chapter 8
Table of Contents
Executive Summary .................................................... 591
8.5 Model Simulations of Extremes ..................... 627
8.1 Introduction and Overview
8.5.1 Extreme
Temperature
.......................................... 627
............................ 594
8.5.2 Extreme
Precipitation
.......................................... 628
8.1.1 What is Meant by Evaluation? ............................. 594
8.5.3 Tropical
Cyclones
................................................ 628
8.1.2 Methods
of
Evaluation
......................................... 594
8.5.4 Summary
............................................................. 629
8.1.3 How Are Models Constructed? ........................... 596
8.6 Climate Sensitivity and Feedbacks
8.2 Advances in Modelling

............... 629
.................................... 596
8.6.1 Introduction
......................................................... 629
8.2.1 Atmospheric
Processes
....................................... 602
8.6.2 Interpreting the Range of Climate Sensitivity
8.2.2 Ocean
Processes
................................................ 603

Estimates Among General Circulation Models .... 629
8.2.3 Terrestrial
Processes
........................................... 604
Box 8.1: Upper-Tropospheric Humidity and Water
8.2.4 Cryospheric
Processes........................................ 606

Vapour Feedback .............................................. 632
8.2.5 Aerosol Modelling and Atmospheric
8.6.3 Key Physical Processes Involved in
Chemistry
............................................................ 607
Climate
Sensitivity
............................................... 633
8.2.6 Coupling
Advances
............................................. 607
8.6.4 How to Assess Our Relative Conﬁ dence in

Feedbacks Simulated by Different Models?........ 639
8.2.7 Flux Adjustments and Initialisation ...................... 607
8.3 Evaluation of Contemporary Climate as
8.7 Mechanisms Producing Thresholds and

Simulated by Coupled Global Models

Abrupt Climate Change ................................... 640
........ 608
8.7.1 Introduction
......................................................... 640
8.3.1 Atmosphere
......................................................... 608
8.7.2 Forced Abrupt Climate Change ........................... 640
8.3.2 Ocean
.................................................................. 613
8.7.3 Unforced Abrupt Climate Change ....................... 643
8.3.3 Sea
Ice
................................................................. 616
8.3.4 Land
Surface
....................................................... 617
8.8 Representing the Global System with
Simpler
Models
8.3.5 Changes in Model Performance .......................... 618
.................................................... 643
8.4 Evaluation of Large-Scale Climate
8.8.1 Why
Lower
Complexity?
..................................... 643

Variability as Simulated by Coupled
8.8.2 Simple Climate Models....................................... 644
Global
Models ..................................................... 620
8.8.3 Earth System Models of Intermediate
Complexity........................................................... 644
8.4.1
Northern and Southern Annular Modes ............. 620
8.4.2 Paciﬁ c Decadal Variability .................................. 621
Frequently Asked Question
8.4.3 Paciﬁ c-North American Pattern ......................... 622
FAQ 8.1: How Reliable Are the Models Used to Make

Projections of Future Climate Change? .................. 600
8.4.4
Cold Ocean-Warm Land Pattern ........................ 622
8.4.5
Atmospheric Regimes and Blocking .................. 623
References ........................................................................ 648
8.4.6
Atlantic Multi-decadal Variability ........................ 623
8.4.7
El Niño-Southern Oscillation .............................. 623
Supplementary Material
8.4.8 Madden-Julian
Oscillation
.................................. 625
8.4.9 Quasi-Biennial
Oscillation
.................................. 625
The following supplementary material is available on CD-ROM and
in on-line versions of this report.
8.4.10 Monsoon Variability
............................................ 626
Figures S8.1–S8.15: Model Simulations for Different Climate Variables
8.4.11 Shorter-Term Predictions Using
Table S8.1: MAGICC Parameter Values
Climate
Models
.................................................. 626
590

Chapter 8
Climate Models and Their Evaluation
Executive Summary
At the same time, there have been improvements in
the simulation of many aspects of present climate. The
uncertainty associated with the use of ﬂ ux adjustments
This chapter assesses the capacity of the global climate
has therefore decreased, although biases and long-term
models used elsewhere in this report for projecting future
trends remain in AOGCM control simulations.
climate change. Conﬁ dence in model estimates of future climate
evolution has been enhanced via a range of advances since the
• Progress in the simulation of important modes of climate
IPCC Third Assessment Report (TAR).
variability has increased the overall conﬁ dence in the
Climate models are based on well-established physical
models’ representation of important climate processes.
principles and have been demonstrated to reproduce observed
As a result of steady progress, some AOGCMs can now
features of recent climate (see Chapters 8 and 9) and past climate
simulate important aspects of the El Niño-Southern
changes (see Chapter 6). There is considerable conﬁ dence that
Oscillation (ENSO). Simulation of the Madden-Julian
Atmosphere-Ocean General Circulation Models (AOGCMs)
Oscillation (MJO) remains unsatisfactory.
provide credible quantitative estimates of future climate
change, particularly at continental and larger scales. Conﬁ dence
• The ability of AOGCMs to simulate extreme events,
in these estimates is higher for some climate variables (e.g.,
especially hot and cold spells, has improved. The
temperature) than for others (e.g., precipitation). This summary
frequency and amount of precipitation falling in intense
highlights areas of progress since the TAR:
events are underestimated.
• Enhanced scrutiny of models and expanded diagnostic
• Simulation of extratropical cyclones has improved. Some
analysis of model behaviour have been increasingly
models used for projections of tropical cyclone changes
facilitated by internationally coordinated efforts to
can simulate successfully the observed frequency and
collect and disseminate output from model experiments
distribution of tropical cyclones.
performed under common conditions. This has encouraged
a more comprehensive and open evaluation of models.
• Systematic biases have been found in most models’
The expanded evaluation effort, encompassing a diversity
simulation of the Southern Ocean. Since the Southern
of perspectives, makes it less likely that signiﬁ cant model
Ocean is important for ocean heat uptake, this results in
errors are being overlooked.
some uncertainty in transient climate response.
• Climate models are being subjected to more
• The possibility that metrics based on observations might
comprehensive tests, including, for example, evaluations
be used to constrain model projections of climate change
of forecasts on time scales from days to a year. This more
has been explored for the ﬁ rst time, through the analysis
diverse set of tests increases conﬁ dence in the ﬁ delity
of ensembles of model simulations. Nevertheless, a
with which models represent processes that affect climate
proven set of model metrics that might be used to narrow
projections.
the range of plausible climate projections has yet to be
developed.
• Substantial progress has been made in understanding the
inter-model differences in equilibrium climate sensitivity.
• To explore the potential importance of carbon cycle
Cloud feedbacks have been conﬁ rmed as a primary source
feedbacks in the climate system, explicit treatment of
of these differences, with low clouds making the largest
the carbon cycle has been introduced in a few climate
contribution. New observational and modelling evidence
AOGCMs and some Earth System Models of Intermediate
strongly supports a combined water vapour-lapse rate
Complexity (EMICs).
feedback of a strength comparable to that found in
General Circulation Models (approximately 1 W m–2 °C–1,
• Earth System Models of Intermediate Complexity
corresponding to around a 50% ampliﬁ cation of global
have been evaluated in greater depth than previously.
mean warming). The magnitude of cryospheric feedbacks
Coordinated intercomparisons have demonstrated that
remains uncertain, contributing to the range of model
these models are useful in addressing questions involving
climate responses at mid- to high latitudes.
long time scales or requiring a large number of ensemble
simulations or sensitivity experiments.
• There have been ongoing improvements to resolution,
computational methods and parametrizations, and
additional processes (e.g., interactive aerosols) have been
included in more of the climate models.
• Most AOGCMs no longer use ﬂ ux adjustments, which
were previously required to maintain a stable climate.
591

Climate Models and Their Evaluation
Chapter 8
Developments in model formulation
Some common model biases in the Southern Ocean have been
identiﬁ ed, resulting in some uncertainty in oceanic heat uptake
Improvements in atmospheric models include reformulated
and transient climate response. Simulations of the thermocline,
dynamics and transport schemes, and increased horizontal
which was too thick, and the Atlantic overturning and heat
and vertical resolution. Interactive aerosol modules have been
transport, which were both too weak, have been substantially
incorporated into some models, and through these, the direct and
improved in many models.
the indirect effects of aerosols are now more widely included.
Despite notable progress in improving sea ice formulations,
Signiﬁ cant developments have occurred in the representation
AOGCMs have typically achieved only modest progress in
of terrestrial processes. Individual components continue to be
simulations of observed sea ice since the TAR. The relatively
improved via systematic evaluation against observations and
slow progress can partially be explained by the fact that
against more comprehensive models. The terrestrial processes
improving sea ice simulation requires improvements in both
that might signiﬁ cantly affect large-scale climate over the next
the atmosphere and ocean components in addition to the sea ice
few decades are included in current climate models. Some
component itself.
processes important on longer time scales are not yet included.
Since the TAR, developments in AOGCM formulation have
Development of the oceanic component of AOGCMs has
improved the representation of large-scale variability over a
continued. Resolution has increased and models have generally
wide range of time scales. The models capture the dominant
abandoned the ‘rigid lid’ treatment of the ocean surface.
extratropical patterns of variability including the Northern and
New physical parametrizations and numerics include true
Southern Annular Modes, the Paciﬁ c Decadal Oscillation, the
freshwater ﬂ uxes, improved river and estuary mixing schemes
Paciﬁ c-North American and Cold Ocean-Warm Land Patterns.
and the use of positive deﬁ nite advection schemes. Adiabatic
AOGCMs simulate Atlantic multi-decadal variability, although
isopycnal mixing schemes are now widely used. Some of
the relative roles of high- and low-latitude processes appear to
these improvements have led to a reduction in the uncertainty
differ between models. In the tropics, there has been an overall
associated with the use of less sophisticated parametrizations
improvement in the AOGCM simulation of the spatial pattern
(e.g., virtual salt ﬂ ux).
and frequency of ENSO, but problems remain in simulating its
Progress in developing AOGCM cryospheric components is
seasonal phase locking and the asymmetry between El Niño
clearest for sea ice. Almost all state-of-the-art AOGCMs now
and La Niña episodes. Variability with some characteristics of
include more elaborate sea ice dynamics and some now include
the MJO is simulated by most AOGCMs, but the events are
several sea ice thickness categories and relatively advanced
typically too infrequent and too weak.
thermodynamics. Parametrizations of terrestrial snow processes
Atmosphere-Ocean General Circulation Models are able
in AOGCMs vary considerably in formulation. Systematic
to simulate extreme warm temperatures, cold air outbreaks
evaluation of snow suggests that sub-grid scale heterogeneity
and frost days reasonably well. Models used in this report for
is important for simulating observations of seasonal snow
projecting tropical cyclone changes are able to simulate present-
cover. Few AOGCMs include ice sheet dynamics; in all of the
day frequency and distribution of cyclones, but intensity is
AOGCMs evaluated in this chapter and used in Chapter 10 for
less well simulated. Simulation of extreme precipitation is
projecting climate change in the 21st century, the land ice cover
dependent on resolution, parametrization and the thresholds
is prescribed.
chosen. In general, models tend to produce too many days with
There is currently no consensus on the optimal way to divide
weak precipitation (<10 mm day–1) and too little precipitation
computer resources among: ﬁ ner numerical grids, which allow
overall in intense events (>10 mm day–1).
for better simulations; greater numbers of ensemble members,
Earth system Models of Intermediate Complexity have
which allow for better statistical estimates of uncertainty; and
been developed to investigate issues in past and future climate
inclusion of a more complete set of processes (e.g., carbon
change that cannot be addressed by comprehensive AOGCMs
feedbacks, atmospheric chemistry interactions).
because of their large computational cost. Owing to the reduced
resolution of EMICs and their simpliﬁ ed representation of some
Developments in model climate simulation
physical processes, these models only allow inferences about
very large scales. Since the TAR, EMICs have been evaluated
The large-scale patterns of seasonal variation in several
via several coordinated model intercomparisons which have
important atmospheric ﬁ elds are now better simulated by
revealed that, at large scales, EMIC results compare well with
AOGCMs than they were at the time of the TAR. Notably,
observational data and AOGCM results. This lends support
errors in simulating the monthly mean, global distribution of
to the view that EMICS can be used to gain understanding
precipitation, sea level pressure and surface air temperature
of processes and interactions within the climate system that
have all decreased. In some models, simulation of marine low-
evolve on time scales beyond those generally accessible to
level clouds, which are important for correctly simulating sea
current AOGCMs. The uncertainties in long-term climate
surface temperature and cloud feedback in a changing climate,
change projections can also be explored more comprehensively
has also improved. Nevertheless, important deﬁ ciencies remain
by using large ensembles of EMIC runs.
in the simulation of clouds and tropical precipitation (with their
important regional and global impacts).
592

Chapter 8
Climate Models and Their Evaluation
Developments in analysis methods
by the strong coupling to polar cloud processes and ocean heat
and freshwater transport. Scarcity of observations in polar
Since the TAR, an unprecedented effort has been initiated
regions also hampers evaluation. New techniques that evaluate
to make available new model results for scrutiny by scientists
surface albedo feedbacks have recently been developed. Model
outside the modelling centres. Eighteen modelling groups
performance in reproducing the observed seasonal cycle of land
performed a set of coordinated, standard experiments, and the
snow cover may provide an indirect evaluation of the simulated
resulting model output, analysed by hundreds of researchers
snow-albedo feedback under climate change.
worldwide, forms the basis for much of the current IPCC
Systematic model comparisons have helped establish the
assessment of model results. The beneﬁ ts of coordinated model
key processes responsible for differences among models in
intercomparison include increased communication among the response of the ocean to climate change. The importance
modelling groups, more rapid identiﬁ cation and correction of
of feedbacks from surface ﬂ ux changes to the meridional
errors, the creation of standardised benchmark calculations and
overturning circulation has been established in many models. At
a more complete and systematic record of modelling progress.
present, these feedbacks are not tightly constrained by available
A few climate models have been tested for (and shown)
observations.
capability in initial value predictions, on time scales from
The analysis of processes contributing to climate feedbacks
weather forecasting (a few days) to seasonal forecasting
in models and recent studies based on large ensembles of models
(annual). The capability demonstrated by models under these
suggest that in the future it may be possible to use observations
conditions increases conﬁ dence that they simulate some of the
to narrow the current spread in model projections of climate
key processes and teleconnections in the climate system.
change.
Developments in evaluation of climate feedbacks
Water vapour feedback is the most important feedback
enhancing climate sensitivity. Although the strength of this
feedback varies somewhat among models, its overall impact on
the spread of model climate sensitivities is reduced by lapse
rate feedback, which tends to be anti-correlated. Several new
studies indicate that modelled lower- and upper-tropospheric
humidity respond to seasonal and interannual variability,
volcanically induced cooling and climate trends in a way
that is consistent with observations. Recent observational and
modelling evidence thus provides strong additional support for
the combined water vapour-lapse rate feedback being around
the strength found in AOGCMs.
Recent studies reafﬁ rm that the spread of climate sensitivity
estimates among models arises primarily from inter-model
differences in cloud feedbacks. The shortwave impact of
changes in boundary-layer clouds, and to a lesser extent mid-
level clouds, constitutes the largest contributor to inter-model
differences in global cloud feedbacks. The relatively poor
simulation of these clouds in the present climate is a reason
for some concern. The response to global warming of deep
convective clouds is also a substantial source of uncertainty
in projections since current models predict different responses
of these clouds. Observationally based evaluation of cloud
feedbacks indicates that climate models exhibit different
strengths and weaknesses, and it is not yet possible to determine
which estimates of the climate change cloud feedbacks are the
most reliable.
Despite advances since the TAR, substantial uncertainty
remains in the magnitude of cryospheric feedbacks within
AOGCMs. This contributes to a spread of modelled climate
response, particularly at high latitudes. At the global scale,
the surface albedo feedback is positive in all the models,
and varies between models much less than cloud feedbacks.
Understanding and evaluating sea ice feedbacks is complicated
593

Climate Models and Their Evaluation
Chapter 8
8.1
Introduction and Overview
8.1.2.1
Model Intercomparisons and Ensembles
The global model intercomparison activities that began in
The goal of this chapter is to evaluate the capabilities and
the late 1980s (e.g., Cess et al., 1989), and continued with the
limitations of the global climate models used elsewhere in
Atmospheric Model Intercomparison Project (AMIP), have now
this assessment. A number of model evaluation activities
proliferated to include several dozen model intercomparison
are described in various chapters of this report. This section
projects covering virtually all climate model components
provides a context for those studies and a guide to direct the
and various coupled model conﬁ gurations (see http://www.
reader to the appropriate chapters.
clivar.org/science/mips.php for a summary). By far the most
ambitious organised effort to collect and analyse Atmosphere-
8.1.1
What is Meant by Evaluation?
Ocean General Circulation Model (AOGCM) output from
standardised experiments was undertaken in the last few
A speciﬁ c prediction based on a model can often be
years (see http://www-pcmdi.llnl.gov/ipcc/about_ipcc.php). It
demonstrated to be right or wrong, but the model itself should
differed from previous model intercomparisons in that a more
always be viewed critically. This is true for both weather
complete set of experiments was performed, including unforced
prediction and climate prediction. Weather forecasts are control simulations, simulations attempting to reproduce
produced on a regular basis, and can be quickly tested against
observed climate change over the instrumental period and
what actually happened. Over time, statistics can be accumulated
simulations of future climate change. It also differed in that, for
that give information on the performance of a particular model
each experiment, multiple simulations were performed by some
or forecast system. In climate change simulations, on the other
individual models to make it easier to separate climate change
hand, models are used to make projections of possible future
signals from internal variability within the climate system.
changes over time scales of many decades and for which there
Perhaps the most important change from earlier efforts was
are no precise past analogues. Conﬁ dence in a model can be
the collection of a more comprehensive set of model output,
gained through simulations of the historical record, or of
hosted centrally at the Program for Climate Model Diagnosis
palaeoclimate, but such opportunities are much more limited
and Intercomparison (PCMDI). This archive, referred to here
than are those available through weather prediction. These and
as ‘The Multi-Model Data set (MMD) at PCMDI’, has allowed
other approaches are discussed below.
hundreds of researchers from outside the modelling groups to
scrutinise the models from a variety of perspectives.
8.1.2 Methods
of
Evaluation
The enhancement in diagnostic analysis of climate model
results represents an important step forward since the Third
A climate model is a very complex system, with many
Assessment Report (TAR). Overall, the vigorous, ongoing
components. The model must of course be tested at the system
intercomparison activities have increased communication
level, that is, by running the full model and comparing the
among modelling groups, allowed rapid identiﬁ cation and
results with observations. Such tests can reveal problems, but
correction of modelling errors and encouraged the creation
their source is often hidden by the model’s complexity. For this
of standardised benchmark calculations, as well as a more
reason, it is also important to test the model at the component
complete and systematic record of modelling progress.
level, that is, by isolating particular components and testing
Ensembles of models represent a new resource for studying
them independent of the complete model.
the range of plausible climate responses to a given forcing. Such
Component-level evaluation of climate models is common.
ensembles can be generated either by collecting results from a
Numerical methods are tested in standardised tests, organised
range of models from different modelling centres (‘multi-model
through activities such as the quasi-biennial Workshops
ensembles’ as described above), or by generating multiple
on Partial Differential Equations on the Sphere. Physical
model versions within a particular model structure, by varying
parametrizations used in climate models are being tested through
internal model parameters within plausible ranges (‘perturbed
numerous case studies (some based on observations and some
physics ensembles’). The approaches are discussed in more
idealised), organised through programs such as the Atmospheric
detail in Section 10.5.
Radiation Measurement (ARM) program, EUROpean Cloud
Systems (EUROCS) and the Global Energy and Water cycle
8.1.2.2
Metrics of Model Reliability
Experiment (GEWEX) Cloud System Study (GCSS). These
activities have been ongoing for a decade or more, and a large
What does the accuracy of a climate model’s simulation
body of results has been published (e.g., Randall et al., 2003).
of past or contemporary climate say about the accuracy of its
System-level evaluation is focused on the outputs of the full
projections of climate change? This question is just beginning
model (i.e., model simulations of particular observed climate
to be addressed, exploiting the newly available ensembles of
variables) and particular methods are discussed in more detail
models. A number of different observationally based metrics
below.
have been used to weight the reliability of contributing models
when making probabilistic projections (see Section 10.5.4).
594

Chapter 8
Climate Models and Their Evaluation
For any given metric, it is important to assess how good
Differences between model and observations should be
a test it is of model results for making projections of future
considered insigniﬁ cant if they are within:
climate change. This cannot be tested directly, since there are no
observed periods with forcing changes exactly analogous to those
1. unpredictable internal variability (e.g., the observational
expected over the 21st century. However, relationships between
period contained an unusual number of El Niño events);
observable metrics and the predicted quantity of interest (e.g.,
2. expected differences in forcing (e.g., observations for the
climate sensitivity) can be explored across model ensembles.
1990s compared with a ‘pre-industrial’ model control run);
Shukla et al. (2006) correlated a measure of the ﬁ delity of the
or
simulated surface temperature in the 20th century with simulated
3. uncertainties in the observed ﬁ elds.
21st-century temperature change in a multi-model ensemble.
They found that the models with the smallest 20th-century error
While space does not allow a discussion of the above issues
produced relatively large surface temperature increases in the
in detail for each climate variable, they are taken into account in
21st century. Knutti et al. (2006), using a different, perturbed
the overall evaluation. Model simulation of present-day climate
physics ensemble, showed that models with a strong seasonal
at a global to sub-continental scale is discussed in this chapter,
cycle in surface temperature tended to have larger climate
while more regional detail can be found in Chapter 11.
sensitivity. More complex metrics have also been developed
Models have been extensively used to simulate observed
based on multiple observables in present day climate, and have
climate change during the 20th century. Since forcing changes
been shown to have the potential to narrow the uncertainty in
are not perfectly known over that period (see Chapter 2), such
climate sensitivity across a given model ensemble (Murphy et
tests do not fully constrain future response to forcing changes.
al., 2004; Piani et al., 2005). The above studies show promise
Knutti et al. (2002) showed that in a perturbed physics ensemble
that quantitative metrics for the likelihood of model projections
of Earth System Models of Intermediate Complexity (EMICs),
may be developed, but because the development of robust
simulations from models with a range of climate sensitivities
metrics is still at an early stage, the model evaluations presented
are consistent with the observed surface air temperature and
in this chapter are based primarily on experience and physical
ocean heat content records, if aerosol forcing is allowed to
reasoning, as has been the norm in the past.
vary within its range of uncertainty. Despite this fundamental
An important area of progress since the TAR has been in
limitation, testing of 20th-century simulations against historical
establishing and quantifying the feedback processes that observations does place some constraints on future climate
determine climate change response. Knowledge of these response (e.g., Knutti et al., 2002). These topics are discussed
processes underpins both the traditional and the metric-
in detail in Chapter 9.
based approaches to model evaluation. For example, Hall
and Qu (2006) developed a metric for the feedback between
8.1.2.4
Other Methods of Evaluation
temperature and albedo in snow-covered regions, based on the
simulation of the seasonal cycle. They found that models with
Simulations of climate states from the more distant past
a strong feedback based on the seasonal cycle also had a strong
allow models to be evaluated in regimes that are signiﬁ cantly
feedback under increased greenhouse gas forcing. Comparison
different from the present. Such tests complement the ‘present
with observed estimates of the seasonal cycle suggested that
climate’ and ‘instrumental period climate’ evaluations, since
most models in the MMD underestimate the strength of this
20th-century climate variations have been small compared with
feedback. Section 8.6 discusses the various feedbacks that
the anticipated future changes under forcing scenarios derived
operate in the atmosphere-land surface-sea ice system to
from the IPCC Special Report on Emission Scenarios (SRES).
determine climate sensitivity, and Section 8.3.2 discusses some
The limitations of palaeoclimate tests are that uncertainties in
processes that are important for ocean heat uptake (and hence
both forcing and actual climate variables (usually derived from
transient climate response).
proxies) tend to be greater than in the instrumental period, and
that the number of climate variables for which there are good
8.1.2.3
Testing Models Against Past and Present Climate
palaeo-proxies is limited. Further, climate states may have
been so different (e.g., ice sheets at last glacial maximum) that
Testing models’ ability to simulate ‘present climate’ processes determining quantities such as climate sensitivity
(including variability and extremes) is an important part of
were different from those likely to operate in the 21st century.
model evaluation (see Sections 8.3 to 8.5, and Chapter 11 for
Finally, the time scales of change were so long that there
speciﬁ c regional evaluations). In doing this, certain practical
are difﬁ culties in experimental design, at least for General
choices are needed, for example, between a long time series or
Circulation Models (GCMs). These issues are discussed in
mean from a ‘control’ run with ﬁ xed radiative forcing (often
depth in Chapter 6.
pre-industrial rather than present day), or a shorter, transient
Climate models can be tested through forecasts based on
time series from a ‘20th-century’ simulation including historical
initial conditions. Climate models are closely related to the
variations in forcing. Such decisions are made by individual
models that are used routinely for numerical weather prediction,
researchers, dependent on the particular problem being studied.
and increasingly for extended range forecasting on seasonal
to interannual time scales. Typically, however, models used
595

Climate Models and Their Evaluation
Chapter 8
for numerical weather prediction are run at higher resolution
studies are available that formally address the question. If
than is possible for climate simulations. Evaluation of such
the model has been tuned to give a good representation of
forecasts tests the models’ representation of some key processes
a particular observed quantity, then agreement with that
in the atmosphere and ocean, although the links between these
observation cannot be used to build conﬁ dence in that
processes and long-term climate response have not always
model. However, a model that has been tuned to give a
been established. It must be remembered that the quality of an
good representation of certain key observations may have a
initial value prediction is dependent on several factors beyond
greater likelihood of giving a good prediction than a similar
the numerical model itself (e.g., data assimilation techniques,
model (perhaps another member of a ‘perturbed physics’
ensemble generation method), and these factors may be less
ensemble) that is less closely tuned (as discussed in Section
relevant to projecting the long-term, forced response of the
8.1.2.2 and Chapter 10).
climate system to changes in radiative forcing. There is a large
body of literature on this topic, but to maintain focus on the goal
Given sufﬁ cient computer time, the tuning procedure can
of this chapter, discussions here are conﬁ ned to the relatively
in principle be automated using various data assimilation
few studies that have been conducted using models that are
procedures. To date, however, this has only been feasible for
very closely related to the climate models used for projections
EMICs (Hargreaves et al., 2004) and low-resolution GCMs
(see Section 8.4.11).
(Annan et al., 2005b; Jones et al., 2005; Severijns and Hazeleger,
2005). Ensemble methods (Murphy et al., 2004; Annan et al.,
8.1.3
How Are Models Constructed?
2005a; Stainforth et al., 2005) do not always produce a unique
‘best’ parameter setting for a given error measure.
The fundamental basis on which climate models are
constructed has not changed since the TAR, although there
8.1.3.2
Model Spectra or Hierarchies
have been many speciﬁ c developments (see Section 8.2).
Climate models are derived from fundamental physical laws
The value of using a range of models (a ‘spectrum’ or
(such as Newton’s laws of motion), which are then subjected
‘hierarchy’) of differing complexity is discussed in the TAR
to physical approximations appropriate for the large-scale
(Section 8.3), and here in Section 8.8. Computationally cheaper
climate system, and then further approximated through models such as EMICs allow a more thorough exploration
mathematical discretization. Computational constraints restrict
of parameter space, and are simpler to analyse to gain
the resolution that is possible in the discretized equations, and
understanding of particular model responses. Models of
some representation of the large-scale impacts of unresolved
reduced complexity have been used more extensively in this
processes is required (the parametrization problem).
report than in the TAR, and their evaluation is discussed in
Section 8.8. Regional climate models can also be viewed as
8.1.3.1
Parameter Choices and ‘Tuning’
forming part of a climate modelling hierarchy.
Parametrizations are typically based in part on simpliﬁ ed
physical models of the unresolved processes (e.g.,
8.2
Advances in Modelling
entraining plume models in some convection schemes). The
parametrizations also involve numerical parameters that must be
speciﬁ ed as input. Some of these parameters can be measured,
Many modelling advances have occurred since the TAR.
at least in principle, while others cannot. It is therefore common
Space does not permit a comprehensive discussion of all major
to adjust parameter values (possibly chosen from some prior
changes made over the past several years to the 23 AOGCMs
distribution) in order to optimise model simulation of particular
used widely in this report (see Table 8.1). Model improvements
variables or to improve global heat balance. This process is
can, however, be grouped into three categories. First, the
often known as ‘tuning’. It is justiﬁ able to the extent that two
dynamical cores (advection, etc.) have been improved, and the
conditions are met:
horizontal and vertical resolutions of many models have been
increased. Second, more processes have been incorporated into
1. Observationally based constraints on parameter ranges are
the models, in particular in the modelling of aerosols, and of
not exceeded. Note that in some cases this may not provide
land surface and sea ice processes. Third, the parametrizations
a tight constraint on parameter values (e.g., Heymsﬁ eld and
of physical processes have been improved. For example, as
Donner, 1990).
discussed further in Section 8.2.7, most of the models no longer
use ﬂ ux adjustments (Manabe and Stouffer, 1988; Sausen et
2. The number of degrees of freedom in the tuneable
al., 1988) to reduce climate drift. These various improvements,
parameters is less than the number of degrees of freedom in
developed across the broader modelling community, are well
the observational constraints used in model evaluation. This
represented in the climate models used in this report.
is believed to be true for most GCMs – for example, climate
Despite the many improvements, numerous issues remain.
models are not explicitly tuned to give a good representation
Many of the important processes that determine a model’s
of North Atlantic Oscillation (NAO) variability – but no
response to changes in radiative forcing are not resolved by
596

Chapter 8
Climate Models and Their Evaluation

and
te type
outing
outing
outing
outing
outing
outing
outing
outing
, r
, r
, r
, r
, r
,r
, r
, r
odini, 1992
, 2001
‘bucket’ vs multi-
lusion of ice leads),
ences
don et al., 2002
rseghy et al., 1993
rseghy et al., 1993
e
e
Land
Soil, Plants, Routing
Refer
layers, canopy
CSMD, 2005
Layers, canopy
Mahfouf et al., 1995;
Douville et al., 1995;
Oki and Sud, 1998
layers, canopy
Oleson et al., 2004;
Branstetter
layers, canopy
V
layers, canopy
V
layers, canopy
Mahfouf et al., 1995;
Douville et al., 1995;
Oki and Sud, 1998
layers, canopy
Gor
bucket, canopy
Hagemann, 2002;
Hagemann and
Dümenil-Gates, 2001
bucket, canopy
Roeckner et al., 1996;
Dümenil and T
tion of results from each model.

rst publica
as well as the oceanic vertical coordina
eshwater
eshwater
eshwater
ences
tion of soil moisture (single-layer

don et al., 2002
‘free drift’ assumption and inc
and Zhang,
evik et al., 2003
ge’) of the ﬁ
u
erray et al., 1998
Coupling
Flux Adjustments
Refer
heat, momentum
Y
2000;
CSMD, 2005
no adjustments
Fur
no adjustments
Collins et al., 2006
heat, fr
Flato, 2005
heat, fr
Flato, 2005
no adjustments
T
no adjustments
Gor
no adjustments
Jungclaus et al.,
2005
heat, fr
Min et al., 2005
y vs
rheolog
,
der
tmosphere and ocean models,
tures such as the representa
, leads
, leads
, leads
, leads
, leads
, leads
, leads
, 1976
, leads
ences
, 1979; Har
, 1979; Flato and
, 1992
, 1979; Flato and
, 1992
ell, 1998
, 1979;
f et al., 1997
olf
ynamics/structure (e.g.,
Land fea
Sea Ice
Dynamics, Leads
Refer
no rheology or leads
Xu et al., 2005
rheology
Hibler
1996
rheology
Briegleb et al., 2004
rheology
Hibler
Hibler
rheology
Hibler
Hibler
rheology
Hunke-Dukowicz, 1997;
Salas-Mélia, 2002
rheology
O’Farr
rheology
Hibler
Semtner
rheology
W
) along with the calendar year (‘vinta
tion (ID

mponents.

ca
g details of these aspects of the models are cited.
, 2001;
b op BC
rtical resolution of the model a
ee surface
ee surface
ee surface
ee surface
ee surface
acteristics of sea ice d

d., T
, fr
ences
don et al., 2002
f et al., 1997
lf
o
Ocean
Resolution
Z Coor
Refer
1.9° x 1.9° L30
depth, fr
Jin et al., 1999
0.5°–1.5° x 1.5° L35
density
Bleck et al., 1992
0.3°–1° x 1° L40
depth, fr
Smith and Gent, 2002
1.9° x 1.9° L29
depth, rigid lid
Pacanowski et al.,
1993
0.9° x 1.4° L29
depth, rigid lid
Flato and Boer
Kim et al., 2002
0.5°–2° x 2° L31
depth, rigid lid
Madec et al., 1998
0.8° x 1.9° L31
depth, rigid lid
Gor
1.5° x 1.5° L40
depth, fr
Marsland et al., 2003
0.5°–2.8° x 2.8° L20
depth, fr
W
ocean and sea ice co
the horizontal and ve
tmosphere,
Relevant references describin
t
PCMDI are listed by IPCC identiﬁ

Also listed are the char
e
a
ences
tmospheric model,
don et al., 2002
op
ting in the MMD a
Atmospher
T
Resolution
Refer
top = 25 hPa
T63 (1.9° x 1.9°) L16
Dong et al., 2000; CSMD,
2005; Xu et al., 2005
top = 10 hPa
T63 (1.9° x 1.9°) L31
Déqué et al., 1994
top = 2.2 hPa
T85 (1.4° x 1.4°) L26
Collins et al., 2004
top = 1 hPa
T47 (~2.8° x 2.8°) L31
McFarlane et al., 1992;
Flato, 2005
top = 1 hPa
T63 (~1.9° x 1.9°) L31
McFarlane et al., 1992;
Flato 2005
top = 0.05 hPa
T63 (~1.9° x 1.9°) L45
Déqué et al., 1994
top = 4.5 hPa
T63 (~1.9° x 1.9°) L18
Gor
top = 10 hPa
T63 (~1.9° x 1.9°) L31
Roeckner et al., 2003
top = 10 hPa
T30 (~3.9° x 3.9°) L19
Roeckner et al., 1996
pplied in coupling the a
free surface or rigid lid).
t
the top of the a
c

uxes are a
, China
ch
ch, USA
e
ch
ch,
ea
AOGCMs participa
ter ﬂ
ches
ea
y condition (BC:
e for Climate
e for Climate
the pressure a
, Germany
tures of the
t or freshwa
tion canopy or a river routing scheme also are noted.
ch, Norway
ologiques, France
ology
ological Institute
ological Resear
ological Administration
hea
ganisation (CSIRO)
oup, Germany/Kor
Salient fea
Sponsor(s), Country
Beijing Climate Center
Bjerknes Centr
Resear
National Center for
Atmospheric Resear
Canadian Centr
Modelling and Analysis,
Canada
Météo-France/Centr
National de Recher
Météor
Commonwealth Scientiﬁ
and Industrial Resear
Or
Atmospheric Resear
Australia
Max Planck Institute for
Meteor
Meteor
of the University of Bonn,
Meteor
Institute of the Kor
Meteor
(KMA), and Model and Data
Gr
tures.

nitions) and upper boundar
intage
Selected model fea

es (2004) for deﬁ
able 8.1.
see Grifﬁ
Model ID, V
1: BCC-CM1, 2005
2: BCCR-BCM2.0, 2005
3: CCSM3, 2005
4: CGCM3.1(T47), 2005
5: CGCM3.1(T63), 2005
6: CNRM-CM3, 2004
7: CSIRO-MK3.0, 2001
8: ECHAM5/MPI-OM, 2005
9: ECHO-G, 1999
T
Also listed are the respective sponsoring institutions,
(Z:
whether adjustments of surface momentum,
layered scheme) and the presence of a vegeta
597

Climate Models and Their Evaluation
Chapter 8
f, 1998
outing
outing
outing
outing
outing
outing
outing
outing
outing
, r
, r
, r
, 2004
, 2004
, r
, r
, r
, no
, r
, r
, r
ykosof
ences
olodin and L
Land
Soil, Plants, Routing
Refer
layers, canopy
Bonan et al., 2002
bucket, canopy
Milly and Shmakin, 2002;
GFDL GAMDT
bucket, canopy
Milly and Shmakin, 2002;
GFDL GAMDT
layers, canopy
Abramopoulos et al.,
1988; Miller et al., 1994
layers, canopy
Friend and Kiang, 2005
layers, canopy
Friend and Kiang, 2005
layers, canopy
r
outing
Alekseev et al., 1998;
V
layers, canopy
Krinner et al., 2005
layers, canopy
K-1 Developers, 2004;
Oki and Sud, 1998
layers, canopy
K-1 Developers, 2004;
Oki and Sud, 1998

olodin,
eshwater
olodin and
, 2004
ences
et al., 2002,
u
Coupling
Flux Adjustments
Refer
no adjustments
Y
2004
no adjustments
Delworth et al.,
2006
no adjustments
Delworth et al.,
2006
no adjustments
Russell, 2005
no adjustments
Schmidt et al., 2006
no adjustments
Schmidt et al., 2006
r
egional fr
Diansky and V
2002; V
Diansky
no adjustments
Marti et al., 2005
no adjustments
K-1 Developers,
2004
no adjustments
K-1 Developers,
2004
, 1992;
, leads
, leads
, leads
, leads
, leads
, leads
, leads
, leads
, leads
ences
Sea Ice
Dynamics, Leads
Refer
rheology
Briegleb et al., 2004
rheology
Winton, 2000;
Delworth et al., 2006
rheology
Winton, 2000; Delworth
et al., 2006
rheology
Flato and Hibler
Russell, 2005
rheology
Liu et al., 2003;
Schmidt et al., 2004
rheology
Liu et al., 2003;
Schmidt et al., 2004
no rheology or leads
Diansky et al., 2002
rheology
Fichefet and Morales
Maqueda, 1997; Goosse
and Fichefet, 1999
rheology
K-1 Developers, 2004
rheology
K-1 Developers, 2004

ee
ee
ee

ee surface
b
op BC
ee surface
ee surface
ee surface
ea, fr
ee surface

d., T
ea, fr
, fr
ences
ee surface
Ocean
Resolution
Z Coor
Refer
1.0° x 1.0° L16
eta, fr
Jin et al., 1999;
Liu et al., 2004
0.3°–1.0° x 1.0°
depth, fr
Gnanadesikan et al.,
2004
0.3°–1.0° x 1.0°
depth, fr
Gnanadesikan et al.,
2004
3° x 4° L16
mass/ar
Russell et al., 1995;
Russell, 2005
2° x 2° L16
density
Bleck, 2002
4° x 5° L13
mass/ar
surface
Russell et al., 1995
2° x 2.5° L33
sigma, rigid lid
Diansky et al., 2002
2° x 2° L31
depth, fr
Madec et al., 1998
0.2° x 0.3° L47
sigma/depth, fr
surface
K-1 Developers, 2004
0.5°–1.4° x 1.4° L43
sigma/depth, fr
surface
K-1 Developers, 2004
L56
, 2004
, 2004
1.1°)
e
a
ences
din et al., 2006
(~1.1° x
op
ang et al., 2004
Atmospher
T
Resolution
Refer
top = 2.2 hPa
T42 (~2.8° x 2.8°) L26
W
top = 3 hPa
2.0° x 2.5° L24
GFDL GAMDT
top = 3 hPa
2.0° x 2.5° L24
GFDL GAMDT
with semi-Lagrangian
transports
top = 10 hPa
3° x 4° L12
Russell et al., 1995;
Russell, 2005
top = 0.1 hPa
4° x 5° L20
Schmidt et al., 2006
top = 0.1 hPa
4° x 5° L20
Schmidt et al., 2006
top = 10 hPa
4° x 5° L21
Alekseev et al., 1998;
Galin et al., 2003
top = 4 hPa
2.5° x 3.75° L19
Hour
top = 40 km
T106
K-1 Developers, 2004
top = 30 km
T42 (~2.8° x 2.8°) L20
K-1 Developers, 2004
ce/
ch Center for
onautics and
e Simon Laplace,
d Institute for Space
ch (University of
onmental Studies, and
okyo), National Institute for
ontier Resear
Sponsor(s), Country
National Key Laboratory
of Numerical Modeling for
Atmospheric Sciences and
Geophysical Fluid Dynamics
(LASG)/Institute of Atmospheric
Physics, China
U.S. Department of Commer
National Oceanic and
Atmospheric Administration
(NOAA)/Geophysical Fluid
Dynamics Laboratory (GFDL),
USA
National Aer
Space Administration (NASA)/
Goddar
Studies (GISS), USA
NASA/GISS, USA
Institute for Numerical
Mathematics, Russia
Institut Pierr
France
Center for Climate System
Resear
T
Envir
Fr
Global Change (JAMSTEC),
Japan
es),
es), 2004
intage
MIROC3.2(medr
2004
ble 8.1 (continued)
a
Model ID, V
10: FGOALS-g1.0, 2004
11: GFDL-CM2.0, 2005
12: GFDL-CM2.1, 2005
13: GISS-AOM, 2004
14: GISS-EH, 2004
15: GISS-ER, 2004
16: INM-CM3.0, 2004
17: IPSL-CM4, 2005
18: MIROC3.2(hir
19:
T
598

Chapter 8
Climate Models and Their Evaluation
outing
outing
outing
, r
, no
, r
, r
ences
Land
Soil, Plants, Routing
Refer
layers, canopy
Sellers et al., 1986; Sato
et al., 1989
layers, canopy
r
outing
Bonan, 1998
layers, canopy
Cox et al., 1999
layers, canopy
Essery et al., 2001; Oki
and Sud, 1998
,
esolution (L) is the number of
ertical r
eshwater
ukimoto and
ences
shington et al.,
don et al., 2000
ukimoto et al.,
a
Coupling
Flux Adjustments
Refer
heat, fr
momentum
(12°S–12°N)
Y
2001; Y
Noda, 2003
no adjustments
W
2000
no adjustments
Gor
no adjustments
Johns et al., 2006

,
, 1976;
ossley
ees latitude and longitude. V
, leads
, leads
ences
ee drift, leads
ee drift, leads
Sea Ice
Dynamics, Leads
Refer
fr
Mellor and Kantha, 1989
rheology
Hunke and Dukowicz 1997,
2003; Zhang et al., 1999
fr
Cattle and Cr
1995
rheology
Hunke and Dukowicz,
1997; Semtner
Lipscomb, 2001

ough translation to degr
evels.
b op BC
ee surface
ee surface

d., T
ences
don et al., 2000
ukimoto et al., 2001
Ocean
Resolution
Z Coor
Refer
0.5°–2.0° x 2.5° L23
depth, rigid lid
Y
0.5°–0.7° x 1.1° L40
depth, fr
Maltrud et al., 1998
1.25° x 1.25° L20
depth, rigid lid
Gor
0.3°–1.0° x 1.0° L40
depth, fr
Roberts, 2004
e
a
esolution (L) is the number of vertical l
ences
op
Atmospher
T
Resolution
Refer
top = 0.4 hPa
T42 (~2.8° x 2.8°) L30
Shibata et al., 1999
top = 2.2 hPa
T42 (~2.8° x 2.8°) L26
Kiehl et al., 1998
top = 5 hPa
2.5° x 3.75° L19
Pope et al., 2000
top = 39.2 km
~1.3° x 1.9° L38
Martin et al., 2004
ch
ch/Met
ch, USA
ees latitude by longitude or as a triangular (T) spectral truncation with a r
e for Climate
ees latitude by longitude, while vertical r
ological Resear
UK
ediction and Resear
ce,
Sponsor(s), Country
Meteor
Institute, Japan
National Center for
Atmospheric Resear
Hadley Centr
Pr
Ofﬁ
essed either as degr
essed as degr
intage
esolution is expr
esolution is expr
UKMO-HadGEM1,
2004
Horizontal r
vertical levels.
Horizontal r
ble 8.1 (continued)
a
Model ID, V
20: MRI-CGCM2.3.2, 2003
21: PCM, 1998
22: UKMO-HadCM3, 1997
23:
T
Notes:

a
b
599

Climate Models and Their Evaluation
Chapter 8
Frequently Asked Question 8.1
How Reliable Are the Models Used to Make Projections
of Future Climate Change?
There is considerable conﬁ dence that climate models provide
increasing skill in representing many important mean climate
credible quantitative estimates of future climate change, particularly
features, such as the large-scale distributions of atmospheric
at continental scales and above. This conﬁ dence comes from the
temperature, precipitation, radiation and wind, and of oceanic
foundation of the models in accepted physical principles and from
temperatures, currents and sea ice cover. Models can also simu-
their ability to reproduce observed features of current climate and
late essential aspects of many of the patterns of climate vari-
past climate changes. Conﬁ dence in model estimates is higher
for some climate variables (e.g., temperature) than for others
ability observed across a range of time scales. Examples include
(e.g., precipitation). Over several decades of development, models
the advance and retreat of the major monsoon systems, the
have consistently provided a robust and unambiguous picture of
seasonal shifts of temperatures, storm tracks and rain belts, and
signiﬁ cant climate warming in response to increasing greenhouse
the hemispheric-scale seesawing of extratropical surface pres-
gases.
sures (the Northern and Southern ‘annular modes’). Some cli-
Climate models are mathematical representations of the cli-
mate models, or closely related variants, have also been tested
mate system, expressed as computer codes and run on powerful
by using them to predict weather and make seasonal forecasts.
computers. One source of confidence in models comes from the
These models demonstrate skill in such forecasts, showing they
fact that model fundamentals are based on established physi-
can represent important features of the general circulation
cal laws, such as conservation of mass, energy and momentum,
across shorter time scales, as well as aspects of seasonal and
along with a wealth of observations.
interannual variability. Models’ ability to represent these and
A second source of confidence comes from the ability of
other important climate features increases our confidence that
models to simulate important aspects of the current climate.
they represent the essential physical processes important for
Models are routinely and extensively assessed by comparing
the simulation of future climate change. (Note that the limita-
their simulations with observations of the atmosphere, ocean,
tions in climate models’ ability to forecast weather beyond a
cryosphere and land surface. Unprecedented levels of evaluation
few days do not limit their ability to predict long-term climate
have taken place over the last decade in the form of organised
changes, as these are very different types of prediction – see
multi-model ‘intercomparisons’. Models show significant and
FAQ 1.2.)
(continued)
FAQ 8.1, Figure 1. Global mean
near-surface temperatures over the 20th
century from observations (black) and as
obtained from 58 simulations produced
by 14 different climate models driven by
both natural and human-caused factors
that inﬂ uence climate (yellow). The
mean of all these runs is also shown
(thick red line). Temperature anomalies
are shown relative to the 1901 to 1950
mean. Vertical grey lines indicate the
timing of major volcanic eruptions.
(Figure adapted from Chapter 9, Figure
9.5. Refer to corresponding caption for
further details.)
600

Chapter 8
Climate Models and Their Evaluation
A third source of confidence comes from the ability of mod-
tion of substantial climate warming under greenhouse gas in-
els to reproduce features of past climates and climate changes.
creases, and this warming is of a magnitude consistent with
Models have been used to simulate ancient climates, such as
independent estimates derived from other sources, such as from
the warm mid-Holocene of 6,000 years ago or the last gla-
observed climate changes and past climate reconstructions.
cial maximum of 21,000 years ago (see Chapter 6). They can
Since confidence in the changes projected by global models
reproduce many features (allowing for uncertainties in recon-
decreases at smaller scales, other techniques, such as the use of
structing past climates) such as the magnitude and broad-scale
regional climate models, or downscaling methods, have been
pattern of oceanic cooling during the last ice age. Models can
specifically developed for the study of regional- and local-scale
also simulate many observed aspects of climate change over the
climate change (see FAQ 11.1). However, as global models con-
instrumental record. One example is that the global temperature
tinue to develop, and their resolution continues to improve,
trend over the past century (shown in Figure 1) can be mod-
they are becoming increasingly useful for investigating impor-
elled with high skill when both human and natural factors that
tant smaller-scale features, such as changes in extreme weather
influence climate are included. Models also reproduce other ob-
events, and further improvements in regional-scale representa-
served changes, such as the faster increase in nighttime than
tion are expected with increased computing power. Models are
in daytime temperatures, the larger degree of warming in the
also becoming more comprehensive in their treatment of the
Arctic and the small, short-term global cooling (and subsequent
climate system, thus explicitly representing more physical and
recovery) which has followed major volcanic eruptions, such
biophysical processes and interactions considered potentially
as that of Mt. Pinatubo in 1991 (see FAQ 8.1, Figure 1). Model
important for climate change, particularly at longer time scales.
global temperature projections made over the last two decades
Examples are the recent inclusion of plant responses, ocean
have also been in overall agreement with subsequent observa-
biological and chemical interactions, and ice sheet dynamics in
tions over that period (Chapter 1).
some global climate models.
Nevertheless, models still show significant errors. Although
In summary, confidence in models comes from their physical
these are generally greater at smaller scales, important large-
basis, and their skill in representing observed climate and past
scale problems also remain. For example, deficiencies re-
climate changes. Models have proven to be extremely important
main in the simulation of tropical precipitation, the El Niño-
tools for simulating and understanding climate, and there is
Southern Oscillation and the Madden-Julian Oscillation (an
considerable confidence that they are able to provide credible
observed variation in tropical winds and rainfall with a time
quantitative estimates of future climate change, particularly at
scale of 30 to 90 days). The ultimate source of most such
larger scales. Models continue to have significant limitations,
errors is that many important small-scale processes cannot be
such as in their representation of clouds, which lead to uncer-
represented explicitly in models, and so must be included in
tainties in the magnitude and timing, as well as regional details,
approximate form as they interact with larger-scale features.
of predicted climate change. Nevertheless, over several decades
This is partly due to limitations in computing power, but also
of model development, they have consistently provided a robust
results from limitations in scientific understanding or in the
and unambiguous picture of significant climate warming in re-
availability of detailed observations of some physical processes.
sponse to increasing greenhouse gases.
Significant uncertainties, in particular, are associated with the
representation of clouds, and in the resulting cloud responses
to climate change. Consequently, models continue to display a
substantial range of global temperature change in response to
specified greenhouse gas forcing (see Chapter 10). Despite such
uncertainties, however, models are unanimous in their predic-
601

Climate Models and Their Evaluation
Chapter 8
the model’s grid. Instead, sub-grid scale parametrizations are
8.2.1.3 Parametrizations
used to parametrize the unresolved processes, such as cloud
formation and the mixing due to oceanic eddies. It continues
The climate system includes a variety of physical processes,
to be the case that multi-model ensemble simulations generally
such as cloud processes, radiative processes and boundary-layer
provide more robust information than runs of any single model.
processes, which interact with each other on many temporal and
Table 8.1 summarises the formulations of each of the AOGCMs
spatial scales. Due to the limited resolutions of the models, many
used in this report.
of these processes are not resolved adequately by the model grid
There is currently no consensus on the optimal way to divide
and must therefore be parametrized. The differences between
computer resources among ﬁ ner numerical grids, which allow
parametrizations are an important reason why climate model
for better simulations; greater numbers of ensemble members,
results differ. For example, a new boundary-layer parametrization
which allow for better statistical estimates of uncertainty; and
(Lock et al., 2000; Lock, 2001) had a strong positive impact on
inclusion of a more complete set of processes (e.g., carbon
the simulations of marine stratocumulus cloud produced by the
feedbacks, atmospheric chemistry interactions).
Geophysical Fluid Dynamics Laboratory (GFDL) and the Hadley
Centre climate models, but the same parametrization had less
8.2.1 Atmospheric
Processes
positive impact when implemented in an earlier version of the
Hadley Centre model (Martin et al., 2006). Clearly, parametrizations
8.2.1.1 Numerics
must be understood in the context of their host models.
Cloud processes affect the climate system by regulating the
In the TAR, more than half of the participating atmospheric
ﬂ ow of radiation at the top of the atmosphere, by producing
models used spectral advection. Since the TAR, semi-Lagrangian
precipitation, by accomplishing rapid and sometimes deep
advection schemes have been adopted in several atmospheric
redistributions of atmospheric mass and through additional
models. These schemes allow long time steps and maintain
mechanisms too numerous to list here (Arakawa and Schubert,
positive values of advected tracers such as water vapour, but
1974; Arakawa, 2004). Cloud parametrizations are based
they are diffusive, and some versions do not formally conserve
on physical theories that aim to describe the statistics of
mass. In this report, various models use spectral, semi-
the cloud ﬁ eld (e.g., the fractional cloudiness or the area-
Lagrangian, and Eulerian ﬁ nite-volume and ﬁ nite-difference
averaged precipitation rate) without describing the individual
advection schemes, although there is still no consensus on
cloud elements. In an increasing number of climate models,
which type of scheme is best.
microphysical parametrizations that represent such processes
as cloud particle and raindrop formation are used to predict the
8.2.1.2
Horizontal and Vertical Resolution
distributions of liquid and ice clouds. These parametrizations
improve the simulation of the present climate, and affect climate
The horizontal and vertical resolutions of AOGCMs have
sensitivity (Iacobellis et al., 2003). Realistic parametrizations of
increased relative to the TAR. For example, HadGEM1 has
cloud processes are a prerequisite for reliable current and future
eight times as many grid cells as HadCM3 (the number of cells
climate simulation (see Section 8.6).
has doubled in all three dimensions). At the National Center for
Data from ﬁ eld experiments such as the Global Atmospheric
Atmospheric Research (NCAR), a T85 version of the Climate
Research Program (GARP) Atlantic Tropical Experiment
System Model (CSM) is now routinely used, while a T42 version
(GATE, 1974), the Monsoon Experiment (MONEX, 1979),
was standard at the time of the TAR. The Center for Climate
ARM (1993) and the Tropical Ocean Global Atmosphere
System Research (CCSR), National Institute for Environmental
(TOGA) Coupled Ocean-Atmosphere Response Experiment
Studies (NIES) and Frontier Research Center for Global
(COARE, 1993) have been used to test and improve
Change (FRCGC) have developed a high-resolution climate
parametrizations of clouds and convection (e.g., Emanuel and
model (MIROC-hi, which consists of a T106 L56 Atmospheric
Zivkovic-Rothmann, 1999; Sud and Walker, 1999; Bony and
GCM (AGCM) and a 1/4° by 1/6° L48 Ocean GCM (OGCM)),
Emanuel, 2001). Systematic research such as that conducted
and The Meteorological Research Institute (MRI) of the Japan
by the GCSS (Randall et al., 2003) has been organised to test
Meteorological Agency (JMA) has developed a TL959 L60
parametrizations by comparing results with both observations
spectral AGCM (Oouchi et al., 2006), which is being used in
and the results of a cloud-resolving model. These efforts have
time-slice mode. The projections made with these models are
inﬂ uenced the development of many of the recent models.
presented in Chapter 10.
For example, the boundary-layer cloud parametrization of
Due to the increased horizontal and vertical resolution, both
Lock et al. (2000) and Lock (2001) was tested via the GCSS.
regional- and global-scale climate features are better simulated.
Parametrizations of radiative processes have been improved
For example, a far-reaching effect of the Hawaiian Islands in
and tested by comparing results of radiation parametrizations
the Paciﬁ c Ocean (Xie et al., 2001) has been well simulated
used in AOGCMs with those of much more detailed ‘line-
(Sakamoto et al., 2004) and the frequency distribution of
by-line’ radiation codes (Collins et al., 2006). Since the TAR,
precipitation associated with the Baiu front has been improved
improvements have been made in several models to the physical
(Kimoto et al., 2005).
coupling between cloud and convection parametrizations, for
example, in the Max Planck Institute (MPI) AOGCM using
602

Chapter 8
Climate Models and Their Evaluation
Tompkins (2002), in the IPSL-CM4 AOGCM using Bony and
To better resolve the equatorial waveguide, several models
Emanuel (2001) and in the GFDL model using Tiedtke (1993).
use enhanced meridional resolution in the tropics. Resolution
These are examples of component-level testing.
high enough to allow oceanic eddies, eddy permitting, has not
In parallel with improvement in parametrizations, a non-
been used in a full suite of climate scenario integrations due to
hydrostatic model has been used for downscaling. A model with
computational cost, but since the TAR it has been used in some
a 5 km grid on a domain of 4,000 x 3,000 x 22 km centred over
idealised and scenario-based climate experiments as discussed
Japan has been run by MRI/JMA, using the time-slice method for
below. A limited set of integrations using the eddy-permitting
the Fourth Assessment Report (AR4) (Yoshizaki et al., 2005).
MIROC3.2 (hires) model is used here and in Chapter 10. Some
Aerosols play an important role in the climate system.
modelling centres have also increased vertical resolution since
Interactive aerosol parametrizations are now used in some models
the TAR.
(HADGEM1, MIROC-hi, MIROC-med). Both the ‘direct’ and
A few coupled climate models with eddy-permitting ocean
‘indirect’ aerosol effects (Chapter 2) have been incorporated in
resolution (1/6° to 1/3°) have been developed (Roberts et al.,
some cases (e.g., IPSL-CM4). In addition to sulphates, other
2004; Suzuki et al., 2005), and large-scale climatic features
types of aerosols such as black and organic carbon, sea salt
induced by local air-sea coupling have been successfully
and mineral dust are being introduced as prognostic variables
simulated (e.g., Sakamoto et al., 2004).
(Takemura et al., 2005; see Chapter 2). Further details are given
Roberts et al. (2004) found that increasing the ocean
in Section 8.2.5.
resolution of the HadCM3 model from about 1° to 0.33° by
0.33° by 40 levels (while leaving the atmospheric component
8.2.2 Ocean
Processes
unchanged) resulted in many improvements in the simulation
of ocean circulation features. However, the impact on the
8.2.2.1 Numerics
atmospheric simulation was relatively small and localised. The
climate change response was similar to the standard resolution
Recently, isopycnic or hybrid vertical coordinates have been
model, with a slightly faster rate of warming in the Northern
adopted in some ocean models (GISS-EH and BCCR-BCM2.0).
Europe-Atlantic region due to differences in the Atlantic
Tests show that such models can produce solutions for complex
Meridional Overturning Circulation (MOC) response. The
regional ﬂ ows that are as realistic as those obtained with the
adjustment time scale of the Atlantic Basin freshwater budget
more common depth coordinate (e.g., Drange et al., 2005).
decreased from being of order 400 years to being of order 150
Issues remain over the proper treatment of thermobaricity
years with the higher resolution ocean, suggesting possible
(nonlinear relationship of temperature, salinity and pressure
differences in transient MOC response on those time scales, but
to density), which means that in some isopycnic coordinate
the mechanisms and the relative roles of horizontal and vertical
models the relative densities of, for example, Mediterranean
resolution are not clear.
and Antarctic Bottom Water masses are distorted. The merits of
The Atlantic MOC is inﬂ uenced by freshwater as well as
these vertical coordinate systems are still being established.
thermal forcing. Besides atmospheric freshwater forcing,
An explicit representation of the sea surface height is being
freshwater transport by the ocean itself is also important. For
used in many models, and real freshwater ﬂ ux is used to force
the Atlantic MOC, the fresh Paciﬁ c water coming through the
those models instead of a ‘virtual’ salt ﬂ ux. The virtual salt
Bering Strait could be poorly simulated on its transit to the
ﬂ ux method induces a systematic error in sea surface salinity
Canadian Archipelago and the Labrador Sea (Komuro and
prediction and causes a serious problem at large river basin
Hasumi, 2005). These aspects have been improved since the
mouths (Hasumi, 2002a,b; Grifﬁ es, 2004).
TAR in many of the models evaluated here.
Generalised curvilinear horizontal coordinates with bipolar
Changes around continental margins are very important for
or tripolar grids (Murray, 1996) have become widely used in the
regional climate change. Over these areas, climate is inﬂ uenced
oceanic component of AOGCMs. These are strategies used to
by the atmosphere and open ocean circulation. High-resolution
deal with the North Pole coordinate singularity, as alternatives
climate models contribute to the improvement of regional
to the previously common polar ﬁ lter or spherical coordinate
climate simulation. For example, the location of the Kuroshio
rotation. The newer grids have the advantage that the singular
separation from the Japan islands is well simulated in the
points can be shifted onto land while keeping grid points
MIROC3.2 (hires) model (see Figure 8.1), which makes it
aligned on the equator. The older methods of representing the
possible to study a change in the Kuroshio axis in the future
ocean surface, surface water ﬂ ux and North Pole are still in use
climate (Sakamoto et al., 2005).
in several AOGCMs.
Guilyardi et al. (2004) suggested that ocean resolution
may play only a secondary role in setting the time scale of
8.2.2.2
Horizontal and Vertical Resolution
model El Niño-Southern Oscillation (ENSO) variability, with
the dominant time scales being set by the atmospheric model
There has been a general increase in resolution since the TAR,
provided the basic speeds of the equatorial ocean wave modes
with a horizontal resolution of order one to two degrees now
are adequately represented.
commonly used in the ocean component of most climate models.
603

Climate Models and Their Evaluation
Chapter 8
Figure 8.1. Long-term mean ocean current velocities at 100 m depth (vectors, unit: m s–1) and sea surface temperature (colours, °C) around the Kuroshio and the Kuroshio
Extension obtained from a control experiment forced by pre-industrial conditions (CO concentration 295.9 ppm) using MIROC3.2 (hires).
2
8.2.2.3 Parametrizations
(2004) studied the impact of the very simple scheme used in
the HadCM3 model to control mixing of overﬂ ow waters from
In the tracer equations, isopycnal diffusion (Redi, 1982)
the Nordic Seas into the North Atlantic. Although the scheme
with isopycnal layer thickness diffusion (Gent et al., 1995),
does result in a change of the subpolar water mass properties, it
including its modiﬁ cation by Visbeck et al. (1997), has become
appears to have little impact on the simulation of the strength of
a widespread choice instead of a simple horizontal diffusion.
the large-scale MOC, or its response to global warming.
This has led to improvements in the thermocline structure and
meridional overturning (Böning et al., 1995; see Section 8.3.2).
8.2.3 Terrestrial Processes
For vertical mixing of tracers, a wide variety of parametrizations
is currently used, such as turbulence closures (e.g., Mellor and
Few multi-model analyses of terrestrial processes included
Yamada, 1982), non-local diffusivity proﬁ les (Large et al.,
in the models in Table 8.1 have been conducted. However,
1994) and bulk mixed-layer models (e.g., Kraus and Turner,
signiﬁ cant advances since the TAR have been reported based
1967). Representation of the surface mixed layer has been
on climate models that are similar to these models. Analysis of
much improved due to developments in these parametrizations
these models provides insight on how well terrestrial processes
(see Section 8.3.2). Observations have shown that deep ocean
are included in the AR4 models.
vertical mixing is enhanced over rough bottoms, steep slopes and
where stratiﬁ cation is weak (Kraus, 1990; Polzin et al., 1997;
8.2.3.1 Surface
Processes
Moum et al., 2002). While there have been modelling studies
indicating the signiﬁ cance of such inhomogeneous mixing for
The addition of the terrestrial biosphere models that simulate
the MOC (e.g., Marotzke, 1997; Hasumi and Suginohara, 1999;
changes in terrestrial carbon sources and sinks into fully coupled
Otterå et al., 2004; Oliver et al., 2005, Saenko and Merryﬁ eld
climate models is at the cutting edge of climate science. The
2005), comprehensive parametrizations of the effects and their
major advance in this area since the TAR is the inclusion of
application in coupled climate models are yet to be seen.
carbon cycle dynamics including vegetation and soil carbon
Many of the dense waters formed by oceanic convection,
cycling, although these are not yet incorporated routinely into
which are integral to the global MOC, must ﬂ ow over ocean
the AOGCMs used for climate projection (see Chapter 10).
ridges or down continental slopes. The entrainment of ambient
The inclusion of the terrestrial carbon cycle introduces a new
water around these topographic features is an important process
and potentially important feedback into the climate system
determining the ﬁ nal properties and quantity of the deep waters.
on time scales of decades to centuries (see Chapters 7 and
Parametrizations for such bottom boundary-layer processes
10). These feedbacks include the responses of the terrestrial
have come into use in some AOGCMs (e.g., Winton et al., 1998;
biosphere to increasing carbon dioxide (CO2), climate change
Nakano and Suginohara, 2002). However, the impact of the
and changes in climate variability (see Chapter 7). However,
bottom boundary-layer representation on the coupled system is
many issues remain to be resolved. The magnitude of the sink
not fully understood (Tang and Roberts, 2005). Thorpe et al.
remains uncertain (Cox et al., 2000; Friedlingstein et al., 2001;
604

Chapter 8
Climate Models and Their Evaluation
Dufresne et al., 2002) because it depends on climate sensitivity
evaluated separately. Pitman et al. (2004) explored the impact
as well as on the response of vegetation and soil carbon to
of the level of complexity used to parametrize the surface
increasing CO2 (Friedlingstein et al., 2003). The rate at which
energy balance on differences found among the AMIP-2
CO2 fertilization saturates in terrestrial systems dominates results. They found that quite large variations in surface energy
the present uncertainty in the role of biospheric feedbacks. A
balance complexity did not lead to systematic differences in the
series of studies have been conducted to explore the present
simulated mean, minimum or maximum temperature variance
modelling capacity of the response of the terrestrial biosphere
at the global scale, or in the zonal averages, indicating that these
rather than the response of just one or two of its components
variables are not limited by uncertainties in how to parametrize
(Friedlingstein et al., 2006). This work has built on systematic
the surface energy balance. This adds conﬁ dence to the use of
efforts to evaluate the capacity of terrestrial biosphere models
the models in Table 8.1, as most include surface energy balance
to simulate the terrestrial carbon cycle (Cramer et al., 2001)
modules of more complexity than the minimum identiﬁ ed by
via intercomparison exercises. For example, Friedlingstein et
Pitman et al. (2004).
al. (2006) found that in all models examined, the sink decreases
While little work has been performed to assess the capability
in the future as the climate warms.
of the land surface models used in coupled climate models,
Other individual components of land surface processes have
the upgrading of the land surface models is gradually taking
been improved since the TAR, such as root parametrization
place and the inclusion of carbon in these models is a major
(Arora and Boer, 2003; Kleidon, 2004) and higher-resolution
conceptual advance. In the simulation of the present-day
river routing (Ducharne et al., 2003). Cold land processes have
climate, the limitations of the standard bucket hydrology model
received considerable attention with multi-layer snowpack
are increasingly clear (Milly and Shmakin, 2002; Henderson-
models now more common (e.g., Oleson et al., 2004) as is the
Sellers et al., 2004; Pitman et al., 2004), including evidence
inclusion of soil freezing and thawing (e.g., Boone et al., 2000;
that it overestimates the likelihood of drought (Seneviratne et
Warrach et al., 2001). Sub-grid scale snow parametrizations
al., 2002). Relatively small improvements to the land surface
(Liston, 2004), snow-vegetation interactions (Essery et al.,
model, for example, the inclusion of spatially variable water-
2003) and the wind redistribution of snow (Essery and Pomeroy,
holding capacity and a simple canopy conductance, lead to
2004) are more commonly considered. High-latitude organic
signiﬁ cant improvements (Milly and Shmakin, 2002). Since
soils are included in some models (Wang et al., 2002). A recent
most models in Table 8.1 represent the continental-scale land
advance is the coupling of groundwater models into land
surface more realistically than the standard bucket hydrology
surface schemes (Liang et al., 2003; Maxwell and Miller, 2005;
scheme, and include spatially variable water-holding capacity,
Yeh and Eltahir, 2005). These have only been evaluated locally
canopy conductance, etc. (Table 8.1), most of these models
but may be adaptable to global scales. There is also evidence
likely capture the key contribution made by the land surface
emerging that regional-scale projection of warming is sensitive
to current large-scale climate simulations. However, it is not
to the simulation of processes that operate at ﬁ ner scales than
clear how well current climate models can capture the impact of
current climate models resolve (Pan et al., 2004). In general, the
future warming on the terrestrial carbon balance. A systematic
improvements in land surface models since the TAR are based
evaluation of AOGCMs with the carbon cycle represented
on detailed comparisons with observational data. For example,
would help increase conﬁ dence in the contribution of the
Boone et al. (2004) used the Rhone Basin to investigate how land
terrestrial surface resulting from future warming.
surface models simulate the water balance for several annual
cycles compared to data from a dense observation network.
8.2.3.2
Soil Moisture Feedbacks in Climate Models
They found that most land surface schemes simulate very similar
total runoff and evapotranspiration but the partitioning between
A key role of the land surface is to store soil moisture and
the various components of both runoff and evaporation varies
control its evaporation. An important process, the soil moisture-
greatly, resulting in different soil water equilibrium states and
precipitation feedback, has been explored extensively since the
simulated discharge. More sophisticated snow parametrizations
TAR, building on regionally speciﬁ c studies that demonstrated
led to superior simulations of basin-scale runoff.
links between soil moisture and rainfall. Recent studies (e.g.,
An analysis of results from the second phase of AMIP
Gutowski et al., 2004; Pan et al., 2004) suggest that summer
(AMIP-2) explored the land surface contribution to climate
precipitation strongly depends on surface processes, notably in
simulation. Henderson-Sellers et al. (2003) found a clear
the simulation of regional extremes. Douville (2001) showed
chronological sequence of land surface schemes (early models
that soil moisture anomalies affect the African monsoon while
that excluded an explicit canopy, more recent biophysically
Schär et al. (2004) suggested that an active soil moisture-
based models and very recent biophysically based models).
precipitation feedback was linked to the anomalously hot
Statistically signiﬁ cant differences in annually averaged European summer in 2003.
evaporation were identiﬁ ed that could be associated with the
The soil moisture-precipitation feedback in climate models
parametrization of canopy processes. Further improvements in
had not been systematically assessed at the time of the TAR.
land surface models depends on enhanced surface observations,
It is associated with the strength of coupling between the land
for example, the use of stable isotopes (e.g., Henderson-Sellers
and atmosphere, which is not directly measurable at the large
et al., 2004) that allow several components of evaporation to be
scale in nature and has only recently been quantiﬁ ed in models
605

Climate Models and Their Evaluation
Chapter 8
(Dirmeyer, 2001). Koster et al. (2004) provided an assessment
include processes associated with ice streams or grounding
of where the soil moisture-precipitation feedback is regionally
line migration, which may permit rapid dynamical changes
important during the Northern Hemisphere (NH) summer by
in the ice sheets. Glaciers and ice caps, due to their relatively
quantifying the coupling strength in 12 atmospheric GCMs.
small scales and low likelihood of signiﬁ cant climate feedback
Some similarity was seen among the model responses, enough
at large scales, are not currently included interactively in any
to produce a multi-model average estimate of where the global
AOGCMs. See Chapters 4 and 10 for further detail. For a
precipitation pattern during the NH summer was most strongly
discussion of terrestrial snow, see Section 8.3.4.1.
affected by soil moisture variations. These ‘hot spots’ of strong
coupling are found in transition regions between humid and dry
8.2.4.2 Sea
Ice
areas. The models, however, also show strong disagreement in
the strength of land-atmosphere coupling. A few studies have
Sea ice components of current AOGCMs usually predict ice
explored the differences in coupling strength. Seneviratne et al.
thickness (or volume), fractional cover, snow depth, surface and
(2002) highlighted the importance of differing water-holding
internal temperatures (or energy) and horizontal velocity. Some
capacities among the models while Lawrence and Slingo (2005)
models now include prognostic sea ice salinity (Schmidt et al.,
explored the role of soil moisture variability and suggested
2004). Sea ice albedo is typically prescribed, with only crude
that frequent soil moisture saturation and low soil moisture
dependence on ice thickness, snow cover and puddling effects.
variability could partially explain the weak coupling strength in
Since the TAR, most AOGCMs have started to employ
the HadAM3 model (note that ‘weak’ does not imply ‘wrong’
complex sea ice dynamic components. The complexity of sea
since the real strength of the coupling is unknown).
ice dynamics in current AOGCMs varies from the relatively
Overall, the uncertainty in surface-atmosphere coupling has
simple ‘cavitating ﬂ uid’ model (Flato and Hibler, 1992) to the
implications for the reliability of the simulated soil moisture-
viscous-plastic model (Hibler, 1979), which is computationally
atmosphere feedback. It tempers our interpretation of the
expensive, particularly for global climate simulations.
response of the hydrologic cycle to simulated climate change in
The elastic-viscous-plastic model (Hunke and Dukowicz,
‘hot spot’ regions. Note that no assessment has been attempted
1997) is being increasingly employed, particularly due to its
for seasons other than NH summer.
efﬁ ciency for parallel computers. New numerical approaches
Since the TAR, there have been few assessments of the
for solving the ice dynamics equations include more accurate
capacity of climate models to simulate observed soil moisture.
representations on curvilinear model grids (Hunke and
Despite the tremendous effort to collect and homogenise soil
Dukowicz, 2002; Marsland et al., 2003; Zhang and Rothrock,
moisture measurements at global scales (Robock et al., 2000),
2003) and Lagrangian methods for solving the viscous-plastic
discrepancies between large-scale estimates of observed soil
equations (Lindsay and Stern, 2004; Wang and Ikeda, 2004).
moisture remain. The challenge of modelling soil moisture,
Treatment of sea ice thermodynamics in AOGCMs has
which naturally varies at small scales, linked to landscape
progressed more slowly: it typically includes constant
characteristics, soil processes, groundwater recharge, vegetation
conductivity and heat capacities for ice and snow (if represented),
type, etc., within climate models in a way that facilitates
a heat reservoir simulating the effect of brine pockets in the
comparison with observed data is considerable. It is not clear
ice, and several layers, the upper one representing snow. More
how to compare climate-model simulated soil moisture with
sophisticated thermodynamic schemes are being developed,
point-based or remotely sensed soil moisture. This makes
such as the model of Bitz and Lipscomb (1999), which introduces
assessing how well climate models simulate soil moisture, or
salinity-dependent conductivity and heat capacities, modelling
the change in soil moisture, difﬁ cult.
brine pockets in an energy-conserving way as part of a variable-
layer thermodynamic model (e.g., Saenko et al., 2002). Some
8.2.4 Cryospheric
Processes
AOGCMs include snow ice formation, which occurs when
an ice ﬂ oe is submerged by the weight of the overlying snow
8.2.4.1 Terrestrial
Cryosphere
cover and the ﬂ ooded snow layer refreezes. The latter process is
particularly important in the antarctic sea ice system.
Ice sheet models are used in calculations of long-term
Even with ﬁ ne grid scales, many sea ice models incorporate
warming and sea level scenarios, though they have not
sub-grid scale ice thickness distributions (Thorndike et al., 1975)
generally been incorporated in the AOGCMs used in Chapter
with several thickness ‘categories’, rather than considering
10. The models are generally run in ‘off-line’ mode (i.e., forced
the ice as a uniform slab with inclusions of open water. An
by atmospheric ﬁ elds derived from high-resolution time-slice
ice thickness distribution enables more accurate simulation of
experiments), although Huybrechts et al. (2002) and Fichefet
thermodynamic variations in growth and melt rates within a
et al. (2003) reported early efforts at coupling ice sheet models
single grid cell, which can have signiﬁ cant consequences for
to AOGCMs. Ice sheet models are also included in some
ice-ocean albedo feedback processes (e.g., Bitz et al., 2001;
EMICs (e.g., Calov et al., 2002). Ridley et al. (2005) pointed
Zhang and Rothrock, 2001). A well-resolved ice thickness
out that the time scale of projected melting of the Greenland
distribution enables a more physical formulation for ice ridging
Ice Sheet may be different in coupled and off-line simulations.
and rafting events, based on energetic principles. Although
Presently available thermomechanical ice sheet models do not
parametrizations of ridging mechanics and their relationship
606

Chapter 8
Climate Models and Their Evaluation
with the ice thickness distribution have improved (Babko et al.,
Coupling frequency is an important issue, because ﬂ uxes are
2002; Amundrud et al., 2004; Toyota et al., 2004), inclusion of
averaged during a coupling interval. Typically, most AOGCMs
advanced ridging parametrizations has lagged other aspects of
evaluated here pass ﬂ uxes and other variables between the
sea ice dynamics (rheology, in particular) in AOGCMs. Better
component parts once per day. The K-Proﬁ le Parametrization
numerical algorithms used for the ice thickness distribution
ocean vertical scheme (Large et al., 1994), used in several
(Lipscomb, 2001) and ice strength (Hutchings et al., 2004) have
models, is very sensitive to the wind energy available for
also been developed for AOGCMs.
mixing. If the models are coupled at a frequency lower than
once per ocean time step, nonlinear quantities such as wind
8.2.5
Aerosol Modelling and Atmospheric
mixing power (which depends on the cube of the wind speed)
Chemistry
must be accumulated over every time step before passing to
the ocean. Improper averaging therefore could lead to too
Climate simulations including atmospheric aerosols with
little mixing energy and hence shallower mixed-layer depths,
chemical transport have greatly improved since the TAR.
assuming the parametrization is not re-tuned. However, high
Simulated global aerosol distributions are better compared
coupling frequency can bring new technical issues. In the
with observations, especially satellite data (e.g., Advanced
MIROC model, the coupling interval is three hours, and in this
Very High Resolution Radar (AVHRR), Moderate Resolution
case, a poorly resolved internal gravity wave is excited in the
Imaging Spectroradiometer (MODIS), Multi-angle Imaging
ocean so some smoothing is necessary to damp this numerical
Spectroradiometer (MISR), Polarization and Directionality
problem. It should also be noted that the AOGCMs used here
of the Earth’s Reﬂ ectance (POLDER), Total Ozone Mapping
have relatively thick top oceanic grid boxes (typically 10 m or
Spectrometer (TOMS)), the ground-based network (Aerosol
more), limiting the sea surface temperature (SST) response to
Robotic Network; AERONET) and many measurement frequent coupling (Bernie et al., 2005).
campaigns (e.g., Chin et al., 2002; Takemura et al., 2002).
The global Aerosol Model Intercomparison project, AeroCom,
8.2.7
Flux Adjustments and Initialisation
has also been initiated in order to improve understanding of
uncertainties of model estimates, and to reduce them (Kinne
Since the TAR, more climate models have been developed
et al., 2003). These comparisons, combined with cloud that do not adjust the surface heat, water and momentum ﬂ uxes
observations, should result in improved conﬁ dence in the
artiﬁ cially to maintain a stable control climate. As noted by
estimation of the aerosol direct and indirect radiative forcing
Stouffer and Dixon (1998), the use of such ﬂ ux adjustments
(e.g., Ghan et al., 2001a,b; Lohmann and Lesins, 2002;
required relatively long integrations of the component models
Takemura et al., 2005). Interactive aerosol sub-component
before coupling. In these models, normally the initial conditions
models have been incorporated in some of the climate models
for the coupled integrations were obtained from long spin ups
used in Chapter 10 (HadGEM1 and MIROC). Some models
of the component models.
also include indirect aerosol effects (e.g., Takemura et al.,
In AOGCMs that do not use ﬂ ux adjustments (see Table 8.1),
2005); however, the formulation of these processes is still the
the initialisation methods tend to be more varied. The oceanic
subject of much research.
components of many models are initialised using values
Interactive atmospheric chemistry components are not obtained either directly from an observationally based, gridded
generally included in the models used in this report. However,
data set (Levitus and Boyer, 1994; Levitus and Antonov, 1997;
CCSM3 includes the modiﬁ
cation of greenhouse gas Levitus et al., 1998) or from short ocean-only integrations that
concentrations by chemical processes and conversion of sulphur
used an observational analysis for their initial conditions. The
dioxide and dimethyl sulphide to sulphur aerosols.
initial atmospheric component data are usually obtained from

atmosphere-only integrations using prescribed SSTs.
8.2.6 Coupling
Advances
To obtain initial data for the pre-industrial control integrations
discussed in Chapter 10, most AOGCMs use variants of the
In an advance since the TAR, a number of groups have
Stouffer et al. (2004) scheme. In this scheme, the coupled model
developed software allowing easier coupling of the various
is initialised as discussed above. The radiative forcing is then
components of a climate model (e.g., Valcke et al., 2006). An
set back to pre-industrial conditions. The model is integrated for
example, the Ocean Atmosphere Sea Ice Soil (OASIS) coupler,
a few centuries using constant pre-industrial radiative forcing,
developed at the Centre Europeen de Recherche et de Formation
allowing the coupled system to partially adjust to this forcing.
Avancee en Calcul Scientiﬁ c (CERFACS) (Terray et al., 1998),
The degree of equilibration in the real pre-industrial climate
has been used by many modelling centres to synchronise the
to the pre-industrial radiative forcing is not known. Therefore,
different models and for the interpolation of the coupling ﬁ elds
it seems unnecessary to have the pre-industrial control fully
between the atmosphere and ocean grids. The schemes for
equilibrated. After this spin-up integration, the pre-industrial
interpolation between the ocean and the atmosphere grids have
control is started and perturbation integrations can begin. An
been revised. The new schemes ensure both a global and local
important next step, once the start of the control integration is
conservation of the various ﬂ uxes at the air-sea interface, and
determined, is the assessment of the control integration climate
track terrestrial, ocean and sea ice ﬂ uxes individually.
drift. Large climate drifts can distort both the natural variability
607

Climate Models and Their Evaluation
Chapter 8
(e.g., Inness et al., 2003) and the climate response to changes in
those elements that can critically affect societies and natural
radiative forcing (Spelman and Manabe, 1984).
ecosystems and that are most likely to respond to changes in
In earlier IPCC reports, the initialisation methods were
radiative forcing.
quite varied. In some cases, the perturbation integrations were
initialised using data from control integrations where the SSTs
8.3.1 Atmosphere
were near present-day values and not pre-industrial. Given that
many climate models now use some variant of the Stouffer et
8.3.1.1
Surface Temperature and the Climate System’s
al. (2004) method, this situation has improved.
Energy Budget
For models to simulate accurately the global distribution of
8.3
Evaluation of Contemporary Climate
the annual and diurnal cycles of surface temperature, they must,
as Simulated by Coupled Global
in the absence of compensating errors, correctly represent a
Models
variety of processes. The large-scale distribution of annual mean
surface temperature is largely determined by the distribution
of insolation, which is moderated by clouds, other surface heat
Due to nonlinearities in the processes governing climate,
ﬂ uxes and transport of energy by the atmosphere and to a lesser
the climate system response to perturbations depends to
extent by the ocean. Similarly, the annual and diurnal cycles
some extent on its basic state (Spelman and Manabe, 1984).
of surface temperature are governed by seasonal and diurnal
Consequently, for models to predict future climatic conditions
changes in these factors, respectively, but they are also damped
reliably, they must simulate the current climatic state with some
by storage of energy in the upper layers of the ocean and to a
as yet unknown degree of ﬁ delity. Poor model skill in simulating
lesser degree the surface soil layers.
present climate could indicate that certain physical or dynamical
processes have been misrepresented. The better a model
8.3.1.1.1 Temperature
simulates the complex spatial patterns and seasonal and diurnal
Figure 8.2a shows the observed time mean surface
cycles of present climate, the more conﬁ dence there is that all
temperature as a composite of surface air temperature over
the important processes have been adequately represented.
regions of land and SST elsewhere. Also shown is the difference
Thus, when new models are constructed, considerable effort is
between the multi-model mean ﬁ eld and the observed ﬁ eld.
devoted to evaluating their ability to simulate today’s climate
With few exceptions, the absolute error (outside polar regions
(e.g., Collins et al., 2006; Delworth et al., 2006).
and other data-poor regions) is less than 2°C. Individual models
Some of the assessment of model performance presented
typically have larger errors, but in most cases still less than 3°C,
here is based on the 20th-century simulations that constitute
except at high latitudes (see Figure 8.2b and Supplementary
a part of the MMD archived at PCMDI. In these simulations,
Material, Figure S8.1). Some of the larger errors occur in
modelling groups initiated the models from pre-industrial (circa
regions of sharp elevation changes and may result simply
1860) ‘control’ simulations and then imposed the natural and
from mismatches between the model topography (typically
anthropogenic forcing thought to be important for simulating
smoothed) and the actual topography. There is also a tendency
the climate of the last 140 years or so. The 23 models considered
for a slight, but general, cold bias. Outside the polar regions,
here (see Table 8.1) are those relied on in Chapters 9 and 10 to
relatively large errors are evident in the eastern parts of the
investigate historical and future climate changes. Some ﬁ gures
tropical ocean basins, a likely symptom of problems in the
in this section are based on results from a subset of the models
simulation of low clouds. The extent to which these systematic
because the data set is incomplete.
model errors affect a model’s response to external perturbations
In order to identify errors that are systematic across models,
is unknown, but may be signiﬁ cant (see Section 8.6).
the mean of ﬁ elds available in the MMD, referred to here as the
In spite of the discrepancies discussed here, the fact is that
‘multi-model mean ﬁ eld’, will often be shown. The multi-model
models account for a very large fraction of the global temperature
mean ﬁ eld results are augmented by results from individual
pattern: the correlation coefﬁ cient between the simulated and
models available as Supplementary Material (see Figures S8.1
observed spatial patterns of annual mean temperature is typically
to S8.15). The multi-model averaging serves to ﬁ lter out biases
about 0.98 for individual models. This supports the view that
of individual models and only retains errors that are generally
major processes governing surface temperature climatology are
pervasive. There is some evidence that the multi-model mean
represented with a reasonable degree of ﬁ delity by the models.
ﬁ eld is often in better agreement with observations than any of the
An additional opportunity for evaluating models is afforded
ﬁ elds simulated by the individual models (see Section 8.3.1.1.2),
by the observed annual cycle of surface temperature. Figure
which supports continued reliance on a diversity of modelling
8.3 shows the standard deviation of monthly mean surface
approaches in projecting future climate change and provides some
temperatures, which is dominated by contributions from the
further interest in evaluating the multi-model mean results.
amplitudes of the annual and semi-annual components of the
Faced with the rich variety of climate characteristics that
annual cycle. The difference between the mean of the model
could potentially be evaluated here, this section focuses on
results and the observations is also shown. The absolute
608

Chapter 8
Climate Models and Their Evaluation
Like the annual range of temperature, the diurnal range (the
difference between daily maximum and minimum surface air
temperature) is much smaller over oceans than over land, where
it is also better observed, so the discussion here is restricted
to continental regions. The diurnal temperature range, zonally
and annually averaged over the continents, is generally too
small in the models, in many regions by as much as 50% (see
Supplementary Material, Figure S8.3). Nevertheless, the models
simulate the general pattern of this ﬁ eld, with relatively high
values over the clearer, drier regions. It is not yet known why
models generally underestimate the diurnal temperature range;
it is possible that in some models it is in part due to shortcomings
of the boundary-layer parametrizations or in the simulation of
freezing and thawing soil, and it is also known that the diurnal
cycle of convective cloud, which interacts strongly with surface
temperature, is rather poorly simulated.
Surface temperature is strongly coupled with the atmosphere
above it. This is especially evident at mid-latitudes, where
migrating cold fronts and warm fronts can cause relatively large
swings in surface temperature. Given the strong interactions
between the surface temperature and the temperature of the
air above, it is of special interest to evaluate how well models
simulate the vertical proﬁ le of atmospheric temperature. The
multi-model mean absolute error in the zonal mean, annual mean
air temperature is almost everywhere less than 2°C (compared
with the observed range of temperatures, which spans more
than 100°C when the entire troposphere is considered; see
Supplementary Material, Figure S8.4). It is notable, however,
that near the tropopause at high latitudes the models are generally
biased cold. This bias is a problem that has persisted for many
Figure 8.2. (a) Observed climatological annual mean SST and, over land, surface
air temperature (labelled contours) and the multi-model mean error in these
years, but in general is now less severe than in earlier models.
temperatures, simulated minus observed (colour-shaded contours). (b) Size of the
In a few of the models, the bias has been eliminated entirely, but
typical model error, as gauged by the root-mean-square error in this temperature,
computed over all AOGCM simulations available in the MMD at PCMDI. The Hadley
Centre Sea Ice and Sea Surface Temperature (HadISST; Rayner et al., 2003) climatol-
ogy of SST for 1980 to 1999 and the Climatic Research Unit (CRU; Jones et al., 1999)
climatology of surface air temperature over land for 1961 to 1990 are shown here.
The model results are for the same period in the 20th-century simulations. In the
presence of sea ice, the SST is assumed to be at the approximate freezing point of
seawater (–1.8°C). Results for individual models can be seen in the Supplementary
Material, Figure S8.1.
differences are in most regions less than 1°C. Even over extensive
land areas of the NH where the standard deviation generally
exceeds 10°C, the models agree with observations within 2°C
almost everywhere. The models, as a group, clearly capture
the differences between marine and continental environments
and the larger magnitude of the annual cycle found at higher
latitudes, but there is a general tendency to underestimate the
annual temperature range over eastern Siberia. In general, the
Figure 8.3. Observed standard deviation (labelled contours) of SST and, over land,
largest fractional errors are found over the oceans (e.g., over
surface air temperature, computed over the climatological monthly mean annual
much of tropical South America and off the east coasts of
cycle, and the multi-model mean error in the standard deviations, simulated minus
North America and Asia). These exceptions to the overall good
observed (colour-shaded contours). In most regions, the standard deviation provides
a measure of the amplitude of the seasonal range of temperature. The observational
agreement illustrate a general characteristic of current climate
data sets, the model results and the climatological periods are as described in Figure
models: the largest-scale features of climate are simulated more
8.2. Results for individual models can be seen in the Supplementary Material, Figure
accurately than regional- and smaller-scale features.
S8.2.
609

Climate Models and Their Evaluation
Chapter 8
compensating errors may be responsible. It is known
that the tropopause cold bias is sensitive to several
factors, including horizontal and vertical resolution,
non-conservation of moist entropy, and the treatment
of sub-grid scale vertical convergence of momentum
(‘gravity wave drag’). Although the impact of the
tropopause temperature bias on the model’s response
to radiative forcing changes has not been deﬁ nitively
quantiﬁ ed, it is almost certainly small relative to other
uncertainties.
8.3.1.1.2
The balance of radiation at the top of the
atmosphere
The primary driver of latitudinal and seasonal
variations in temperature is the seasonally varying
pattern of incident sunlight, and the fundamental
driver of the circulation of the atmosphere and ocean is
the local imbalance between the shortwave (SW) and
longwave (LW) radiation at the top of the atmosphere.
The impact on temperature of the distribution of
insolation can be strongly modiﬁ ed by the distribution
of clouds and surface characteristics.
Considering ﬁ rst the annual mean SW ﬂ ux at
the ‘top’ of the atmosphere (TOA)1, the insolation
is determined by well-known orbital parameters
that ensure good agreement between models and
observations. The annual mean insolation is strongest
in the tropics, decreasing to about half as much at the
poles. This largely drives the strong equator-to-pole
temperature gradient. As for outgoing SW radiation,
the Earth, on average, reﬂ ects about the same amount
of sunlight (~100 W m–2 in the annual mean) at all
latitudes. At most latitudes, the difference between
the multi-model mean zonally averaged outgoing
SW radiation and observations is in the annual mean
less than 6 W m–2 (i.e., an error of about 6%; see
Supplementary Material, Figure S8.5). Given that Figure 8.4. Root-mean-square (RMS) model error, as a function of latitude, in simulation of (a)
outgoing SW radiation reﬂ ected to space and (b) outgoing LW radiation. The RMS error is calculated
clouds are responsible for about half the outgoing SW
over all longitudes and over all 12 months of a climatology formed from several years of data.
radiation, these errors are not surprising, for it is known
The RMS statistic labelled ‘Mean Model’ is computed by ﬁ rst calculating the multi-model monthly
that cloud processes are among the most difﬁ cult to
mean ﬁ elds, and then calculating the RMS error (i.e., it is not the mean of the individual model RMS
errors). The Earth Radiation Budget Experiment (ERBE; Barkstrom et al., 1989) observational esti-
simulate with models (see Section 8.6.3.2.3).
mates used here are for the period 1985 to 1989 from satellite-based radiometers, and the model
There are additional errors in outgoing SW results are for the same period in the 20th-century simulations in the MMD at PCMDI. See Table 8.1
radiation due to variations with longitude and season,
for model descriptions. Results for individual models can be seen in the Supplementary Material,
and these can be quantiﬁ ed by means of the root-mean-
Figures S8.5 to S8.8.
square (RMS) error, calculated for each latitude over
all longitudes and months and plotted in Figure 8.4a (see also
model ﬁ elds. In the case of outgoing SW radiation, this is true at
Supplementary Material, Figure S8.6). Errors in the complete
nearly all latitudes. Calculation of the global mean RMS error,
two-dimensional ﬁ elds (see Supplementary Material, Figure
based on the monthly mean ﬁ elds and area-weighted over all
S8.6) tend to be substantially larger than the zonal mean errors
grid cells, indicates that the individual model errors are in the
of about 6 W m–2, an example of the common result that model
range 15 to 22 W m–2, whereas the error in the multi-model
errors tend to increase as smaller spatial scales and shorter time
mean climatology is only 13.1 W m–2. Why the multi-model
scales are considered. Figure 8.4a also illustrates a common
mean ﬁ eld turns out to be closer to the observed than the
result that the errors in the multi-model average of monthly
ﬁ elds in any of the individual models is the subject of ongoing
mean ﬁ elds are often smaller than the errors in the individual
research; a superﬁ cial explanation is that at each location and
1 The atmosphere clearly has no identiﬁ able ‘top’, but the term is used here to refer to an altitude above which the absorption of SW and LW radiation is negligibly small.
610

Chapter 8
Climate Models and Their Evaluation
for each month, the model estimates tend to scatter around
various kinds and by the ﬂ ow of air over orographic features.
the correct value (more or less symmetrically), with no single
For models to simulate accurately the seasonally varying
model consistently closest to the observations. This, however,
pattern of precipitation, they must correctly simulate a number
does not explain why the results should scatter in this way.
of processes (e.g., evapotranspiration, condensation, transport)
At the TOA, the net SW radiation is everywhere partially
that are difﬁ cult to evaluate at a global scale. Some of these are
compensated by outgoing LW radiation (i.e., infrared emissions)
discussed further in Sections 8.2 and 8.6. In this subsection, the
emanating from the surface and the atmosphere. Globally and
focus is on the distribution of precipitation and water vapour.
annually averaged, this compensation is nearly exact. The
Figure 8.5a shows observation-based estimates of annual
pattern of LW radiation emitted by earth to space depends most
mean precipitation and Figure 8.5b shows the multi-model
critically on atmospheric temperature, humidity, clouds and
mean ﬁ eld. At the largest scales, the lower precipitation rates
surface temperature. With a few exceptions, the models can
at higher latitudes reﬂ ect both reduced local evaporation at
simulate the observed zonal mean of the annual mean outgoing
lower temperatures and a lower saturation vapour pressure of
LW within 10 W m–2 (an error of around 5%; see Supplementary
cooler air, which tends to inhibit the transport of vapour from
Material, Figure S8.7). The models reproduce the relative
other regions. In addition to this large-scale pattern, captured
minimum in this ﬁ eld near the equator where the relatively high
well by models, is a local minimum in precipitation near the
humidity and extensive cloud cover in the tropics raises the
equator in the Paciﬁ c, due to a tendency for the Inter-Tropical
effective height (and lowers the effective temperature) at which
Convergence Zone (ITCZ)2 to reside off the equator. There
LW radiation emanates to space.
are local maxima at mid-latitudes, reﬂ ecting the tendency for
The seasonal cycle of the outgoing LW radiation pattern is
subsidence to suppress precipitation in the subtropics and for
also reasonably well simulated by models (see Figure 8.4b). The
storm systems to enhance precipitation at mid-latitudes. The
RMS error for most individual models varies from about 3% of
models capture these large-scale zonal mean precipitation
the outgoing LW radiation (OLR) near the poles to somewhat
differences, suggesting that they can adequately represent these
less than 10% in the tropics. The errors for the multi-model mean
features of atmospheric circulation. Moreover, there is some
simulation, ranging from about 2 to 6% across all latitudes, are
evidence provided in Section 8.3.5 that models have improved
again generally smaller than those in the individual models.
over the last several years in simulating the annual cycle of the
For a climate in equilibrium, any local annual mean imbalance
precipitation patterns.
in the net TOA radiative ﬂ ux (SW plus LW) must be balanced
Models also simulate some of the major regional
by a vertically integrated net horizontal divergence of energy
characteristics of the precipitation ﬁ eld, including the major
carried by the ocean and atmosphere. The fact that the TOA
convergence zones and the maxima over tropical rain forests,
SW and LW ﬂ uxes are well simulated implies that the models
although there is a tendency to underestimate rainfall over the
must also be properly accounting for poleward transport of
Amazon. When considered in more detail, however, there are
total energy by the atmosphere and ocean. This proves to be the
deﬁ ciencies in the multi-model mean precipitation ﬁ eld. There
case, with most models correctly simulating poleward energy
is a distinct tendency for models to orient the South Paciﬁ c
transport within about 10%. Although superﬁ cially this would
convergence zone parallel to latitudes and to extend it too far
seem to provide an important check on models, it is likely that
eastward. In the tropical Atlantic, the precipitation maximum
in current models compensating errors improve the agreement
is too weak in most models with too much rain south of the
of the simulations with observations. There are theoretical
equator. There are also systematic east-west positional errors
and model studies that suggest that if the atmosphere fails to
in the precipitation distribution over the Indo-Paciﬁ c Warm
transport the observed portion of energy, the ocean will tend to
Pool in most models, with an excess of precipitation over the
largely compensate (e.g., Shaffrey and Sutton, 2004).
western Indian Ocean and over the Maritime Continent. These
lead to systematic biases in the location of the major rising
8.3.1.2
Moisture and Precipitation
branches of the Walker Circulation and can compromise major
teleconnection3 pathways, in particular those associated with El
Water is fundamental to life, and if regional seasonal
Niño (e.g., Turner et al., 2005). Systematic dry biases over the
precipitation patterns were to change, the potential impacts
Bay of Bengal are related to errors in the monsoon simulations.
could be profound. Consequently, it is of real practical interest
Despite the apparent skill suggested by the multi-model
to evaluate how well models can simulate precipitation, not
mean (Figure 8.5), many models individually display substantial
only at global scales, but also regionally. Unlike seasonal
precipitation biases, especially in the tropics, which often
variation in temperature, which at large scales is strongly
approach the magnitude of the mean observed climatology
determined by the insolation pattern and conﬁ guration of the
(e.g., Johns et al., 2006; see also the Supplementary Material,
continents, precipitation variations are also strongly inﬂ uenced
Figures S8.9 and S8.10). Although some of these biases can be
by vertical movement of air due to atmospheric instabilities of
attributed to errors in the SST ﬁ eld of the coupled model, even
2 The ITCZ is manifested as a band of relatively intense convective precipitation, accompanied by surface convergence of moisture, which tends to locate seasonally over the
warmest surface temperatures and circumnavigates the earth in the tropics (though not continuously).
3 Teleconnection describes the process through which changes in one part of the climate system affect a remote location via changes in atmospheric circulation patterns.
611

Climate Models and Their Evaluation
Chapter 8
in the failure of models to capture the regional rainfall
patterns across the Indo-Paciﬁ c Warm Pool (Neale
and Slingo, 2003). Over the oceans, the precipitation
distribution along the ITCZ results from organised
convection associated with weather systems occurring
on synoptic and intra-seasonal time scales (e.g., the
Madden-Julian Oscillation (MJO); see Section 8.4.8).
These systems are frequently linked to convectively
coupled equatorial wave structures (e.g., Yang et al.,
2003), but these are poorly represented in models
(e.g., Lin et al., 2006; Ringer et al., 2006). Thus
the rain-bearing systems, which establish the mean
precipitation climatology, are not well simulated,
contributing also to the poor temporal characteristics
of daily rainfall (e.g., Dai, 2006) in which many
models simulate rain too frequently but with reduced
intensity.
Precipitation patterns are intimately linked to
atmospheric humidity, evaporation, condensation and
transport processes. Good observational estimates of
the global pattern of evaporation are not available, and
condensation and vertical transport of water vapour
can often be dominated by sub-grid scale convective
processes which are difﬁ cult to evaluate globally. The
best prospect for assessing water vapour transport
processes in humid regions, especially at annual and
longer time scales, may be to compare modelled and
observed streamﬂ ow, which must nearly balance
atmospheric transport since terrestrial water storage
variations on longer time scales are small (Milly et
Figure 8.5. Annual mean precipitation (cm), observed (a) and simulated (b), based on the multi-
al., 2005; see Section 8.3.4.2).
model mean. The Climate Prediction Center Merged Analysis of Precipitation (CMAP; Xie and Arkin,
Although an analysis of runoff in the MMD at
1997) observation-based climatology for 1980 to 1999 is shown, and the model results are for
PCMDI has not yet been performed, the net result of
the same period in the 20th-century simulations in the MMD at PCMDI. In (a), observations were
evaporation, transport and condensation processes
not available for the grey regions. Results for individual models can be seen in Supplementary
Material, Figure S8.9.
can be seen in the atmospheric humidity distribution.
Models reproduce the large-scale decrease in humidity
with both latitude and altitude (see Supplementary
atmosphere-only versions of the models show similarly large
Material, Figure S8.11), although this is not truly an independent
errors (e.g., Slingo et al., 2003). This may be one factor leading
check of models, since it is almost a direct consequence of their
to a lack of consensus among models even as to the sign of
reasonably realistic simulation of temperature. The multi-model
future regional precipitation changes predicted in parts of the
mean bias in humidity, zonally and annually averaged, is less
tropics (see Chapter 10).
than 10% throughout most of the lower troposphere compared
At the heart of understanding what determines the regional
with reanalyses, but model evaluation in the upper troposphere
distribution of precipitation over land and oceans in the tropics
is considerably hampered by observational uncertainty.
is atmospheric convection and its interaction with large-scale
Any errors in the water vapour distribution should affect
circulation. Convection occurs on a wide range of spatial
the outgoing LW radiation (see Section 8.3.1.1.2), which was
and temporal scales, and there is increasing evidence that
seen to be free of systematic zonal mean biases. In fact, the
interactions across all scales may be crucial for determining
observed differences in outgoing LW radiation between the
the mean tropical climate and its regional rainfall distributions
moist and dry regions are reproduced by the models, providing
(e.g., Khairoutdinov et al., 2005). Over tropical land, the
some conﬁ dence that any errors in humidity are not critically
diurnal cycle dominates, and yet many models have difﬁ culty
affecting the net ﬂ uxes at the TOA. However, the strength of
simulating the early evening maximum in rainfall. Instead, they
water vapour feedback, which strongly affects global climate
systematically tend to simulate rain before noon (Yang and
sensitivity, is primarily determined by fractional changes in
Slingo, 2001; Dai, 2006), which compromises the energy budget
water vapour in response to warming, and the ability of models
of the land surface. Similarly, the land-sea breezes around the
to correctly represent this feedback is perhaps better assessed
complex system of islands in Indonesia have been implicated
with process studies (see Section 8.6).
612

Chapter 8
Climate Models and Their Evaluation
8.3.1.3 Extratropical
Storms
8.3.2.1
Simulation of Mean Temperature and Salinity
Structure
The impact of extratropical cyclones on global climate
derives primarily from their role in transporting heat,
Before discussing the oceanic variables directly involved
momentum and humidity. Regionally and individually, these
in determining the climatic response, it is important to discuss
mid-latitude storms often provide beneﬁ cial precipitation, but
the ﬂ uxes between the ocean and atmosphere. Modelling
also occasionally produce destructive ﬂ ooding and high winds.
experience shows that the surface ﬂ uxes play a large part in
For these reasons, the effect of climate change on extratropical
determining the ﬁ delity of the oceanic simulation. Since the
cyclones is of considerable importance and interest.
atmosphere and ocean are coupled, the ﬁ delity of the oceanic
Among the several approaches used to characterise cyclone
simulation feeds back to the atmospheric simulation, affecting
activity (e.g., Paciorek et al., 2002), analysis methods that
the surface ﬂ uxes.
identify and track extratropical cyclones can provide the most
Unfortunately, the total surface heat and water ﬂ uxes (see
direct information concerning their frequency and movement
Supplementary Material, Figure S8.14) are not well observed.
(Hoskins and Hodges, 2002, 2005). Climatologies for the
Normally, they are inferred from observations of other ﬁ elds,
distribution and properties of cyclones found in models can be
such as surface temperature and winds. Consequently, the
compared with reanalysis products (Chapter 3), which provide
uncertainty in the observational estimate is large – of the order
the best observation-constrained data.
of tens of watts per square metre for the heat ﬂ ux, even in the
Results from a systematic analysis of AMIP-2 simulations
zonal mean. An alternative way of assessing the surface ﬂ uxes
(Hodges, 2004; Stratton and Pope, 2004) indicate that models
is by looking at the horizontal transports in the ocean. In the
run with observed SSTs are capable of producing storm tracks
long-term average, the heat and water storage in the ocean
located in about the right locations, but nearly all show some
are small so that the horizontal transports have to balance the
deﬁ ciency in the distribution and level of cyclone activity. In
surface ﬂ uxes. Since the heat transport seems better constrained
particular, simulated storm tracks are often more zonally oriented
by the available observations, it is presented here.
than is observed. A study by Lambert and Fyfe (2006), based on
North of 45°N, most model simulations transport too much
the MMD at PCMDI, ﬁ nds that as a group, the recent models,
heat northward when compared to the observational estimates
which include interactive oceans, tend to underestimate slightly
used here (Figure 8.6), but there is uncertainty in the observations.
the total number of cyclones in both hemispheres. However, the
At 45°N, for example, the model simulations lie much closer to
number of intense storms is slightly overestimated in the NH,
the estimate of 0.6 x 1015 W obtained by Ganachaud and Wunsch
but underestimated in the Southern Hemisphere (SH), although
(2003). From 45°N to the equator, most model estimates lie near
observations are less certain there.
or between the observational estimates shown. In the tropics
Increases in model resolution (characteristic of models
and subtropical zone of the SH, most models underestimate
over the last several years) appear to improve some aspects of
the southward heat transport away from the equator. At middle
extratropical cyclone climatology (Bengtsson et al., 2006),
particularly in the NH where observations are most reliable
(Hodges et al., 2003; Hanson et al., 2004; Wang et al.,
2006). Improvements to the dynamical core and physics of
models have also led to better agreement with reanalyses
(Ringer et al., 2006; Watterson, 2006).
Our assessment is that although problems remain,
climate models are improving in their simulation of
extratropical cyclones.
8.3.2 Ocean
As noted earlier, this chapter focuses only on those
variables important in determining the transient response of
climate models (see Section 8.6). Due to space limitations,
much of the analysis performed for this section is found in
the Supplementary Material (Figures S8.12 to S8.15). An
assessment of the modes of natural, internally generated
Figure 8.6. Annual mean, zonally averaged oceanic heat transport implied by net heat
ﬂ ux imbalances at the sea surface, under an assumption of negligible changes in oceanic
variability can be found in Section 8.4. Comparisons of the
heat content. The observationally based estimate, taken from Trenberth and Caron (2001)
type performed here need to be made with an appreciation
for the period February 1985 to April 1989, derives from reanalysis products from the
of the uncertainties in the historical estimates of radiative
National Centers for Environmental Prediction (NCEP)/NCAR (Kalnay et al., 1996) and
forcing and various sampling issues in the observations
European Centre for Medium Range Weather Forecasts 40-year reanalysis (ERA40; Uppala
et al., 2005). The model climatologies are derived from the years 1980 to 1999 in the
(see Chapters 2 and 5). Unless otherwise noted, all results
20th-century simulations in the MMD at PCMDI. The legend identifying individual models
discussed here are based on the MMD at PCMDI.
appears in Figure 8.4.
613

Climate Models and Their Evaluation
Chapter 8
and high latitudes of the SH, the observational estimates are
more uncertain and the model-simulated heat transports tend to
surround the observational estimates.
The oceanic heat ﬂ uxes have large seasonal variations
which lead to large variations in the seasonal storage of heat
by the oceans, especially in mid-latitudes. The oceanic heat
storage tends to damp and delay the seasonal cycle of surface
temperature. The model simulations evaluated here agree well
with the observations of seasonal heat storage by the oceans
(Gleckler et al., 2006a). The most notable problem area for the
models is in the tropics, where many models continue to have
biases in representing the ﬂ ow of heat from the tropics into
middle and high latitudes.
The annually averaged zonal component of surface wind
Figure 8.7. Annual mean east-west component of wind stress zonally averaged
stress, zonally averaged over the oceans, is reasonably well
over the oceans. The observationally constrained estimate is from the years 1980
simulated by the models (Figure 8.7). At most latitudes, the
to 1999 in the European Centre for Medium Range Weather Forecasts 40-year re-
reanalysis estimates (based on atmospheric models constrained
analysis (ERA40; Uppala et al., 2005), and the model climatologies are calculated for
by observations) lie within the range of model results. At middle
the same period in the 20th-century simulations in the MMD at PCMDI. The legend
identifying individual models appears in Figure 8.4.
to low latitudes, the model spread is relatively small and all
the model results lie fairly close to the reanalysis. At middle to
high latitudes, the model-simulated wind stress maximum tends
to lie equatorward of the reanalysis. This error is particularly
large in the SH, a region where there is more uncertainty in
the reanalysis. Almost all model simulations place the SH wind
stress maximum north of the reanalysis estimate. The Southern
Ocean wind stress errors in the control integrations may
adversely affect other aspects of the simulation and possibly the
oceanic heat uptake under climate change, as discussed below.
The largest individual model errors in the zonally averaged
SST (Figure 8.8) are found at middle and high latitudes,
particularly the mid-latitudes of the NH where the model-
simulated temperatures are too cold. Almost every model
has some tendency for this cold bias. This error seems to be
associated with poor simulation of the path of the North
Atlantic Current and seems to be due to an ocean component
Figure 8.8. Annual mean, zonally averaged SST error, simulated minus observed
problem rather than a problem with the surface ﬂ uxes. In the
climatology. The Hadley Centre Sea Ice and Sea Surface Temperature (HadISST; Rayner
zonal averages near 60°S, there is a warm bias in the multi-
et al., 2003) observational climatology for 1980 to 1999 is the reference used here, and
the model results are for the same period in the 20th-century simulations in the MMD
model mean results. Many models suffer from a too-warm bias
at PCMDI. In the presence of sea ice, the SST is assumed to be at the freezing point of
in the Southern Ocean SSTs.
seawater. The legend identifying individual models appears in Figure 8.4.
In the individual model SST error maps (see Supplementary
Material, Figure S8.1), it is apparent that most models have a
large warm bias in the eastern parts of the tropical ocean basins,
zonally averaged SST error is less than 2°C, which is fairly small
near the continental boundaries. This is also evident in the
considering that most models do not use ﬂ ux adjustments in these
multi-model mean result (Figure 8.2a) and is associated with
simulations. The model mean local SST errors are also less than
insufﬁ cient resolution, which leads to problems in the simulation
2°C over most regions, with only relatively small areas exceeding
of the local wind stress, oceanic upwelling and under-prediction
this value. Even relatively small SST errors, however, can
of the low cloud amounts (see Sections 8.2 and 8.3.1). These
adversely affect the simulation of variability and teleconnections
are also regions where there is a relatively large spread among
(Section 8.4).
the model simulations, indicating a relatively wide range in
Over most latitudes, at depths ranging from 200 to 3,000
the magnitude of these errors. Another area where the model
m, the multi-model mean zonally averaged ocean temperature
error spread is relatively large is found in the North Atlantic
is too warm (see Figure 8.9). The maximum warm bias (about
Ocean. As noted above, this is an area where many models have
2°C) is located in the region of the North Atlantic Deep Water
problems properly locating the North Atlantic Current, a region
(NADW) formation. Above 200 m, however, the multi-model
of large SST gradients.
mean is too cold, with maximum cold bias (more than 1°C)
In spite of the errors, the model simulation of the SST ﬁ eld is
near the surface at mid-latitudes of the NH, as discussed above.
fairly realistic overall. Over all latitudes, the multi-model mean
Most models generally have an error pattern similar to the
614

Chapter 8
Climate Models and Their Evaluation
Figure 8.9. Time-mean observed potential temperature (°C), zonally averaged over all ocean basins (labelled contours) and multi-model mean error in this ﬁ eld, simulated
minus observed (colour-ﬁ lled contours). The observations are from the 2004 World Ocean Atlas compiled by Levitus et al. (2005) for the period 1957 to 1990, and the model
results are for the same period in the 20th-century simulations in the MMD at PCMDI. Results for individual models can be seen in the Supplementary Material, Figure S8.12.
multi-model mean (see Supplementary Material, Figure S8.12)
depth. These waters then ﬂ ow southward towards the Southern
except for CNRM-CM3 and MRI-CGCM2.3.2, which are too
Ocean where they mix with the rest of the World Ocean waters
cold throughout most of the mid- and low-latitude ocean (see
(see Supplementary Material, Figure S8.15).
Supplementary Material, Figure S8.12). The GISS-EH model
The models simulate this major aspect of the MOC and
is much too cold throughout the subtropical thermocline and
also simulate a number of distinct wind-driven surface cells
only the NH part of the FGOALS-g1.0 error pattern is similar
(see Supplementary Material, Figure S8.15). In the tropics and
to the model mean error described here. The magnitude of these
subtropics, these cells are quite shallow, but at the latitude of the
errors, especially in the deeper parts of the ocean, depends on
Drake Passage (55°S) the wind-driven cell extends to a much
the AOGCM initialisation method (Section 8.2.7).
greater depth (2 to 3 km). Most models in the multi-model
The error pattern, in which the upper 200 m of the ocean tend
data set have some manifestation of the wind-driven cells. The
to be too cold while the layers below are too warm, indicates
strength and pattern of the overturning circulation varies greatly
that the thermocline in the multi-model mean is too diffuse.
from model to model (see Supplementary Material, Figure
This error, which was also present at the time of the TAR, seems
S8.15). The GISS-AOM exhibits the strongest overturning
partly related to the wind stress errors in the SH noted above
circulation, almost 40 to 50 Sv (106 m3 s–1). The CGCM
and possibly to errors in formation and mixing of NADW. The
(T47 and T63) and FGOALS have the weakest overturning
multi-model mean errors in temperature (too warm) and salinity
circulations, about 10 Sv. The observed value is about 18 Sv
(too salty; see Supplementary Material, Figure S8.13) at middle
(Ganachaud and Wunsch 2000).
and low latitudes near the base of the thermocline tend to cancel
In the Atlantic, the MOC, extending to considerable depth,
in terms of a density error and appear to be associated with
is responsible for a large fraction of the northward oceanic
the problems in the formation of Antarctic Intermediate Water
heat transport in both observations and models (e.g., Hall and
(AAIW), as discussed below.
Bryden, 1982; Gordon et al., 2000). Figure 10.15 contains an
index of the Atlantic MOC at 30°N for the suite of AOGCM
8.3.2.2
Simulation of Circulation Features Important for
20th-century simulations. While the majority of models show
Climate Response
an MOC strength that is within observational uncertainty, some
show higher and lower values and a few show substantial drifts
8.3.2.2.1
Meridional overturning circulation
which could make interpretation of MOC projections using
The MOC is an important component of present-day climate
those models very difﬁ cult.
and many models indicate that it will change in the future
Overall, some aspects of the simulation of the MOC have
(Chapter 10). Unfortunately, many aspects of this circulation
improved since the TAR. This is due in part to improvements
are not well observed. The MOC transports large amounts of
in mixing schemes, the use of higher resolution ocean models
heat and salt into high latitudes of the North Atlantic Ocean,
(see Section 8.2) and better simulation of the surface ﬂ uxes.
where the relatively warm, salty surface waters are cooled
This improvement can be seen in the individual model MOC
by the atmosphere, making the water dense enough to sink to
sections (see Supplementary Material, Figure S8.15) by the
615

Climate Models and Their Evaluation
Chapter 8
fact that (1) the location of the deep-water formation is more
8.3.3 Sea
Ice
realistic, with more sinking occurring in the Greenland-Iceland-
Norwegian and Labrador Seas as evidenced by the larger stream
The magnitude and spatial distribution of the high-
function values north of the sill located at 60°N (e.g., Wood et
latitude climate changes can be strongly affected by sea ice
al., 1999) and (2) deep waters are subjected to less spurious
characteristics, but evaluation of sea ice in models is hampered
mixing, resulting in better water mass properties (Thorpe et
by insufﬁ cient observations of some key variables (e.g., ice
al., 2004) and a larger fraction of the water that sinks in the
thickness) (see Section 4.4). Even when sea ice errors can be
northern part of the North Atlantic Ocean exiting the Atlantic
quantiﬁ ed, it is difﬁ cult to isolate their causes, which might
Ocean near 30°S (Danabasoglu et al., 1995). There is still room
arise from deﬁ ciencies in the representation of sea ice itself,
for improvement in the models’ simulation of these processes,
but could also be due to ﬂ awed simulation of the atmospheric
but there is clear evidence of improvement in many of the
and oceanic ﬁ elds at high latitudes that drive ice movement (see
models analysed here.
Sections 8.3.1, 8.3.2 and 11.3.8).
Although sea ice treatment in AOGCMs has become
8.3.2.2.2
Southern Ocean circulation
more sophisticated, including better representation of both
The Southern Ocean wind stress error has a particularly large
the dynamics and thermodynamics (see Section 8.2.4),
detrimental impact on the Southern Ocean simulation by the
improvement in simulating sea ice in these models, as a group,
models. Partly due to the wind stress error identiﬁ ed above, the
is not obvious (compare Figure 8.10 with TAR Figure 8.10;
simulated location of the Antarctic Circumpolar Current (ACC)
or Kattsov and Källén, 2005, Figure 4.11). In some models,
is also too far north in most models (Russell et al., 2006). Since
however, the geographic distribution and seasonality of sea ice
the AAIW is formed on the north side of the ACC, the water
is now better reproduced.
mass properties of the AAIW are distorted (typically too warm
For the purposes of model evaluation, the most reliably
and salty: Russell et al., 2006). The relatively poor AAIW
measured characteristic of sea ice is its seasonally varying extent
simulation contributes to the multi-model mean error identiﬁ ed
(i.e., the area enclosed by the ice edge, operationally deﬁ ned as
above where the thermocline is too diffuse, because the waters
the 15% contour; see Section 4.4). Despite the wide differences
near the base of thermocline are too warm and salty.
among the models, the multi-model mean of sea ice extent is
It is likely that the relatively poor Southern Ocean simulation
in reasonable agreement with observations. Based on 14 of the
will inﬂ uence the transient climate response to increasing
15 AOGCMs available at the time of analysis (one model was
greenhouse gases by affecting the oceanic heat uptake. When
excluded because of unrealistically large ice extents; Arzel et
forced by increases in radiative forcing, models with too
al., 2006), the mean extent of simulated sea ice exceeds that
little Southern Ocean mixing will probably underestimate the
observed in the NH by up to roughly 1 x 106 km2 throughout the
ocean heat uptake; models with too much mixing will likely
year, whereas in the SH the annual cycle is exaggerated, with too
exaggerate it. These errors in oceanic heat uptake will also have
much sea ice in September (~2 x 106 km2) and too little in March
a large impact on the reliability of the sea level rise projections.
by a lesser amount. In many models the regional distribution of
See Chapter 10 for more discussion of this subject.
sea ice is poorly simulated, even if the hemispheric areal extent is
approximately correct (Arzel et al., 2006; Holland and Raphael,
8.3.2.3
Summary of Oceanic Component Simulation
2006; Zhang and Walsh, 2006). The spread of simulated sea ice
extents, measured as the multi-model standard deviation from
Overall, the improvements in the simulation of the observed
the model mean, is generally narrower in the NH than in the
time mean ocean state noted in the TAR (McAvaney et al., 2001)
SH (Arzel et al., 2006). Even in the best case (NH winter), the
have continued in the models evaluated here. It is notable that
range of simulated sea ice extent exceeds 50% of the mean, and
this improvement has continued in spite of the fact that nearly
ice thickness also varies considerably, suggesting that projected
all models no longer use ﬂ ux adjustments. This suggests that
decreases in sea ice cover remain rather uncertain. The model sea
the improvements in the physical parametrizations, increased
ice biases may inﬂ uence global climate sensitivity (see Section
resolution (see Section 8.2) and improved surface ﬂ uxes are
8.6). There is a tendency for models with relatively large sea ice
together having a positive impact on model simulations. The
extent in the present climate to have higher sensitivity. This is
temperature and salinity errors in the thermocline, while still
apparently especially true of models with low to moderate polar
large, have been reduced in many models. In the NH, many
ampliﬁ cation (Holland and Bitz, 2003).
models still suffer from a cold bias in the upper ocean which is
Among the primary causes of biases in simulated sea
at a maximum near the surface and may distort the ice-albedo
ice (especially its distribution) are biases in the simulation
feedback in some models (see Section 8.3.3). In the Southern
of high-latitude winds (Bitz et al., 2002; Walsh et al., 2002;
Ocean, the equatorward bias of the westerly wind stress
Chapman and Walsh, 2007), as well as vertical and horizontal
maximum found in most model simulations is a problem that
mixing in the ocean (Arzel et al., 2006). Also important are
may affect the models’ response to increasing radiative forcing.
surface heat ﬂ ux errors, which in particular may result from
inadequate parametrizations of the atmospheric boundary layer
(under stable conditions commonly occurring at night and in
the winter over sea ice) and generally from poor simulation
616

Chapter 8
Climate Models and Their Evaluation
8.3.4.1 Snow
Cover
Analysis and comparison of AMIP-2
results, available at the time of the TAR,
and more recent AOGCM results in the
present MMD at PCMDI, show that
models are now more consistent in their
simulation of snow cover. Problems
remain, however, and Roesch (2006)
showed that the recent models predict
excessive snow water equivalent (SWE)
in spring, likely because of excessive
winter precipitation. Frei et al. (2005)
found that AMIP-2 models simulate
the seasonal timing and the relative
spatial patterns of SWE over North
America fairly well, but identiﬁ ed
a tendency to overestimate ablation
during spring. At the continental scale,
the highest monthly SWE integrated
over the North American continent in
AMIP-2 models varies within ±50% of
the observed value of about 1,500 km3.
The magnitude of these model errors
is large enough to affect continental
water balances. Snow cover area (SCA)
is well captured by the recent models,
but interannual variability is too low
during melt. Frei et al. (2003) showed
where observations were within the
inter-quartile range of AMIP-2 models
for all months at the hemispheric and
continental scale. Encouragingly, there
was signiﬁ
cant improvement over
earlier AMIP-1 simulations for seasonal
Figure 8.10. Baseline climate (1980–1999) sea ice distribution in the Northern Hemisphere (upper panels) and
and interannual variability of SCA (Frei
Southern Hemisphere (lower panels) simulated by 14 of the AOGCMs listed in Table 8.1 for March (left) and Sep-
tember (right), adapted from Arzel et al. (2006). For each 2.5° x 2.5° longitude-latitude grid cell, the ﬁ gure indicates
et al., 2005). Both the recent AOGCMs
the number of models that simulate at least 15% of the area covered by sea ice. The observed 15% concentration
and AMIP models reproduced the
boundaries (red line) are based on the Hadley Centre Sea Ice and Sea Surface Temperature (HadISST; Rayner et al.,
observed decline in annual SCA over the
2003) data set.
period 1979 to 1995 and most models
captured the observed decadal-scale
of high-latitude cloudiness, which is evident from the striking
variability over the 20th century. Despite these improvements,
inter-model scatter (e.g., Kattsov and Källén, 2005).
a minority of models still exaggerate SCA.
Large discrepancies remain in albedo for forested areas
8.3.4 Land
Surface
under snowy conditions, due to difﬁ culties in determining
the extent of masking of snow by vegetation (Roesch, 2006).
Evaluation of the land surface component in coupled models
The ability of terrestrial models to simulate snow under
is severely limited by the lack of suitable observations. The
observed meteorological forcing has been evaluated via several
terrestrial surface plays key climatic roles in inﬂ uencing the
intercomparisons. At the scale of individual grid cells, for mid-
partitioning of available energy between sensible and latent heat
latitude (Slater et al., 2001) and alpine (Etchevers et al., 2004)
ﬂ uxes, determining whether water drains or remains available
locations, the spread of model simulations usually encompasses
for evaporation, determining the surface albedo and whether
observations. However, grid-box scale simulations of snow
snow melts or remains frozen, and inﬂ uencing surface ﬂ uxes of
over high-latitude river basins identiﬁ ed signiﬁ cant limitations
carbon and momentum. Few of these can be evaluated at large
(Nijssen et al., 2003), due to difﬁ culties relating to calculating
spatial or long temporal scales. This section therefore evaluates
net radiation, fractional snow cover and interactions with
those quantities for which some observational data exist.
vegetation.
617

Climate Models and Their Evaluation
Chapter 8
8.3.4.2 Land
Hydrology
absorbed solar and infrared radiation at the surface leads
inevitably to uncertainty in the simulation of surface sensible
The evaluation of the hydrological component of climate
and latent heat ﬂ uxes.
models has mainly been conducted uncoupled from AOGCMs
(Bowling et al., 2003; Nijssen et al., 2003; Boone et al., 2004).
8.3.4.4 Carbon
This is due in part to the difﬁ culties of evaluating runoff
simulations across a range of climate models due to variations
A major advance since the TAR is some systematic
in rainfall, snowmelt and net radiation. Some attempts have,
assessments of the capability of land surface models to simulate
however, been made. Arora (2001) used the AMIP-2 framework
carbon. Dargaville et al. (2002) evaluated the capacity of four
to show that the Canadian Climate Model’s simulation of the
global vegetation models to simulate the seasonal dynamics and
global hydrological cycle compared well to observations, but
interannual variability of atmospheric CO2 between 1980 and
regional variations in rainfall and runoff led to differences at
1991. Using off-line forcing, they evaluated the capacity of these
the basin scale. Gerten et al. (2004) evaluated the hydrological
models to simulate carbon ﬂ uxes, via an atmospheric transport
performance of the Lund-Potsdam-Jena (LPJ) model and
model, using observed atmospheric CO2 concentration. They
showed that the model performed well in the simulation
found that the terrestrial models tended to underestimate the
of runoff and evapotranspiration compared to other global
amplitude of the seasonal cycle and simulated the spring uptake
hydrological models, although the version of LPJ assessed had
of CO2 approximately one to two months too early. Of the four
been enhanced to improve the simulation of hydrology over the
models, none was clearly superior in its capacity to simulate
versions used by Sitch et al. (2003).
the global carbon budget, but all four reproduced the main
Milly et al. (2005) made use of the MMD, which contains
features of the observed seasonal cycle in atmospheric CO2. A
results from recent models, to investigate whether observed 20th-
further off-line evaluation of the LPJ global vegetation model
century trends in regional land hydrology could be attributed to
by Sitch et al. (2003) provided conﬁ dence that the model could
variations in atmospheric composition and solar irradiance. Their
replicate the observed vegetation pattern, seasonal variability in
analysis, based on an ensemble of 26 integrations of 20th-century
net ecosystem exchange and local soil moisture measurements
climate from nine climate models, showed that at regional scales
when forced by observed climatologies.
these models simulated observed streamﬂ ow measurements
The only systematic evaluation of carbon models that were
with good qualitative skill. Further, the models demonstrated
interactively coupled to climate models occurred as part of the
highly signiﬁ cant quantitative skill in identifying the regional
Coupled Climate-Carbon Cycle Model Intercomparison Project
runoff trends indicated by 165 long-term stream gauges. They
(C4MIP), where Friedlingstein et al. (2006) compared the ability
concluded that the impact of changes in atmospheric composition
of a suite of models to simulate historical atmospheric CO2
and solar irradiance on observed streamﬂ ow was, at least in part,
concentration forced by observed emissions. Issues relating
predictable. This is an important scientiﬁ c advance: it suggests
to the magnitude of the fertilization effect and the partitioning
that despite limitations in the hydrological parametrizations
between land and ocean uptake were identiﬁ ed in individual
included in climate models, these models can capture observed
models, but it is only under increasing CO2 in the future (see
changes in 20th-century streamﬂ ow associated with atmospheric
Chapter 10) that the differences become large. Several other
composition and solar irradiance changes. This enhances groups have evaluated the impact of coupling speciﬁ c models
conﬁ dence in the use of these models for future projection.
of carbon to climate models but clear results are difﬁ cult to
obtain because of inevitable biases in both the terrestrial and
8.3.4.3 Surface
Fluxes
atmospheric modules (e.g., Delire et al., 2003).
Despite considerable effort since the TAR, uncertainties
8.3.5
Changes in Model Performance
remain in the representation of solar radiation in climate
models (Potter and Cess, 2004). The AMIP-2 results and the
Standard experiments, agreed upon by the climate modelling
recent model results in the MMD provide an opportunity for
community to facilitate model intercomparison (see Section
a major systematic evaluation of model ability to simulate
8.1.2.2), have produced archives of model output that make
solar radiation. Wild (2005) and Wild et al. (2006) evaluated
it easier to track historical changes in model performance.
these models and found considerable differences in the global
Most of the modelling groups that contributed output to the
annual mean solar radiation absorbed at the Earth’s surface.
current MMD at PCMDI also archived simulations from their
In comparison to global surface observations, Wild (2005)
earlier models (circa 2000) as part of the Coupled Model
concluded that many climate models overestimate surface
Intercomparison Project (CMIP1&2). The TAR largely relied
absorption of solar radiation partly due to problems in the
on the earlier generation of models in its assessment.
parametrizations of atmospheric absorption, clouds and aerosols.
Based on the archived model output, it is possible to quantify
Similar uncertainties exist in the simulation of downwelling
changes in performance of evolving models.4 This can be done
infrared radiation (Wild et al., 2001). Difﬁ culties in simulating
most straightforwardly by only considering the 14 modelling
4 One modelling group participating in CMIP1&2 did not contribute to the MMD, and four groups providing output to the MMD did not do so for CMIP1&2. Results from these ﬁ ve
groups are therefore not considered in this subsection. Some modelling groups contributed results from more than one version of their model (sometimes, simply running it at two
different resolutions), and in these cases the mean of the two model results is considered here.
618

Chapter 8
Climate Models and Their Evaluation
groups that contributed output from both their earlier and more
The models in Figure 8.11 are categorised based on whether
recent models. One important aspect of model skill is how well
or not ﬂ ux adjustments were applied (see Section 8.2.7). Of
the models simulate the seasonally varying global pattern of
the earlier generation models, 8 of the 14 models were ﬂ ux
climatically important ﬁ elds. The only monthly mean ﬁ elds
adjusted, but only two of these groups continue this practice.
available in the CMIP1&2 archive are surface air temperature,
Several conclusions can be drawn from the ﬁ gure: 1) although
precipitation and mean sea level pressure, so these are the
ﬂ ux-adjusted models on average have smaller errors than those
focus of this analysis. Although the simulation conditions
without (in both generations), the smallest errors in simulating
in the MMD 20th-century simulations were not identical to
sea level pressure and surface temperature are found in models
those in the CMIP1&2 control runs, the differences do not
without ﬂ ux adjustment; 2) despite the elimination of ﬂ ux
alter the conclusions summarised below because the large-
adjustment in all but two of the recent models, the mean error
scale climatological features dominate, not the relatively small
obtained from the recent suite of 14 models is smaller than
perturbations resulting from climate change.
errors found in the corresponding earlier suite of models; and
A summary of the ability of AOGCMs to simulate the
3) models without ﬂ ux adjustment have improved on average,
seasonally varying climate state is provided by Figure 8.11,
as have the ﬂ ux-adjusted models. An exception to this last
which displays error measures that gauge how well recent
statement is the slight increase in mean RMS error for sea
models simulate precipitation, sea level pressure and surface
level pressure found in non-ﬂ ux-adjusted models. Despite no
temperature, compared with their predecessors. The normalised
apparent improvement in the mean in this case, three of the
RMS error shown is a so-called space-time statistic, computed
recent generation models have smaller sea level pressure errors
from squared errors, summed over all 12 climatological months
than any of the earlier models.
and over the entire globe, with grid cell values weighted by
These results demonstrate that the models now being used in
the corresponding grid cell area. This statistic can be used
applications by major climate modelling groups better simulate
to assess the combined contributions of both spatial pattern
seasonally varying patterns of precipitation, mean sea level
errors and seasonal cycle errors. The RMS error is divided by
pressure and surface air temperature than the models relied on
the corresponding observed standard deviation of the ﬁ eld to
by these same groups at the time of the TAR.
provide a relative measure of the error. In Figure 8.11 this scaling
implies that pressure is better simulated than precipitation, and
that surface temperature is simulated best of all.
Figure 8.11. Normalised RMS error in simulation of climatological patterns of monthly precipitation, mean sea level pressure and surface air temperature. Recent AOGCMs
(circa 2005) are compared to their predecessors (circa 2000 and earlier). Models are categorised based on whether or not any ﬂ ux adjustments were applied. The models are
gauged against the following observation-based datasets: Climate Prediction Center Merged Analysis of Precipitation (CMAP; Xie and Arkin, 1997) for precipitation (1980–1999),
European Centre for Medium Range Weather Forecasts 40-year reanalysis (ERA40; Uppala et al., 2005) for sea level pressure (1980–1999) and Climatic Research Unit (CRU;
Jones et al., 1999) for surface temperature (1961–1990). Before computing the errors, both the observed and simulated ﬁ elds were mapped to a uniform 4° x 5° latitude-longi-
tude grid. For the earlier generation of models, results are based on the archived output from control runs (speciﬁ cally, the ﬁ rst 30 years, in the case of temperature, and the ﬁ rst
20 years for the other ﬁ elds), and for the recent generation models, results are based on the 20th-century simulations with climatological periods selected to correspond with
observations. (In both groups of models, results are insensitive to the period selected.)
619

Climate Models and Their Evaluation
Chapter 8
8.4
Evaluation of Large-Scale Climate
Like its NH counterpart, the NAM, the Southern Annular
Variability as Simulated by Coupled
Mode (SAM; see Chapters 3 and 9) has signatures in the
Global Models
tropospheric circulation, the stratospheric polar vortex, mid-
latitude storm tracks, ocean circulation and sea ice. AOGCMs
generally simulate the SAM realistically (Fyfe et al., 1999;
The atmosphere-ocean coupled climate system shows various
Cai et al., 2003; Miller et al., 2006). For example, Figure 8.12
modes of variability that range widely from intra-seasonal to
compares the austral winter SAM simulated in the MMD at
inter-decadal time scales. Successful simulation and prediction
PCMDI to the observed SAM as represented in the National
over a wide range of these phenomena increase conﬁ dence in
Centers for Environmental Prediction (NCEP) reanalysis. The
the AOGCMs used for climate predictions of the future.
main elements of the pattern, the low-pressure anomaly over
Antarctica and the high-pressure anomalies equatorward of 60°S
8.4.1
Northern and Southern Annular Modes
are captured well by the AOGCMs. In all but two AOGCMs, the
spatial correlation between the observed and simulated SAM is
There is evidence (e.g., Fyfe et al., 1999; Shindell et al.,
greater than 0.95. Further analysis shows that the SAM signature
1999) that the simulated response to greenhouse gas forcing in
in surface temperature, such as the surface warm anomaly over
AOGCMs has a pattern that resembles the models’ Northern
the Antarctic Peninsula associated with a positive SAM event,
Annular Mode (NAM), and thus it would appear important
is also captured by some AOGCMs (e.g., Delworth et al., 2006;
that the NAM (see Chapters 3 and 9) is realistically simulated.
Otto-Bliesner et al., 2006). This follows from the realistic
Analyses of individual AOGCMs (e.g., Fyfe et al., 1999;
simulation of the SAM-related circulation shown in Figure
Shindell et al., 1999) have demonstrated that they are capable
8.12, because the surface temperature signatures of the SAM
of simulating many aspects of the NAM and NAO patterns
typically reﬂ ect advection of the climatological temperature
including linkages between circulation and temperature. Multi-
distribution by the SAM-related circulation (Thompson and
model comparisons of winter atmospheric pressure (Osborn,
Wallace, 2000).
2004), winter temperature (Stephenson and Pavan, 2003) and
Although the spatial structure of the SAM is well simulated
atmospheric pressure across all months of the year (AchutaRao
by the AOGCMs in the MMD at PCMDI, other features of the
et al., 2004), including assessments of the MMD at PCMDI
SAM, such as the amplitude, the detailed zonal structure and the
(Miller et al., 2006) conﬁ rm the overall skill of AOGCMs but
temporal spectra, do not always compare well with the NCEP
also identify that teleconnections between the Atlantic and
reanalysis SAM (Miller et al., 2006; Raphael and Holland,
Paciﬁ c Oceans are stronger in many models than is observed
2006). For example, Figure 8.12 shows that the simulated SAM
(Osborn, 2004). In some models this is related to a bias towards a
variance (the square of the SAM amplitude) ranges between 0.9
strong polar vortex in all winters so that their simulations nearly
and 2.4 times the NCEP reanalysis SAM variance. However,
always reﬂ ect behaviour that is only observed at times with
such features vary considerably among different realisations of
strong vortices (when a stronger Atlantic-Paciﬁ c correlation is
multiple-member ensembles (Raphael and Holland, 2006), and
observed; Castanheira and Graf, 2003).
the temporal variability of the NCEP reanalysis SAM does not
Most AOGCMs organise too much sea level-pressure variability
compare well to station data (Marshall, 2003). Thus, it is difﬁ cult
into the NAM and NAO (Miller et al., 2006). The year-to-year
to assess whether these discrepancies between the simulated
variance of the NAM or NAO is correctly simulated by some
SAM and the NCEP reanalysis SAM point to shortcomings in
AOGCMs, while other simulations are signiﬁ cantly too variable
the models or to shortcomings in the observed analysis.
(Osborn, 2004); for the models that simulate stronger variability,
Resolving these issues may require a better understanding
the persistence of anomalous states is greater than is observed
of SAM dynamics. Although the SAM exhibits clear signatures
(AchutaRao et al., 2004). The magnitude of multi-decadal variability
in the ocean and stratosphere, its tropospheric structure can be
(relative to sub-decadal variability) is lower in AOGCM control
simulated, for example, in atmospheric GCMs with a poorly
simulations than is observed, and cannot be reproduced in current
resolved stratosphere and driven by prescribed SSTs (e.g.,
model simulations with external forcings (Osborn, 2004; Gillett,
Limpasuvan and Hartmann, 2000; Cai et al., 2003). Even
2005). However, Scaife et al. (2005) show that the observed multi-
much simpler atmospheric models with one or two vertical
decadal trend in the surface NAM and NAO can be reproduced in
levels produce SAM-like variability (Vallis et al., 2004). These
an AOGCM if observed trends in the lower stratospheric circulation
relatively simple models capture the dynamics that underlie
are prescribed in the model. Troposphere-stratosphere coupling
SAM variability – namely, interactions between the tropospheric
processes may therefore need to be included in models to fully
jet stream and extratropical weather systems (Limpasuvan and
simulate NAM variability. The response of the NAM and NAO to
Hartmann, 2000; Lorenz and Hartmann, 2001). Nevertheless,
volcanic aerosols (Stenchikov et al., 2002), sea surface temperature
the ocean and stratosphere might still inﬂ uence SAM variability
variability (Hurrell et al., 2004) and sea ice anomalies (Alexander et
in important ways. For example, AOGCM simulations suggest
al., 2004) demonstrate some compatibility with observed variations,
strong SAM-related impacts on ocean temperature, ocean heat
though the difﬁ culties in determining cause and effect in the coupled
transport and sea ice distribution (Watterson, 2001; Hall and
system limit the conclusions that can be drawn with regards to the
Visbeck, 2002), suggesting a potential for air-sea interactions
trustworthiness of model behaviour.
to inﬂ uence SAM dynamics. Furthermore, observational
620

Chapter 8
Climate Models and Their Evaluation
Figure 8.12. Ensemble mean leading Empirical
Orthogonal Function (EOF) of summer (November
through February) Southern Hemisphere sea level
pressure (hPa) for 1950 to 1999. The EOFs are scaled
so that the associated principal component has unit
variance over this period. The percentage of variance
accounted for by the leading mode is listed at the
upper left corner of each panel. The spatial correlation
(r) with the observed pattern is given at the upper right
corner. At the lower right is the ratio of the EOF spatial
variance to the observed value. “Canadian CC” refers
to CGCM3.1 (T47), and “Russell GISS” refers to the
GISS AOM. Adapted from Miller et al. (2006).
and modelling studies (e.g., Thompson and Solomon, 2002;
2006). Some of the inter-decadal variability in the tropics also
Baldwin et al., 2003; Gillett and Thompson, 2003) suggest that
has an extratropical origin (e.g., Barnett et al., 1999; Hazeleger
the stratosphere might also inﬂ uence the tropospheric SAM, at
et al., 2001) and this might give the IPO a predictable component
least in austral spring and summer. Thus, an accurate simulation
(Power et al., 2006).
of stratosphere-troposphere and ocean-atmosphere coupling
Atmosphere-Ocean General Circulation Models do not seem
may still be necessary to accurately simulate the SAM.
to have difﬁ culty in simulating IPO-like variability (e.g., Yeh
and Kirtman, 2004; Meehl and Hu, 2006), even AOGCMs that
8.4.2 Paciﬁ c Decadal Variability
are too coarse to properly resolve equatorially trapped waves
important for ENSO dynamics. Some studies have provided
Recent work suggests that the Paciﬁ c Decadal Oscillation
objective measures of the realism of the modelled decadal
(PDO, see Chapters 3 and 9) is the North Paciﬁ c expression of
variability. For example, Pierce et al. (2000) found that the
a near-global ENSO-like pattern of variability called the Inter-
ENSO-like decadal SST mode in the Paciﬁ c Ocean of their
decadal Paciﬁ c Oscillation or IPO (Power et al., 1999; Deser et
AOGCM had a pattern that gave a correlation of 0.56 with
al., 2004). The appearance of the IPO as the leading Empirical
its observed counterpart. This compared with a correlation
Orthogonal Function (EOF) of SST in AOGCMs that do not
coefﬁ cient of 0.79 between the modelled and observed
include inter-decadal variability in natural or external forcing
interannual ENSO mode. The reduced agreement on decadal
indicates that the IPO is an internally generated, natural form
time scales was attributed to lower than observed variability
of variability. Note, however, that some AOGCMs exhibit an
in the North Paciﬁ c subpolar gyre, over the southwest Paciﬁ c
El Niño-like response to global warming (Cubasch et al., 2001)
and along the western coast of North America. The latter was
that can take decades to emerge (Cai and Whetton, 2000).
attributed to poor resolution of the coastal waveguide in the
Therefore some, though certainly not all, of the variability seen
AOGCM. The importance of properly resolving coastally
in the IPO and PDO indices might be anthropogenic in origin
trapped waves in the context of simulating decadal variability
(Shiogama et al., 2005). The IPO and PDO can be partially
in the Paciﬁ c has been raised in a number studies (e.g., Meehl
understood as the residual of random inter-decadal changes
and Hu, 2006). Finally, there has been little work evaluating the
in ENSO activity (e.g., Power et al., 2006), with their spectra
amplitude of Paciﬁ c decadal variability in AOGCMs. Manabe
reddened (i.e., increasing energy at lower frequencies) by the
and Stouffer (1996) showed that the variability has roughly
integrating effect of the upper ocean mixed layer (Newman et
the right magnitude in their AOGCM, but a more detailed
al., 2003; Power and Colman, 2006) and the excitation of low
investigation using recent AOGCMs with a speciﬁ c focus on
frequency off-equatorial Rossby waves (Power and Colman,
IPO-like variability would be useful.
621

Climate Models and Their Evaluation
Chapter 8
8.4.3 Paciﬁ c-North American Pattern
the strength of the ensemble mean signal remains low when
compared with the statistical spread due to sampling ﬂ uctuations
The Paciﬁ c-North American (PNA) pattern (see Chapter 3) is
among different models, and among different realisations of a
commonly associated with the response to anomalous boundary
given model. The model skill is notably lower for other seasons
forcing. However, PNA-like patterns have been simulated in
and longer lead times. Empirical Orthogonal Function analyses
atmospheric GCM experiments subjected to constant boundary
of the geopotential height data produced by individual member
conditions. Hence, both external and internal processes may
models conﬁ rm that the PNA pattern is a leading spatial mode of
contribute to the formation of this pattern. Particular attention
atmospheric variability in these models.
has been paid to the external inﬂ uences due to SST anomalies
Multi-century integrations have also been conducted at
related to ENSO episodes in the tropical Paciﬁ c, as well as those
various institutions using the current generation of AOGCMs.
situated in the extratropical North Paciﬁ c. Internal mechanisms
Unlike the hindcasting or forecasting experiments mentioned
that might play a role in the formation of the PNA pattern
above, these climate simulations are not aimed at reproducing
include interactions between the slowly varying component
speciﬁ c ENSO events in the observed system. Diagnosis of
of the circulation and high-frequency transient disturbances,
the output from one such AOGCM integration indicates that
and instability of the climatological ﬂ ow pattern. Trenberth et
the modelled ENSO events are linked to a PNA-like pattern in
al. (1998) reviewed the myriad observational and modelling
the upper troposphere (Wittenberg et al., 2006). The centres of
studies on various processes contributing to the PNA pattern.
action of the simulated patterns are systematically displaced 20
The ability of GCMs to replicate various aspects of the
to 30 degrees of longitude west of the observed positions. This
PNA pattern has been tested in coordinated experiments.
discrepancy is evidently linked to a corresponding spatial shift
Until several years ago, such experiments were conducted
in the ENSO-related SST and precipitation anomaly centres
by prescribing observed SST anomalies as lower boundary
simulated in the tropical Paciﬁ c. This ﬁ nding illustrates that
conditions for atmospheric GCMs. Particularly noteworthy are
the spatial conﬁ guration of the PNA pattern in AOGCMs is
the ensembles of model runs performed under the auspices of
crucially dependent on the accuracy of ENSO simulations in
the European Prediction of Climate Variations on Seasonal to
the tropics.
Interannual Time Scales (PROVOST) and the US Dynamical
Seasonal Prediction (DSP) projects. The skill of seasonal
8.4.4
Cold Ocean-Warm Land Pattern
hindcasts of the participating models’ atmospheric anomalies
in different regions of the globe (including the PNA sector)
The Cold Ocean-Warm Land (COWL) pattern indicates that
was summarised in a series of articles edited by Palmer and
the oceans are relatively cold and the continents are relatively
Shukla (2000). These results demonstrate that the prescribed
warm poleward of 40°N when the NH is relatively warm. The
SST forcing exerts a notable impact on the model atmospheres.
COWL pattern results from the contrast in thermal inertia
The hindcast skill for the winter extratropical NH is particularly
between the continents and oceans, which allows continental
high during the largest El Niño and La Niña episodes. However,
temperature anomalies to have greater amplitude, and thus
these experiments indicate considerable variability of the
more strongly inﬂ uence hemispheric mean temperature.
responses in individual models, and among ensemble members
The COWL pattern has been simulated in climate models of
of a given model. This large scatter of model responses suggests
varying degrees of complexity (e.g., Broccoli et al., 1998), and
that atmospheric changes in the extratropics are only weakly
similar patterns have been obtained from cluster analysis (Wu
constrained by tropical SST forcing.
and Straus, 2004a) and EOF analysis (Wu and Straus, 2004b)
The performance of the dynamical seasonal forecast system
of reanalysis data. In a number of studies, cold season trends
at the US NCEP in predicting the atmospheric anomalies given
in NH temperature and sea level pressure during the late 20th
prescribed anomalous SST forcing (in the PNA sector) was
century have been associated with secular trends in indices of
assessed by Kanamitsu et al. (2002). During the large El Niño
the COWL pattern (Wallace et al., 1996; Lu et al., 2004).
event of 1997 to 1998, the forecasts based on this system with one-
In their analysis of AOGCM simulations, Broccoli et al. (1998)
month lead time are in good agreement with the observed changes
found that the original method for extracting the COWL pattern
in the PNA sector, with anomaly correlation scores of 0.8 to 0.9
could yield potentially misleading results when applied to a
(for 200 mb height), 0.6 to 0.8 (surface temperature) and 0.4 to 0.5
simulation forced by past and future variations in anthropogenic
(precipitation). More recently, hindcast experiments have been
forcing (as is the case with most other patterns, or modes, of
launched using AOGCMs. The European effort was supported
climate variability). The resulting spatial pattern was a mixture
by the Development of a European Multimodel Ensemble
of the patterns associated with unforced climate variability and
System for Seasonal to Interannual Prediction (DEMETER)
the anthropogenic ﬁ ngerprint. Broccoli et al. (1998) also noted
programme (Palmer et al., 2004). For the boreal winter season,
that temperature anomalies in the two continental centres of the
and with hindcasts initiated in November, the model-generated
COWL pattern are virtually uncorrelated, suggesting that different
PNA indices exhibit statistically signiﬁ cant temporal correlations
atmospheric teleconnections are involved in producing this pattern.
with the corresponding observations. The ﬁ delity of the PNA
Quadrelli and Wallace (2004) recently showed that the COWL
simulations is evident in both the multi-model ensemble means, as
pattern can be reconstructed as a linear combination of the ﬁ rst two
well as in the output from individual member models. However,
EOFs of monthly mean December to March sea level pressure.
622

Chapter 8
Climate Models and Their Evaluation
These two EOFs are the NAM and a mode closely resembling
8.4.6 Atlantic
Multi-decadal
Variability
the PNA pattern. A linear combination of these two fundamental
patterns can also account for a substantial fraction of the winter
The Atlantic Ocean exhibits considerable multi-decadal
trend in NH sea level pressure during the late 20th century.
variability with time scales of about 50 to 100 years (see Chapter
3). This multi-decadal variability appears to be a robust feature
8.4.5
Atmospheric Regimes and Blocking
of the surface climate in the Atlantic region, as shown by tree
ring reconstructions for the last few centuries (e.g., Mann et al.,
Weather, or climate, regimes are important factors in
1998). Atlantic multi-decadal variability has a unique spatial
determining climate at various locations around the world
pattern in the SST anomaly ﬁ eld, with opposite changes in the
and they can have a large impact on day-to-day variability
North and South Atlantic (e.g., Mestas-Nunez and Enﬁ eld, 1999;
(e.g., Plaut and Simonnet, 2001; Trigo et al., 2004; Yiou and
Latif et al., 2004), and this dipole pattern has been shown to be
Nogaj, 2004). General Circulation Models have been found to
signiﬁ cantly correlated with decadal changes in Sahelian rainfall
simulate hemispheric climate regimes quite similar to those
(Folland et al., 1986). Decadal variations in hurricane activity
found in observations (Robertson, 2001; Achatz and Opsteegh,
have also been linked to the multi-decadal SST variability in the
2003; Selten and Branstator, 2004). Simulated regional climate
Atlantic (Goldenberg et al., 2001). Atmosphere-Ocean General
regimes over the North Atlantic strongly similar to the observed
Circulation Models simulate Atlantic multi-decadal variability
regimes were reported by Cassou et al. (2004), while the North
(e.g., Delworth et al., 1993; Latif, 1998 and references therein;
Paciﬁ c regimes simulated by Farrara et al. (2000) were broadly
Knight et al., 2005), and the simulated space-time structure is
consistent with those in observations. Since the TAR, agreement
consistent with that observed (Delworth and Mann, 2000). The
between different studies has improved regarding the number
multi-decadal variability simulated by the AOGCMs originates
and structure of both hemispheric and sectoral atmospheric
from variations in the MOC (see Section 8.3). The mechanisms,
regimes, although this remains a subject of research (e.g.,
however, that control the variations in the MOC are fairly
Wu and Straus, 2004a) and the statistical signiﬁ cance of the
different across the ensemble of AOGCMs. In most AOGCMs,
regimes has been discussed and remains an unresolved issue
the variability can be understood as a damped oceanic eigenmode
(e.g., Hannachi and O’Neill, 2001; Hsu and Zwiers, 2001;
that is stochastically excited by the atmosphere. In a few other
Stephenson et al., 2004; Molteni et al., 2006).
AOGCMs, however, coupled interactions between the ocean
Blocking events are an important class of sectoral weather
and the atmosphere appear to be more important. The relative
regimes (see Chapter 3), associated with local reversals
roles of high- and low-latitude processes differ also from model
of the mid-latitude westerlies. The most recent systematic
to model. The variations in the Atlantic SST associated with
intercomparison of atmospheric GCM simulations of NH
the multi-decadal variability appear to be predictable a few
blocking (D’Andrea et al., 1998) was reported in the TAR.
decades ahead, which has been shown by potential (diagnostic)
Consistent with the conclusions of this earlier study, recent
and classical (prognostic) predictability studies. Atmospheric
studies have found that GCMs tend to simulate the location
quantities do not exhibit predictability at decadal time scales
of NH blocking more accurately than frequency or duration:
in these studies, which supports the picture of stochastically
simulated events are generally shorter and rarer than observed
forced variability.
events (e.g., Pelly and Hoskins, 2003b). An analysis of one
of the AOGCMs from the MMD at the PCMDI found that
8.4.7 El
Niño-Southern
Oscillation
increased horizontal resolution combined with better physical
parametrizations has led to improvements in simulations of
During the last decade, there has been steady progress in
NH blocking and synoptic weather regimes over Europe.
simulating and predicting ENSO (see Chapters 3 and 9) and
Finally, both GCM simulations and analyses of long data sets
the related global variability using AOGCMs (Latif et al.,
suggest the existence of considerable interannual to inter-
2001; Davey et al., 2002; AchutaRao and Sperber, 2002).
decadal variability in blocking frequency (e.g., Stein, 2000;
Over the last several years the parametrized physics have
Pelly and Hoskins, 2003a), highlighting the need for caution
become more comprehensive (Gregory et al., 2000; Collins et
when assessing blocking climatologies derived from short
al., 2001; Kiehl and Gent, 2004), the horizontal and vertical
records (either observed or simulated). Blocking events also
resolutions, particularly in the atmospheric component models,
occur in the SH mid-latitudes (Sinclair, 1996); no systematic
have markedly increased (Guilyardi et al., 2004) and the
intercomparison of observed and simulated SH blocking application of observations in initialising forecasts has become
climatologies has been carried out. There is also evidence of
more sophisticated (Alves et al., 2004). These improvements
connections between North and South Paciﬁ c blocking and
in model formulation have led to a better representation of
ENSO variability (e.g., Renwick, 1998; Chen and Yoon, 2002),
the spatial pattern of the SST anomalies in the eastern Paciﬁ c
and between North Atlantic blocks and sudden stratospheric
(AchutaRao and Sperber, 2006). In fact, as an indication of
warmings (e.g., Kodera and Chiba, 1995; Monahan et al., 2003)
recent model improvements, some IPCC class models are being
but these connections have not been systematically explored in
used for ENSO prediction (Wittenberg et al., 2006). Despite this
AOGCMs.
progress, serious systematic errors in both the simulated mean
climate and the natural variability persist. For example, the
623

Climate Models and Their Evaluation
Chapter 8
considerably faster than observed (AchutaRao
and Sperber, 2002), although there has been
some notable progress in this regard over
the last decade (AchutaRao and Sperber,
2006) in that more models are consistent
with the observed time scale for ENSO (see
Figure 8.13). The models also have difﬁ culty
capturing the correct phase locking between
the annual cycle and ENSO. Further, some
AOGCMs fail to represent the spatial and
temporal structure of the El Niño-La Niña
asymmetry (Monahan and Dai, 2004). Other
weaknesses in the simulated amplitude and
structure of ENSO variability are discussed
in Davey et al. (2002) and van Oldenborgh
et al. (2005).
Current research points to some promise
in addressing some of the above problems.
For example, increasing the atmospheric
resolution in both the horizontal (Guilyardi
et al., 2004) and vertical (NCEP Coupled
Forecast System) may improve the simulated
spectral characteristics of the variability,
ocean parametrized physics have also been
shown to signiﬁ cantly inﬂ uence the coupled
variability (Meehl et al., 2001) and continued
methodical numerical experimentation into
the sources of model error (e.g., Schneider,
2001) will ultimately suggest model
improvement strategies.
In terms of ENSO prediction, the two
biggest recent advances are: (i) the recognition
that forecasts must include quantitative
information regarding uncertainty (i.e.,
probabilistic prediction) and that veriﬁ cation
must include skill measures for probability
Figure 8.13. Maximum entropy power spectra of surface air temperature averaged over the NINO region
3
forecasts (Kirtman, 2003); and (ii) that a
(i.e., 5°N to 5°S, 150°W to 90° W) for (a) the MMD at the PCMDI and (b) the CMIP2 models. Note the differ-
multi-model ensemble strategy may be the
ing scales on the vertical axes and that ECMWF reanalysis in (b) refers to the European Centre for Medium
Range Weather Forecasts (ECMWF) 15-year reanalysis (ERA15) as in (a). The vertical lines correspond to
best current approach for adequately dealing
periods of two and seven years. The power spectra from the reanalyses and for SST from the Hadley Centre
with forecast uncertainty, for example, Palmer
Sea Ice and Sea Surface Temperature (HadISST) version 1.1 data set are given by the series of solid, dashed
et al. (2004), in which Figure 2 demonstrates
and dotted black curves. Adapted from AchutaRao and Sperber (2006).
that a multi-model ensemble forecast has
better skill than a comparable ensemble

so-called ‘double ITCZ’ problem noted by Mechoso et al.
based on a single model. Improvements in the use of data,
(1995; see Section 8.3.1) remains a major source of error in
particularly in the ocean, for initialising forecasts continues
simulating the annual cycle in the tropics in most AOGCMs,
to yield enhancements in forecast skill (Alves et al., 2004);
which ultimately affects the ﬁ delity of the simulated ENSO.
moreover, other research indicates that forecast initialisation
Along the equator in the Paciﬁ c the models fail to adequately
strategies that are implemented within the framework of the
capture the zonal SST gradient, the equatorial cold tongue
coupled system as opposed to the individual component models
structure is equatorially conﬁ ned and extends too far too to
may also lead to substantial improvements in skill (Chen et al.,
the west (Cai et al., 2003), and the simulations typically have
1995). However, basic questions regarding the predictability
thermoclines that are far too diffuse (Davey et al., 2002). Most
of SST in the tropical Paciﬁ c remain open challenges in the
AOGCMs fail to capture the meridional extent of the anomalies
forecast community. For instance, it is unclear how westerly
in the eastern Paciﬁ c and tend to produce anomalies that extend
wind bursts, intra-seasonal variability or atmospheric weather
too far into the western tropical Paciﬁ c. Most, but not all,
noise in general limit the predictability of ENSO (e.g.,
AOGCMs produce ENSO variability that occurs on time scales
Thompson and Battisti, 2001; Kleeman et al., 2003; Flugel et
624

Chapter 8
Climate Models and Their Evaluation
al., 2004; Kirtman et al., 2005). There are also apparent decadal
MJO temporal and spatial scales. The interaction with an active
variations in ENSO forecast skill (Balmaseda et al., 1995; Ji et
ocean is important especially in the suppressed convective
al., 1996; Kirtman and Schopf, 1998), and the sources of these
phase when SSTs are warming and the atmospheric boundary
variations are the subject of some debate. Finally, it remains
layer is recovering (e.g., Hendon, 2005). Thus, the most realistic
unclear how changes in the mean climate will ultimately affect
simulation of the MJO is anticipated to be with AOGCMs.
ENSO predictability (Collins et al., 2002).
However, coupling, in general, has not been a panacea. While
coupling in some models improves some aspects of the MJO,
8.4.8
Madden-Julian
Oscillation
especially eastward propagation and coherence of convective
anomalies across the Indian and western Paciﬁ c Oceans (e.g.,
The MJO (Madden and Julian, 1971) refers to the dominant
Kemball-Cook et al., 2002; Inness and Slingo, 2003), problems
mode of intra-seasonal variability in the tropical troposphere.
with the horizontal structure and seasonality remain. Typically,
It is characterised by large-scale regions of enhanced and
models that show the most beneﬁ cial impact of coupling on the
suppressed convection, coupled to a deep baroclinic, primarily
propagation characteristics of the MJO are also the models that
zonal circulation anomaly. Together, they propagate slowly
possess the most unrealistic seasonal variation of MJO activity
eastward along the equator from the western Indian Ocean to the
(e.g., Zhang, 2005). Unrealistic simulation of the seasonal
central Paciﬁ c and exhibit local periodicity in a broad 30- to 90-
variation of MJO activity implies that the simulated MJO will
day range. Simulation of the MJO in contemporary coupled and
improperly interact with climate phenomena that are tied to the
uncoupled climate models remains unsatisfactory (e.g., Zhang,
seasonal cycle (e.g., the monsoons and ENSO).
2005; Lin et al., 2006), partly because more is now demanded
Simulation of the MJO is also adversely affected by biases
from the model simulations, as understanding of the role of the
in the mean state (see Section 8.4.7). These biases include the
MJO in the coupled atmosphere-ocean climate system expands.
tendency for coupled models to exaggerate the double ITCZ
For instance, simulations of the MJO in models at the time of
in the Indian and western Paciﬁ c Oceans, under-predict the
the TAR were judged using gross metrics (e.g., Slingo et al.,
eastward extent of surface monsoonal westerlies into the
1996). The spatial phasing of the associated surface ﬂ uxes, for
western Paciﬁ c, and over-predict the westward extension of the
instance, are now recognised as critical for the development
Paciﬁ c cold tongue. Together, these ﬂ aws limit development,
of the MJO and its interaction with the underlying ocean (e.g.,
maintenance and the eastward extent of convection associated
Hendon, 2005; Zhang, 2005). Thus, while a model may simulate
with the MJO, thereby reducing its overall strength and
some gross characteristics of the MJO, the simulation may be
coherence (e.g., Inness et al., 2003). To date, simulation of
deemed unsuccessful when the detailed structure of the surface
the MJO has proven to be most sensitive to the convective
ﬂ uxes is examined (e.g., Hendon, 2000).
parametrization employed in climate models (e.g., Wang and
Variability with MJO characteristics (e.g., convection and
Schlesinger, 1999; Maloney and Hartmann, 2001; Slingo et al.,
wind anomalies of the correct spatial scale that propagate
2005). A consensus, although with exceptions (e.g., Liu et al.,
coherently eastward with realistic phase speeds) is simulated in
2005), appears to be emerging that convective schemes based on
many contemporary models (e.g., Sperber et al., 2005; Zhang,
local vertical stability and that include some triggering threshold
2005), but this variability is typically not simulated to occur
produce more realistic MJO variability than those that convect
often enough or with sufﬁ cient strength so that the MJO stands
too readily. However, some sophisticated models, with arguably
out realistically above the broadband background variability (Lin
the most physically based convective parametrizations, are
et al., 2006). This underestimation of the strength and coherence
unable to simulate reasonable MJO activity (e.g., Slingo et al.,
of convection and wind variability at MJO temporal and spatial
2005).
scales means that contemporary climate models still simulate
poorly many of the important climatic effects of the MJO
8.4.9
Quasi-Biennial
Oscillation
(e.g., its impact on rainfall variability in the monsoons or the
modulation of tropical cyclone development). Simulation of the
The Quasi-Biennial Oscillation (QBO; see Chapter 3) is
spatial structure of the MJO as it evolves through its life cycle is
a quasi-periodic wave-driven zonal mean wind reversal that
also problematic, with tendencies for the convective anomaly to
dominates the low-frequency variability of the lower equatorial
split into double ITCZs in the Paciﬁ c and for erroneously strong
stratosphere (3 to 100 hPa) and affects a variety of extratropical
convective signals to sometimes develop in the eastern Paciﬁ c
phenomena including the strength and stability of the winter
ITCZ (e.g., Inness and Slingo, 2003). It has also been suggested
polar vortex (e.g., Baldwin et al., 2001). Theory and observations
that inadequate representation in climate models of cloud-
indicate that a broad spectrum of vertically propagating waves
radiative interactions and/or convection-moisture interactions
in the equatorial atmosphere must be considered to explain
may explain some of the difﬁ culties in simulating the MJO (e.g.,
the QBO. Realistic simulation of the QBO in GCMs therefore
Lee et al., 2001; Bony and Emanuel, 2005).
depends on three important conditions: (i) sufﬁ cient vertical
Even though the MJO is probably not fundamentally a
resolution in the stratosphere to allow the representation
coupled ocean-atmosphere mode (e.g., Waliser et al., 1999),
of equatorial waves at the horizontally resolved scales of a
air-sea coupling does appear to promote more coherent GCM, (ii) a realistic excitation of resolved equatorial waves
eastward, and, in northern summer, northward propagation at
by simulated tropical weather and (iii) parametrization of the
625

Climate Models and Their Evaluation
Chapter 8
effects of unresolved gravity waves. Due to the computational
climate in North Africa in the MMD at PCMDI. They found that
cost associated with the requirement of a well-resolved the simulation of North African summer precipitation was less
stratosphere, the models employed for the current assessment
realistic than the simulation of summer precipitation over North
do not generally include the QBO.
America or Europe. In short, most AOGCMs do not simulate
The inability of resolved wave driving to induce a the spatial or intra-seasonal variation of monsoon precipitation
spontaneous QBO in GCMs has been a long-standing issue
accurately. See Chapter 11 for a more detailed regional evaluation
(Boville and Randel, 1992). Only recently (Takahashi, 1996,
of simulated monsoon variability.
1999; Horinouchi and Yoden, 1998; Hamilton et al., 2001) have
two necessary conditions been identiﬁ ed that allow resolved
8.4.11 Shorter-Term Predictions Using Climate
waves to induce a QBO: high vertical resolution in the lower
Models
stratosphere (roughly 0.5 km), and a parametrization of deep
cumulus convection with sufﬁ ciently large temporal variability.
This subsection focuses on the few results of initial value
However, recent analysis of satellite and radar observations
predictions made using models that are identical, or very
of deep tropical convection (Horinouchi, 2002) indicates that
close to, the models used in other chapters of this report for
the forcing of a QBO by resolved waves alone requires a
understanding and predicting climate change.
parametrization of deep convection with an unrealistically large
amount of temporal variability. Consequently, it is currently
Weather prediction
thought that a combination of resolved and parametrized
Since the TAR, it has been shown that climate models can be
waves is required to properly model the QBO. The utility of
integrated as weather prediction models if they are initialised
parametrized non-orographic gravity wave drag to force a QBO
appropriately (Phillips et al., 2004). This advance appears to be
has now been demonstrated by a number of studies (Scaife et
due to: (i) improvements in the forecast model analyses and (ii)
al., 2000; Giorgetta et al., 2002, 2006). Often an enhancement
increases in the climate model spatial resolution. An advantage
of input momentum ﬂ ux in the tropics relative to that needed
of testing a model’s ability to predict weather is that some of
in the extratropics is required. Such an enhancement, however,
the sub-grid scale physical processes that are parametrized in
depends implicitly on the amount of resolved waves and in
models (e.g., cloud formation, convection) can be evaluated
turn, the spatial and temporal properties of parametrized deep
on time scales characteristic of those processes, without the
convection employed in each model (Horinouchi et al., 2003;
complication of feedbacks from these processes altering the
Scinocca and McFarlane, 2004).
underlying state of the atmosphere (Pope and Stratton, 2002;
Boyle et al., 2005; Williamson et al., 2005; Martin et al., 2006).
8.4.10
Monsoon
Variability
Full use can be made of the plentiful meteorological data sets
and observations from specialised ﬁ eld experiments. According
Monsoon variability (see Chapters 3, 9 and 11) occurs over
to these studies, some of the biases found in climate simulations
a range of temporal scales from intra-seasonal to inter-decadal.
are also evident in the analysis of their weather forecasts. This
Since the TAR, the ability of AOGCMs to simulate monsoon
suggests that ongoing improvements in model formulation
variability on intra-seasonal as well as interannual time scales
driven primarily by the needs of weather forecasting may lead
has been examined. Lambert and Boer (2001) compared the
also to more reliable climate predictions.
AOGCMs that participated in CMIP, ﬁ nding large errors in the
simulated precipitation in the equatorial regions and in the Asian
Seasonal prediction
monsoon region. Lin et al. (2006) evaluated the intra-seasonal
Veriﬁ cation of seasonal-range predictions provides a direct
variation of precipitation in the MMD at PCMDI. They found
test of a model’s ability to represent the physical and dynamical
that the intra-seasonal variance of precipitation simulated by
processes controlling (unforced) ﬂ uctuations in the climate
most AOGCMs was smaller than observed. The space-time
system. Satisfactory prediction of variations in key climate
spectra of most model simulations have much less power than is
signals such as ENSO and its global teleconnections provides
observed, especially at periods shorter than six days. The speed
evidence that such features are realistically represented in long-
of the equatorial waves is too fast, and the persistence of the
term forced climate simulations.
precipitation is too long, in most of the AOGCM simulations.
A version of the HadCM3 AOGCM (known as GloSea) has
Annamalai et al (2004) examined the ﬁ delity of precipitation
been assessed for skill in predicting observed seasonal climate
simulation in the Asian monsoon region in the MMD at
variations (Davey et al., 2002; Graham et al., 2005). Graham
PCMDI. They found that just 6 of the 18 AOGCMs considered
et al. (2005) analysed 43 years of retrospective six-month
realistically simulated climatological monsoon precipitation
forecasts (‘hindcasts’) with GloSea, run from observed ocean-
for the 20th century. For the former set of models, the spatial
land-atmosphere initial conditions. A nine-member ensemble
correlation of the patterns of monsoon precipitation between the
was used to sample uncertainty in the initial conditions.
models exceeded 0.6, and the seasonal cycle of monsoon rainfall
Conclusions relevant to HadCM3 include: (i) the model is able
was simulated well. Among these models, only four exhibited a
to reproduce observed large-scale lagged responses to ENSO
robust ENSO-monsoon contemporaneous teleconnection. Cook
events in the tropical Atlantic and Indian Ocean SSTs; and (ii)
and Vizy (2006) evaluated the simulation of the 20th-century
the model can realistically predict anomaly patterns in North
626

Chapter 8
Climate Models and Their Evaluation
Atlantic SSTs, shown to have important links with the NAO
8.5.1 Extreme
Temperature
and seasonal temperature anomalies over Europe.

The GFDL-CM2.0 AOGCM has also been assessed
Kiktev et al. (2003) compared station observations of
for seasonal prediction. Twelve-month retrospective and extreme events with the simulations of an atmosphere-only
contemporaneous forecasts were produced using a six-member
GCM (Hadley Centre Atmospheric Model version 3; HadAM3)
ensemble over 15 years starting in 1991. The forecasts were
forced by prescribed oceanic forcing and anthropogenic
initialised using global ocean data assimilation (Derber and
radiative forcing during 1950 to 1995. The indices of extreme
Rosati, 1989; Rosati et al., 1997) and observed atmospheric
events they used were those proposed by Frich et al. (2002).
forcing, combined with atmospheric initial conditions derived
They found that inclusion of anthropogenic radiative forcing
from the atmospheric component of the model forced with
was required to reproduce observed changes in temperature
observed SSTs. Results indicated considerable model skill out
extremes, particularly at large spatial scales. The decrease in
to 12 months for ENSO prediction (see http://www.gfdl.noaa.
the number of frost days in Southern Australia simulated by
gov/~rgg/si_workdir/Forecasts.html). Global teleconnections,
HadAM3 with anthropogenic forcing is in good agreement with
as diagnosed from the NCEP reanalysis (GFDL GAMDT,
the observations. The increase in the number of warm nights
2004), were evident throughout the 12-month forecasts.
over Eurasia is poorly simulated when anthropogenic forcing
is not included, but the inclusion of anthropogenic forcing
improves the modelled trend patterns over western Russia and
8.5 Model Simulations of Extremes
reproduces the general increase in the occurrence of warm
nights over much of the NH.
Meehl et al. (2004) compared the number of frost days
Society’s perception of climate variability and climate
simulated by the PCM model with observations. The 20th-
change is largely formed by the frequency and the severity of
century simulations include the variations in solar, volcano,
extremes. This is especially true if the extreme events have
sulphate aerosol, ozone and greenhouse gas forcing. Both
large and negative impacts on lives and property. As climate
model simulations and observations show that the number of
models’ resolution and the treatment of physical processes
frost days decreased by two days per decade in the western USA
have improved, the simulation of extremes has also improved.
during the 20th century. The model simulations do not agree
Mainly because of increased data availability (e.g., daily data,
with observations in the southeastern USA, where the model
various indices, etc.), the modelling community has now
simulates a decrease in the number of frost days in this region in
examined the model simulations in greater detail and presented
the 20th century, while observations indicate an increase in this
a comprehensive description of extreme events in the coupled
region. Meehl et al. (2004) argue that this discrepancy could be
models used for climate change projections.
due to the model’s inability to simulate the impact of El Niño
Some extreme events, by their very nature of being smaller in
events on the number of frost days in the southeastern USA.
scale and shorter in duration, are manifestations of either a rapid
Meehl and Tebaldi (2004) compared the heat waves simulated
ampliﬁ cation, or an equilibration at a higher amplitude, of naturally
by the PCM with observations. They deﬁ ned a heat wave as
occurring local instabilities. Large-scale and long-duration the three consecutive warmest nights during the year. During
extreme events are generally due to persistence of weather patterns
the period 1961 to 1990, there is good agreement between the
associated with air-sea and air-land interactions. A reasonable
model and observations (NCEP reanalysis).
hypothesis might be that the coarse-resolution AOGCMs might
Kharin et al. (2005) examined the simulations of temperature
not be able to simulate local short-duration extreme events,
and precipitation extremes for AMIP-2 models, some of which
but that is not the case. Our assessment of the recent scientiﬁ c
are atmospheric components of coupled models used in this
literature shows, perhaps surprisingly, that the global statistics of
assessment. They found that models simulate the temperature
the extreme events in the current climate, especially temperature,
extremes, especially the warm extremes, reasonably well.
are generally well simulated by the current models (see Section
Models have serious deﬁ ciencies in simulating precipitation
8.5.1). These models have been more successful in simulating
extremes, particularly in the tropics. Vavrus et al. (2006) used
temperature extremes than precipitation extremes.
daily values of 20th-century integrations from seven models.
The assessment of extremes, especially for temperature,
They deﬁ ned a cold air outbreak as ‘an occurrence of two
has been done by examining the amplitude, frequency and
or more consecutive days during which the local mean daily
persistence of the following quantities: daily maximum and
surface air temperature is at least two standard deviations
minimum temperature (e.g., hot days, cold days, frost days), daily
below the local winter mean temperature’. They found that the
precipitation intensity and frequency, seasonal mean temperature
climate models reproduce the location and magnitude of cold
and precipitation and frequency and tracks of tropical cyclones.
air outbreaks in the current climate.
For precipitation, the assessment has been done either in terms of
Researchers have also established relationships between
return values or extremely high rates of precipitation.
large-scale circulation features and cold air outbreaks or heat

waves. For example, Vavrus et al. (2006) found that ‘the favored
regions of cold air outbreaks are located near and downstream
627

Climate Models and Their Evaluation
Chapter 8
from preferred locations of atmosphere blocking’. Likewise,
in the number of wet days (the number of days in a year with
Meehl and Tebaldi (2004) found that heat waves over Europe
precipitation greater than 10 mm).
and North America were associated with changes in the 500
Using the Palmer Drought Severity Index (PDSI), Dai at
hPa circulation pattern.
al. (2004) concluded that globally very dry or wet areas (PDSI

above +3 or below –3) have increased from 20% to 38% since
8.5.2 Extreme
Precipitation
1972. In addition to simulating the short-duration events like heat

waves, frost days and cold air outbreaks, models have also shown
Sun et al. (2006) investigated the intensity of daily success in simulating long time-scale anomalies. For example,
precipitation simulated by 18 AOGCMs, including several used
Burke et al. (2006) showed that the HadCM3 model, on a global
in this report. They found that most of the models produce light
basis and at decadal time scales, ‘reproduces the observed drying
precipitation (<10 mm day–1) more often than observed, too few
trend’ as deﬁ ned by the PDSI if the anthropogenic forcing is
heavy precipitation events and too little precipitation in heavy
included, although the model does not always simulate correctly
events (>10 mm day–1). The errors tend to cancel, so that the
the regional distributions of wet and dry areas.
seasonal mean precipitation is fairly realistic (see Section 8.3).

Since the TAR, many simulations have been made with
8.5.3 Tropical
Cyclones
high-resolution GCMs. Iorio et al. (2004) examined the impact

of model resolution on the simulation of precipitation in the
The spatial resolution of the coupled ocean-atmosphere
USA using the Community Climate Model version 3 (CCM3).
models used in the IPCC assessment is generally not high
They found that the high-resolution simulation produces enough to resolve tropical cyclones, and especially to simulate
more realistic daily precipitation statistics. The coarse-
their intensity. A common approach to investigate the effects
resolution model had too many days with weak precipitation
of global warming on tropical cyclones has been to utilise the
and not enough with intense precipitation. This tendency was
SST boundary conditions from a global change scenario run
partially eliminated in the high-resolution simulation, but,
to force a high-resolution AGCM. That model run is then
in the simulation at the highest resolution (T239), the high-
compared with a control run using the high-resolution AGCM
percentile daily precipitation was still too low. This problem
forced with speciﬁ ed observed SST for the current climate
was eliminated when a cloud-resolving model was embedded
(Sugi et al., 2002; Camargo et al., 2005; McDonald et al., 2005;
in every grid point of the GCM.
Bengtsson et al., 2006; Oouchi et al., 2006; Yoshimura et al.,
Kimoto et al. (2005) compared the daily precipitation over
2006). There are also several idealised model experiments in
Japan in an AOGCM with two different resolutions (high
which a high-resolution AGCM is integrated with and without
res. and med res. of MIROC 3.2) and found more realistic
a ﬁ xed global warming or cooling of SST. Another method is to
precipitation distributions with the higher resolution. Emori et al.
embed a high-resolution regional model in the lower-resolution
(2005) showed that a high-resolution AGCM (the atmospheric
climate model (Knutson and Tuleya, 1999; Walsh et al., 2004).
part of high res. MIROC 3.2) can simulate the extreme daily
Projections using these methods are discussed in Chapter 10.
precipitation realistically if there is provision in the model to
Bengtsson et al. (2006) showed that the global metrics of
suppress convection when the ambient relative humidity is
tropical cyclones (tropical or hemispheric averages) are broadly
below 80%, suggesting that modelled extreme precipitation
reproduced by the ECHAM5 model, even as a function of
can be strongly parametrization dependent. Kiktev et al. (2003)
intensity. However, varying degrees of errors (in some cases
compared station observations of rainfall with the simulations
substantial) in simulated tropical storm frequency and intensity
of the atmosphere-only GCM HadAM3 forced by prescribed
have been noted in some models (e.g., GFDL GAMDT, 2004;
oceanic forcing and anthropogenic radiative forcing. They
Knutson and Tuleya, 2004; Camargo et al., 2005). The tropical
found that this model shows little skill in simulating changing
cyclone simulation has been shown to be sensitive to the choice
precipitation extremes. May (2004) examined the variability and
of convection parametrization in some cases.
extremes of daily rainfall in the simulation of present day climate
Oouchi et al. (2006) used one of the highest-resolution (20
by the ECHAM4 GCM. He found that this model simulates the
km) atmospheric models to simulate the frequency, distribution
variability and extremes of rainfall quite well over most of India
and intensity of tropical cyclones in the current climate.
when compared to satellite-derived rainfall, but has a tendency
Although there were some deﬁ ciencies in simulating the
to overestimate heavy rainfall events in central India. Durman
geographical distribution of tropical cyclones (over-prediction
et al. (2001) compared the extreme daily European precipitation
of tropical cyclones between 0° to 10°S in the Indian Ocean,
simulated by the HadCM2 GCM with station observations. They
and under-prediction between 0° to 10°N in the western
found that the GCM’s ability to simulate daily precipitation
Paciﬁ c), the overall simulation of geographical distribution and
events exceeding 15 mm per day was good but its ability to
frequency was remarkably good. The model could not simulate
simulate events exceeding 30 mm per day was poor. Kiktev et
the strongest observed maximum wind speeds, and central
al. (2003) showed that HadAM3 was able to simulate the natural
pressures were not as low as observed, suggesting that even
variability of the precipitation intensity index (annual mean
higher resolution may be required to simulate the most intense
precipitation divided by number of days with precipitation less
tropical cyclones.
than 1 mm) but was not able to simulate accurately the variability

628

Chapter 8
Climate Models and Their Evaluation
8.5.4 Summary
treatment of these processes in the global climate models used

to make projections of future climate change (Section 8.6.3).
Because most AOGCMs have coarse resolution and large-
Finally we discuss how we can assess our relative conﬁ dence in
scale systematic errors, and extreme events tend to be short lived
the different climate sensitivity estimates derived from climate
and have smaller spatial scales, it is somewhat surprising how
models (Section 8.6.4). Note that climate feedbacks associated
well the models simulate the statistics of extreme events in the
with chemical or biochemical processes are not discussed in this
current climate, including the trends during the 20th century
section (they are addressed in Chapters 7 and 10), nor are local-
(see Chapter 9 for more detail). This is especially true for the
scale feedbacks (e.g., between soil moisture and precipitation;
temperature extremes, but intensity, frequency and distribution
see Section 8.2.3.2).
of extreme precipitation are less well simulated. The higher-
resolution models used for projections of tropical cyclone changes
8.6.2
Interpreting the Range of Climate Sensitivity
(Chapter 10) produce generally good simulation of the frequency
Estimates Among General Circulation Models
and distribution of tropical cyclones, but less good simulation of
their intensity. Improvements in the simulation of the intensity
8.6.2.1 Deﬁ nition of Climate Sensitivity
of precipitation and tropical cyclones with increases in the
resolution of AGCMs (Oouchi et al., 2006) suggest that when
As deﬁ ned in previous assessments (Cubasch et al., 2001)
climate models have sufﬁ cient resolution to explicitly resolve at
and in the Glossary, the global annual mean surface air
least the large convective systems without using parametrizations
temperature change experienced by the climate system after
for deep convection, it is likely that simulation of precipitation
it has attained a new equilibrium in response to a doubling of
and intensity of tropical cyclones will improve.
atmospheric CO2 concentration is referred to as the ‘equilibrium
climate sensitivity’ (unit is °C), and is often simply termed the
‘climate sensitivity’. It has long been estimated from numerical
8.6 Climate Sensitivity and Feedbacks
experiments in which an AGCM is coupled to a simple non-
dynamic model of the upper ocean with prescribed ocean heat
transports (usually referred to as ‘mixed-layer’ or ‘slab’ ocean
8.6.1 Introduction
models) and the atmospheric CO2 concentration is doubled.
In AOGCMs and non-steady-state (or transient) simulations,
Climate sensitivity is a metric used to characterise the
the ‘transient climate response’ (TCR; Cubasch et al., 2001) is
response of the global climate system to a given forcing. It
deﬁ ned as the global annual mean surface air temperature change
is broadly deﬁ ned as the equilibrium global mean surface
(with respect to a ‘control’ run) averaged over a 20-year period
temperature change following a doubling of atmospheric
centred at the time of CO2 doubling in a 1% yr–1 compound CO2
CO
increase scenario. That response depends both on the sensitivity
2 concentration (see Box 10.2). Spread in model climate
sensitivity is a major factor contributing to the range in
and on the ocean heat uptake. An estimate of the equilibrium
projections of future climate changes (see Chapter 10) along
climate sensitivity in transient climate change integrations is
with uncertainties in future emission scenarios and rates of
obtained from the ‘effective climate sensitivity’ (Murphy, 1995).
oceanic heat uptake. Consequently, differences in climate
It corresponds to the global temperature response that would
sensitivity between models have received close scrutiny in all
occur if the AOGCM was run to equilibrium with feedback
four IPCC reports. Climate sensitivity is largely determined
strengths held ﬁ xed at the values diagnosed at some point of the
by internal feedback processes that amplify or dampen
transient climate evolution. It is computed from the oceanic heat
the inﬂ uence of radiative forcing on climate. To assess the
storage, the radiative forcing and the surface temperature change
reliability of model estimates of climate sensitivity, the ability
(Cubasch et al., 2001; Gregory et al., 2002).
of climate models to reproduce different climate changes
The climate sensitivity depends on the type of forcing agents
induced by speciﬁ c forcings may be evaluated. These include
applied to the climate system and on their geographical and
the Last Glacial Maximum and the evolution of climate over
vertical distributions (Allen and Ingram, 2002; Sausen et al.,
the last millennium and the 20th century (see Section 9.6). The
2002; Joshi et al., 2003). As it is inﬂ uenced by the nature and
compilation and comparison of climate sensitivity estimates
the magnitude of the feedbacks at work in the climate response,
derived from models and from observations are presented in
it also depends on the mean climate state (Boer and Yu, 2003).
Box 10.2. An alternative approach, which is followed here,
Some differences in climate sensitivity will also result simply
is to assess the reliability of key climate feedback processes
from differences in the particular radiative forcing calculated by
known to play a critical role in the models’ estimate of climate
different radiation codes (see Sections 10.2.1 and 8.6.2.3). The
sensitivity.
global annual mean surface temperature change thus presents
This section explains why the estimates of climate sensitivity
limitations regarding the description and the understanding of
and of climate feedbacks differ among current models (Section
the climate response to an external forcing. Indeed, the regional
8.6.2), summarises understanding of the role of key radiative
temperature response to a uniform forcing (and even more to
feedback processes associated with water vapour and lapse rate,
a vertically or geographically distributed forcing) is highly
clouds, snow and sea ice in climate sensitivity, and assesses the
inhomogeneous. In addition, climate sensitivity only considers
629

Climate Models and Their Evaluation
Chapter 8
the surface mean temperature and gives no indication of the
largely cancel each other. In addition, the parametrization
occurrence of abrupt changes or extreme events. Despite its
changes can interact nonlinearly with each other so that the
limitations, however, the climate sensitivity remains a useful
sum of change A and change B does not produce the same as
concept because many aspects of a climate model scale well
the change in A plus B (e.g., Stainforth et al., 2005). Finally,
with global average temperature (although not necessarily
the interaction among the different parametrizations of a model
across models), because the global mean temperature of the
explains why the inﬂ uence on climate sensitivity of a given
Earth is fairly well measured, and because it provides a simple
change is often model dependent (see Section 8.2). For instance,
way to quantify and compare the climate response simulated
the introduction of the Lock boundary-layer scheme (Lock et
by different models to a speciﬁ ed perturbation. By focusing on
al., 2000) to HadCM3 had a minimal impact on the climate
the global scale, climate sensitivity can also help separate the
sensitivity, in contrast to the introduction of the scheme to the
climate response from regional variability.
GFDL atmospheric model (Soden et al., 2004; Johns et al.,
2006).
8.6.2.2
Why Have the Model Estimates Changed Since
the TAR?
8.6.2.3
What Explains the Current Spread in Models’
Climate Sensitivity Estimates?
The current generation of GCMs5 covers a range of
equilibrium climate sensitivity from 2.1°C to 4.4°C (with a
As discussed in Chapter 10 and throughout the last three
mean value of 3.2°C; see Table 8.2 and Box 10.2), which is
IPCC assessments, climate models exhibit a wide range of
quite similar to the TAR. Yet most climate models have
climate sensitivity estimates (Table 8.2). Webb et al. (2006),
undergone substantial developments since the TAR (probably
investigating a selection of the slab versions of models in Table
more than between the Second Assessment Report and the
8.1, found that differences in feedbacks contribute almost
TAR) that generally involve improved parametrizations of
three times more to the range in equilibrium climate sensitivity
speciﬁ c processes such as clouds, boundary layer or convection
estimates than differences in the models’ radiative forcings (the
(see Section 8.2). In some cases, developments have also
spread of models’ forcing is discussed in Section 10.2).
concerned numerics, dynamical cores or the coupling to
Several methods have been used to diagnose climate
new components (ocean, carbon cycle, etc.). Developing
feedbacks in GCMs, whose strengths and weaknesses are
new versions of a model to improve the physical basis of
reviewed in Stephens (2005) and Bony et al. (2006). These
parametrizations or the simulation of the current climate is at
methods include the ‘partial radiative perturbation’ approach
the heart of modelling group activities. The rationale for these
and its variants (e.g., Colman, 2003a; Soden and Held, 2006),
changes is generally based upon a combination of process-level
the use of radiative-convective models and the ‘cloud radiative
tests against observations or against cloud-resolving or large-
forcing’ method (e.g., Webb et al., 2006). Since the TAR,
eddy simulation models (see Section 8.2), and on the overall
there has been progress in comparing the feedbacks produced
quality of the model simulation (see Sections 8.3 and 8.4).
by climate models in doubled atmospheric CO2 equilibrium
These developments can, and do, affect the climate sensitivity
experiments (Colman, 2003a; Webb et al., 2006) and in
of models.
transient climate change integrations (Soden and Held, 2006).
The equilibrium climate sensitivity estimates from the
Water vapour, lapse rate, cloud and surface albedo feedback
latest model version used by modelling groups have increased
parameters, as estimated by Colman (2003a), Soden and Held
(e.g., CCSM3 vs CSM1.0, ECHAM5/MPI-OM vs ECHAM3/
(2006) and Winton (2006a) are shown in Figure 8.14.
LSG, IPSL-CM4 vs IPSL-CM2, MRI-CGCM2.3.2 vs MRI2,
In AOGCMs, the water vapour feedback constitutes by far
UKMO-HadGEM1 vs UKMO-HadCM3), decreased (e.g., the strongest feedback, with a multi-model mean and standard
CSIRO-MK3.0 vs CSIRO-MK2, GFDL-CM2.0 vs GFDL_
deviation for the MMD at PCMDI of 1.80 ± 0.18 W m–2
R30_c, GISS-EH and GISS-ER vs GISS2, MIROC3.2(hires)
°C–1, followed by the (negative) lapse rate feedback (–0.84 ±
and MIROC3.2(medres) vs CCSR/NIES2) or remained 0.26 W m–2 °C–1) and the surface albedo feedback (0.26 ±
roughly unchanged (e.g., CGCM3.1(T47) vs CGCM1, GFDL-
0.08 W m–2 °C–1). The cloud feedback mean is 0.69 W m–2 °C–1
CM2.1 vs GFDL_R30_c) compared to the TAR. In some
with a very large inter-model spread of ±0.38 W m–2 °C–1
models, changes in climate sensitivity are primarily ascribed to
(Soden and Held, 2006).
changes in the cloud parametrization or in the representation of
A substantial spread is apparent in the strength of water
cloud-radiative properties (e.g., CCSM3, MRI-CGCM2.3.2,
vapour feedback that is smaller in Soden and Held (2006)
MIROC3.2(medres) and MIROC3.2(hires)). However, in most
than in Colman (2003a). It is not known whether this smaller
models the change in climate sensitivity cannot be attributed to
spread indicates a closer consensus among current AOGCMs
a speciﬁ c change in the model. For instance, Johns et al. (2006)
than among older models, differences in the methodology or
showed that most of the individual changes made during the
differences in the nature of climate change integrations between
development of HadGEM1 have a small impact on the climate
the two studies. In both studies, the lapse rate feedback also
sensitivity, and that the global effects of the individual changes
shows a substantial spread among models, which is explained
5 Unless explicitly stated, GCM here refers both to AOGCM (used to estimate TCR) and AGCM coupled to a slab ocean (used to estimate equilibrium climate sensitivity).
630

Chapter 8
Climate Models and Their Evaluation
Table 8.2. Climate sensitivity estimates from the AOGCMs assessed in this report
the difference in mean lapse rate feedback between the two
(see Table 8.1 for model details). Transient climate response (TCR) and equilibrium
studies is unclear, but may relate to inappropriate inclusion of
climate sensitivity (ECS) were calculated by the modelling groups (using atmosphere
stratospheric temperature response in some feedback analyses
models coupled to slab ocean for equilibrium climate sensitivity), except those in ital-
(Soden and Held, 2006).
ics, which were calculated from simulations in the MMD at PCMDI. The ocean heat
uptake efﬁ ciency (W m–2 °C–1), discussed in Chapter 10, may be roughly estimated
The three studies, using different methodologies to estimate
as F2x x (TCR–1 – ECS–1), where F2x is the radiative forcing for doubled atmospheric
the global surface albedo feedback associated with snow and
CO2 concentration (see Supplementary Material, Table 8.SM.1)
sea ice changes, all suggest that this feedback is positive in all
the models, and that its range is much smaller than that of cloud
Equilibrium climate
Transient climate
AOGCM
feedbacks. Winton (2006a) suggests that about three-quarters
sensitivity (°C)
response (°C)
of the global surface albedo feedback arises from the NH (see
1: BCC-CM1
n.a.
n.a.
Section 8.6.3.3).
2: BCCR-BCM2.0
n.a.
n.a.
The diagnosis of global radiative feedbacks allows better
3: CCSM3
2.7
1.5
understanding of the spread of equilibrium climate sensitivity
estimates among current GCMs. In the idealised situation that the
4: CGCM3.1(T47)
3.4
1.9
climate response to a doubling of atmospheric CO2 consisted of
5: CGCM3.1(T63)
3.4
n.a.
a uniform temperature change only, with no feedbacks operating
6: CNRM-CM3
n.a.
1.6
(but allowing for the enhanced radiative cooling resulting from
7: CSIRO-MK3.0
3.1
1.4
the temperature increase), the global warming from GCMs
8: ECHAM5/MPI-OM
3.4
2.2
would be around 1.2°C (Hansen et al., 1984; Bony et al., 2006).
The water vapour feedback, operating alone on top of this,
9: ECHO-G
3.2
1.7
would at least double the response.6 The water vapour feedback
10: FGOALS-g1.0
2.3 1.2
is, however, closely related to the lapse rate feedback (see
11: GFDL-CM2.0
2.9
1.6
above), and the two combined result in a feedback parameter of
12: GFDL-CM2.1
3.4
1.5
approximately 1 W m–2 °C–1, corresponding to an ampliﬁ cation
13: GISS-AOM
n.a.
n.a.
of the basic temperature response by approximately 50%. The
14: GISS-EH
2.7
1.6
15: GISS-ER
2.7
1.5
16: INM-CM3.0
2.1
1.6
17: IPSL-CM4
4.4
2.1
18: MIROC3.2(hires)
4.3
2.6
19: MIROC3.2(medres)
4.0
2.1
20: MRI-CGCM2.3.2
3.2
2.2
21: PCM
2.1
1.3
22: UKMO-HadCM3
3.3
2.0
23: UKMO-HadGEM1
4.4
1.9
by inter-model differences in the relative surface warming of
low and high latitudes (Soden and Held, 2006). Because the
water vapour and temperature responses are tightly coupled
in the troposphere (see Section 8.6.3.1), models with a larger
(negative) lapse rate feedback also have a larger (positive) water
Figure 8.14. Comparison of GCM climate feedback parameters for water vapour
vapour feedback. These act to offset each other (see Box 8.1). As
(WV), cloud (C), surface albedo (A), lapse rate (LR) and the combined water vapour
a result, it is more reasonable to consider the sum of water vapour
plus lapse rate (WV + LR) in units of W m–2 °C–1. ‘ALL’ represents the sum of all feed-
and lapse rate feedbacks as a single quantity when analysing
backs. Results are taken from Colman (2003a; blue, black), Soden and Held (2006;
red) and Winton (2006a; green). Closed blue and open black symbols from Colman
the causes of inter-model variability in climate sensitivity.
(2003a) represent calculations determined using the partial radiative perturbation
This makes inter-model differences in the combination of
(PRP) and the radiative-convective method (RCM) approaches respectively. Crosses
water vapour and lapse rate feedbacks a substantially smaller
represent the water vapour feedback computed for each model from Soden and Held
contributor to the spread in climate sensitivity estimates than
(2006) assuming no change in relative humidity. Vertical bars depict the estimated
uncertainty in the calculation of the feedbacks from Soden and Held (2006).
differences in cloud feedback (Figure 8.14). The source of
1
6 Under these simplifying assumptions the ampliﬁ cation of the global warming from a feedback parameter λ (in W m-2 °C–1) with no other feedbacks operating is , where
1 + λ/ λp
λp is the ‘uniform temperature’ radiative cooling response (of value approximately –3.2 W m–2 °C–1; Bony et al., 2006). If n independent feedbacks operate, λ is replaced by ( λ +
1
λ +... ).
2
λn
631

Climate Models and Their Evaluation
Chapter 8
Box 8.1: Upper-Tropospheric Humidity and Water Vapour Feedback
Water vapour is the most important greenhouse gas in the atmosphere. Tropospheric water vapour concentration diminishes
rapidly with height, since it is ultimately limited by saturation-specifi c humidity, which strongly decreases as temperature decreases.
Nevertheless, these relatively low upper-tropospheric concentrations contribute disproportionately to the ‘natural’ greenhouse eff ect,
both because temperature contrast with the surface increases with height, and because lower down the atmosphere is nearly opaque
at wavelengths of strong water vapour absorption.
In the stratosphere, there are potentially important radiative impacts due to anthropogenic sources of water vapour, such as from
methane oxidation (see Section 2.3.7). In the troposphere, the radiative forcing due to direct anthropogenic sources of water vapour
(mainly from irrigation) is negligible (see Section 2.5.6). Rather, it is the response of tropospheric water vapour to warming itself – the
water vapour feedback – that matters for climate change. In GCMs, water vapour provides the largest positive radiative feedback (see
Section 8.6.2.3): alone, it roughly doubles the warming in response to forcing (such as from greenhouse gas increases). There are also
possible stratospheric water vapour feedback eff ects due to tropical tropopause temperature changes and/or changes in deep con-
vection (see Sections 3.4.2 and 8.6.3.1.1).
The radiative eff ect of absorption by water vapour is roughly proportional to the logarithm of its concentration, so it is the frac-
tional change in water vapour concentration, not the absolute change, that governs its strength as a feedback mechanism. Calcula-
tions with GCMs suggest that water vapour remains at an approximately constant fraction of its saturated value (close to unchanged
relative humidity (RH)) under global-scale warming (see Section 8.6.3.1). Under such a response, for uniform warming, the largest frac-
tional change in water vapour, and thus the largest contribution to the feedback, occurs in the upper troposphere. In addition, GCMs
fi nd enhanced warming in the tropical upper troposphere, due to changes in the lapse rate (see Section 9.4.4). This further enhances
moisture changes in this region, but also introduces a partially off setting radiative response from the temperature increase, and the
net eff ect of the combined water vapour/lapse rate feedback is to amplify the warming in response to forcing by around 50% (Section
8.6.2.3). The close link between these processes means that water vapour and lapse rate feedbacks are commonly considered together.
The strength of the combined feedback is found to be robust across GCMs, despite signifi cant inter-model diff erences, for example, in
the mean climatology of water vapour (see Section 8.6.2.3).
Confi dence in modelled water vapour feedback is thus aff ected by uncertainties in the physical processes controlling upper-tro-
pospheric humidity, and confi dence in their representation in GCMs. One important question is what the relative contribution of
large-scale advective processes (in which confi dence in GCMs’ representation is high) is compared with microphysical processes (in
which confi dence is much lower) for determining the distribution and variation in water vapour. Although advection has been shown
to establish the general distribution of tropical upper-tropospheric humidity in the present climate (see Section 8.6.3.1), a signifi cant
role for microphysics in humidity response to climate change cannot yet be ruled out.
Diffi
culties in observing water vapour in the upper troposphere have long hampered both observational and modelling studies,
and signifi cant limitations remain in coverage and reliability of observational humidity data sets (see Section 3.4.2). To reduce the im-
pact of these problems, in recent years there has been increased emphasis on the use of satellite data (such as 6.3 to 6.7 μm thermal
radiance measurements) for inferring variations or trends in humidity, and on direct simulation of satellite radiances in models as a
basis for model evaluation (see Sections 3.4.2 and 8.6.3.1.1).
Variations in upper-tropospheric water vapour have been observed across time scales from seasonal and interannual to decadal,
as well as in response to external forcing (see Section 3.4.2.2). At tropics-wide scales, they correspond to roughly unchanged RH (see
Section 8.6.3.1), and GCMs are generally able to reproduce these observed variations. Both column-integrated (see Section 3.4.2.1)
and upper-tropospheric (see Section 3.4.2.2) specifi c humidity have increased over the past two decades, also consistent with roughly
unchanged RH. There remains substantial disagreement between diff erent observational estimates of lapse rate changes over recent
decades, but some of these are consistent with GCM simulations (see Sections 3.4.1 and 9.4.4).
Overall, since the TAR, confi dence has increased in the conventional view that the distribution of RH changes little as climate
warms, particularly in the upper troposphere. Confi dence has also increased in the ability of GCMs to represent upper-tropospheric
humidity and its variations, both free and forced. Together, upper-tropospheric observational and modelling evidence provide strong
support for a combined water vapour/lapse rate feedback of around the strength found in GCMs (see Section 8.6.3.1.2).
632

Chapter 8
Climate Models and Their Evaluation
surface albedo feedback ampliﬁ es the basic response by about
rate feedback. To a ﬁ rst approximation, GCM simulations
10%, and the cloud feedback does so by 10 to 50% depending
indeed maintain a roughly unchanged distribution of RH under
on the GCM. Note, however, that because of the inherently
greenhouse gas forcing. More precisely, a small but widespread
nonlinear nature of the response to feedbacks, the ﬁ nal impact
RH decrease in GCM simulations typically reduces feedback
on sensitivity is not simply the sum of these responses. The
strength slightly compared with a constant RH response
effect of multiple positive feedbacks is that they mutually
(Colman, 2004; Soden and Held, 2006; Figure 8.14).
amplify each other’s impact on climate sensitivity.
In the planetary boundary layer, humidity is controlled by
Using feedback parameters from Figure 8.14, it can be
strong coupling with the surface, and a broad-scale quasi-
estimated that in the presence of water vapour, lapse rate and
unchanged RH response is uncontroversial (Wentz and Schabel,
surface albedo feedbacks, but in the absence of cloud feedbacks,
2000; Trenberth et al., 2005; Dai, 2006). Conﬁ dence in GCMs’
current GCMs would predict a climate sensitivity (±1 standard
water vapour feedback is also relatively high in the extratropics,
deviation) of roughly 1.9°C ± 0.15°C (ignoring spread from
because large-scale eddies, responsible for much of the
radiative forcing differences). The mean and standard deviation
moistening throughout the troposphere, are explicitly resolved,
of climate sensitivity estimates derived from current GCMs are
and keep much of the atmosphere at a substantial fraction of
larger (3.2°C ± 0.7°C) essentially because the GCMs all predict
saturation throughout the year (Stocker et al., 2001). Humidity
a positive cloud feedback (Figure 8.14) but strongly disagree
changes in the tropical middle and upper troposphere, however,
on its magnitude.
are less well understood and have more TOA radiative impact
The large spread in cloud radiative feedbacks leads to the
than do other regions of the atmosphere (e.g., Held and Soden,
conclusion that differences in cloud response are the primary
2000; Colman, 2001). Therefore, much of the research since
source of inter-model differences in climate sensitivity (see
the TAR has focused on the RH response in the tropics with
discussion in Section 8.6.3.2.2). However, the contributions
emphasis on the upper troposphere (see Bony et al., 2006 for a
of water vapour/lapse rate and surface albedo feedbacks to
review), and conﬁ dence in the humidity response of this region
sensitivity spread are non-negligible, particularly since their
is central to conﬁ dence in modelled water vapour feedback.
impact is reinforced by the mean model cloud feedback being
The humidity distribution within the tropical free troposphere
positive and quite strong.
is determined by many factors, including the detrainment of
vapour and condensed water from convective systems and
8.6.3
Key Physical Processes Involved in Climate
the large-scale atmospheric circulation. The relatively dry
Sensitivity
regions of large-scale descent play a major role in tropical
LW cooling, and changes in their area or humidity could
The traditional approach in assessing model sensitivity
potentially have a signiﬁ cant impact on water vapour feedback
has been to consider water vapour, lapse rate, surface albedo
strength (Pierrehumbert, 1999; Lindzen et al., 2001; Peters
and cloud feedbacks separately. Although this division can be
and Bretherton, 2005). Given the complexity of processes
regarded as somewhat artiﬁ cial because, for example, water
controlling tropical humidity, however, simple convincing
vapour, clouds and temperature interact strongly, it remains
physical arguments about changes under global-scale warming
conceptually useful, and is consistent in approach with previous
are difﬁ cult to sustain, and a combination of modelling and
assessments. Accordingly, and because of the relationship
observational studies are needed to assess the reliability of
between lapse rate and water vapour feedbacks, this subsection
model water vapour feedback.
separately addresses the water vapour/lapse rate feedbacks and
In contrast to cloud feedback, a strong positive water vapour
then the cloud and surface albedo feedbacks.
feedback is a robust feature of GCMs (Stocker et al., 2001),
being found across models with many different schemes for
8.6.3.1
Water Vapour and Lapse Rate
advection, convection and condensation of water vapour. High-
resolution mesoscale (Larson and Hartmann, 2003) and cloud-
Absorption of LW radiation increases approximately with the
resolving models (Tompkins and Craig, 1999) run on limited
logarithm of water vapour concentration, while the Clausius-
tropical domains also display humidity responses consistent
Clapeyron equation dictates a near-exponential increase in
with strong positive feedback, although with differences in
moisture-holding capacity with temperature. Since tropospheric
the details of upper-tropospheric RH (UTRH) trends with
and surface temperatures are closely coupled (see Section
temperature. Experiments with GCMs have found water vapour
3.4.1), these constraints predict a strongly positive water vapour
feedback strength to be insensitive to large changes in vertical
feedback if relative humidity (RH) is close to unchanged.
resolution, as well as convective parametrization and advection
Furthermore, the combined water vapour-lapse rate feedback is
schemes (Ingram, 2002). These modelling studies provide
relatively insensitive to changes in lapse rate for unchanged RH
evidence that the free-tropospheric RH response of global
(Cess, 1975) due to the compensating effects of water vapour
coupled models under climate warming is not simply an artefact
and temperature on the OLR (see Box 8.1). Understanding
of GCMs or of coarse GCM resolution, since broadly similar
processes determining the distribution and variability in RH
changes are found in a range of models of different complexity
is therefore central to understanding of the water vapour-lapse
and scope. Indirect supporting evidence for model water vapour
633

Climate Models and Their Evaluation
Chapter 8
feedback strength also comes from experiments which show
8.6.3.1.1
Evaluation of water vapour/lapse rate feedback
that suppressing humidity variation from the radiation code in
processes in models
an AOGCM produces unrealistically low interannual variability
Evaluation of the humidity distribution and its variability in
(Hall and Manabe, 1999).
GCMs, while not directly testing their climate change feedbacks,
Conﬁ dence in modelled water vapour feedback is dependent
can assess their ability to represent key physical processes
upon understanding of the physical processes important for
controlling water vapour and therefore affect conﬁ dence
controlling UTRH, and conﬁ dence in their representation
in their water vapour feedback. Limitations in coverage or
in GCMs. The TAR noted a sensitivity of UTRH to the
accuracy of radiosonde measurements or reanalyses have long
representation of cloud microphysical processes in several
posed a problem for UTRH evaluation in models (Trenberth et
simple modelling studies. However, other evidence suggests
al., 2001; Allan et al., 2004), and recent emphasis has been on
that the role of microphysics is limited. The observed RH
assessments using satellite measurements, along with increasing
ﬁ eld in much of the tropics can be well simulated without
efforts to directly simulate satellite radiances in models (so as to
microphysics, but simply by observed winds while imposing
reduce errors in converting to model-level RH) (e.g., Soden et
an upper limit of 100% RH on parcels (Pierrehumbert and
al., 2002; Allan et al., 2003; Iacono et al., 2003; Brogniez et al.,
Roca, 1998; Gettelman et al., 2000; Dessler and Sherwood,
2005; Huang et al., 2005).
2000), or by determining a detrainment proﬁ le from clear-
Major features of the mean humidity distribution are
sky radiative cooling (Folkins et al., 2002). Evaporation of
reasonably simulated by GCMs, along with the consequent
detrained cirrus condensate also does not play a major part in
distribution of OLR (see Section 8.3.1). In the important
moistening the tropical upper troposphere (Soden, 2004; Luo
subtropical subsidence regions, models show a range of skill
and Rossow, 2004), although cirrus might be important as
in representing the mean UTRH. Some large regional biases
a water vapour sink (Luo and Rossow, 2004). Overall, these
have been found (Iacono et al., 2003; Chung et al., 2004),
studies increase conﬁ dence in GCM water vapour feedback,
although good agreement of distribution and variability with
since they emphasise the importance of large-scale advective
satellite data has also been noted in some models (Allan et al.,
processes, or radiation, in which conﬁ dence in representation
2003; Brogniez et al., 2005). Uncertainties in satellite-derived
by GCMs is high, compared with microphysical processes, in
data sets further complicate such comparisons, however. Skill
which conﬁ dence is much lower. However, a signiﬁ cant role
in the reproduction of ‘bimodality’ in the humidity distribution
for microphysics in determining the distribution of changes in
at different time scales has also been found to differ between
water vapour under climate warming cannot yet be ruled out.
models (Zhang et al., 2003; Pierrehumbert et al., 2007),
Observations provide ample evidence of regional-scale
possibly associated with mixing processes and resolution. Note,
increases and decreases in tropical UTRH in response to
however, that given the near-logarithmic dependence of LW
changes in convection (Zhu et al., 2000; Bates and Jackson,
radiation on humidity, errors in the control climate humidity
2001; Blankenship and Wilheit, 2001; Wang et al., 2001; Chen
have little direct effect on climate sensitivity: it is the fractional
et al., 2002; Chung et al., 2004; Sohn and Schmetz, 2004).
change of humidity as climate changes that matters (Held and
Such changes, however, provide little insight into large-scale
Soden, 2000).
thermodynamic relationships (most important for the water
A number of new tests of large-scale variability of UTRH
vapour feedback) unless considered over entire circulation
have been applied to GCMs since the TAR, and have generally
systems. Recent observational studies of the tropical mean
found skill in model simulations. Allan et al. (2003) found
UTRH response to temperature have found results consistent
that an AGCM forced by observed SSTs simulated interannual
with that of near-unchanged RH at a variety of time scales (see
changes in tropical mean 6.7 μm radiance (sensitive to UTRH
Section 3.4.2.2). These include responses from interannual
and temperature) in broad agreement with High Resolution
variability (Bauer et al., 2002; Allan et al., 2003; McCarthy and
Infrared Radiation Sounder (HIRS) observations over the
Toumi, 2004), volcanic forcing (Soden et al., 2002; Forster and
last two decades. Minschwaner et al. (2006) analysed the
Collins, 2004) and decadal trends (Soden et al., 2005), although
interannual response of tropical mean 250 hPa RH to the mean
modest RH decreases are noted at high levels on interannual
SST of the most convectively active region in 16 AOGCMs
time scales (Minschwaner and Dessler, 2004; Section 3.4.2.3).
from the MMD at PCMDI. The mean model response (a
Seasonal variations in observed global LW radiation trapping
small decrease in RH) was statistically consistent with the
are also consistent with a strong positive water vapour feedback
215 hPa response inferred from satellite observations, when
(Inamdar and Ramanathan, 1998; Tsushima et al., 2005). Note,
uncertainties from observations and model spread were taken
however, that humidity responses to variability or shorter time-
into account. AGCMs have been able to reproduce global or
scale forcing must be interpreted cautiously, as they are not
tropical mean variations in clear sky OLR (sensitive to water
direct analogues to that from greenhouse gas increases, because
vapour and temperature distributions) over seasonal (Tsushima
of differences in patterns of warming and circulation changes.
et al., 2005) as well as interannual and decadal (Soden, 2000;
Allan and Slingo, 2002) time scales (although aerosol or
greenhouse gas uncertainties and sampling differences can
affect these latter comparisons; Allan et al., 2003). In the lower
634

Chapter 8
Climate Models and Their Evaluation
troposphere, GCMs can simulate global-scale interannual gradient (Soden and Held, 2006). There has been extensive
moisture variability well (e.g., Allan et al., 2003). At a smaller
testing of GCM tropospheric temperature response against
scale, a number of GCMs have also shown skill in reproducing
observational trends for climate change detection purposes (see
regional changes in UTRH in response to circulation changes
Section 9.4.4). Although some recent studies have suggested
such as from seasonal or interannual variability (e.g., Soden,
consistency between modelled and observed changes (e.g.,
1997; Allan et al., 2003; Brogniez et al., 2005).
Fu et al., 2004; Santer et al., 2005), debate continues as to the
A further test of the response of free tropospheric temperature
level of agreement, particularly in the tropics (Section 9.4.4).
and humidity to surface temperature in models is how well
Regardless, if RH remains close to unchanged, the combined
they can reproduce interannual correlations between surface
lapse rate and water vapour feedback is relatively insensitive
temperature and vertical humidity proﬁ les. Although GCMs are
to differences in lapse rate response (Cess, 1975; Allan et al.,
only partially successful in reproducing regional (Ross et al.,
2002; Colman, 2003a).
2002) and mean tropical (Bauer et al., 2002) correlations, the
In the stratosphere, GCM water vapour response is sensitive
marked disagreement found in previous studies (Sun and Held,
to the location of initial radiative forcing (Joshi et al., 2003;
1996; Sun et al., 2001) has been shown to be in large part an
Stuber et al., 2005). Forcing concentrated in the lower
artefact of sampling techniques (Bauer et al., 2002).
stratosphere, such as from ozone changes, invoked a positive
There have also been efforts since the TAR to test feedback involving increased stratospheric water vapour and
GCMs’ water vapour response against that from global-scale
tropical cold point temperatures in one study (Stuber et al.,
temperature changes of recent decades. One recent study used
2005). However, for more homogenous forcing, such as from
a long period of satellite data (1982–2004) to infer trends in
CO2, the stratospheric water vapour contribution to model
UTRH, and found that an AGCM, forced by observed SSTs,
sensitivity appears weak (Colman, 2001; Stuber et al., 2001,
was able to capture the observed global and zonal humidity
2005). There is observational evidence of possible long-term
trends well (Soden et al., 2005). A second approach uses the
increases in stratospheric water vapour (Section 3.4.2.3),
cooling following the eruption of Mt Pinatubo. Using estimated
although it is not yet clear whether this is a feedback process.
aerosol forcing, Soden et al. (2002) found a model-simulated
If there is a signiﬁ cant global mean trend associated with
response of HIRS 6.7 μm radiance consistent with satellite
feedback mechanisms, however, this could imply a signiﬁ cant
observations. They also found a model global temperature
stratospheric water vapour feedback (Forster and Shine, 2002).
response similar to that observed, but not if the water vapour
feedback was switched off (although the study neglected changes
8.6.3.1.2
Summary of water vapour and lapse fate
in cloud cover and potential heat uptake by the deep ocean).
feedbacks
Using radiation calculations based on humidity observations,
Signiﬁ cant progress has been made since the TAR in
Forster and Collins (2004) found consistency in inferred water
understanding and evaluating water vapour and lapse rate
vapour feedback strength with an ensemble of coupled model
feedbacks. New tests have been applied to GCMs, and have
integrations, although the latitude-height pattern of the observed
generally found skill in the representation of large-scale free
humidity response did not closely match any single realisation.
tropospheric humidity responses to seasonal and interannual
They deduced a water vapour feedback of 0.9 to 2.5 W m–2 °C–1,
variability, volcano-induced cooling and climate trends. New
a range which covers that of models under greenhouse gas
evidence from both observations and models has reinforced
forcing (see Figure 8.14). An important caveat to these studies
the conventional view of a roughly unchanged RH response to
is that the climate perturbation from Mt Pinatubo was small,
warming. It has also increased conﬁ dence in the ability of GCMs
not sitting clearly above natural variability (Forster and Collins,
to simulate important features of humidity and temperature
2004). Caution is also required when comparing with feedbacks
response under a range of different climate perturbations.
from increased greenhouse gases, because radiative forcing
Taken together, the evidence strongly favours a combined
from volcanic aerosol is differently distributed and occurs
water vapour-lapse rate feedback of around the strength found
over shorter time scales, which can induce different changes in
in global climate models.
circulation and bias the relative land/ocean response (although
a recent AOGCM study found similar global LW radiation clear
8.6.3.2 Clouds
sky feedbacks between the two forcings; Yokohata et al., 2005).
Nevertheless, comparing observed and modelled water vapour
By reﬂ ecting solar radiation back to space (the albedo
response to the eruption of Mt. Pinatubo constitutes one way to
effect of clouds) and by trapping infrared radiation emitted by
test model ability to simulate humidity changes induced by an
the surface and the lower troposphere (the greenhouse effect
external global-scale forcing.
of clouds), clouds exert two competing effects on the Earth’s
At low latitudes, GCMs show negative lapse rate feedback
radiation budget. These two effects are usually referred to as the
because of their tendency towards a moist adiabatic lapse rate,
SW and LW components of the cloud radiative forcing (CRF).
producing ampliﬁ ed warming aloft. At middle to high latitudes,
The balance between these two components depends on many
enhanced low-level warming, particuarly in winter, contributes
factors, including macrophysical and microphysical cloud
a positive feedback (e.g., Colman, 2003b), and global properties. In the current climate, clouds exert a cooling effect
feedback strength is dependent upon the meridional warming
on climate (the global mean CRF is negative). In response to
635

Climate Models and Their Evaluation
Chapter 8
global warming, the cooling effect of clouds on climate might be
Lindzen et al. (2001) speculated that the tropical area covered
enhanced or weakened, thereby producing a radiative feedback
by anvil clouds could decrease with rising temperature, and that
to climate warming (Randall et al., 2006; NRC, 2003; Zhang,
would lead to a negative climate feedback (iris hypothesis).
2004; Stephens, 2005; Bony et al., 2006).
Numerous objections have been raised about various aspects
In many climate models, details in the representation of
of the observational evidence provided so far (Chambers et al.,
clouds can substantially affect the model estimates of cloud
2002; Del Genio and Kovari, 2002; Fu et al., 2002; Harrison,
feedback and climate sensitivity (e.g., Senior and Mitchell,
2002; Hartmann and Michelsen, 2002; Lin et al., 2002, 2004),
1993; Le Treut et al., 1994; Yao and Del Genio, 2002; Zhang,
leading to a vigorous debate with the authors of the hypothesis
2004; Stainforth et al., 2005; Yokohata et al., 2005). Moreover,
(Bell et al., 2002; Chou et al., 2002; Lindzen et al., 2002). Other
the spread of climate sensitivity estimates among current
observational studies (Del Genio and Kovari, 2002; Del Genio
models arises primarily from inter-model differences in cloud
et al., 2005b) suggest an increase in the convective cloud cover
feedbacks (Colman, 2003a; Soden and Held, 2006; Webb et al.,
with surface temperature.
2006; Section 8.6.2, Figure 8.14). Therefore, cloud feedbacks
Boundary-layer clouds have a strong impact on the net
remain the largest source of uncertainty in climate sensitivity
radiation budget (e.g., Harrison et al., 1990; Hartmann et al.,
estimates.
1992) and cover a large fraction of the global ocean (e.g., Norris,
This section assesses the evolution since the TAR in the
1998a,b). Understanding how they may change in a perturbed
understanding of the physical processes involved in cloud
climate is thus a vital part of the cloud feedback problem. The
feedbacks (see Section 8.6.3.2.1), in the interpretation of the
observed relationship between low-level cloud amount and a
range of cloud feedback estimates among current climate
particular measure of lower tropospheric stability (Klein and
models (see Section 8.6.3.2.2) and in the evaluation of model
Hartmann, 1993), which has been used in some simple climate
cloud feedbacks using observations (see Section 8.6.3.2.3).
models and in some GCMs’ parametrizations of boundary-
layer cloud amount (e.g., CCSM3, FGOALS), led to the
8.6.3.2.1
Understanding of the physical processes involved
suggestion that a global climate warming might be associated
in cloud feedbacks
with an increased low-level cloud cover, which would produce
The Earth’s cloudiness is associated with a large spectrum
a negative cloud feedback (e.g., Miller, 1997; Zhang, 2004).
of cloud types, ranging from low-level boundary-layer clouds
However, variants of the lower-tropospheric stability measure,
to deep convective clouds and anvils. Understanding cloud
which may predict boundary-layer cloud amount as well as the
feedbacks requires an understanding of how a change in climate
Klein and Hartmann (1993) measure, would not necessarily
may affect the spectrum and the radiative properties of these
predict an increase in low-level clouds in a warmer climate
different clouds, and an estimate of the impact of these changes
(e.g., Williams et al., 2006). Moreover, observations indicate
on the Earth’s radiation budget. Moreover, since cloudy regions
that in regions covered by low-level clouds, the cloud optical
are also moist regions, a change in the cloud fraction matters for
depth decreases and the SW CRF weakens as temperature rises
both the water vapour and the cloud feedbacks (Pierrehumbert,
(Tselioudis and Rossow, 1994; Greenwald et al., 1995; Bony et
1995; Lindzen et al., 2001). Since the TAR, there have been some
al., 1997; Del Genio and Wolf, 2000; Bony and Dufresne, 2005),
advances in the analysis of physical processes involved in cloud
but the different factors that may explain these observations are
feedbacks, thanks to the combined analysis of observations,
not well established. Therefore, understanding of the physical
simple conceptual models, cloud-resolving models, mesoscale
processes that control the response of boundary-layer clouds
models and GCMs (reviewed in Bony et al., 2006). Major issues
and their radiative properties to a change in climate remains
are presented below.
very limited.
Several climate feedback mechanisms involving convective
At mid-latitudes, the atmosphere is organised in synoptic
anvil clouds have been examined. Hartmann and Larson (2002)
weather systems, with prevailing thick, high-top frontal clouds
proposed that the emission temperature of tropical anvil clouds is
in regions of synoptic ascent and low-level or no clouds in
essentially independent of the surface temperature (Fixed Anvil
regions of synoptic descent. In the NH, several climate models
Temperature hypothesis), and that it will thus remain unchanged
report a decrease in overall extratropical storm frequency and an
during climate change. This suggestion is consistent with cloud-
increase in storm intensity in response to climate warming (e.g.,
resolving model simulations showing that in a warmer climate,
Carnell and Senior, 1998; Geng and Sugi, 2003) and a poleward
the vertical proﬁ les of mid- and upper-tropospheric cloud
shift of the storm tracks (Yin, 2005). Using observations and
fraction, condensate and RH all tend to be displaced upward
reanalyses to investigate the impact that dynamical changes
in height together with the temperature (Tompkins and Craig,
such as those found by Carnell and Senior (1998) would have
1999). However, this hypothesis has not yet been tested with
on the NH radiation budget, Tselioudis and Rossow (2006)
observations or with cloud-resolving model simulations having
suggested that the increase in storm strength would have a larger
a ﬁ ne vertical resolution in the upper troposphere. The response
radiative impact than the decrease in storm frequency, and that
of the anvil cloud fraction to a change in temperature remains a
this would produce increased reﬂ ection of SW radiation and
subject of debate. Assuming that an increase with temperature
decreased emission of LW radiation. However, the poleward
in the precipitation efﬁ ciency of convective clouds could
shift of the storm tracks may decrease the amount of SW
decrease the amount of water detrained in the upper troposphere,
radiation reﬂ ected (Tsushima et al., 2006). In addition, several
636

Chapter 8
Climate Models and Their Evaluation
studies have used observations to investigate the dependence
SW cloud feedback component, and that the responses to global
of mid-latitude cloud radiative properties on temperature. Del
warming of both deep convective clouds and low-level clouds
Genio and Wolf (2000) showed that the physical thickness of
differ among GCMs. Recent analyses suggest that the response
low-level continental clouds decreases with rising temperature,
of boundary-layer clouds constitutes the largest contributor to
resulting in a decrease in the cloud water path and optical
the range of climate change cloud feedbacks among current
thickness as temperature rises, and Norris and Iacobellis (2005)
GCMs (Bony and Dufresne, 2005; Webb et al., 2006; Wyant et
suggested that over the NH ocean, a uniform change in surface
al., 2006). It is due both to large discrepancies in the radiative
temperature would result in decreased cloud amount and optical
response simulated by models in regions dominated by low-
thickness for a large range of dynamical conditions. The sign
level cloud cover (Figure 8.15), and to the large areas of the
of the climate change radiative feedback associated with the
globe covered by these regions. However, the response of other
combined effects of dynamical and temperature changes on
cloud types is also important because for each model it either
extratropical clouds is still unknown.
reinforces or partially cancels the radiative response from low-
The role of polar cloud feedbacks in climate sensitivity
level clouds. The spread of model cloud feedbacks is substantial
has been emphasized by Holland and Bitz (2003) and Vavrus
at all latitudes, and tends to be larger in the tropics (Bony et
(2004). However, these feedbacks remain poorly understood.
al., 2006; Webb et al., 2006). Differences in the representation
of mixed-phase clouds and in the degree of latitudinal shift
8.6.3.2.2 Interpretation of the range of cloud feedbacks
of the storm tracks predicted by the models also contribute to
among climate models
inter-model differences in the CRF response to climate change,
In doubled atmospheric CO2 equilibrium experiments particularly in the extratropics (Tsushima et al., 2006).
performed by mixed-layer ocean-atmosphere models as well

as in transient climate change integrations performed by fully
8.6.3.2.3
Evaluation of cloud feedbacks produced by climate
coupled ocean-atmosphere models, models exhibit a large range
models
of global cloud feedbacks, with roughly half of the climate
The evaluation of clouds in climate models has long been
models predicting a more negative CRF in response to global
based on comparisons of observed and simulated climatologies of
warming, and half predicting the opposite (Soden and Held,
TOA radiative ﬂ uxes and total cloud amount (see Section 8.3.1).
2006; Webb et al., 2006). Several studies
suggest that the sign of cloud feedbacks may
not be necessarily that of CRF changes (Zhang
et al., 1994; Colman, 2003a; Soden et al., 2004),
due to the contribution of clear-sky radiation
changes (i.e., of water vapour, temperature
and surface albedo changes) to the change in
CRF. The Partial Radiative Perturbation (PRP)
method, that excludes clear-sky changes from
the deﬁ nition of cloud feedbacks, diagnoses a
positive global net cloud feedback in virtually
all the models (Colman, 2003a; Soden and
Held, 2006). However, the cloud feedback
estimates diagnosed from either the change in
CRF or the PRP method are well correlated
(i.e., their relative ranking is similar), and they
exhibit a similar spread among GCMs.
By decomposing the GCM feedbacks into
regional components or dynamical regimes,
substantial progress has been made in the
interpretation of the range of climate change
Figure 8.15. Sensitivity (in W m–2 °C–1) of the tropical net cloud radiative forcing (CRF) to SST changes
associated with global warming (simulations in which CO
cloud feedbacks. The comparison of coupled
2 increases by 1% yr–1). The inset shows the
tropically averaged sensitivity Σ predicted by 15 AOGCMs used in this report: 7 models predict Σ < 0 and
AOGCMs used for the climate projections
8 models predict Σ > 0. The main panel compares the CRF sensitivity to SST predicted by the two groups
presented in Chapter 10 (Bony and Dufresne,
of models in different regimes of the large-scale tropical circulation (the 500 hPa vertical pressure velocity
2005), of atmospheric or slab ocean versions
is used as a proxy for large-scale motions, with negative values corresponding to large-scale ascending
motions, and positive values to sinking motions).Thick lines and vertical lines represent the mean and the
of current GCMs (Webb et al., 2006; Williams
standard deviation of model sensitivities within each group; dotted lines represent the minimum and maxi-
et al., 2006; Wyant et al., 2006), or of slightly
mum values of model sensitivities within each dynamical regime. The discrepancy between the two groups
older models (Williams et al., 2003; Bony
of models is greatest in regimes of large-scale subsidence. These regimes, which have a large statistical
et al., 2004; Volodin, 2004; Stowasser et al.;
weight in the tropics, are primarily covered by boundary-layer clouds. As a result, the spread of tropical
cloud feedbacks among the models (inset) primarily arises from inter-model differences in the radiative re-
2006) show that inter-model differences in
sponse of low-level clouds in regimes of large-scale subsidence. Adapted from Bony and Dufresne (2005).
cloud feedbacks are mostly attributable to the
637

Climate Models and Their Evaluation
Chapter 8
However, a good agreement with these observed quantities may
et al. (2006) found for instance that by considering the CRF
result from compensating errors. Since the TAR, and partly
response to a change in large-scale vertical velocity and in
due to the use of an International Satellite Cloud Climatology
lower-tropospheric stability, a component of the local mean
Project (ISCCP) simulator (Klein and Jakob, 1999; Webb et al.,
climate change cloud response can be related to the present-day
2001), the evaluation of simulated cloud ﬁ elds is increasingly
variability, and thus evaluated using observations. Bony and
done in terms of cloud types and cloud optical properties (Klein
Dufresne (2005) and Stowasser and Hamilton (2006) examined
and Jakob, 1999; Webb et al., 2001; Williams et al., 2003; Lin
the ability of the AOGCMs of Chapter 10 to simulate the change
and Zhang, 2004; Weare, 2004; Zhang et al., 2005; Wyant et
in tropical CRF to a change in SST, in large-scale vertical velocity
al., 2006). It has thus become more powerful and constrains
and in lower-tropospheric RH. They showed that the models are
the models more. In addition, a new class of observational tests
most different and least realistic in regions of subsidence, and
has been applied to GCMs, using clustering or compositing
to a lesser extent in regimes of deep convective activity. This
techniques, to diagnose errors in the simulation of particular
emphasizes the necessity to improve the representation and the
cloud regimes or in speciﬁ c dynamical conditions (Tselioudis et
evaluation of cloud processes in climate models, and especially
al., 2000; Norris and Weaver, 2001; Jakob and Tselioudis, 2003;
those of boundary-layer clouds.
Williams et al., 2003; Bony et al., 2004; Lin and Zhang, 2004;
Ringer and Allan, 2004; Bony and Dufresne, 2005; Del Genio
8.6.3.2.4
Conclusion on cloud feedbacks
et al., 2005a; Gordon et al., 2005; Bauer and Del Genio, 2006;
Despite some advances in the understanding of the physical
Williams et al., 2006; Wyant et al., 2006). An observational test
processes that control the cloud response to climate change
focused on the global response of clouds to seasonal variations
and in the evaluation of some components of cloud feedbacks
has been proposed to evaluate model cloud feedbacks (Tsushima
in current models, it is not yet possible to assess which of
et al., 2005), but has not yet been applied to current models.
the model estimates of cloud feedback is the most reliable.
These studies highlight some common biases in the However, progress has been made in the identiﬁ cation of the
simulation of clouds by current models (e.g., Zhang et al., 2005).
cloud types, the dynamical regimes and the regions of the globe
This includes the over-prediction of optically thick clouds
responsible for the large spread of cloud feedback estimates
and the under-prediction of optically thin low and middle-top
among current models. This is likely to foster more speciﬁ c
clouds. However, uncertainties remain in the observational
observational analyses and model evaluations that will improve
determination of the relative amounts of the different cloud
future assessments of climate change cloud feedbacks.
types (Chang and Li, 2005). For mid-latitudes, these biases have
been interpreted as the consequence of the coarse resolution
8.6.3.3 Cryosphere
Feedbacks
of climate GCMs and their resulting inability to simulate the
right strength of ageostrophic circulations (Bauer and Del
A number of feedbacks that signiﬁ cantly contribute to the
Genio, 2006) and the right amount of sub-grid scale variability
global climate sensitivity are due to the cryosphere. A robust
(Gordon et al., 2005). Although the errors in the simulation of
feature of the response of climate models to increases in
the different cloud types may eventually compensate and lead to
atmospheric concentrations of greenhouse gases is the poleward
a prediction of the mean CRF in agreement with observations
retreat of terrestrial snow and sea ice, and the polar ampliﬁ cation
(see Section 8.3), they cast doubts on the reliability of the model
of increases in lower-tropospheric temperature. At the same
cloud feedbacks. For instance, given the nonlinear dependence
time, the high-latitude response to increased greenhouse
of cloud albedo on cloud optical depth, the overestimate of the
gas concentrations is highly variable among climate models
cloud optical thickness implies that a change in cloud optical
(e.g., Holland and Bitz, 2003) and does not show substantial
depth, even of the right sign and magnitude, would produce a
convergence in the latest generation of AOGCMs (Chapman
too small radiative signature. Similarly, the under-prediction of
and Walsh, 2007; see also Section 11.8). The possibility of
low- and mid-level clouds presumably affects the magnitude
threshold behaviour also contributes to the uncertainty of how
of the radiative response to climate warming in the widespread
the cryosphere may evolve in future climate scenarios.
regions of subsidence. Modelling assumptions controlling the
Arguably, the most important simulated feedback associated
cloud water phase (liquid, ice or mixed) are known to be critical
with the cryosphere is an increase in absorbed solar radiation
for the prediction of climate sensitivity. However, the evaluation
resulting from a retreat of highly reﬂ ective snow or ice cover
of these assumptions is just beginning (Doutriaux-Boucher and
in a warmer climate. Since the TAR, some progress has been
Quaas, 2004; Naud et al., 2006). Tsushima et al. (2006) suggested
made in quantifying the surface albedo feedback associated
that observations of the distribution of each phase of cloud water
with the cryosphere. Hall (2004) found that the albedo feedback
in the current climate would provide a substantial constraint on
was responsible for about half the high-latitude response to a
the model cloud feedbacks at middle and high latitudes.
doubling of atmospheric CO2. However, an analysis of long
As an attempt to assess some components of the cloud
control simulations showed that it accounted for surprisingly
response to a change in climate, several studies have investigated
little internal variability. Hall and Qu (2006) show that biases of
the ability of GCMs to simulate the sensitivity of clouds and
a number of MMD models in reproducing the observed seasonal
CRF to interannual changes in environmental conditions. When
cycle of land snow cover (especially the spring melt) are tightly
examining atmosphere-mixed-layer ocean models, Williams
related to the large variations in snow albedo feedback strength
638

Chapter 8
Climate Models and Their Evaluation
latitude atmosphere and ocean, particularly to polar cloud
processes and ocean heat and freshwater transport. Additionally,
while impressive advances have occurred in developing sea ice
components of the AOGCMs since the TAR, particularly by the
inclusion of more sophisticated dynamics in most of them (see
Section 8.2.4), evaluation of cryospheric feedbacks through
the testing of model parametrizations against observations is
hampered by the scarcity of observational data in the polar
regions. In particular, the lack of sea ice thickness observations
is a considerable problem.
The role of sea ice dynamics in climate sensitivity has
remained uncertain for years. Some recent results with AGCMs
coupled to slab ocean models (Hewitt et al., 2001; Vavrus and
Harrison, 2003) support the hypothesis that a representation of
sea ice dynamics in climate models has a moderating impact on
climate sensitivity. However, experiments with full AOGCMs
(Holland and Bitz, 2003) show no compelling relationship
between the transient climate response and the presence or
absence of ice dynamics, with numerous model differences
presumably overwhelming whatever signal might be due to
Figure 8.16. Scatter plot of simulated springtime Δα /
values in climate
s ΔTs
ice dynamics. A substantial connection between the initial (i.e.,
change (ordinate) vs simulated springtime Δα /
values in the seasonal cycle
s ΔTs
control) simulation of sea ice and the response to greenhouse gas
(abscissa) in transient climate change experiments with 17 AOGCMs used in this
forcing (Holland and Bitz, 2003; Flato, 2004) further hampers
report (Δα and T are surface albedo and surface air temperature, respectively). The
s
s
climate change
/
values are the reduction in springtime surface albedo aver-
‘clean’ experiments aimed at identifying or quantifying the role
Δαs ΔTs
aged over Northern Hemisphere continents between the 20th and 22nd centuries
of sea ice dynamics.
divided by the increase in surface air temperature in the region over the same time
A number of processes, other than surface albedo feedback,
period. Seasonal cycle Δα /
values are the difference between 20th-century
s ΔTs
have been shown to also contribute to the polar ampliﬁ cation
mean April and May α averaged over Northern Hemisphere continents divided by
s
the difference between April and May T averaged over the same area and time
of warming in models (Alexeev, 2003, 2005; Holland and Bitz,
s
period. A least-squares ﬁ t regression line for the simulations (solid line) and the ob-
2003; Vavrus, 2004; Cai, 2005; Winton, 2006b). An important
served seasonal cycle Δα /
value based on ISCCP and ERA40 reanalysis (dashed
s ΔTs
one is additional poleward energy transport, but contributions
vertical line) are also shown. The grey bar gives an estimate of statistical error,
from local high-latitude water vapour, cloud and temperature
according to a standard formula for error in the estimate of the mean of a time series
(in this case the observed time series of
/
) given the time series’ length and
feedbacks have also been found. The processes and their
Δαs ΔTs
variance. If this statistical error only is taken into account, the probability that the
interactions are complex, however, with substantial variation
actual observed value lies outside the grey bar is 5%. Each number corresponds to a
between models (Winton, 2006b), and their relative importance
particular AOGCM (see Table 8.1). Adapted from Hall and Qu (2006).
contributing to or dampening high-latitude ampliﬁ cation has
not yet been properly resolved.
simulated by the same models in climate change scenarios.
Addressing the seasonal cycle biases would therefore provide a
8.6.4
How to Assess Our Relative Conﬁ dence in
constraint that would reduce divergence in simulations of snow
Feedbacks Simulated by Different Models?
albedo feedback under climate change. However, possible use of
seasonal snow albedo feedback to evaluate snow albedo feedback
Assessments of our relative conﬁ
dence in climate
under climate change conditions is of course dependent upon the
projections from different models should ideally be based on
realism of the correlation between the two feedbacks suggested
a comprehensive set of observational tests that would allow us
by GCMs (Figure 8.16). A new result found independently by
to quantify model errors in simulating a wide variety of climate
Winton (2006a) and Qu and Hall (2005) is that surface processes
statistics, including simulations of the mean climate and
are the main source of divergence in climate simulations of
variability and of particular climate processes. The collection
surface albedo feedback, rather than simulated differences in
of measures that quantify how well a model performs in an
cloud ﬁ elds in cryospheric regions.
ensemble of tests of this kind are referred to as ‘climate metrics’.
Understanding of other feedbacks associated with the
To have the ability to constrain future climate projections, they
cryosphere (e.g., ice insulating feedback, MOC/SST-sea ice
would ideally have strong connections with one or several
feedback, ice thickness/ice growth feedback) has improved
aspects of climate change: climate sensitivity, large-scale
since the TAR (NRC, 2003; Bony et al., 2006). However, the
patterns of climate change (inter-hemispheric symmetry, polar
relative inﬂ uence on climate sensitivity of these feedbacks has
ampliﬁ cation, vertical patterns of temperature change, land-
not been quantiﬁ ed.
sea contrasts), regional patterns or transient aspects of climate
Understanding and evaluating sea ice feedbacks is change. For example, to assess conﬁ dence in model projections
complicated by their strong coupling to processes in the high-
of the Australian climate, the metrics would need to include
639

Climate Models and Their Evaluation
Chapter 8
some measures of the quality of ENSO simulation because
quantitative measures to identify these points in a time series
the Australian climate depends much on this variability (see
of a given variable (e.g., Lanzante, 1996; Seidel and Lanzante,
Section 11. 7).
2004; Tomé and Miranda, 2004). The most common way to
To better assess conﬁ dence in the different model estimates
identify thresholds and abrupt changes is by linearly de-trending
of climate sensitivity, two kinds of observational tests are
the input time series and looking for large deviations from the
available: tests related to the global climate response associated
trend line. More statistically rigorous methods are usually based
with speciﬁ ed external forcings (discussed in Chapters 6, 9
on Bayesian statistics.
and 10; Box 10.2) and tests focused on the simulation of key
This section explores the potential causes and mechanisms
feedback processes.
for producing thresholds and abrupt climate change and
Based on the understanding of both the physical processes that
addresses the issue of how well climate models can simulate
control key climate feedbacks (see Section 8.6.3), and also the
these changes. The following discussion is split into two main
origin of inter-model differences in the simulation of feedbacks
areas: forcing changes that can result in abrupt changes and
(see Section 8.6.2), the following climate characteristics appear
abrupt climate changes that result from large natural variability
to be particularly important: (i) for the water vapour and lapse
on long time scales. Formally, the latter abrupt changes do not
rate feedbacks, the response of upper-tropospheric RH and lapse
ﬁ t the deﬁ nition of thresholds and abrupt changes, because
rate to interannual or decadal changes in climate; (ii) for cloud
the forcing (at least radiative forcing – the external boundary
feedbacks, the response of boundary-layer clouds and anvil
condition) is not changing in time. However these changes
clouds to a change in surface or atmospheric conditions and the
have been discussed in the literature and popular press and are
change in cloud radiative properties associated with a change
worthy of assessment here.
in extratropical synoptic weather systems; (iii) for snow albedo
feedbacks, the relationship between surface air temperature and
8.7.2
Forced Abrupt Climate Change
snow melt over northern land areas during spring and (iv) for
sea ice feedbacks, the simulation of sea ice thickness.
8.7.2.1
Meridional Overturning Circulation Changes
A number of diagnostic tests have been proposed since the
TAR (see Section 8.6.3), but few of them have been applied to
As the radiative forcing of the planet changes, the climate
a majority of the models currently in use. Moreover, it is not yet
system responds on many different time scales. For the
clear which tests are critical for constraining future projections.
physical climate system typically simulated in coupled models
Consequently, a set of model metrics that might be used to
(atmosphere, ocean, land, sea ice), the longest response time
narrow the range of plausible climate change feedbacks and
scales are found in the ocean (Stouffer, 2004). In terms of
climate sensitivity has yet to be developed.
thresholds and abrupt climate changes on decadal and longer
time scales, the ocean has also been a focus of attention. In
particular, the ocean’s Atlantic MOC (see Box 5.1 for deﬁ nition
and description) is a main area of study.

8.7 Mechanisms
Producing
The MOC transports large amounts of heat (order of

Thresholds and Abrupt Climate

1015 Watts) and salt into high latitudes of the North Atlantic.
Change
There, the heat is released to the atmosphere, cooling the
surface waters. The cold, relatively salty waters sink to depth
and ﬂ ow southward out of the Atlantic Basin. The complete
8.7.1 Introduction
set of climatic drivers of this circulation remains unclear but
it is likely that both density (e.g., Stommel 1961; Rooth 1982)
This discussion of thresholds and abrupt climate change is
and wind stress forcings (e.g., Wunsch, 2002; Timmermann
based on the deﬁ nitions of ‘threshold’ and ‘abrupt’ proposed
and Goosse, 2004) are important. Both palaeoclimate studies
by Alley et al. (2002). The climate system tends to respond
(e.g., Broecker, 1997; Clark et al., 2002) and modelling studies
to changes in a gradual way until it crosses some threshold:
(e.g., Manabe and Stouffer, 1988, 1997; Vellinga and Wood,
thereafter any change that is deﬁ ned as abrupt is one where
2002) suggest that disruptions in the MOC can produce abrupt
the change in the response is much larger than the change in
climate changes. A systematic model intercomparison study
the forcing. The changes at the threshold are therefore abrupt
(Rahmstorf et al., 2005) found that all 11 participating EMICs
relative to the changes that occur before or after the threshold
had a threshold where the MOC shuts down (see Section 8.8.3).
and can lead to a transition to a new state. The spatial scales for
Due to the high computational cost, such a search for thresholds
these changes can range from global to local. In this deﬁ nition,
has not yet been performed with AOGCMs.
the magnitude of the forcing and response are important. In
It is important to note the distinction between the equilibrium
addition to the magnitude, the time scale being considered
and transient or time-dependent responses of the MOC to
is also important. This section focuses mainly on decadal to
changes in forcing. Due to the long response time scales found
centennial time scales.
in the ocean (some longer than 1 kyr), it is possible that the short-
Because of the somewhat subjective nature of the deﬁ nitions
term response to a given forcing change may be very different
of threshold and abrupt, there have been efforts to develop
from the equilibrium response. Such behaviour of the coupled
640

Chapter 8
Climate Models and Their Evaluation
system has been documented in at least one AOGCM (Stouffer
to test the models’ response on decadal to centennial time scales
and Manabe, 2003) and suggested in the results of a few other
remains to be accomplished.
AOGCM studies (e.g., Hirst, 1999; Senior and Mitchell, 2000;
The processes determining MOC response to increasing
Bryan et al., 2006). In these AOGCM experiments, the MOC
greenhouse gases have been studied in a number of models. In
weakens as the greenhouse gases increase in the atmosphere.
many models, initial MOC response to increasing greenhouse
When the CO2 concentration is stabilised, the MOC slowly
gases is dominated by thermal effects. In most models, this is
returns to its unperturbed value.
enhanced by changes in salinity driven by, among other things,
As discussed in section 10.3.4, the MOC typically weakens
the expected strengthening of the hydrological cycle (Gregory
as greenhouse gases increase due to the changes in surface heat
et al., 2005; Chapter 10). Melt water runoff from a melting of the
and freshwater ﬂ uxes at high latitudes (Manabe et al., 1991). The
Greenland Ice Sheet is a potentially major source of freshening
surface ﬂ ux changes reduce the surface density, hindering the
not yet included in the models found in the MMD (see Section
vertical movement of water and slowing the MOC. As the MOC
8.7.2.2). More complex feedbacks, associated with wind and
slows, it could approach a threshold where the circulation can
hydrological changes, are also important in many models. These
no longer sustain itself. Once the MOC crosses this threshold,
include local surface ﬂ ux anomalies in deep-water formation
it could rapidly change states, causing abrupt climate change
regions (Gent, 2001) and oceanic teleconnections driven by
where the North Atlantic and surrounding land areas would
changes to the freshwater budget of the tropical and South
cool relative to the case where the MOC is active. This cooling
Atlantic (e.g., Latif et al., 2000; Thorpe et al., 2001; Vellinga
is the result of the loss of heat transport from low latitudes in
et al., 2002; Hu et al., 2004). The magnitudes of the climate
the Atlantic and the feedbacks associated with the reduction in
factors causing the MOC to weaken, along with the feedbacks
the vertical mixing of high-latitude waters.
and the associated restoring factors, are all uncertain at this time.
A common misunderstanding is that the MOC weakening
Evaluation of these processes in AOGCMs is mainly restricted
could cause the onset of an ice age. However, no model has
by lack of observations, but some early progress has been made
supported this speculation when forced with realistic estimates
in individual studies (e.g., Schmittner et al., 2000; Pardaens et
of future climate forcings (see Section 10.3.4). In addition, in
al., 2003; Wu et al., 2005; Chapter 9). Model intercomparison
idealised modelling studies where the MOC was forced to shut
studies (e.g., Gregory et al., 2005; Rahmstorf et al., 2005;
down through very large sources of freshwater (not changes
Stouffer et al., 2006) were developed to identify and understand
in greenhouse gases), the surface temperature changes do not
the causes for the wide range of MOC responses in the coupled
support the idea that an ice age could result from a MOC shut
models used here (see Chapters 4, 6 and 10).
down, although the impacts on climate would be large (Manabe
and Stouffer, 1988, 1997; Schiller et al., 1997; Vellinga and Wood,
8.7.2.2
Rapid West Antarctic and/or Greenland Ice Sheet
2002; Stouffer et al., 2006). In a recent intercomparison involving
Collapse and Meridional Overturning Circulation
11 coupled atmosphere-ocean models (Gregory et al., 2005), the
Changes
MOC decreases by only 10 to 50% during a 140-year period (as
atmospheric CO2 quadruples), and in no model is there a land
Increased inﬂ ux of freshwater to the ocean from the ice
cooling anywhere (as the global-scale heating due to increasing
sheets is a potential forcing for abrupt climate changes. For
CO2 overwhelms the local cooling effect due to reduced MOC).
Antarctica in the present climate, these ﬂ uxes chieﬂ y arise from
Because of the large amount of heat and salt transported
melting below the ice shelves and from melting of icebergs
northward and its sensitivity to surface ﬂ uxes, the changes
transported by the ocean; both ﬂ uxes could increase signiﬁ cantly
in the MOC are able to produce abrupt climate change on
in a warmer climate. Ice sheet runoff and iceberg calving, in
decadal to centennial time scales (e.g., Manabe and Stouffer,
roughly equal shares, currently dominate the freshwater ﬂ ux
1995; Stouffer et al., 2006). Idealised studies using present-
from the Greenland Ice Sheet (Church et al., 2001; Chapter 4).
day simulations have shown that models can simulate many
In a warming climate, runoff is expected to quickly increase and
of the variations seen in the palaeoclimate record on decadal
become much larger than the calving rate, the latter of which
to centennial time scales when forced by ﬂ uxes of freshwater
in turn is likely to decrease as less and thinner ice borders the
water at the ocean surface. However, the quantitative response
ocean; basal melting from below the grounded ice will remain
to freshwater inputs varies widely among models (Stouffer et
several orders of magnitude smaller than the other ﬂ uxes
al., 2006), which led the CMIP and Paleoclimate Modelling
(Huybrechts et al., 2002). For a discussion of the likelihood
Intercomparison Project (PMIP) panels to design and support a
of these ice sheet changes and the effects on sea level, see the
set of coordinated experiments to study this issue (http://www.
discussion in Chapter 10.
gfdl.noaa.gov/~kd/CMIP.html and http://www.pmip2.cnrs-gif.
Changes in the surface forcing near the deep-water
fr/pmip2/design/experiments/waterhosing.shtml).
production areas seem to be most capable of producing rapid
In addition to the amount of the freshwater input, the exact
climate changes on decadal and longer time scales due to
location of that input may also be important (Rahmstorf 1996,
changes in the ocean circulation and mixing. If there are large
Manabe and Stouffer, 1997; Rind et al., 2001). Designing
changes in the ice volume over Greenland, it is likely that
experiments and determining the realistic past forcings needed
much of this melt water will freshen the surface waters in the
641

Climate Models and Their Evaluation
Chapter 8
high-latitude North Atlantic, slowing down the MOC (see
Section 4.7.2.4). Methane is also stored in the soils in areas of
Section 8.7.2.1; Chapter 10). Rind et al. (2001) found that
permafrost and warming increases the likelihood of a positive
changes in the NADW formation rate could instigate changes
feedback in the climate system via permafrost melting and the
in the deep-water formation around Antarctica.
release of trapped methane into the atmosphere. The likelihood
The response of the Atlantic MOC to changes in the Antarctic
of methane release from methane hydrates found in the oceans or
Ice Sheet is less well understood. Experiments with ocean-only
methane trapped in permafrost layers is assessed in Chapter 7.
models where the melt water changes are imposed as surface
This subsection considers the potential usefulness of models
salinity changes indicate that the Atlantic MOC will intensify as
in determining if those releases could trigger an abrupt climate
the waters around Antarctica become less dense (Seidov et al.,
change. Both forms of methane release represent a potential
2001). Weaver et al. (2003) showed that by adding freshwater in
threshold in the climate system. As the climate warms, the
the Southern Ocean, the MOC could change from an ‘off’ state
likelihood of the system crossing a threshold for a sudden
to a state similar to present day. However, in an experiment
release increases (see Chapters 4, 7 and 10). Since these changes
with an AOGCM, Seidov et al. (2005) found that an external
produce changes in the radiative forcing through changes in the
source of freshwater in the Southern Ocean resulted in a surface
greenhouse gas concentrations, the climatic impacts of such
freshening throughout the world ocean, weakening the Atlantic
a release are the same as an increase in the rate of change in
MOC. In these model results, the SH MOC associated with
the radiative forcing. Therefore, the models’ ability to simulate
Antarctic Bottom Water (AABW) formation weakened, causing
any abrupt climate change should be similar to their ability to
a cooling around Antarctica. See Chapters 4, 6 and 10 for more
simulate future abrupt climate changes due to changes in the
discussion about the likelihood of large melt water ﬂ uxes from
greenhouse gas forcing.
the ice sheets affecting the climate.
In summary, there is a potential for rapid ice sheet changes to
8.7.2.5 Biogeochemical
produce rapid climate change both through sea level changes and
ocean circulation changes. The ocean circulation changes result
Two questions concerning biogeochemical aspects of the
from increased freshwater ﬂ ux over the particularly sensitive
climate system are addressed here. First, can biogeochemical
deep-water production sites. In general, the possible climate
changes lead to abrupt climate change? Second, can abrupt
changes associated with future evolution of the Greenland
changes in the MOC further affect radiative forcing through
Ice Sheet are better understood than are those associated with
biogeochemical feedbacks?
changes in the Antarctic Ice Sheets.
Abrupt changes in biogeochemical systems of relevance to
our capacity to simulate the climate of the 21st century are not
8.7.2.3 Volcanoes
well understood (Friedlingstein et al., 2003). The potential for
major abrupt change exists in the uptake and storage of carbon
Volcanoes produce abrupt climate responses on short time
by terrestrial systems. While abrupt change within the climate
scales. The surface cooling effect of the stratospheric aerosols,
system is beginning to be seriously considered (Rial et al., 2004;
the main climatic forcing factor, decays in one to three years
Schneider, 2004), the potential for abrupt change in terrestrial
after an eruption due to the lifetime of the aerosols in the
systems, such as loss of soil carbon (Cox et al., 2000) or die
stratosphere. It is possible for one large volcano or a series of
back of the Amazon forests (Cox et al., 2004) remains uncertain.
large volcanic eruptions to produce climate responses on longer
In part this is due to lack of understanding of processes (see
time scales, especially in the subsurface region of the ocean
Friedlingstein et al., 2003; Chapter 7) and in part it results from
(Delworth et al., 2005; Gleckler et al., 2006b).
the impact of differences in the projected climate sensitivities
The models’ ability to simulate any possible abrupt response
in the host climate models (Joos et al., 2001; Govindasamy et
of the climate system to volcanic eruptions seems conceptually
al., 2005; Chapter 10) where changes in the physical climate
similar to their ability to simulate the climate response to future
system affect the biological response.
changes in greenhouse gases in that both produce changes in the
There is some evidence of multiple equilibria within
radiative forcing of the planet. However, mechanisms involved
vegetation-soil-climate systems. These include North Africa
in the exchange of heat between the atmosphere and ocean may
and Central East Asia where Claussen (1998), using an EMIC
be different in response to volcanic forcing when compared
with a land vegetation component, showed two stable equilibria
to the response to increase greenhouse gases. Therefore, the
for rainfall, dependent on initial land surface conditions.
feedbacks involved may be different (see Section 9.6.2.2 for
Kleidon et al. (2000), Wang and Eltahir (2000) and Renssen et
more discussion).
al. (2003) also found evidence for multiple equilibria. These are
preliminary assessments using relatively simple physical climate
8.7.2.4
Methane Hydrate Instability/Permafrost Methane
models that highlight the possibility of irreversible change in
the Earth system but require extensive further research to assess
Methane hydrates are stored on the seabed along continental
the reliability of the phenomena found.
margins where they are stabilised by high pressures and low
There have only been a few preliminary studies of the impact
temperatures, implying that ocean warming may cause hydrate
of abrupt climate changes such as the shutdown of the MOC on
instability and release of methane into the atmosphere (see
the carbon cycle. The ﬁ ndings of these studies indicate that the
642

Chapter 8
Climate Models and Their Evaluation
shutdown of the MOC would tend to increase the amount of
is their high computational cost. To date, unless modest-
greenhouse gases in the atmosphere (Joos et al., 1999; Plattner
resolution models are executed on an exceptionally large-
et al., 2001; Chapter 6). In both of these studies, only the effect
scale distributed computed system, as in the climateprediction.
of the oceanic component of the carbon cycle changes was
net project (http://climateprediction.net; Stainforth et al.,
considered.
2005), only a limited number of multi-decadal experiments
can be performed with AOGCMs, which hinders a systematic
8.7.3
Unforced Abrupt Climate Change
exploration of uncertainties in climate change projections and
prevents studies of the long-term evolution of climate.
Formally, as noted above, the changes discussed here do
At the other end of the spectrum of climate system model
not fall into the deﬁ nition of abrupt climate change. In the
complexity are the so-called simple climate models (see Harvey
literature, unforced abrupt climate change falls into two general
et al., 1997 for a review of these models). The most advanced
categories. One is just a red noise time series, where there is
simple climate models contain modules that calculate in a highly
power at decadal and longer time scales. A second category is a
parametrized way (1) the abundances of atmospheric greenhouse
bimodal or multi-modal distribution. In practice, it can be very
gases for given future emissions, (2) the radiative forcing
difﬁ cult to distinguish between the two categories unless the
resulting from the modelled greenhouse gas concentrations
time series are very long – long enough to eliminate sampling
and aerosol precursor emissions, (3) the global mean surface
as an issue – and the forcings are fairly constant in time. In
temperature response to the computed radiative forcing and
observations, neither of these conditions is normally met.
(4) the global mean sea level rise due to thermal expansion of
Models, both AOGCMs and less complex models, have
sea water and the response of glaciers and ice sheets. These
produced examples of large abrupt climate change (e.g., Hall
models are much more computationally efﬁ cient than AOGCMs
and Stouffer 2001; Goosse et al., 2002) without any changes
and thus can be utilised to investigate future climate change in
in forcing. Typically, these events are associated with changes
response to a large number of different scenarios of greenhouse
in the ocean circulation, mainly in the North Atlantic. An
gas emissions. Uncertainties from the modules can also be
abrupt event can last for several years to a few centuries. They
concatenated, potentially allowing the climate and sea level
bear some similarities with the conditions observed during a
results to be expressed as probabilistic distributions, which is
relatively cold period in the recent past in the Arctic (Goosse
harder to do with AOGCMs because of their computational
et al., 2003)
expense. A characteristic of simple climate models is that climate
Unfortunately, the probability of such an event is difﬁ cult
sensitivity and other subsystem properties must be speciﬁ ed
to estimate as it requires a very long experiment and is
based on the results of AOGCMs or observations. Therefore,
certainly dependent on the mean state simulated by the model.
simple climate models can be tuned to individual AOGCMs
Furthermore, comparison with observations is nearly impossible
and employed as a tool to emulate and extend their results (e.g.,
since it would require a very long period with constant forcing
Cubasch et al., 2001; Raper et al., 2001). They are useful mainly
which does not exist in nature. Nevertheless, if an event such as
for examining global-scale questions.
the one of those mentioned above were to occur in the future,
To bridge the gap between AOGCMs and simple climate
it would make the detection and attribution of climate changes
models, EMICs have been developed. Given that this gap is
very difﬁ cult.
quite large, there is a wide range of EMICs (see the reviews of
Saltzman, 1978 and Claussen et al., 2002). Typically, EMICs use
a simpliﬁ ed atmospheric component coupled to an OGCM or
8.8
Representing the Global System
simpliﬁ ed atmospheric and oceanic components. The degree of
with Simpler Models
simpliﬁ cation of the component models varies among EMICs.
Earth System Models of Intermediate Complexity
are reduced-resolution models that incorporate most of
8.8.1 Why
Lower
Complexity?
the processes represented by AOGCMs, albeit in a more
parametrized form. They explicitly simulate the interactions
An important concept in climate system modelling is that
between various components of the climate system. Similar to
of a spectrum of models of differing levels of complexity,
AOGCMs, but in contrast to simple climate models, the number
each being optimum for answering speciﬁ c questions. It is not
of degrees of freedom of an EMIC exceeds the number of
meaningful to judge one level as being better or worse than
adjustable parameters by several orders of magnitude. However,
another independently of the context of analysis. What is
these models are simple enough to permit climate simulations
important is that each model be asked questions appropriate for
over several thousand of years or even glacial cycles (with a
its level of complexity and quality of its simulation.
period of some 100 kyr), although not all are suitable for this
The most comprehensive models available are AOGCMs.
purpose. Moreover, like simple climate models, EMICs can
These models, which include more and more components of the
explore the parameter space with some completeness and
climate system (see Section 8.2), are designed to provide the best
are thus appropriate for assessing uncertainty. They can also
representation of the system and its dynamics, thereby serving
be utilised to screen the phase space of climate or the history
as the most realistic laboratory of nature. Their major limitation
of climate in order to identify interesting time slices, thereby
643

Climate Models and Their Evaluation
Chapter 8
providing guidance for more detailed studies to be undertaken
climate model are the effective climate sensitivity, the ocean
with AOGCMs. In addition, EMICs are invaluable tools for
effective vertical diffusivity, and the equilibrium land-ocean
understanding large-scale processes and feedbacks acting warming ratio. Values speciﬁ c to each AOGCM for the radiative
within the climate system. Certainly, it would not be sensible to
forcing for CO2 doubling were used in the tuning procedure
apply an EMIC to studies that require high spatial and temporal
where available (from Forster and Taylor, 2006, supplemented
resolution. Furthermore, model assumptions and restrictions,
with values provided directly from the modelling groups).
hence the limit of applicability of individual EMICs, must be
Otherwise, a default value of 3.71 W m–2 was chosen (Myhre
carefully studied. Some EMICs include a zonally averaged
et al., 1998). Default values of 1 W m–2 °C–1, 1 W m–2 °C–1 and
atmosphere or zonally averaged oceanic basins. In a number
8°C were used for the land-ocean heat exchange coefﬁ cient, the
of EMICs, cloudiness and/or wind ﬁ elds are prescribed and
inter-hemispheric heat exchange coefﬁ cient and the magnitude
do not evolve with changing climate. In still other EMICs, the
of the warming that would result in a collapse of the MOC,
atmospheric synoptic variability is not resolved explicitly, but
respectively (see Appendix 9.1 of the TAR).
diagnosed by using a statistical-dynamical approach. A priori,
The obtained best-ﬁ t climate sensitivity estimates differ for
it is not obvious how the reduction in resolution or dynamics/
various reasons from other estimates that were derived with
physics affects the simulated climate. As shown in Section 8.8.3
alternative methods. Such alternative methods include, for
and in Chapters 6, 9 and 10, at large scales most EMIC results
example, regression estimates that use a global energy balance
compare well with observational or proxy data and AOGCM
equation around the year of atmospheric CO2 doubling or the
results. Therefore, it is argued that there is a clear advantage in
analysis of slab ocean equilibrium warmings. The resulting
having available a spectrum of climate system models.
differences in climate sensitivity estimates can be partially
explained by the non-time constant effective climate sensitivities
8.8.2
Simple Climate Models
in many of the AOGCM runs. Furthermore, tuning results of a
simple climate model will be affected by the model structure,
As in the TAR, a simple climate model is utilised in this report
although simple, and other default parameter settings that affect
to emulate the projections of future climate change conducted
the simple model transient response.
with state-of-the-art AOGCMs, thus allowing the investigation
of the temperature and sea level implications of all relevant
8.8.3
Earth System Models of Intermediate
emission scenarios (see Chapter 10). This model is an updated
Complexity
version of the Model for the Assessment of Greenhouse-Gas
Induced Climate Change (MAGICC) model (Wigley and Raper,
Pictorially, EMICs can be deﬁ ned in terms of the components
1992, 2001; Raper et al., 1996). The calculation of the radiative
of a three-dimensional vector (Claussen et al., 2002): the
forcings from emission scenarios closely follows that described in
number of interacting components of the climate system
Chapter 2, and the feedback between climate and the carbon cycle
explicitly represented in the model, the number of processes
is treated consistently with Chapter 7. The atmosphere-ocean
explicitly simulated and the detail of description. Some basic
module consists of an atmospheric energy balance model coupled
information on the EMICs used in Chapter 10 of this report
to an upwelling-diffusion ocean model. The atmospheric energy
is presented in Table 8.3. A comprehensive description of all
balance model has land and ocean boxes in each hemisphere, and
EMICs in operation can be found in Claussen (2005). Actually,
the upwelling-diffusion ocean model in each hemisphere has 40
there is a broad range of EMICs, reﬂ ecting the differences in
layers with inter-hemispheric heat exchange in the mixed layer.
scope. In some EMICs, the number of processes and the detail
This simple climate model has been tuned to outputs from 19
of description are reduced to simulate feedbacks between as
of the AOGCMs described in Table 8.1, with resulting parameter
many components of the climate system as feasible. Others,
values as given in the Supplementary Material, Table S8.1. The
with fewer interacting components, are utilised in long-term
applied tuning procedure involves an iterative optimisation
ensemble experiments to investigate speciﬁ c aspects of climate
to derive least-square optimal ﬁ ts between the simple model
variability. The gap between some of the most complicated
results and the AOGCM outputs for temperature time series
EMICs and AOGCMs is not very large. In fact, this particular
and net oceanic heat uptake. This procedure attempts to match
class of EMICs is derived from AOGCMs. On the other hand,
not only the global mean temperature but also the hemispheric
EMICs and simple climate models differ much more. For
land and ocean surface temperature changes of the AOGCM
instance, EMICs as well as AOGCMs realistically represent the
results by adjusting the equilibrium land-ocean warming ratio.
large-scale geographical structures of the Earth, like the shape
Where data availability allowed, the tuning procedure took
of continents and ocean basins, which is certainly not the case
simultaneous account of low-pass ﬁ ltered AOGCM data for
for simple climate models.
two scenarios, namely a 1% per year compounded increase in
Since the TAR, EMICs have intensively been used to
atmospheric CO2 concentration to twice and quadruple the pre-
study past and future climate changes (see Chapters 6, 9 and
industrial level, with subsequent stabilisation. Before tuning,
10). Furthermore, a great deal of effort has been devoted
the AOGCM temperature and heat uptake data was de-drifted
to the evaluation of those models through coordinated
by subtracting the respective low-pass ﬁ ltered pre-industrial
intercomparisons.
control run segments. The three tuned parameters in the simple
644

Chapter 8
Climate Models and Their Evaluation
Figure 8.17 compares the results from
some of the EMICs utilised in Chapter
10 (see Table 8.3) with observation-
based estimates and results of GCMs that
took part in AMIP and CMIP1 (Gates
et al., 1999; Lambert and Boer, 2001).
The EMIC results refer to simulations
in which climate is in equilibrium with
an atmospheric CO2 concentration
of 280 ppm. Figures 8.17a and 8.17b
show that the simulated latitudinal
distributions of the zonally averaged
surface air temperature for boreal
winter and boreal summer are in good
-1
-1
agreement with observations, except at
northern and southern high latitudes.
Interestingly, the GCM results also
exhibit a larger scatter in these regions,
and they somewhat deviate from data
there. Figures 8.17c and 8.17d indicate
that EMICs satisfactorily reproduce the
general structure of the observed zonally
averaged precipitation. Here again, at
most latitudes, the scatter in the EMIC
results seems to be as large as the scatter
in the GCM results, and both EMIC and
GCM results agree with observational
estimates. When these EMICs are
allowed to adjust to a doubling of
atmospheric CO2 concentration, they all
Figure 8.17. Latitudinal distributions of the zonally averaged surface air temperature (a, b) and precipitation
simulate an increase in globally averaged
rate (c, d) for boreal winter (DJF) (a, c) and boreal summer (JJA) (b, d) as simulated at equilibrium by some of
annual mean surface temperature and
the EMICs used in Chapter 10 (see Table 8.3) for an atmospheric CO concentration of 280 ppm. In (a) and (b),
2
precipitation that falls largely within the
observational data merged from Jennings (1975), Jones (1988), Schubert et al. (1992), da Silva et al. (1994) and
range of GCM results (Petoukhov et al.,
Fiorino (1997) are shown by crosses. In (c) and (d), observation-based estimates from Jaeger (1976; crosses) and
Xie and Arkin (1997; open circles) are shown. The vertical grey bars indicate the range of GCM results from AMIP
2005).
and CMIP1 (see text). Note that the model versions used in this intercomparison have no interactive biosphere
The responses of the North Atlantic
and ice sheet components. The MIT-UW model is an earlier version of MIT-IGSM2.3. Adapted from Petoukhov et
MOC to increasing atmospheric CO
al., 2005.
2
concentration and idealised freshwater
perturbations as simulated by EMICs have also been compared
bifurcation point, beyond which NADW formation cannot be
to those obtained by AOGCMs (Gregory et al., 2005; Petoukhov
sustained, varies from less than 0.1 Sv to over 0.5 Sv.
et al., 2005; Stouffer et al., 2006). These studies reveal no
A ﬁ nal example of EMIC intercomparison is discussed in
systematic difference in model behaviour, which gives added
Brovkin et al. (2006). Earth System Models of Intermediate
conﬁ dence to the use of EMICs.
Complexity that explicitly simulate the interactions between
In a further intercomparison, Rahmstorf et al. (2005)
atmosphere, ocean and land surface were forced by a
compared results from 11 EMICs in which the North Atlantic
reconstruction of land cover changes during the last millennium.
Ocean was subjected to a slowly varying change in freshwater
In response to historical deforestation of about 18 x 106 km2, all
input. All the models analysed show a characteristic hysteresis
models exhibited a decrease in globally averaged annual mean
response of the North Atlantic MOC to freshwater forcing,
surface temperature in the range of 0.13°C to 0.25°C, mainly
which can be explained by Stommel’s (1961) salt advection
due to the increase in land surface albedo. Further experiments
feedback. The width of the hysteresis curve varies between 0.2
with the models forced by the historical atmospheric CO2 trend
and 0.5 Sv in the models. Major differences are found in the
reveal that, for the whole last millennium, the biogeophysical
location of the present-day climate on the hysteresis diagram.
cooling due to land cover changes is less pronounced than
In seven of the models, the present-day climate for standard
the warming induced by the elevated atmospheric CO2 level
parameter choices is found in the bi-stable regime, while in the
(0.27°C–0.62°C). During the 19th century, the cooling effect of
other four models, this climate is situated in the mono-stable
deforestation appears to counterbalance, albeit not completely,
regime. The proximity of the present-day climate to Stommel’s
the warming effect of increasing CO2 concentration.
645

Climate Models and Their Evaluation
Chapter 8

x
fect
g
), 1.8°
), 0.5°
λ
,
eaver et
ϕ
and
ϕ
echts,
x
, 2002)
esolution is
Ice Sheets
ger
ough translation
ization of the ef
TM, 3-D, 0.75°
1.5°, L20* (Calov
et al., 2005)
TM, 3-D, 10 km
x 10 km, L30
(Huybr
2002)
M, 1-D (
(Cruciﬁ
Ber
M, 2-D (
x 3.6°* (W
al., 2001)
t this component or parametri-
-
) = zonally and vertically
f e
ϕ
L = non-interactive cloudiness;

x, 2005), BT*
tion means tha
esolutions: the horizontal r
chal et al., 1998),
Biospher
ekh et al., 2005),
fusion; MESO = parametr
ovkin et al., 2002),
ovkin et al., 2002),
ovkin et al., 2002)
, 1996), BT*
eaver et al., 2001),
ovkin et al., 2002), BV*
ovkin et al., 2002)
ovkin et al., 2002), BV
ovkin et al., 2002)
ovkin et al., 2002), BV
ovkin et al., 2002)
ophic model; 1-D (
BO (Mar
BT (Sitch et al., 2003;
Gerber et al., 2003),
BV (Sitch et al., 2003;
Gerber et al., 2003)
BO (Br
BT (Br
BV (Br
BO* (Six and Maier
Reimer
(Br
(Br
BO (Mouchet and
François, 1996), BT
(Br
(Br
BO (Par
BT (Felzer et al., 2005),
BV* (Felzer et al., 2005)
BO* (Cruciﬁ
(Br
(Br
BO (W
BT (Cox, 2001), BV (Cox,
2001)
e
ehensive radiation scheme; NC
, 1999)
ents; horizontal and vertical r
ds and Marsh,
, CSM (Bonan
, CSM, RIV
, CSM, RIV
, BSM, RIV
, BSM (Gallée
, CSM, RIV
ees latitude x longitude or as spectral truncation with a r
ee surface; ISO = isopycnal dif
Land Surface
, NSM
, NSM, RIV
l model; QG = quasi-geostr
NST
(Schmittner and
Stocker
NST
(Edwar
2005)
1-LST
(Petoukhov et al.,
2000)
1-LST
(Petoukhov et al.,
2000)
1-LST
(Opsteegh et al.,
1998)
10-LST
et al., 2002)
1-LST
et al., 1991)
1-LST
(Meissner et al., 2003)
.

An asterisk after a component or parametriza
x
n-sloping curr
d
eaver
sed either as degr
viscous-plastic rheology; 2-LIT = two-level ice thickness distribution (level ice and leads).
ls.

es
W
W
W
, 1999)
ds and Marsh,
Coupling/Flux
Adjustments
PM, NH, NW
(Stocker et al., 1992;
Schmittner and
Stocker
GM, NH, R
(Edwar
2005)
NM, NH, NW
(Petoukhov et al.,
2000)
AM, NH, R
(Montoya et al., 2005)
NM, NH, R
(Driesschaert., 2005)
AM, GH, GW
(Sokolov et al., 2005)
NM, NH, NW (Cruciﬁ
et al., 2002)
AM, NH, NW (W
et al., 2001)
esolution is expr
ee-dimensional; RL = rigid lid; FS = fr
c
, 1993)
greed by all modelling groups involved
ds and

x et al.,
Sea Ice
,
2-LIT
,
DOC, 2-LIT
,
DOC, 2-LIT
,
R, 2-
,
R, 2-
,
2-LIT
,
PD, 2-LIT
,
R, 2-LIT
e m is the number of vertical leve
ents; R = viscous-plastic or elastic-
T
right and
T
T
T
T
T
T
T
eaver et al.,
0-L
(W
Stocker
0-L
(Edwar
Marsh, 2005)
0-L
(Petoukhov et
al., 2000)
2-L
LIT (Fichefet
and Morales
Maqueda, 1997)
3-L
LIT (Fichefet
and Morales
Maqueda, 1997)
3-L
(Winton, 2000)
0-L
(Cruciﬁ
2002)
0-L
(W
2001)
ee-dimensional; LRAD = linearized radiation scheme; CRAD = compr
esolutions: the horizontal r
e m is the number of vertical levels.
e scheme; DC = parametrization of density-driven dow
, 1992)
ds and
3.75°,
b
x
, z) = zonally averaged; 3-D = thr
ϕ
essed as ‘Lm’, wher
3°, L30
e gradient,
e gradient,
e gradient,
eaver et al.,
e balance model including some dynamics; SD = statistical-dynamica
right and
x
Ocean
essur
essur
, 1992)
essur
, z), 3 basins, RL,
, z), 3 basins, RL,
, z), 3 basins, RL,
3.6° (W
essed as ‘Lm’, wher
ϕ
10°, L8 (Edwar
ϕ
4°, L15 (Marshall et
ϕ
x
gy-moistur
right and Stocker
x
x
esolution is expr
pter 10.
FG with parametrized
zonal pr
2-D (
ISO, MESO, 7.5°x15°, L14
(W
FG, 3-D, RL, ISO, MESO,
5°
Marsh, 2005)
FG with parametrized
zonal pr
2-D (
2.5°, L21 (W
Stocker
PE, 3-D, FS, ISO, MESO,
TCS, DC*, 3.75°
L24 (Montoya et al., 2005)
PE, 3-D, FS, ISO, MESO,
TCS, DC, 3°
(Goosse and Fichefet,
1999)
PE, 3-D, FS, ISO, MESO,
4°
al., 1997)
PE with parametrized
zonal pr
2-D (
DC, 5°, L15 (Hovine and
Fichefet, 1994)
PE, 3-D, RG, ISO, MESO,
1.8°
2001)

The naming convention for the models is as a

, z) = zonally averaged; 3-D = thr
escribed drift; DOC = drift with oceanic curr
),
ϕ
a
λ
pter 10.
e

,
esolution is expr
), NCL,
)
, NCL,
ds and
ϕ
ϕ
,
λ
, 1999)
(
ϕ
3.6°
51°, L10
5.6°), L3
, z), CRAD,
x
x
x
, z), CRAD,
ϕ
ϕ
15° (Schmittner
22.5°, L10
Atmospher
x
10° (Edwar
x
x
eaver et al., 2001)
EMBM, 1-D (
7.5°
and Stocker
EMBM, 2-D
5°
Marsh, 2005)
SD, 3-D, CRAD,
ICL, 10°
(Petoukhov et al.,
2000)
SD, 3-D, CRAD, ICL,
7.5°
(Petoukhov et al.,
2000)
QG, 3-D, LRAD, NCL,
T21 (5.6°
(Opsteegh et al.,
1998)
SD, 2-D (
ICL, 4°, L11 (Sokolov
and Stone, 1998),
CHEM* (Mayer et al.,
2000)
QG, 2-D (
NCL, 5°, L2 (Gallée et
al., 1991)
DEMBM, 2-D (
NCL, 1.8°
(W
e balance model; DEMBM = ener
ophic model; PE = primitive equation model; 2-D (
) = vertically averaged; 2-D(
ees latitude x longitude; the vertical r
ted in the experiments discussed in Cha
λ
,
gy-moistur
ϕ
Description of the EMICs used in Cha
Name
ees latitude x longitude; the vertical r
ds and Marsh,
-IGSM2.3

x et al., 2002)
essed as degr
T = n-layer thermodynamic scheme; PD = pr
eaver et al., 2001)
EMBM = ener
averaged; 2-D(
ICL = interactive cloudiness; CHEM = chemistry module; horizontal and vertical r
to degr
FG = frictional geostr
of mesoscale eddies on tracer distribution; TCS = complex turbulence closur
expr
n-L
able 8.3.
tion was not activa
E1: BERN2.5CC
(Plattner et al., 2001;
Joos et al., 2001)
E2: C-GOLDSTEIN
(Edwar
2005)
E3: CLIMBER-2
(Petoukhov et al.,
2000)
E4: CLIMBER-3a
(Montoya et al., 2005)
E5: LOVECLIM
(Driesschaert, 2005)
E6: MIT
(Sokolov et al., 2005)
E7: MOBIDIC
(Cruciﬁ
E8: UVIC
(W

T
za

Notes:
a
b
c
646

Chapter 8
Climate Models and Their Evaluation
e; RIV =
ux
adjustment.
ux
eshwater ﬂ
ee-dimensional; horizontal and vertical
e; CSM = complex model for soil moistur
e m is the number of vertical levels.

ux adjustment; NW = no fr
essed as ‘Lm’, wher
) = vertically averaged; 3-D = thr
e computed and added to climatological data; NM = no momentum ﬂ
λ
,
eshwater ﬂ
ϕ
ol run ar
BSM = bucket model for soil moistur

le; 2-D (
esolution is expr
egional fr
oﬁ
= r
ical r
W
e storage in soil;
es; the vert
elative to the contr

ux adjustment; R
es x kilometr

ux anomalies r
eshwater ﬂ
e scheme; NSM = no moistur
) = vertically averaged with east-west parabolic pr
ϕ
ees latitude x longitude or kilometr

ux adjustment; GW = global fr

ux adjustment; AM = momentum ﬂ
estrial carbon dynamics; BV = dynamical vegetation model.
e; n-LST = n-layer soil temperatur
essed either as degr

ux; GM = global momentum ﬂ

ux adjustment; NH = no heat ﬂ
esolution is expr
escribed momentum ﬂ
outing scheme.
PM = pr
adjustment; GH = global heat ﬂ
NST = no explicit computation of soil temperatur
river r
BO = model of oceanic carbon dynamics; BT = model of terr
TM = thermomechanical model; M = mechanical model (isothermal); 1-D (
r
esolutions: the horizontal r

Notes (continued):
d
e
f
g
647

Climate Models and Their Evaluation
Chapter 8
References
Arakawa, A., and W.H. Schubert, 1974: Interaction of a cumulus cloud
ensemble with the large-scale environment, Part I. J. Atmos. Sci., 31,
674–701.
Abramopoulos, F., C. Rosenzweig, and B. Choudhury, 1988: Improved
Arora, V.K., 2001: Assessment of simulated water balance for continental-
ground hydrology calculations for global climate models (GCMs): Soil
scale river basins in an AMIP 2 simulation. J. Geophys. Res., 106,
water movement and evapotranspiration. J. Clim., 1, 921–941.
14827–14842.
Achatz, U., and J.D. Opsteegh, 2003: Primitive-equation-based low-order
Arora, V.K., and G.J. Boer, 2003: A representation of variable root
models with seasonal cycle, Part II: Application to complexity and
distribution in dynamic vegetation models. Earth Interactions, 7, 1–19.
nonlinearity of large-scale atmospheric dynamics. J. Atmos. Sci., 60,
Arzel, O., T. Fichefet, and H. Goosse, 2006: Sea ice evolution over the
478–490.
20th and 21st centuries as simulated by the current AOGCMs. Ocean
AchutaRao, K., and K.R. Sperber, 2002: Simulation of the El Niño
Modelling, 12, 401–415.
Southern Oscillation: Results from the coupled model intercomparison
Babko, O., D.A. Rothrock, and G.A. Maykut, 2002: Role of rafting in the
project. Clim. Dyn., 19, 191–209.
mechanical redistribution of sea ice thickness. J. Geophys. Res., 107,
AchutaRao, K., and K.R. Sperber, 2006: ENSO simulation in coupled
3113, doi:10.1029/1999JC000190.
ocean-atmosphere models: Are the current models better? Clim. Dyn.,
Baldwin, M.P., et al., 2001: The quasi-biennial oscillation. Rev. Geophys.,
27, 1–15.
39, 179–229.
AchutaRao, K., et al., 2004: An Appraisal of Coupled Climate Model
Baldwin, M.P., et al., 2003: Stratospheric memory and skill of extended-
Simulations. UCRL-TR-202550, Lawrence Livermore National
range weather forecasts. Science, 301, 636–640.
Laboratory, Livermore, CA, 197 pp.
Balmaseda, M.A., M.K. Davey, and D.L.T. Anderson, 1995: Decadal and
Alexander, M.A., et al., 2004: The atmospheric response to realistic Arctic
seasonal dependence of ENSO prediction skill. J. Clim., 8, 2705–2715.
sea ice anomalies in an AGCM during winter. J. Clim., 17, 890–905.
Barkstrom, B., et al., 1989: Earth Radiation Budget Experiment (ERBE)
Alexeev, V.A., 2003: Sensitivity to CO
archival and April 1985 results. Bull. Am. Meteorol. Soc., 70, 1254–
2 doubling of an atmospheric GCM
coupled to an oceanic mixed layer: a linear analysis. Clim. Dyn., 20,
1262.
775–787.
Barnett, T.P., et al., 1999: Origins of midlatitude Paciﬁ c decadal variability.
Alexeev, V.A., P.L. Langen, and J.R. Bates, 2005: Polar ampliﬁ cation
Geophys. Res. Lett., 26, 1453–1456.
of surface warming on an aquaplanet in “ghost forcing” experiments
Bates, J.J., and D.L. Jackson, 2001: Trends in upper-tropospheric humidity.
without sea ice feedbacks. Clim. Dyn., 24, 655–666.
Geophys. Res. Lett., 28, 1695–1698.
Alexeev, V.A., et al., 1998: Modelling of the present-day climate by the
Bauer, M., and A.D. Del Genio, 2006: Composite analysis of winter
INM RAS atmospheric model “DNM GCM”. Institute of Numerical
cyclones in a GCM: Inﬂ uence on climatological humidity. J. Clim., 19,
Mathematics, Moscow, Russia, 200 pp.
1652–1672. .
Allan, R.P., and A. Slingo, 2002: Can current climate forcings explain the
Bauer, M., A.D. Del Genio, and J.R. Lanzante, 2002: Observed and
spatial and temporal signatures of decadal OLR variations? Geophys.
simulated temperature humidity relationships: sensitivity to sampling
Res. Lett., 29(7), 1141, doi:10.1029/2001GL014620.
and analysis. J. Clim., 15, 203–215.
Allan, R.P., V. Ramaswamy, and A. Slingo, 2002: A diagnostic analysis
Bell, T.L., M.-D. Chou, R.S. Lindzen, and A.Y. Hou, 2002: Comments
of atmospheric moisture and clear-sky radiative feedback in the Hadley
on “Does the Earth have an adaptive infrared iris?” Reply. Bull. Am.
Centre and Geophysical Fluid Dynamics Laboratory (GFDL) climate
Meteorol. Soc., 83, 598–600.
models. J. Geophys. Res., 107(D17), 4329, doi:10.1029/2001JD001131.
Bengtsson, L.K., I. Hodges, and E. Roeckner, 2006: Storm tracks and
Allan, R.P., M.A. Ringer, and A. Slingo, 2003: Evaluation of moisture
climate change. J. Clim., 19, 3518–3543.
in the Hadley Centre Climate Model using simulations of HIRS water
Bernie, D., S.J. Woolnough, J.M. Slingo, and E. Guilyardi, 2005: Modelling
vapour channel radiances. Q. J. R. Meteorol. Soc., 129, 3371–3389.
diurnal and intraseasonal variability of the ocean mixed layer. J. Clim.,
Allan, R.P., M.A. Ringer, J.A. Pamment, and A. Slingo, 2004: Simulation
15, 1190–1202.
of the Earth’s radiation budget by the European Centre for Medium
Bitz, C.M., and W.H. Lipscomb, 1999: An energy-conserving
Range Weather Forecasts 40-year Reanalysis (ERA40). J. Geophys. Res.,
thermodynamic sea ice model for climate study. J. Geophys. Res., 104,
109, D18107, doi:10.1029/2004JD004816.
15669–15677.
Allen, M.R., and W.J. Ingram, 2002: Constraints on future changes in
Bitz, C.M., G. Flato, and J. Fyfe, 2002: Sea ice response to wind forcing
climate and the hydrologic cycle. Nature, 419, 224–231.
from AMIP models. J. Clim., 15, 523–535.
Alley, R.B., et al., 2002: Abrupt Climate Changes: Inevitable Surprises.
Bitz, C.M., M.M., Holland, A.J. Weaver, and M. Eby, 2001: Simulating the
National Research Council, National Academy Press, Washington, DC,
ice-thickness distribution in a coupled climate model. J. Geophys. Res.,
221 pp.
106, 2441–2463.
Alves, O., M.A. Balmaseda, D. Anderson, and T. Stockdale, 2004:
Blankenship, C.B., and T.T Wilheit, 2001: SSM/T-2 measurements of
Sensitivity of dynamical seasonal forecast to ocean initial conditions. Q.
regional changes in three-dimensional water vapour ﬁ elds during ENSO
J. R. Meteorol. Soc., 130, 647–667.
events. J. Geophys. Res., 106, 5239–5254.
Amundrud, T.L., H. Mailing, and R.G. Ingram, 2004: Geometrical
Bleck, R., 2002: An oceanic general circulation model framed in hybrid
constraints on the evolution of ridged sea ice. J. Geophys. Res., 109,
isopycnic-Cartesian coordinates. Ocean Modelling, 4, 55–88.
C06005, doi:10.1029/2003JC002251.
Bleck, R., C. Rooth, D. Hu, and L.T. Smith, 1992: Salinity-driven
Annamalai, H., K. Hamilton, and K.R. Sperber, 2007: South Asian summer
thermocline transients in a wind- and thermohaline-forced isopycnic
monsoon and its relationship with ENSO in the IPCC AR4 simulations.
coordinate model of the North Atlantic. J. Phys. Oceanogr., 22, 1486–
J. Clim., 20, 1071-1083.
1505.
Annan, J.D., J.C. Hargreaves, N.R. Edwards, and R. Marsh, 2005a:
Boer, G.J., and B. Yu, 2003: Climate sensitivity and climate state. Clim.
Parameter estimation in an intermediate complexity Earth System Model
Dyn., 21, 167–176.
using an ensemble Kalman ﬁ lter. Ocean Modelling, 8, 135–154.
Bonan, G.B., 1998: The land surface climatology of the NCAR land
Annan, J.D., et al., 2005b: Efﬁ ciently constraining climate sensitivity with
surface model (LSM 1.0) coupled to the NCAR Community Climate
palaeoclimate observations. Scientiﬁ c Online Letters on the Atmosphere,
Model (CCM3). J. Clim., 11, 1307–1326.
1, 181–184.
Bonan, G.B., K.W. Oleson, M. Vertenstein, and S. Levis, 2002: The land
Arakawa, A., 2004: The cumulus parameterization problem: Past, present,
surface climatology of the Community Land Model coupled to the NCAR
and future. J. Clim., 17, 2493–2525.
Community Climate Model. J. Clim., 15, 3123–3149.
648

Chapter 8
Climate Models and Their Evaluation
Böning, C.W., et al., 1995: An overlooked problem in model simulations
Cai, W.J., and P.H. Whetton, 2000: Evidence for a time-varying pattern of
of the thermohaline circulation and heat transports in the Atlantic Ocean.
greenhouse warming in the Paciﬁ c Ocean. Geophys. Res. Lett., 27(16),
J. Clim., 8, 515–523.
2577–2580.
Bony, S., and K.A. Emanuel, 2001: A parameterization of the cloudiness
Cai, W.J., P.H. Whetton, and D.J. Karoly, 2003: The response of the
associated with cumulus convection: Evaluation using TOGA COARE
Antarctic Oscillation to increasing and stabilized atmospheric CO2. J.
data. J. Atmos. Sci., 58, 3158–3183.
Clim., 16, 1525–1538.
Bony, S., and J.-L. Dufresne, 2005: Marine boundary-layer clouds at the
Calov, R., et al., 2002: Large-scale instabilities of the Laurentide ice sheet
heart of tropical cloud feedback uncertainties in climate models. Geophys.
simulated in a fully coupled climate-system model. Geophys. Res. Lett.,
Res. Lett., 32(20), L20806, doi:10.1029/2005GL023851.
29(24), 2216, doi:10.1029/2002GL016078.
Bony, S., and K.A. Emanuel, 2005: On the role of moist processes
Calov, R., et al., 2005: Transient simulation of the last glacial inception.
in tropical intraseasonal variability: cloud-radiation and moisture-
Part I: Glacial inception as a bifurcation of the climate system. Clim.
convection feedbacks. J. Atmos. Sci., 62, 2770–2789.
Dyn., 24(6), 545–561.
Bony, S., K.-M. Lau, and Y.C. Sud, 1997: Sea surface temperature and
Camargo, S., A.G. Barnston, and S.E. Zebiak, 2005: A statistical assessment
large-scale circulation inﬂ uences on tropical greenhouse effect and cloud
of tropical cyclone activity in atmospheric general circulation models.
radiative forcing. J. Clim., 10, 2055–2077.
Tellus, 57A, 589–604.
Bony, S., et al., 2004: On dynamic and thermodynamic components of
Carnell, R., and C. Senior, 1998: Changes in mid-latitude variability due
cloud changes. Clim. Dyn., 22, 71–86.
to increasing greenhouse gases and sulphate aerosols. Clim. Dyn., 14,
Bony, S., et al., 2006: How well do we understand and evaluate climate
369–383.
change feedback processes? J. Clim., 19, 3445–3482.
Cassou, C., L. Terray, J.W. Hurrell, and C. Deser, 2004: North Atlantic
Boone, A., V. Masson, T. Meyers, and J. Noilhan, 2000: The inﬂ uence of the
winter climate regimes: Spatial asymmetry, stationarity with time, and
inclusion of soil freezing on simulations by a soil-vegetation-atmosphere
oceanic forcing. J. Clim., 17, 1055–1068.
transfer scheme. J. Appl. Meteorol., 39(9), 1544–1569.
Castanheira, J.M., and H.-F. Graf, 2003: North Paciﬁ c–North Atlantic
Boone, A., et al., 2004: The Rhone-Aggregation land surface scheme
relationships under stratospheric control? J. Geophys. Res., 108, 4036,
intercomparison project: An overview. J. Clim., 17, 187–208.
doi:10.1029/2002JD002754.
Boville, B.A., and W.J. Randel, 1992: Equatorial waves in a stratospheric
Cattle, H., and J. Crossley, 1995: Modelling Arctic climate change. Philos.
GCM: Effects of resolution. J. Atmos. Sci., 49, 785–801.
Trans. R. Soc. London Ser. A, 352, 201–213.
Bowling, L.C., et al., 2003: Simulation of high latitude hydrological
Cess, R.D., 1975: Global climate change: an investigation of atmospheric
processes in the Torne-Kalix basin: PILPS Phase 2(e) 1: Experiment
feedback mechanisms. Tellus, 27, 193–198.
description and summary intercomparisons. Global Planet. Change, 38,
Cess, R.D., et al., 1989: Interpretation of cloud-climate feedback as
1–30.
produced by 14 atmospheric general circulation models. Science, 245,
Boyle, J.S., et al., 2005: Diagnosis of Community Atmospheric Model
513–516.
2 (CAM2) in numerical weather forecast conﬁ guration at Atmospheric
Chambers, L.H., B. Lin, and D.F. Young, 2002: Examination of new CERES
Radiation Measurement (ARM) sites. J. Geophys. Res., 110, doi:10.1029/
data for evidence of tropical Iris feedback. J. Clim., 15, 3719–3726.
2004JD005042.
Chang, F.-L., and Z. Li, 2005: A comparison of the global surveys of
Branstetter, M.L., 2001: Development of a Parallel River Transport
high, mid and low clouds from satellite and general circulation models.
Algorithm and Application to Climate Studies. PhD Dissertation,
In: Proceedings of the Fifteenth Atmospheric Radiation Measurement
University of Texas, Austin, TX.
(ARM) Science Team Meeting, Daytona Beach, Florida, 14–18 March
Briegleb, B.P., et al., 2004: Scientiﬁ c Description of the Sea Ice Component
2005. Atmospheric Radiation Measurement Program, US Department of
in the Community Climate System Model, Version Three. Technical Note
Energy, Washington, DC, http://www.arm.gov/publications/proceedings/
TN-463STR, NTIS #PB2004-106574, National Center for Atmospheric
conf15/
Research, Boulder, CO, 75 pp.
Chapman, W.L., and J. E. Walsh, 2007: Simulations of arctic temperature
Broccoli, A.J., N.-C. Lau, and M.J. Nath, 1998: The cold ocean-warm land
and pressure by global coupled models. J. Clim., 20, 609-632.
pattern: Model simulation and relevance to climate change detection. J.
Chen, D., S.E. Zebiak, A.J. Busalacchi, and M.A. Cane, 1995: An improved
Clim., 11, 2743–2763.
procedure for El Niño forecasting. Science, 269, 1699–1702.
Broecker, W.S., 1997: Thermohaline circulation, the Achilles heel of our
Chen, J., B.E. Carlson, and A.D. Del Genio, 2002: Evidence for
climate system: will man-made CO2 upset the current balance? Science,
strengthening of the tropical general circulation in the 1990s. Science,
278, 1582–1588.
295, 838–841.
Brogniez, H., R. Roca, and L. Picon, 2005: Evaluation of the distribution
Chen, T.-C., and J.-H. Yoon, 2002: Interdecadal variation of the North
of subtropical free tropospheric humidity in AMIP-2 simulations using
Paciﬁ c wintertime blocking. Mon. Weather Rev., 130, 3136–3143.
METEOSAT water vapour channel data. Geophys. Res. Lett., 32, L19708,
Chin, M., et al., 2002: Tropospheric aerosol optical thickness from
doi:10.1029/2005GL024341.
GOCART model and comparisons with satellite and sun photometer
Brovkin, V., et al., 2002: Carbon cycle, vegetation and climate dynamics
measurements. J. Atmos. Sci., 59, 461–483.
in the Holocene: Experiments with the CLIMBER-2 model. Global
Chou, M.-D., R.S. Lindzen, and A.Y. Hou, 2002: Reply to: “Tropical cirrus
Biogeochem. Cycles, 16(4), 1139, doi:10.1029/2001GB001662.
and water vapor: An effective Earth infrared iris feedback?”. Atmos.
Brovkin, V., et al., 2006: Biogeophysical effects of historical land cover
Chem. Phys., 2, 99–101.
changes simulated by six Earth system models of intermediate complexity.
Chung, E.S., B.J. Sohn, and V. Ramanathan, 2004: Moistening processes in
Clim. Dyn., 26, 587–600, doi:10.1007/s00382-005-0092-6.
the upper troposphere by deep convection: a case study over the tropical
Bryan, F.O., et al., 2006: Response of the North Atlantic thermohaline
Indian Ocean. J. Meteorol. Soc. Japan, 82, 959–965.
circulation and ventilation to increasing carbon dioxide in CCSM3. J.
Church, J.A., et al., 2001: Changes in sea level. In: Climate Change 2001:
Clim., 19, 2382–2397.
The Scientiﬁ c Basis. Contribution of Working Group I to the Third
Burke, E.J., S.J. Brown, and N. Christidis, 2006: Modelling the recent
Assessment Report of the Intergovernmental Panel on Climate Change
evolution of global drought and projections for the 21st century with the
[Houghton, J.T., et al. (eds.)]. Cambridge University Press, Cambridge,
Hadley Centre climate model. J. Hydrometeorol., 7, 1113–1125.
United Kingdom and New York, NY, USA, pp. 663–693.
Cai, M., 2005: Dynamical ampliﬁ cation of polar warming. Geophys. Res.
Clark, P.U., N.G. Pisias, T.F. Stocker, and A.J. Weaver, 2002: The role
Lett., 32, L22710, doi:10.1029/2005GL024481.
of the thermohaline circulation in abrupt climate change. Nature, 415,
863–869.
649

Climate Models and Their Evaluation
Chapter 8
Claussen, M., 1998: On multiple solutions of the atmosphere-vegetation
Dai, A., K.E. Trenberth, and T. Qian, 2004: A global data set of Palmer
system in present-day climate. Global Change Biol., 4, 549–559.
Drought Severity Index for 1870-2002: Relationship with soil moisture
Claussen, M., 2005: Table of EMICs (Earth System Models of Intermediate
and effects of surface warming. J. Hydrometeorol., 5, 1117–1130. PDSI
Complexity). PIK Report 98, Potsdam-Institut für Klimafolgenforschung,
data: http://www.cgd.ucar.edu/cas/catalog/climind/pdsi.html.
Potsdam, Germany, 55 pp, http://www.pik-potsdam.de/emics.
Danabasoglu, G., J.C. McWilliams, and P.R. Gent, 1995: The role of
Claussen, M., et al., 2002: Earth system models of intermediate complexity:
mesoscale tracer transports in the global ocean circulation. Science, 264,
closing the gap in the spectrum of climate system models. Clim. Dyn., 18,
1123–1126.
579–586.
D’Andrea, F., et al., 1998: Northern Hemisphere atmospheric blocking as
Collins, M., S.F.B. Tett, and C. Cooper, 2001: The internal climate
simulated by 15 atmospheric general circulation models in the period
variability of HadCM3, a version of the Hadley Centre coupled model
1979–1988. Clim. Dyn., 14(6), 385–407.
without ﬂ ux adjustments. Clim. Dyn., 17, 61–81.
Dargaville, R.J., et al., 2002: Evaluation of terrestrial carbon cycle models
Collins, M., D. Frame, B. Sinha, and C. Wilson, 2002: How far ahead could
with atmospheric CO2 measurements: Results from transient simulations
we predict El Niño? Geophys. Res. Lett., 29(10), 1492, doi:10.1029/
considering increasing CO2, climate, and land-use effects. Global
2001GL013919.
Biogeochem. Cycles, 16, 1092, doi:10.1029/2001GB001426.
Collins, W.D., et al., 2004: Description of the NCAR Community
Davey, M., et al., 2002: STOIC: A study of coupled GCM climatology
Atmosphere Model (CAM3.0). Technical Note TN-464+STR, National
and variability in tropical ocean regions. Clim. Dyn., 18, 403–420,
Center for Atmospheric Research, Boulder, CO, 214 pp.
doi:10.1007/s00382-001-0188-6.
Collins, W.D., et al., 2006: The Community Climate System Model:
Del Genio, A.D., and A.B. Wolf, 2000: The temperature dependence of the
CCSM3. J. Clim., 19, 2122–2143.
liquid water path of low clouds in the southern great plains. J. Clim., 13,
Colman, R.A., 2001: On the vertical extent of atmospheric feedbacks.
3465 3486.
Clim. Dyn., 17, 391–405.
Del Genio, A.D., and W. Kovari, 2002: Climatic properties of tropical
Colman, R.A., 2003a: A comparison of climate feedbacks in general
precipitating convection under varying environmental conditions. J.
circulation models. Clim. Dyn., 20, 865–873.
Clim., 15, 2597–2615.
Colman, R.A., 2003b: Seasonal contributions to climate feedbacks. Clim.
Del Genio, A.D., A. Wolf, and M.-S. Yao, 2005a: Evaluation of regional
Dyn., 20, 825–841.
cloud feedbacks using single-column models. J. Geophys. Res., 110,
Colman, R.A., 2004: On the structure of water vapour feedbacks in climate
D15S13, doi:10.1029/2004JD005011.
models. Geophys. Res. Lett., 31, L21109, doi:10.1029/2004GL020708.
Del Genio, A.D., W. Kovari, M.-S. Yao, and J. Jonas, 2005b: Cumulus
Cook, K.H., and E.K. Vizy, 2006: Coupled model simulations of the West
microphysics and climate sensitivity. J. Clim., 18, 2376–2387,
African monsoon system: 20th century simulations and 21st century
doi:10.1175/JCLI3413.1.
predictions. J. Clim., 19, 3681–3703.
Delire, C., J.A. Foley, and S. Thompson, 2003: Evaluating the carbon cycle
Cox, P., 2001: Description of the “TRIFFID” Dynamic Global
of a coupled atmosphere–biosphere model. Global Biogeochem. Cycles,
Vegetation Model. Technical Note 24, Hadley Centre, United Kingdom
17, 1012, doi:10.1029/2002GB001870.
Meteorological Ofﬁ ce, Bracknell, UK.
Delworth, T.L., and M.E. Mann, 2000: Observed and simulated multidecadal
Cox, P.M., et al., 1999: The impact of new land surface physics on the GCM
variability in the Northern Hemisphere. Clim. Dyn., 16(9), 661–676.
simulation of climate and climate sensitivity. Clim. Dyn., 15, 183–203.
Delworth, T., S. Manabe, and R.J. Stouffer, 1993: Interdecadal variations
Cox, P.M., et al., 2000: Acceleration of global warming due to carbon-
of the thermohaline circulation in a coupled ocean-atmosphere model. J.
cycle feedbacks in a coupled climate model. Nature, 408, 184–187.
Clim., 6, 1993–2011.
Cox, P.M., et al., 2004: Amazonian forest dieback under climate-carbon
Delworth, T.L., V. Ramaswamy, and G.L. Stenchikov, 2005: The impact
cycle projections for the 21st century. Theor. Appl. Climatol., 78, 137–
of aerosols on simulated ocean temperature and heat content in the 20th
156, doi:10.1007/s00704-004-0049-4.
century. Geophys. Res. Lett., 32, L24709, doi:10.1029/2005GL024457.
Cramer, W., et al., 2001: Global response of terrestrial ecosystem structure
Delworth, T., et al., 2006: GFDL’s CM2 global coupled climate models
and function to CO2 and climate change: results from six dynamic global
– Part 1: Formulation and simulation characteristics. J. Clim., 19, 643–
vegetation models. Global Change Biol., 7, 357–373.
674.
Cruciﬁ x, M., 2005: Carbon isotopes in the glacial ocean: A model study.
Déqué, M., C. Dreveton, A. Braun, and D. Cariolle, 1994: The ARPEGE/
Paleoceanography, 20, PA4020, doi:10.1029/2005PA001131.
IFS atmosphere model: A contribution to the French community climate
Cruciﬁ x, M., and A. Berger, 2002: Simulation of ocean–ice sheet
modeling. Clim. Dyn., 10, 249–266.
interactions during the last deglaciation. Paleoceanography, 17(4), 1054,
Derber, J., and A. Rosati, 1989: A global oceanic data assimilation system.
doi:10.1029/2001PA000702.
J. Phys. Oceanogr., 19(9), 1333–1347.
Cruciﬁ x, M., et al., 2002: Climate evolution during the Holocene: A study
Deser, C., A.S. Phillips, and J.W. Hurrell, 2004: Paciﬁ c interdecadal climate
with an Earth system model of intermediate complexity. Clim. Dyn., 19,
variability: Linkages between the tropics and North Paciﬁ c during boreal
43–60, doi:10.10007/s00382-001-0208-6.
winter since 1900. J. Clim., 17, 3109–3124.
CSMD (Climate System Modeling Division), 2005: An introduction to the
Dessler, A.E., and S.C. Sherwood, 2000: Simulations of tropical upper
ﬁ rst general operational climate model at the National Climate Center.
tropospheric humidity. J. Geophys. Res., 105, 20155–20163.
Advances in Climate System Modeling, 1, National Climate Center,
Diansky, N.A., and E.M. Volodin, 2002: Simulation of the present-day
China Meteorological Administration, 14 pp (in English and Chinese).
climate with a coupled atmosphere-ocean general circulation model. Izv.
Cubasch, U., et al., 2001: Projections of future climate changes. In: Climate
Atmos. Ocean. Phys., 38, 732–747 (English translation).
Change 2001: The Scientiﬁ c Basis. Contribution of Working Group I to
Diansky, N.A., A.V. Bagno, and V.B. Zalesny, 2002: Sigma model of
the Third Assessment Report of the Intergovernmental Panel on Climate
global ocean circulation and its sensitivity to variations in wind stress.
Change [Houghton, J.T., et al. (eds.)]. Cambridge University Press,
Izv. Atmos. Ocean. Phys., 38, 477–494 (English translation).
Cambridge, United Kingdom and New York, NY, USA, pp. 525–582.
Dirmeyer, P.A., 2001: An evaluation of the strength of land-atmosphere
da Silva, A.M., C.C. Young, and S. Levitus, 1994: Atlas of Surface
coupling. J. Hydrometeorol., 2, 329–344.
Marine Data 1994, NOAA Atlas NESDIS 6. NOAA/NESDIS E/OC21 (6
Dong, M., et al., 2000: Developments and implications of the atmospheric
Volumes). US Department of Commerce, National Oceanographic Data
general circulation model. In: Investigations on the Model System of
Center, User Services Branch, Washington, DC.
the Short-Term Climate Predictions [Ding, Y., et al. (eds.)]. China
Dai, A., 2006: Precipitation characteristics in eighteen coupled climate
Meteorological Press, Beijing, China, pp. 63–69 (in Chinese).
models. J. Clim., 19, 4605–4630.
650

Chapter 8
Climate Models and Their Evaluation
Doutriaux-Boucher, M., and J. Quaas, 2004: Evaluation of cloud
Fiorino, M., 1997: PCMDI IPCC ’95 AMIP Analysis: Observations used in
thermodynamic phase parametrizations in the LMDZ GCM by using
the analysis. PCMDI Web. Rep., Program for Climate Model Diagnosis
POLDER satellite data. Geophys. Res. Lett., 31, L06126, doi:10.1029/
and Intercomparison, Lawrence Livermore National Laboratory,
2003GL019095.
Livermore, CA, http://www-pcmdi.llnl.gov/obs/ipcc/ipcc.obs.dat.htm.
Douville, H., 2001: Inﬂ uence of soil moisture on the Asian and African
Flato, G.M., 2004: Sea-ice and its response to CO2 forcing as simulated by
Monsoons. Part II: interannual variability. J. Clim., 15, 701–720.
global climate models. Clim. Dyn., 23, 229–241, doi:10.1007/s00382-
Douville, H., J.-F. Royer, and J.-F. Mahfouf, 1995: A new snow
004-0436-7.
parameterization for the Meteo-France climate model. Clim. Dyn., 12,
Flato, G.M., 2005: The Third Generation Coupled Global Climate
21–35.
Model (CGCM3) (and included links to the description of the AGCM3
Drange, H., et al., 2005: Ocean general circulation modelling of the Nordic
atmospheric model). http://www.cccma.bc.ec.gc.ca/models/cgcm3.shtml.
Seas. In: The Nordic Seas: An Integrated Perspective [Drange, H., et
Flato, G.M., and W.D. Hibler, 1992: Modeling pack ice as a cavitating
al. (eds.)]. Geophysical Monograph 158, American Geophysical Union,
ﬂ uid. J. Phys. Oceanogr., 22, 626–651.
Washington, DC, pp. 199–220.
Flato, G.M., and G.J. Boer, 2001: Warming asymmetry in climate change
Driesschaert, E., 2005: Climate Change over the Next Millennia Using
simulations. Geophys. Res. Lett., 28, 195–198.
LOVECLIM, a New Earth System Model Including Polar Ice Sheets. PhD
Flugel, M., P. Chang, and C. Penland, 2004: The role of stochastic forcing
Thesis, Université Catholique de Louvain, Louvain-la-Neuve, Belgium,
in modulating ENSO predictability. J. Clim., 17(16), 3125–3140.
214 pp, http://edoc.bib.ucl.ac.be:81/ETD-db/collection/available/
Folkins, I., K.K. Kelly, and E.M. Weinstock, 2002: A simple explanation
BelnUcetd-10172005-185914/.
of the increase in relative humidity between 11 and 14 km in the tropics.
Ducharne, A., et al., 2003: Development of a high resolution runoff routing
J. Geophys. Res., 107, doi:10.1029/2002JD002185.
model, calibration and application to assess runoff from the LMD GCM.
Folland, C.K., T.K. Palmer, and D.E. Parker, 1986: Sahel rainfall and
J. Hydrol., 280, 207–228.
worldwide sea temperatures. Nature, 320, 602–607.
Dufresne, J.-L., et al., 2002: On the magnitude of positive feedback
Forster, P.M. de F., and K.P. Shine, 2002: Assessing the climate impact of
between future climate change and the carbon cycle. Geophys. Res. Lett.,
trends in stratospheric water vapour. Geophys. Res. Lett., 6, doi:10.1029/
29(10), doi:10.1029/2001GL013777.
2001GL013909.
Dümenil, L., and E. Todini, 1992: A rainfall-runoff scheme for use in
Forster, P.M. de F., and M. Collins, 2004: Quantifying the water vapour
the Hamburg climate model. In: Advances in Theoretical Hydrology:
feedback associated with post-Pinatubo cooling. Clim. Dyn., 23, 207–
A Tribute to James Dooge. European Geophysical Society Series on
214.
Hydrological Sciences, Vol. 1 [O’Kane, J.P. (ed.)]. Elsevier Press,
Forster, P.M. de F., and K.E. Taylor, 2006: Climate forcings and climate
Amsterdam, pp. 129–157.
sensitivities diagnosed from coupled climate model integrations. J.
Durman, C.F., et al., 2001: A comparison of extreme European daily
Clim., 19, 6181–6194.
precipitation simulated by a global model and regional climate model for
Frei, A., J. Miller, and D. Robinson, 2003: Improved simulations of snow
present and future climates. Q. J. R. Meteorol. Soc., 127, 1005–1015.
extent in the second phase of the Atmospheric Model Intercomparison
Edwards, N.R., and R.J. Marsh, 2005: Uncertainties due to transport-
Project (AMIP-2). J. Geophys. Res., 108(D12), 4369, doi:10.1029/
parameter sensitivity in an efﬁ cient 3-D ocean-climate model. Clim.
2002JD003030.
Dyn., 24, 415–433, doi:10.1007/s00382-004-0508-8.
Frei, A., J.A. Miller, R. Brown, and D.A. Robinson, 2005: Snow mass
Emanuel, K.A., and M. Zivkovic-Rothman, 1999: Development and
over North America: observations and results from the second phase
evaluation of a convection scheme for use in climate models. J. Atmos.
of the Atmospheric Model Intercomparison Project (AMIP-2). J.
Sci., 56, 1766–1782.
Hydrometeorol., 6, 681–695.
Emori, S., A. Hasegawa, T. Suzuki, and K. Dairaku, 2005: Validation,
Frich, P., et al., 2002: Observed coherent changes in climatic extremes
parameterization dependence and future projection of daily precipitation
during the second half of the twentieth century. Clim. Res., 19, 193–
simulated with an atmospheric GCM. Geophys. Res. Lett., 32, L06708,
212.
doi:10.1029/2004GL022306.
Friedlingstein, P., et al., 2001: Positive feedback between future climate
Essery, R.H., and J. Pomeroy, 2004: Vegetation and topographic control of
change and the carbon cycle. Geophys. Res. Lett., 28(8), 1543–1546.
wind-blown snow distributions in distributed and aggregated simulations.
Friedlingstein, P., J.-L. Dufresne, P.M Cox, and P. Rayner, 2003: How
J. Hydrometeorol., 5(5), 735–744.
positive is the feedback between climate change and the carbon cycle?
Essery, R., M. Best, and P. Cox, 2001: MOSES 2.2 Technical
Tellus, 55B, 692–700.
Documentation. Hadley Centre Technical Note No. 30, Hadley Centre
Friedlingstein, P., et al., 2006: Climate–carbon cycle feedback analysis,
for Climate Prediction and Research, UK Met Ofﬁ ce, Exeter, UK, http://
results from the C4MIP model intercomparison. J. Clim., 19, 3337–
www.metofﬁ ce.gov.uk/research/hadleycentre/pubs/HCTN/index.html.
3353.
Essery, R.H., J. Pomeroy, J. Parvianen, and P. Storck, 2003: Sublimation of
Friend, A.D., and N.Y. Kiang, 2005: Land surface model development for
snow from boreal forests in a climate model. J. Clim., 16, 1855–1864.
the GISS GCM: Effects of improved canopy physiology on simulated
Etchevers, P., et al., 2004: Validation of the energy budget of an alpine
climate. J. Clim., 18, 2883–2902.
snowpack simulated by several snow models (SnowMIP project). Ann.
Fu, Q., M. Baker, and D.L. Hartmann, 2002: Tropical cirrus and water
Glaciol., 38, 150–158.
vapour: an effective Earth infrared iris? Atmos. Chem. Phys., 2, 31–37.
Farrara, J.D., C.R. Mechoso, and A.W. Robertson, 2000: Ensembles of
Fu, Q., C.M. Johanson, S.G. Warren, and D.J. Seidel, 2004: Contribution
AGCM two-tier predictions and simulations of the circulation anomalies
of stratospheric cooling to satellite-inferred tropospheric temperature
during winter 1997–1998. Mon. Weather Rev., 128, 3589–3604.
trends. Nature, 429, 55–58.
Felzer, B., et al., 2005: Global and future implications of ozone on net
Furevik, T., et al., 2003: Description and evaluation of the Bergen climate
primary production and carbon sequestration using a biogeochemical
model: ARPEGE coupled with MICOM. Clim. Dyn., 21, 27–51.
model. Clim. Change, 73, 345–373.
Fyfe, J.C., G.J. Boer, and G.M. Flato, 1999: The Arctic and Antarctic
Fichefet, T., and M.A. Morales Maqueda, 1997: Sensitivity of a global
Oscillations and their projected changes under global warming. Geophys.
sea ice model to the treatment of ice thermodynamics and dynamics. J.
Res. Lett., 11, 1601–1604.
Geophys. Res., 102, 12609–12646.
Galin, V. Ya., E.M. Volodin, and S.P. Smyshliaev, 2003: Atmospheric
Fichefet, T., et al., 2003: Implications of changes in freshwater ﬂ ux from
general circulation model of INM RAS with ozone dynamics. Russ.
the Greenland ice sheet for the climate of the 21st century. Geophys. Res.
Meteorol. Hydrol., 5, 13–22.
Lett., 30(17), 1911, doi:10.1029/2003GL017826.
651

Climate Models and Their Evaluation
Chapter 8
Gallée, H., et al., 1991: Simulation of the last glacial cycle by a coupled,
Gordon, C., et al., 2000: The simulation of SST, sea ice extents and ocean
sectorally averaged climate–ice sheet model. Part I: The climate model.
heat transports in a version of the Hadley Centre coupled model without
J. Geophys. Res., 96, 13139–13161.
ﬂ ux adjustments. Clim. Dyn., 16, 147–168.
Ganachaud, A., and C. Wunsch, 2000: Improved estimates of global ocean
Gordon, H.B., et al., 2002: The CSIRO Mk3 Climate System Model.
circulation, heat transport and mixing from hydrographic data. Nature,
CSIRO Atmospheric Research Technical Paper No. 60, Commonwealth
408, 453–457.
Scientiﬁ c and Industrial Research Organisation Atmospheric Research,
Ganachaud, A., and C. Wunsch, 2003: Large-scale ocean heat and
Aspendale, Victoria, Australia, 130 pp, http://www.cmar.csiro.au/e-print/
freshwater transports during the World Ocean Circulation Experiment. J.
open/gordon_2002a.pdf.
Clim., 16, 696–705.
Gordon, N.D., J.R. Norris, C.P. Weaver, and S.A. Klein, 2005: Cluster
Gates, W.L., et al., 1999: An overview of the results of the Atmospheric
analysis of cloud regimes and characteristic dynamics of midlatitude
Model Intercomparison Project (AMIP I). Bull. Am. Meteorol. Soc., 80,
synoptic systems in observations and a model. J. Geophys. Res., 110,
29–55.
D15S17, doi:10.1029/2004JD005027.
Geng, Q., and M. Sugi, 2003: Possible change of extratropical cyclone
Govindasamy, B., et al., 2005: Increase of the carbon cycle feedback with
activity due to enhanced greenhouse gases and sulfate aerosols–Study
climate sensitivity: results from a coupled and carbon climate and carbon
with a high-resolution AGCM. J. Clim., 16, 2262–2274.
cycle model. Tellus, 57B, 153–163.
Gent, P.R., 2001: Will the North Atlantic Ocean thermohaline circulation
Graham, R.J., et al., 2005: A performance comparison of coupled and
weaken during the 21st century? Geophys. Res. Lett., 28, 1023–1026.
uncoupled versions of the Met Ofﬁ ce seasonal prediction general
Gent, P.R., J. Willebrand, T.J. McDougall, and J.C. McWilliams, 1995:
circulation model. Tellus, 57A, 320–339.
Parameterizing eddy-induced tracer transports in ocean circulation
Greenwald, T.J., G.L. Stephens, S.A. Christopher, and T.H.V. Haar, 1995:
models. J. Phys. Oceanogr., 25, 463–474.
Observations of the global characteristics and regional radiative effects
Gerber, S., et al., 2003: Constraining temperature variations over the last
of marine cloud liquid water. J. Clim., 8, 2928–2946.
millennium by comparing simulated and observed atmospheric CO2.
Gregory, D., et al., 2000: Revision of convection, radiation and cloud
Clim. Dyn., 20, 281–299.
schemes in the ECMWF Integrated Forecasting System. Q. J. R.
Gerten, D., et al., 2004: Terrestrial vegetation and water balance–
Meteorol. Soc., 126, 1685–1710.
hydrological evaluation of a dynamic global vegetation model. J. Hydrol.,
Gregory, J.M., et al., 2002: An observationally based estimate of the
286, 249–270.
climate sensitivity. J. Clim., 15, 3117–3121.
Gettelman, A., J.R. Holton, and A.R. Douglass, 2000: Simulations of water
Gregory, J.M., et al., 2005: A model intercomparison of changes in the
vapor in the lower stratosphere and upper troposphere. J. Geophys. Res.,
Atlantic thermohaline circulation in response to increasing atmospheric
105, 9003–9023.
CO2 concentration. Geophys. Res. Lett., 32, L12703, doi:10.1029/
GFDL GAMDT (The GFDL Global Atmospheric Model Development
2005GL023209.
Team), 2004: The new GFDL global atmosphere and land model AM2-
Grifﬁ es, S.M., 2004: Fundamentals of Ocean Climate Models. Princeton
LM2: Evaluation with prescribed SST simulations. J. Clim., 17, 4641–
University Press, Princeton, NJ, 496 pp.
4673.
Guilyardi, E., et al., 2004: Representing El Niño in coupled ocean-
Ghan, S.J., R. Easter, J. Hudson, and F.-M. Bréon, 2001a: Evaluation of
atmosphere GCMs: the dominant role of the atmospheric component. J.
aerosol indirect radiative forcing in MIRAGE. J. Geophys. Res., 106,
Clim., 17, 4623–4629.
5317–5334.
Gutowski, W.J., et al., 2004: Diagnosis and attribution of a seasonal
Ghan, S.J., et al., 2001b: Evaluation of aerosol direct radiative forcing in
precipitation deﬁ
cit in a US regional climate simulation. J.
MIRAGE. J. Geophys. Res., 106, 5295–5316.
Hydrometeorol., 5(1), 230–242.
Gillett, N.P., 2005: Northern Hemisphere circulation. Nature, 437, 496.
Hagemann, S., 2002: An Improved Land Surface Parameter Dataset
Gillett, N.P., and D.W.J. Thompson, 2003: Simulation of recent Southern
for Global and Regional Climate Models. Max Planck Institute for
Hemisphere climate change. Science, 302, 273–275.
Meteorology Report 162, MPI for Meteorology, Hamburg, Germany, 21
Giorgetta, M.A., E. Manzini, and E. Roeckner, 2002: Forcing of the
pp.
quasi-biennial oscillation from a broad spectrum of atmospheric waves.
Hagemann, S., and L. Dümenil-Gates, 2001: Validation of the hydrological
Geophys. Res. Lett., 29, 1245, doi:10.1029/2002GL014756.
cycle of ECMWF and NCEP reanalyses using the MPI hydrological
Giorgetta M.A., et al., 2006: Climatology and forcing of the quasi-biennial
discharge model. J. Geophys. Res., 106, 1503–1510.
oscillation in the MAECHAM5 model. J. Clim., 19, 3882–3901.
Hall, A., 2004: The role of surface albedo feedback in climate. J. Clim.,
Gleckler, P.J., K.R. Sperber, and K. AchutaRao, 2006a: The annual cycle
17, 1550–1568.
of global ocean heat content: observed and simulated. J. Geophys. Res.,
Hall, A., and S. Manabe, 1999: The role of water vapour feedback in
111, C06008, doi:10.1029/2005JC003223.
unperturbed climate variability and global warming. J. Clim., 12, 2327–
Gleckler, P.J., et al., 2006b: Krakatoa’s signature persists in the ocean.
2346.
Nature, 439, 675, doi:10.1038/439675a.
Hall, A., and R.J. Stouffer, 2001: An abrupt climate event in a coupled
Gnanadesikan, A., et al., 2004: GFDL’s CM2 global coupled climate
ocean-atmosphere simulation without external forcing. Nature,
models–Part 2: The baseline ocean simulation. J. Clim., 19, 675–697.
409(6817), 171–174.
Goldenberg, S.B., C.W. Landsea, A.M. Mestas-Nunez, and W.M. Gray,
Hall, A., and M. Visbeck, 2002: Synchronous variability in the Southern
2001: The recent increase in Atlantic hurricane activity: Causes and
Hemisphere atmosphere, sea ice and ocean resulting from the annular
implications. Science, 293, 474–479.
mode. J. Clim., 15, 3043–3057.
Goosse, H., and T. Fichefet, 1999: Importance of ice-ocean interactions
Hall, A., and X. Qu, 2006: Using the current seasonal cycle to constrain
for the global ocean circulation: A model study. J. Geophys. Res., 104,
snow albedo feedback in future climate change. Geophys. Res. Lett., 33,
23337–23355.
L03502, doi:10.1029/2005GL025127.
Goosse, H., F.M. Selten, R.J. Haarsma, and J.D. Opsteegh, 2003: Large
Hall, M.M., and H.L. Bryden, 1982: Direct estimates and mechanisms of
sea-ice volume anomalies simulated in a coupled climate model. Clim.
ocean heat transport. Deep Sea Res., 29, 339–359.
Dyn., 20, 523–536, doi:10.1007/s00382-002-0290-4.
Hamilton, K., R.J. Wilson, and R.S. Hemler, 2001: Spontaneous
Goosse, H., et al., 2002: Potential causes of abrupt climate events: a
stratospheric QBO-like oscillations simulated by the GFDL SKYHI
numerical study with a three-dimensional climate model. Geophys. Res.
general circulation model. J. Atmos. Sci., 58, 3271–3292.
Lett., 29(18), 1860, doi:10.1029/2002GL014993.
Hannachi, A., and A. O’Neill, 2001: Atmospheric multiple equilibria and
non-Gaussian behaviour in model simulations. Q. J. R. Meteorol. Soc.,
127, 939–958.
652

Chapter 8
Climate Models and Their Evaluation
Hansen, J., et al., 1984: Climate sensitivity: analysis of feedback
Hodges, K.: Feature based diagnostics from ECMWF/NCEP Analyses and
mechanisms. Meteorol. Monogr., 29, 130–163.
AMIP II: Model Climatologies. In: The Second Phase of the Atmospheric
Hanson, C.E., J.P. Palutikof, and T.D. Davies, 2004: Objective cyclone
Model Intercomparison Project (AMIP2) [Gleckler, P. (ed.)]. Proceedings
climatologies of the North Atlantic - a comparison between the ECMWF
of the WCRP/WGNE Workshop, Toulouse, France, pp. 201-204.
and NCEP Reanalyses. Clim. Dyn., 22, 757–769.
Holland, M.M., and C.M. Bitz, 2003: Polar ampliﬁ cation of climate change
Harder, M., 1996: Dynamik, Rauhigkeit und Alter des Meereises
in coupled models. Clim. Dyn., 21, 221–232, doi:10.1007/s00382-003-
in der Arktis. PhD Thesis, Alfred-Wegener-Institut für Polar und
0332-6.
Meeresforschung, Bremerhaven, Germany, 124 pp.
Holland, M.M., and M. Raphael, 2006: Twentieth century simulation of
Hargreaves, J.C., J.D. Annan, N.R. Edwards, and R. Marsh, 2004: An
the Southern Hemisphere climate in coupled models. Part II: sea ice
efﬁ cient climate forecasting method using an intermediate complexity
conditions and variability. Clim. Dyn., 26, 229–245, doi:10.1007/s00382-
Earth System Model and the ensemble Kalman ﬁ lter. Clim. Dyn., 23,
005-0087-3.
745–760.
Horinouchi, T., 2002: Mesoscale variability of tropical precipitation:
Harrison, E.F., et al., 1990: Seasonal variation of cloud radiative forcing
Validation of satellite estimates of wave forcing using TOGA COARE
derived from the Earth Radiation Budget Experiment. J. Geophys. Res.,
radar data. J. Atmos. Sci., 59, 2428–2437.
95, 18687–18703.
Horinouchi, T., and S. Yoden, 1998: Wave-mean ﬂ ow interaction associated
Harrison, H., 2002: Comments on “Does the Earth have an adaptive
with a QBO-like oscillation simulated in a simpliﬁ ed GCM. J. Atmos.
infrared iris?”. Bull. Am. Meteorol. Soc., 83, 597.
Sci., 55, 502–526.
Hartmann, D.L., and K. Larson, 2002: An important constraint on tropical
Horinouchi, T., et al., 2003: Tropical cumulus convection and upward-
cloud-climate feedback. Geophys. Res. Lett., 29(20), 1951–1954.
propagating waves in middle-atmospheric GCMs. J. Atmos. Sci., 60,
Hartmann, D.L., and M.L. Michelsen, 2002: No evidence for iris. Bull. Am.
2765–2782.
Meteorol. Soc., 83, 249–254.
Hoskins, B.J., and K.I. Hodges, 2002: New perspectives on the Northern
Hartmann, D.L., M.E. Ockert-Bell, and M.L. Michelsen, 1992: The effect
Hemisphere winter storm tracks. J. Atmos. Sci., 59, 1041–1061.
of cloud type on Earth’s energy balance: Global analysis. J. Clim., 5,
Hoskins, B.J., and K.I. Hodges, 2005: New perspectives on the Southern
1281–1304.
Hemisphere storm tracks. J. Clim., 18, 4108–4129.
Harvey, D., et al., 1997: An Introduction to Simple Climate Models
Hourdin, F., et al., 2006: The LMDZ4 general circulation model: Climate
Used in the IPCC Second Assessment Report. IPCC Technical Paper 2
performance and sensitivity to parameterized physics with emphasis on
[Houghton, J.T., L.G. Meira Filho, D.J. Griggs, and K. Maskell (eds.)].
tropical convection. Clim. Dyn., 27, 787–813.
IPCC, Geneva, Switzerland, 51 pp.
Hovine, S., and T. Fichefet, 1994: A zonally averaged, three-basin ocean
Hasumi, H., 2002a: Sensitivity of the global thermohaline circulation to
circulation model for climate studies. Clim. Dyn., 15, 1405–1413.
interbasin freshwater transport by the atmosphere and the Bering Strait
Hsu, C.J., and F. Zwiers, 2001: Climate change in recurrent regimes and
throughﬂ ow. J. Clim., 15, 2516–2526.
modes of atmospheric variability. J. Geophys. Res., 106, 20145–20160.
Hasumi, H., 2002b: Modeling the global thermohaline circulation. J.
Hu, A.X., G.A. Meehl, W.M. Washington, and A. Dai, 2004: Response of
Oceanogr., 58, 25–33.
the Atlantic thermohaline circulation to increased atmospheric CO2 in a
Hasumi, H., and N. Suginohara, 1999: Effects of locally enhanced vertical
coupled model. J. Clim., 17, 4267–4279.
diffusivity over rough bathymetry on the world ocean circulation. J.
Huang, X., B.J. Soden, and D.L. Jackson, 2005: Interannual co-variability
Geophys. Res., 104, 23367–23374.
of tropical temperature and humidity: A comparison of model, reanalysis
Hazeleger, W., et al., 2001: Decadal upper ocean temperature variability in
data and satellite observation. Geophys. Res. Lett., 32, L17808,
the tropical Paciﬁ c. J. Geophys. Res., 106(C5), 8971–8988.
doi:10.1029/2005GL023375.
Held, I.M., and B.J. Soden, 2000: Water vapour feedback and global
Hunke, E.C., and J.K. Dukowicz, 1997: An elastic-viscous-plastic model
warming. Annu. Rev. Energy Environ., 25, 441– 475.
for sea ice dynamics. J. Phys. Oceanogr., 27, 1849–1867.
Henderson-Sellers, A., P. Irannejad, K. McGufﬁ e, and A.J. Pitman, 2003:
Hunke, E.C., and J.K. Dukowicz, 2002: The Elastic-Viscous-Plastic sea
Predicting land-surface climates - better skill or moving targets? Geophys.
ice dynamics model in general orthogonal curvilinear coordinates on a
Res. Lett., 30(14), 1777–1780.
sphere–Effect of metric terms. Mon. Weather Rev., 130, 1848–1865.
Henderson-Sellers, A., K. McGufﬁ e, D. Noone, and P. Irannejad, 2004:
Hunke, E.C., and J.K. Dukowicz, 2003: The Sea Ice Momentum Equation
Using stable water isotopes to evaluate basin-scale simulations of surface
in the Free Drift Regime. Technical Report LA-UR-03-2219, Los Alamos
water budgets. J. Hydrometeorol., 5(5), 805–822.
National Laboratory, Los Alamos, NM.
Hendon, H.H., 2000: Impact of air–sea coupling on the Madden–Julian
Hurrell, J.W., M.P. Hoerling, A.S. Phillips, and T. Xu, 2004: Twentieth
oscillation in a general circulation model. J. Atmos. Sci., 57, 3939–
century North Atlantic climate change. Part I: assessing determinism.
3952.
Clim. Dyn., 23, 371–389.
Hendon, H.H., 2005: Air sea interaction. In: Intraseasonal Variability in
Hutchings, J.K., H. Jasak, and S.W. Laxon, 2004: A strength implicit
the Atmosphere-Ocean Climate System [Lau, W.K.M., and D.E. Waliser
correction scheme for the viscous-plastic sea ice model. Ocean Modelling,
(eds.)]. Praxis Publishing, 436 pp.
7, 111–133.
Hewitt, C.D., C.S. Senior, and J.F.B. Mitchell, 2001: The impact of
Huybrechts, P., 2002: Sea-level changes at the LGM from ice-dynamics
dynamic sea-ice on the climate sensitvity of a GCM: a study of past,
reconstructions of the Greenland and Antarctic ice sheets during the
present and future climates. Clim. Dyn., 17, 655–668.
glacial cycles. Quat. Sci. Rev., 21, 203–231.
Heymsﬁ eld, A.J., and L. Donner, 1990: A scheme for parameterizing ice-
Huybrechts, P., I. Janssens, C. Poncin, and T. Fichefet, 2002: The response
cloud water content in general circulation models. J. Atmos. Sci., 47,
of the Greenland ice sheet to climate changes in the 21st century by
1865–1877.
interactive coupling of an AOGCM with a thermomechanical ice sheet
Hibler, W.D., 1979: A dynamic thermodynamic sea ice model. J. Phys.
model. Ann. Glaciol., 35, 409–415.
Oceanogr., 9, 817–846.
Iacobellis, S.F., G.M. McFarquhar, D.L. Mitchell, and R.C.J. Somerville,
Hirst, A.C., 1999: The Southern Ocean response to global warming in the
2003: The sensitivity of radiative ﬂ uxes to parameterized cloud
CSIRO coupled ocean-atmosphere model. Environ. Model. Software, 14,
microphysics. J. Clim., 16, 2979–2996.
227–241.
Iacono, M.J., J.S. Delamere, E.J. Mlawer, and S.A. Clough, 2003:
Hodges, K.I., B.J. Hoskins, J. Boyle, and C. Thorncroft, 2003: A comparison
Evaluation of upper tropospheric water vapor in the NCAR Community
of recent reanalysis data sets using objective feature tracking: storm
Climate Model, CCM3, using modeled and observed HIRS radiances. J.
tracks and tropical easterly waves. Mon. Weather Rev., 131, 2012–2037.
Geophys. Res., 108(D2), 4037, doi:10.1029/2002JD002539.
653

Climate Models and Their Evaluation
Chapter 8
Inamdar, A.K., and V. Ramanathan, 1998: Tropical and global scale
Kiehl, J.T., and P.R. Gent, 2004: The Community Climate System Model,
interactions among water vapour, atmospheric greenhouse effect, and
Version 2. J. Clim., 17, 3666–3682.
surface temperature. J. Geophys. Res., 103, 32177–32194.
Kiehl, J.T., et al., 1998: The National Center for Atmospheric Research
Ingram, W.J., 2002: On the robustness of the water vapor feedback: GCM
Community Climate Model: CCM3. J. Clim., 11, 1131–1149.
vertical resolution and formulation. J. Clim., 15, 917–921.
Kiktev, D., D.M.H. Sexton, L. Alexander, and C.K. Folland, 2003:
Inness, P.M., and J.M. Slingo, 2003: Simulation of the MJO in a coupled
Comparison of modeled and observed trends in indices of daily climate
GCM. I: Comparison with observations and atmosphere-only GCM. J.
extremes. J. Clim., 16(22), 3560–3571.
Clim., 16, 345–364.
Kim, S.-J., G.M. Flato, G.J. Boer, and N.A. McFarlane, 2002: A coupled
Inness, P.M., J.M. Slingo, E. Guilyardi, and J. Cole, 2003: Simulation of
climate model simulation of the Last Glacial Maximum, Part 1: Transient
the MJO in a coupled GCM. II: The role of the basic state. J. Clim., 16,
multi-decadal response. Clim. Dyn., 19, 515–537.
365–382.
Kimoto, M., N. Yasutomi, C. Yokoyama, and S. Emori, 2005: Projected
Iorio, J.P., et al., 2004: Effects of model resolution and subgrid scale physics
changes in precipitation characteristics near Japan under the global
on the simulation of precipitation in the continental United States. Clim.
warming. Scientiﬁ c Online Letters on the Atmosphere, 1, 85–88, doi:
Dyn., 23, 243–258, doi:10.1007/s00382-004-0440-y.
10.2151/sola.2005-023.
Jaeger, L., 1976: Monatskarten des Niederschlags für die Ganze Erde. Ber.
Kinne, S., et al., 2003: Monthly averages of aerosol properties: A global
Deutsche Wetterdienstes 139, Germany, 38 pp.
comparison among models, satellite, and AERONET ground data. J.
Jakob, C., and G. Tselioudis, 2003: Objective identiﬁ cation of cloud
Geophys. Res., 108(D20), 4634, doi:10.1029/2001JD001253.
regimes in the tropical western paciﬁ c. Geophys. Res. Lett., 30,
Kirtman, B.P., 2003: The COLA anomaly coupled model: Ensemble ENSO
doi:10.1029/2003GL018367.
prediction. Mon. Weather Rev., 131, 2324–2341.
Jennings, R.L., 1975: Data Sets for Meteorological Research. NCAR-TN/
Kirtman, B.P., and P.S. Schopf, 1998: Decadal variability in ENSO
1A, National Center for Atmospheric Research, Boulder, CO, 156 pp.
predictability and prediction. J. Clim., 11, 2804–2822.
Ji, M., A. Leetmaa, and V.E. Kousky, 1996: Coupled model predictions
Kirtman, B.P., K. Pegion, and S. Kinter, 2005: Internal atmospheric
of ENSO during the 1980s and the 1990s at the National Centers for
dynamics and tropical indo-paciﬁ c climate variability. J. Atmos. Sci., 62,
Environmental Prediction. J. Clim., 9, 3105–3120.
2220–2233.
Jin, X.Z., X.H. Zhang, and T.J. Zhou, 1999: Fundamental framework
Kleeman, R., Y. Tang, and A.M. Moore, 2003: The calculation of
and experiments of the third generation of the IAP/LASG World Ocean
climatically relevant singular vectors in the presence of weather noise as
General Circulation Model. Adv. Atmos. Sci., 16, 197–215.
applied to the ENSO problem. J. Atmos. Sci., 60, 2856–2868.
Johns, T.C., et al., 2006: The new Hadley Centre climate model HadGEM1:
Kleidon, A., 2004: Global datasets of rooting zone depth inferred from
Evaluation of coupled simulations. J. Clim., 19, 1327–1353.
inverse methods. J. Clim., 17, 2714–2722.
Jones, C.D., et al., 2005: Systematic optimisation and climate simulation
Kleidon, A., K. Fraedrich, and M. Heimann, 2000: A green planet versus
of FAMOUS, a fast version of HadCM3. Clim. Dyn., 25, 189–204.
a desert world: estimating the maximum effect of vegetation on the land
Jones, P.D., 1988: Hemispheric surface air temperature variations: Recent
surface climate. Clim. Change, 44, 471–493.
trends and an update to 1987. J. Clim., 1, 654–660.
Klein, S.A., and D.L. Hartmann, 1993: The seasonal cycle of low stratiform
Jones, P.D., et al., 1999: Surface air temperature and its variations over the
clouds. J. Clim., 6, 1587–1606.
last 150 years. Rev. Geophys., 37, 173–199.
Klein, S.A., and C. Jakob, 1999: Validation and sensitivities of frontal
Joos, F., et al., 1999: Global warming and marine carbon cycle feedbacks
clouds simulated by the ECMWF model. Mon. Weather Rev., 127, 2514–
on future atmospheric CO2. Science, 284, 464–467.
2531.
Joos, F., et al., 2001: Global warming feedbacks on terrestrial carbon
Knight, J.R, et al., 2005: A signature of persistent natural thermohaline
uptake under the IPCC emission scenarios. Global Biogeochem. Cycles,
circulation cycles in observed climate. Geophys. Res. Lett., 32, L20708,
15, 891–907.
doi:10.1029/2005GL024233.
Joshi, M., et al., 2003: A comparison of climate response to different
Knutson, T.R., and R.E. Tuleya, 1999: Increased hurricane intensities with
radiative forcings in three general circulation models: towards an
CO2-induced global warming as simulated using the GFDL hurricane
improved metric of climate change. Clim. Dyn., 20, 843–854.
prediction system. Clim. Dyn., 15(7), 503–519.
Jungclaus, J.H., et al., 2006: Ocean circulation and tropical variability in
Knutson, T.R., and R.E. Tuleya, 2004: Impact of CO2-induced warming
the AOGCM ECHAM5/MPI-OM. J. Clim., 19, 3952–3972.
on simulated hurricane intensity and precipitation: Sensitivity to the
K-1 Model Developers, 2004: K-1 Coupled Model (MIROC) Description.
choice of climate model and convective parameterization. J. Clim., 17,
K-1 Technical Report 1 [Hasumi, H., and S. Emori (eds.)]. Center for
3477–3495.
Climate System Research, University of Tokyo, Tokyo, Japan, 34 pp.,
Knutti, R., T.F. Stocker, F. Joos, and G.K. Plattner, 2002: Constraints
http://www.ccsr.u-tokyo.ac.jp/kyosei/hasumi/MIROC/tech-repo.pdf.
on radiative forcing and future climate change from observations and
Kalnay, E., et al., 1996: The NCEP/NCAR 40-year reanalysis project. Bull.
climate model ensembles. Nature, 416, 719–723.
Am. Meteorol. Soc., 77, 437–471.
Knutti, R., G.A. Meehl, M.R. Allen and D.A. Stainforth, 2006: Constraining
Kanamitsu, M., et al., 2002: NCEP dynamical seasonal forecast system
climate sensitivity from the seasonal cycle in surface temperature. J.
2000. Bull. Am. Meteorol. Soc., 83, 1019–1037.
Clim., 19, 4224–4233.
Kattsov, V., and E. Källén, 2005: Future climate change: Modeling and
Kodera, K., and M. Chiba, 1995: Tropospheric circulation changes
scenarios for the Arctic. In: Arctic Climate Impact Assessment (ACIA).
associated with stratospheric sudden warmings: A case study. J. Geophys.
Cambridge University Press, Cambridge, UK, pp. 99–150.
Res., 100, 11055–11068.
Kemball-Cook, S., B. Wang, and X. Fu, 2002: Simulation of the
Komuro, Y., and H. Hasumi, 2005: Intensiﬁ cation of the Atlantic deep
intraseasonal oscillation in ECHAM-4 model: The impact of coupling
circulation by the Canadian Archipelago throughﬂ ow. J. Phys. Oceanogr.,
with an ocean model. J. Atmos. Sci., 59, 1433–1453.
35, 775–789.
Khairoutdinov, M., D. Randall, and C. DeMott, 2005: Simulations of
Koster, R.D., et al., 2004: Regions of coupling between soil moisture and
the atmospheric general circulation using a cloud-resolving model as a
precipitation. Science, 305, 1138–1140.
superparameterization of physical processes. J. Atmos. Sci., 62, 2136–
Kraus, E.B., 1990: Diapycnal mixing. In: Climate-Ocean Interaction
2154.
[Schlesinger, M.E. (ed.)]. Kluwer, Amsterdam, pp. 269–293.
Kharin, V.V., F.W. Zwiers, and X. Zhang, 2005: Intercomparison of near
Kraus, E.B., and J.S. Turner, 1967: A one-dimensional model of the
surface temperature and precipitation extremes in AMIP-2 simulations,
seasonal thermocline. II. The general theory and its consequences. Tellus,
reanalyses and observations. J. Clim., 18(24), 5201–5223.
19, 98–105.
654

Chapter 8
Climate Models and Their Evaluation
Krinner, G., et al., 2005: A dynamic global vegetation model for studies of
Lindsay, R.W., and H.L. Stern, 2004: A new Lagrangian model of Arctic
the coupled atmosphere-biosphere system. Global Biogeochem. Cycles,
sea ice. J. Phys. Oceanogr., 34, 272–283.
19, GB1015, doi:10.1029/2003GB002199.
Lindzen, R.S., M.-D. Chou, and A.Y. Hou, 2001: Does the Earth have an
Lambert, S.J., and G.J. Boer, 2001: CMIP1 evaluation and intercomparison
adaptative infrared iris? Bull. Am. Meteorol. Soc., 82, 417–432.
of coupled climate models. Clim. Dyn., 17, 83–106.
Lindzen, R.S., M.-D. Chou, and A.Y. Hou, 2002: Comment on “No
Lambert, S.J., and J. Fyfe, 2006: Changes in winter cyclone frequencies
evidence for iris”. Bull. Am. Meteorol. Soc., 83, 1345–1349.
and strengths simulated in enhanced greenhouse gas simulations: Results
Lipscomb, W.H., 2001: Remapping the thickness distribution in sea ice
from the models participating in the IPCC diagnostic exercise. Clim.
models. J. Geophys. Res., 106, 13989–14000.
Dyn., 26, 713–728.
Liston, G., 2004: Representing subgrid snow cover heterogeneities in
Lanzante, J.R., 1996: Resistant, robust and nonparametric techniques for
regional and global models. J. Clim., 17, 1381–1397.
analysis of climate data: Theory and examples, including applications to
Liu, H., et al., 2004: An eddy-permitting oceanic general circulation model
historical radiosonde station data. Int. J. Climatol., 16, 1197–1226.
and its preliminary evaluations. Adv. Atmos. Sci., 21, 675–690.
Large, W.G., J.C. McWilliams, and S.C. Doney, 1994: Oceanic vertical
Liu, J., et al., 2003: Sensitivity of sea ice to physical parameterizations in
mixing: a review and a model with a nonlocal boundary layer
the GISS global climate model. J. Geophys. Res., 108, 3053, doi:10.1029/
parameterization. Rev. Geophys., 32, 363–403.
2001JC001167.
Larson, K., and D.L. Hartmann, 2003: Interactions among cloud, water
Liu, P., et al., 2005: MJO in the NCAR CAM2 with the Tiedtke convective
vapour, radiation and large-scale circulation in the tropical climate. Part
scheme. J. Clim., 18, 3007–3020.
1: sensitivity to uniform sea surface temperature changes. J. Clim., 15,
Lock, A.P., 2001: The numerical representation of entrainment in
1425–1440.
parameterizations of boundary layer turbulent mixing. Mon. Weather
Latif, M., 1998: Dynamics of interdecadal variability in coupled ocean-
Rev., 129, 1148–1163.
atmosphere models. J. Clim., 11, 602–624.
Lock, A.P., et al., 2000: A new boundary layer mixing scheme. Part I:
Latif, M., E. Roeckner, U. Mikolajewicz, and R. Voss, 2000: Tropical
Scheme description and SCM tests. Mon. Weather Rev., 128, 3187–
stabilisation of the thermohaline circulation in a greenhouse warming
3199.
simulation. J. Clim., 13, 1809–1813.
Lohmann, U., and G. Lesins, 2002: Stronger constraints on the
Latif, M., et al., 2001: ENSIP: The El Niño simulation intercomparison
anthropogenic indirect aerosol effect. Science, 298, 1012–1015.
project. Clim. Dyn., 18, 255–276.
Lorenz, D.J., and D.L. Hartmann, 2001: Eddy–zonal ﬂ ow feedback in the
Latif, M., et al., 2004: Reconstructing, monitoring, and predicting
Southern Hemisphere. J. Atmos. Sci., 58, 3312–3327.
multidecadal scale changes in the North Atlantic thermohaline circulation
Lu, J., R.J. Greatbatch, and K.A. Peterson, 2004: Trend in Northern
with sea surface temperatures. J. Clim., 17, 1605–1614.
Hemisphere winter atmospheric circulation during the last half of the
Lawrence, D.M., and J.M. Slingo, 2005: Weak land–atmosphere
twentieth century. J. Clim., 17, 3745–3760.
coupling strength in HadAM3: The role of soil moisture variability. J.
Luo, Z., and W.B. Rossow, 2004: Characterising tropical cirrus life cycle,
Hydrometeorol., 6, 670–680.
evolution and interaction with upper tropospheric water vapour using
Le Treut, H., Z.X. Li, and M. Forichon, 1994: Sensitivity of the LMD
a Lagrangian trajectory analysis of satellite observations. J. Clim., 17,
general circulation model to greenhouse forcing associated with two
4541–4563.
different cloud water parametrizations. J. Clim., 7, 1827–1841.
Madden, R.A., and P.R. Julian, 1971: Detection of a 40-50 day oscillation
Lee, M.-I., I.-S. Kang, J.-K. Kim, and B. E. Mapes, 2001: Inﬂ uence
in the zonal wind in the tropical Paciﬁ c. J. Atmos. Sci., 28, 702–708.
of cloud-radiation interaction on simulating tropical intraseasonal
Madec, G., P. Delecluse, M. Imbard, and C. Lévy, 1998: OPA Version 8.1
oscillation with an atmospheric general circulation model. J. Geophys.
Ocean General Circulation Model Reference Manual. Notes du Pôle de
Res., 106, 14219–14233.
Modélisation No. 11, Institut Pierre-Simon Laplace, Paris, 91 pp., http://
Levitus, S., and T.P. Boyer, 1994: World Ocean Atlas 1994, Volume 4:
www.lodyc.jussieu.fr/opa/Docu_Free/Doc_models/Doc_OPA8.1.pdf.
Temperature. NOAA NESDIS E/OC21, Washington, DC, 117 pp.
Mahfouf, J.-F., et al., 1995: The land surface scheme ISBA within the
Levitus, S., and J. Antonov, 1997: Variability of Heat Storage of and the
Meteo-France climate model ARPEGE. Part 1: Implementation and
Rate of Heat Storage of the World Ocean. NOAA NESDIS Atlas 16, US
preliminary results. J. Clim., 8, 2039–2057.
Government Printing Ofﬁ ce, Washington, DC, 6 pp., 186 ﬁ gures.
Maloney, E.D., and D.L. Hartmann, 2001: The sensitivity of the
Levitus, S., J. Antonov, and T. Boyer, 2005: Warming of the world
intraseasonal variability in the NCAR CCM3 to changes in convective
ocean, 1955-2003. Geophys. Res. Lett., 32, L02604, doi:10.1029/
parameterization. J. Clim., 14, 2015–2034.
2004GLO21592.
Maltrud, M.E., R.D. Smith, A.J. Semtner, and R.C. Malone, 1998: Global
Levitus, S., et al., 1998: World Ocean Database 1998, Volume 1:
eddy-resolving ocean simulations driven by 1985–1995 atmospheric
Introduction. NOAA Atlas NESDIS 18, US Government Printing Ofﬁ ce,
winds. J. Geophys. Res., 103, 30825–30853.
Washington, DC.
Manabe, S., and R.J. Stouffer, 1988: Two stable equilibria of a coupled
Liang, X., Z. Xie, and M. Huang, 2003: A new parameterization for surface
ocean-atmosphere model. J. Clim., 1(9), 841–866.
and groundwater interactions and its impact on water budgets with the
Manabe, S., and R.J. Stouffer, 1995: Simulation of abrupt climate change
variable inﬁ ltration capacity (VIC) land surface model. J. Geophys. Res.,
induced by fresh water input to the North Atlantic Ocean. Nature, 378,
108, 8613, doi:10.1029/2002JD003090.
165–167.
Limpasuvan, V., and D.L. Hartmann, 2000: Wave-maintained annular
Manabe, S., and R.J. Stouffer, 1996: Low-frequency variability of surface
modes of climate variability. J. Clim., 13, 4414–4429.
air temperature in a 1000-year integration of a coupled atmosphere-
Lin, B., T. Wong, B.A. Wielicki, and Y. Hu, 2004: Examination of the
ocean-land surface model. J. Clim., 9, 376–393.
decadal tropical mean ERBS nonscanner radiation data for the iris
Manabe, S., and R.J. Stouffer, 1997: Coupled ocean-atmosphere model
hypothesis. J. Clim., 17, 1239–1246.
response to freshwater input: Comparison to Younger Dryas event.
Lin, B., et al., 2002: The iris hypothesis: A negative or positive cloud
Paleoceanography, 12, 321–336.
feedback? J. Clim., 15, 3–7.
Manabe, S., R.J. Stouffer, M.J. Spelman, and K. Bryan, 1991: Transient
Lin, J.L., et al., 2006: Tropical intraseasonal variability in 14 IPCC AR4
responses of a coupled ocean atmosphere model to gradual changes of
climate models. Part I: Convective signals. J. Clim., 19, 2665–2690.
atmospheric CO2. I: Annual mean response. J. Clim., 4, 785–818.
Lin, W.Y., and M.H. Zhang, 2004: Evaluation of clouds and their radiative
Mann, M.E., R.S. Bradley, and M.K. Hughes, 1998: Global-scale
effects simulated by the NCAR Community Atmospheric Model against
temperature patterns and climate forcing over the past six centuries.
satellite observations. J. Clim., 17, 3302–3318.
Nature, 392, 779–787.
655

Climate Models and Their Evaluation
Chapter 8
Marchal, O., T.F. Stocker, and F. Joos, 1998: A latitude-depth, circulation-
Miller, R.L., 1997: Tropical thermostats and low cloud cover. J. Clim., 10,
biogeochemical ocean model for paleoclimate studies. Tellus, 50B, 290–
409–440.
316.
Miller, R.L., G.A. Schmidt, and D.T. Shindell, 2006: Forced variations of
Marotzke, J., 1997: Boundary mixing and the dynamics of three-dimensional
annular modes in the 20th century IPCC AR4 simulations. J. Geophys.
thermohaline circulation. J. Phys. Oceanogr., 27, 1713–1728.
Res., 111, D18101, doi:10.1029/2005JD006323.
Marshall, G.J., 2003: Trends in the Southern Annular Mode from
Milly, P.C.D., and A.B. Shmakin, 2002: Global modeling of land water
observations and reanalyses. J. Clim., 16, 4134–4143.
and energy balances, Part I: The Land Dynamics (LaD) model. J.
Marshall, J.C., C. Hill, L. Perelman, and A. Adcroft, 1997: Hydrostatic,
Hydrometeorol., 3, 283–299.
quasi-hydrostatic and non-hydrostatic ocean modeling. J. Geophys. Res.,
Milly, P.C.D., K.A. Dunne, and A.V. Vecchia, 2005: Global pattern of
102, 5733–5752.
trends in streamﬂ ow and water availability in a changing climate. Nature,
Marsland, S.J., et al., 2003: The Max-Planck-Institute global ocean/sea
438, 347–350, doi:10.1038/nature04312.
ice model with orthogonal curvilinear coordinates. Ocean Modelling, 5,
Min, S.-K., S. Legutke, A. Hense, and W.-T. Kwon, 2005: Climatology and
91–127.
internal variability in a 1000-year control simulation with the coupled
Marti, O., et al., 2005: The New IPSL Climate System Model: IPSL-CM4.
climate model ECHO-G—I. Near-surface temperature, precipitation and
Note du Pôle de Modélisation No. 26, Institut Pierre Simon Laplace des
mean sea level pressure. Tellus, 57A, 605–621.
Sciences de l’Environnement Global, Paris, http://dods.ipsl.jussieu.fr/
Minschwaner, K., and A.E. Dessler, 2004: Water vapor feedback in the
omamce/IPSLCM4/DocIPSLCM4/FILES/DocIPSLCM4.pdf.
tropical upper troposphere: model results and observations. J. Clim., 17,
Martin, G.M., et al., 2004: Evaluation of the Atmospheric Performance of
1272–1282.
HadGAM/GEM1. Hadley Centre Technical Note No. 54, Hadley Centre
Minschwaner, K., A.E. Dessler, and S. Parnchai, 2006: Multi-model
for Climate Prediction and Research/Met Ofﬁ ce, Exeter, UK, http://www.
analysis of the water vapour feedback in the tropical upper troposphere.
metofﬁ ce.gov.uk/research/hadleycentre/pubs/HCTN/index.html.
J. Clim., 19, 5455–5464.
Martin, G.M., et al., 2006: The physical properties of the atmosphere in
Mitchell, T.D., and P.D. Jones, 2005: An improved method of constructing a
the new Hadley Centre Global Environmental Model, HadGEM1. Part I:
database of monthly climate observations and associated high-resolution
Model description and global climatology. J. Clim., 19, 1274–1301.
grids. Int. J. Climatol., 25, 693 712.
Maxwell, R.M., and N.L. Miller, 2005: Development of a coupled land
Molteni, F., Kuchraski, F., and Corti, S., 2006: On the predictability of
surface and groundwater model. J. Hydrometeorol., 6, 233–247.
ﬂ ow-regime properties on interannual to interdecadal timescales. In:
May, W., 2004: Simulation of the variability and extremes of daily rainfall
Predictability of Weather and Climate [Palmer, T. and R. Hagedorn
during the Indian summer monsoon for present and future times in a
(eds.)]. Cambridge University Press, Cambridge, UK.
global time-slice experiment. Clim. Dyn., 22, 183–204.
Monahan, A.H., and A. Dai, 2004: The spatial and temporal structure of
Mayer, M., C. Wang, M. Webster, and R. Prinn, 2000: Linking air pollution
ENSO nonlinearity. J. Clim., 17, 3026–3036.
to global chemistry and climate. J. Geophys. Res., 105, 22869–22896.
Monahan, A.H., J.C. Fyfe, and L. Pandolfo, 2003: The vertical structure
McAvaney, B.J., et al., 2001: Model evaluation. In: Climate Change
of wintertime climate regimes of the Northern Hemisphere extratropical
2001: The Scientiﬁ c Basis. Contribution of Working Group I to the Third
atmosphere. J. Clim., 16, 2005–2021.
Assessment Report of the Intergovernmental Panel on Climate Change
Montoya, M., et al., 2005: The Earth System Model of Intermediate
[Houghton, J.T., et al. (eds.)]. Cambridge University Press, Cambridge,
Complexity CLIMBER-3α. Part I: Description and performance for
United Kingdom and New York, NY, USA, pp. 471–523.
present day conditions. Clim. Dyn., 25, 237–263, doi:10.1007/s00382-
McCarthy, M.P., and R. Toumi, 2004: Observed interannual variability of
005-0044-1.
tropical troposphere relative humidity. J. Clim., 17, 3181–3191.
Mouchet, A., and L. François, 1996: Sensitivity of a global oceanic
McDonald, R.E., et al., 2005: Tropical storms: representation and diagnosis
carbon cycle model to the circulation and to the fate of organic matter:
in climate models and the impacts of climate change. Clim. Dyn., 25,
Preliminary results. Phys. Chem. Earth, 21, 511–516.
19–36.
Moum, J.N., D.R. Caldwell, J.D. Nash, and G.D. Gunderson, 2002:
McFarlane, N.A., G.J. Boer, J.-P. Blanchet, and M. Lazare, 1992: The
Observations of boundary mixing over the continental slope. J. Phys.
Canadian Climate Centre second-generation general circulation model
Oceanogr., 32, 2113–2130.
and its equilibrium climate. J. Clim., 5, 1013–1044.
Murphy, J.M., 1995: Transient response of the Hadley Centre coupled
Mechoso, C.R., et al., 1995: The seasonal cycle over the tropical Paciﬁ c in
ocean-atmosphere model to increasing carbon dioxide. Part III: analysis
general circulation model. Mon. Weather Rev., 123, 2825–2838.
of global-mean response using simple models. J. Clim., 8, 496–514.
Meehl, G.A., and C. Tebaldi, 2004: More intense, more frequent, and
Murphy, J.M., et al., 2004: Quantiﬁ cation of modelling uncertainties in a
longer lasting heat waves in the 21st century. Science, 305, 994–997.
large ensemble of climate change simulations. Nature, 430, 768–772.
Meehl, G.A., and A. Hu, 2006: Mega droughts in the Indian monsoon and
Murray, R.J., 1996: Explicit generation of orthogonal grids for ocean
southwest North America and a mechanism for associated multi-decadal
models. J. Comput. Phys., 126, 251–273.
sea surface temperature anomalies. J. Clim., 19, 1605–1623.
Myhre, G., E.J. Highwood, K.P. Shine, and F. Stordal, 1998: New estimates
Meehl, G.A., C. Tebaldi, and D. Nychka, 2004: Changes in frost days in
of radiative forcing due to well mixed greenhouse gases. Geophys. Res.
simulations of twenty-ﬁ rst century climate. Clim. Dyn., 23, 495–511.
Lett., 25, 2715–2718.
Meehl, G.A., et al., 2001: Factors that affect the amplitude of El Niño in
Nakano, H., and N. Suginohara, 2002: Effects of bottom boundary layer
global coupled climate models. Clim. Dyn., 17, 515–526.
parameterization on reproducing deep and bottom waters in a World
Meissner, K.J., A.J. Weaver, H.D. Matthews, and P.M. Cox, 2003: The role
Ocean model. J. Phys. Oceanogr., 32, 1209–1227.
of land surface dynamics in glacial inception: A study with the UVic
Naud, C.M., A.D. Del Genio, and M. Bauer, 2006: Observational
Earth System Model. Clim. Dyn., 21, 515–537, doi:10.1007/s00382-003-
constraints on cloud thermodynamic phase in midlatitude storms. J.
0352-2.
Clim., 19, 5273–5288.
Mellor, G.L., and T. Yamada, 1982: Development of a turbulence closure
Neale, R., and J. Slingo, 2003: The maritime continent and its role in the
model for geophysical ﬂ uid problems. Rev. Geophys., 20, 851–875.
global climate: A GCM study. J. Clim., 16, 834–848.
Mellor, G.L., and L. Kantha, 1989: An ice-ocean coupled model. J.
Newman, M., G.P. Compo, and M.A. Alexander, 2003: ENSO-forced
Geophys. Res., 94, 10937–10954.
variability of the PDO. J. Clim., 16, 3853–3857.
Mestas-Nunez, A.M., and D.B. Enﬁ eld, 1999: Rotated global modes of
Nijssen, B., et al., 2003: Simulation of high latitude hydrological processes
non-ENSO sea surface temperature variability. J. Clim., 12, 2734–2745.
in the Torne-Kalix basin: PILPS Phase 2(e) 2: Comparison of model
Miller, J.R., G.L. Russell, and G. Caliri, 1994: Continental-scale river ﬂ ow
results with observations. Global Planet. Change, 38, 31–53.
in climate models. J. Clim., 7, 914–928.
656

Chapter 8
Climate Models and Their Evaluation
Norris, J.R., 1998a: Low cloud type over the ocean from surface
Petoukhov, V., et al., 2000: CLIMBER-2: A climate system model of
observations. Part I: relationship to surface meteorology and the vertical
intermediate complexity. Part I: Model description and performance for
distribution of temperature and moisture. J. Clim., 11, 369–382.
present climate. Clim. Dyn., 16, 1–17.
Norris, J.R., 1998b: Low cloud type over the ocean from surface
Petoukhov, V., et al., 2005: EMIC Intercomparison Project (EMIP-CO2):
observations. Part II: geographical and seasonal variations. J. Clim., 11,
Comparative analysis of EMIC simulations of current climate and
383–403.
equilibrium and transient responses to atmospheric CO2 doubling. Clim.
Norris, J.R., and C.P. Weaver, 2001: Improved techniques for evaluating
Dyn., 25, 363–385, doi:10.1007/s00382-005-0042-3.
GCM cloudiness applied to the NCAR CCM3. J. Clim., 14, 2540–2550.
Phillips, T.J., et al., 2004: Evaluating parameterizations in general
Norris, J.R., and S.F. Iacobellis, 2005: North paciﬁ c cloud feedbacks
circulation models: Climate simulation meets weather prediction. Bull.
inferred from synoptic-scale dynamic and thermodynamic relationships.
Am. Meteorol. Soc., 85, 1903–1915.
J. Clim., 18, 4862–4878.
Piani, C., D.J. Frame, D.A. Stainforth, and M.R. Allen, 2005: Constraints on
NRC (National Research Council), 2003: Understanding Climate Change
climate change from a multi-thousand member ensemble of simulations.
Feedbacks. National Academies Press, Washington, DC, 152 pp.
Geophys. Res. Lett., 32, L23825, doi:10.1029/2005GL024452.
O’Farrell, S.P., 1998: Investigation of the dynamic sea ice component of a
Pierce, D.W., T.P. Barnett, and M. Latif, 2000: Connections between the
coupled atmosphere sea-ice general circulation model. J. Geophys. Res.,
Paciﬁ c Ocean tropics and midlatitudes on decadal time scales. J. Clim.,
103, 15751–15782.
13, 1173–1194.
Oki, T., and Y.C. Sud, 1998: Design of total runoff integrating pathways
Pierrehumbert, R.T., 1995: Thermostats, radiator ﬁ ns, and the local
(TRIP)—A global river channel network. Earth Interactions, 2, 1–37.
runaway greenhouse. J. Atmos. Sci., 52, 1784–180.
Oleson, K.W., et al., 2004: Technical Description of the Community Land
Pierrehumbert, R.T., 1999: Subtropical water vapour as a mediator of
Model (CLM). NCAR Technical Note NCAR/TN-461+STR, National
rapid global climate change. In: Mechanisms of Global Climate Change
Center for Atmospheric Research, Boulder, CO, 173 pp.
at Millennial Timescales. Geophysical Monograph 112, American
Oliver, K.I.C., A.J. Watson, and D.P. Stevens, 2005: Can limited ocean
Geophysical Union, Washington, DC, pp. 339–361.
mixing buffer rapid climate change? Tellus, 57A, 676–690.
Pierrehumbert, R.T., and R. Roca, 1998: Evidence for control of Atlantic
Oouchi, K., et al., 2006: Tropical cyclone climatology in a global-warming
subtropical humidity by large scale advection. Geophys. Res. Lett., 25,
climate as simulated in a 20 km-mesh global atmospheric model:
4537–4540.
Frequency and wind intensity analyses. J. Meteorol. Soc. Japan, 84,
Pierrehumbert, R.T., H. Brogniez, and R. Roca, 2007: On the relative
259–276.
humidity of the Earth’s atmosphere. In: The General Circulation
Opsteegh, J.D., R.J. Haarsma, F.M. Selten, and A. Kattenberg, 1998:
[Schneider, T., and A. Sobel (eds.)]. Princeton University Press,
ECBILT: A dynamic alternative to mixed boundary conditions in ocean
Princeton, NJ, in press.
models. Tellus, 50A, 348–367.
Pitman, A.J., B.J. McAvaney, N. Bagnoud, and B. Cheminat, 2004: Are
Osborn, T.J., 2004: Simulating the winter North Atlantic Oscillation: the
inter-model differences in AMIP-II near surface air temperature means
roles of internal variability and greenhouse gas forcing. Clim. Dyn., 22,
and extremes explained by land surface energy balance complexity?
605–623.
Geophys. Res. Lett., 31, L05205, doi:10.1029/2003GL019233.
Otterå, O.H., et al., 2004: Transient response of the Atlantic meridional
Plattner, G.-K., F. Joos, T.F. Stocker, and O. Marchal, 2001: Feedback
overturning circulation to enhanced freshwater input to the Nordic Seas-
mechanisms and sensitivities of ocean carbon uptake under global
Arctic Ocean in the Bergen Climate Model. Tellus, 56A, 342–361.
warming. Tellus, 53B, 564–592.
Otto-Bliesner, B.L., et al., 2006: Climate sensitivity of moderate- and low-
Plaut, G., and E. Simonnet, 2001: Large–scale circulation classiﬁ cation,
resolution versions of CCSM3 to preindustrial forcings. J. Clim., 19,
weather regimes, and local climate over France, the Alps, and Western
2567–2583.
Europe. Clim. Res., 17, 303–324.
Pacanowski, R.C., K. Dixon, and A. Rosati, 1993: The GFDL Modular
Polzin, K.L., J.M. Toole, J.R. Redwell, and R.W. Schmitt, 1997: Spatial
Ocean Model Users Guide, Version 1.0. GFDL Ocean Group Technical
variability of turbulent mixing in the abyssal ocean. Science, 276, 93–
Report No. 2, Geophysical Fluid Dynamics Laboratory, Princeton, NJ.
96.
Paciorek, C.J., J.S. Risbey, V. Ventura, and R.D. Rosen, 2002: Multiple
Pope, V.D., and R.A. Stratton, 2002: The processes governing horizontal
indices of Northern Hemisphere cyclone activity, winters 1949-99. J.
resolution sensitivity in a climate model. Clim. Dyn., 19, 211–236.
Clim., 15, 1573–1590.
Pope, V.D., M.L. Gallani, P.R. Rowntree, and R.A. Stratton, 2000: The
Palmer, T.N., and J. Shukla, 2000: Editorial (for special issue on DSP/
impact of new physical parametrizations in the Hadley Centre climate
PROVOST). Q. J. R. Meteorol. Soc., 126, 1989–1990.
model: HadAM3. Clim. Dyn., 16, 123–146.
Palmer, T.N., et al., 2004: Development of a European multimodel
Potter, G.L., and R.D. Cess, 2004: Testing the impact of clouds on the
ensemble system for seasonal to interannual prediction (DEMETER).
radiation budgets of 19 atmospheric general circulation models. J.
Bull. Am. Meteorol. Soc., 85, 853–872.
Geophys. Res., 109, doi:10.1029/2003JD004018.
Pan, Z., et al., 2004: Evaluation of uncertainties in regional climate change
Power, S.B., and R. Colman, 2006: Multi-decadal predictability in a
simulations. J. Geophys. Res., 106, 17735–17752.
coupled GCM. Clim. Dyn., 26, 247–272.
Pardaens, A.K., H.T. Banks, J.M. Gregory, and P.R. Rowntree, 2003:
Power, S.B., M.H. Haylock, R. Colman, and X. Wang, 2006: The
Freshwater transports in HadCM3. Clim. Dyn., 21, 177–195.
predictability of interdecadal changes in ENSO activity and ENSO
Parekh, P., M.J. Follows, and E. Boyle, 2005: Decoupling of iron
teleconnections. J. Clim., 19, 4755–4771.
and phosphate in the global ocean. Global Biogeochem. Cycles, 19,
Power, S., et al., 1999: Interdecadal modulation of the impact of ENSO on
doi:10.1029/2004GB002280.
Australia. Clim. Dyn., 15, 319–324.
Pelly, J.L., and B.J. Hoskins, 2003a: A new perspective on blocking. J.
Qu, X., and A. Hall, 2005: Surface contribution to planetary albedo
Atmos. Sci., 60, 743–755.
variability in cryosphere regions. J. Clim., 18, 5239–5252.
Pelly, J.L., and B.J. Hoskins, 2003b: How well does the ECMWF Ensemble
Quadrelli, R., and J.M. Wallace, 2004: A simpliﬁ ed linear framework
Prediction System predict blocking? Q. J. R. Meteorol. Soc., 129, 1683–
for interpreting patterns of northern hemisphere wintertime climate
1702.
variability. J. Clim., 17, 3728–3744.
Peters, M.E., and C.S. Bretherton, 2005: A simpliﬁ ed model of the Walker
Rahmstorf, S., 1996: On the freshwater forcing and transport of the Atlantic
circulation with an interactive ocean mixed layer and cloud-radiative
thermohaline circulation. Clim. Dyn., 12, 799–811.
feedbacks. J. Clim., 18, 4216–4234.
Rahmstorf, S., et al., 2005: Thermohaline circulation hysteresis: A
model intercomparison. Geophys. Res. Lett., 32, L23605, doi:10.1029/
2005GL023655.
657

Climate Models and Their Evaluation
Chapter 8
Randall, D.A., et al., 2003: Confronting models with data: The GEWEX
Ross, R.J., W.P. Elliott, D.J. Seidel, and participating AMIP-II modelling
Cloud Systems Study. Bull. Am. Meteorol. Soc., 84, 455–469.
groups, 2002: Lower tropospheric humidity-temperature relationships in
Randall, D.A., et al., 2006: Cloud feedbacks. In: Frontiers in the Science of
radiosonde observations and atmospheric general circulation models. J.
Climate Modeling [Kiehl, J.T., and V. Ramanathan (eds.)]. Proceedings
Hydrometeorol., 3, 26–38.
of a symposium in honor of Professor Robert D. Cess.
Russell, G.L., 2005: 4x3 Atmosphere-Ocean Model Documentation. http://
Raper, S.C.B., T.M.L. Wigley, and R.A. Warrick, 1996: Global sea-level
aom.giss.nasa.gov/doc4x3.html.
rise: past and future. In: Sea-Level Rise and Coastal Subsidence: Causes,
Russell, G.L., J.R. Miller, and D. Rind, 1995: A coupled atmosphere-ocean
Consequences and Strategies [Milliman, J.D., and B.U. Haq (eds.)].
model for transient climate change studies. Atmos.-Ocean, 33, 683–730.
Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 11–46.
Russell, J.L., R.J. Stouffer, and K.W. Dixon, 2006: Intercomparison of the
Raper, S.C.B., J.M. Gregory, and T.J. Osborn, 2001: Use of an upwelling-
Southern Ocean circulations in IPCC coupled model control simulations.
diffusion energy balance model to simulate and diagnose A/OGCM
J. Clim., 19, 4560–4575.
results. Clim. Dyn., 17, 601–613.
Saenko, O.A., and W.J. Merryﬁ eld, 2005: On the effect of topographically-
Raphael, M.N., and M.M. Holland, 2006: Twentieth century simulation of
enhanced mixing on the global ocean circulation. J. Phys. Oceanogr.,
the Southern Hemisphere climate in coupled models. Part 1: Large scale
35, 826–834.
circulation variability. Clim. Dyn., 26, 217–228, doi:10.1007/s00382-
Saenko, O.A., G.M. Flato, and A.J. Weaver, 2002: Improved representation
005-0082-8.
of sea-ice processes in climate models. Atmos.-Ocean, 40, 21–43.
Rayner, N.A., et al., 2003: Global analyses of sea surface temperature, sea
Sakamoto, T.T., et al., 2004: Far-reaching effects of the Hawaiian Islands
ice, and night marine air temperature since the late nineteenth century. J.
in the CCSR/NIES/FRCGC high-resolution climate model. Geophys.
Geophys. Res., 108(D14), doi:10.1029/2002JD002670.
Res. Lett., 31, doi:10.1029/2004GL020907.
Redi, M.H., 1982: Oceanic isopycnal mixing by coordinate rotation. J.
Sakamoto, T., et al., 2005: Responses of the Kuroshio and the Kuroshio
Phys. Oceanogr., 12, 1154–1158.
Extension to global warming in a high-resolution climate model.
Renssen, H., V. Brovkin, T. Fichefet, and H. Goosse, 2003: Holocene
Geophys. Res. Lett., 32, L14617, doi:10.1029/2005GL023384.
climate instability during the termination of the African Humid Period.
Salas-Mélia, D., 2002: A global coupled sea ice-ocean model. Ocean
Geophys. Res. Lett., 30(4), 1184, doi:10.1029/2002GL016636.
Modelling, 4, 137–172.
Renwick, J.A., 1998: ENSO-related variability in the frequency of South
Saltzman, B., 1978: A survey of statistical-dynamical models of the
Paciﬁ c blocking. Mon. Weather Rev., 126, 3117–3123.
terrestrial climate. Adv. Geophys., 20, 183–295.
Rial, J.A., 2004: Abrupt climate change: chaos and order at orbital and
Santer, B.D., et al., 2005: Ampliﬁ cation of surface temperature trends and
millennial scales. Global Planet. Change, 41, 95–109.
variability in the tropical atmosphere. Science, 309, 1551–1556.
Ridley, J.K., P. Huybrechts, J.M. Gregory, and J.A. Lowe, 2005:
Sato, N., et al., 1989: Effects of implementing the simple biosphere model
Elimination of the Greenland ice sheet in a high CO2 climate. J. Clim.,
in a general circulation model. J. Atmos. Sci., 46, 2757–2782.
18, 3409–3427.
Sausen, R., K. Barthel, and K. Hasselmann, 1988: Coupled ocean-
Rind, D.G., et al., 2001: Effects of glacial meltwater in the GISS Coupled
atmosphere models with ﬂ ux correction. Clim. Dyn., 2, 145–163.
Atmosphere-Ocean model: Part II. A bipolar seesaw in deep water
Sausen, R., et al., 2002: Climate response to inhomogeneously distributed
production. J. Geophys. Res., 106, 27355–27365.
forcing agents. In: Non-CO2 Greenhouse Gases: Scientiﬁ c Understanding,
Ringer, M.A., and R.P. Allan, 2004: Evaluating climate model simulations
Control Options and Policy Aspects [van Ham, J., A.P.M. Baede, R.
of tropical clouds. Tellus, 56A, 308–327.
Guicherit, and J.G.F.M. Williams-Jacobse (eds.)]. Millpress, Rotterdam,
Ringer, M.A., et al., 2006: The physical properties of the atmosphere in the
Netherlands, pp. 377–381.
new Hadley Centre Global Environmental Model (HadGEM1). Part II:
Schär, C., et al., 2004: The role of increasing temperature variability for
Aspects of variability and regional climate. J. Clim., 19, 1302–1326.
European summer heat waves. Nature, 427, 332–336, doi:10.1038/
Roberts, M.J., 2004: The Ocean Component of HadGEM1. GMR Report
nature02300.
Annex IV.D.3, Met Ofﬁ ce, Exeter, UK.
Scaife, A.A., J.R. Knight, C.K. Folland, and G.K. Vallis, 2005: A
Roberts, M., et al., 2004: Impact of an eddy-permitting ocean resolution
stratospheric inﬂ uence on the winter NAO and North Atlantic surface
on control and climate change simulations with a global coupled GCM.
climate. Geophys. Res. Lett., 32, L18715.
J. Clim., 17, 3–20.
Scaife, A.A., et al., 2000: Realistic quasi-biennial oscillations in a
Robertson, A.W., 2001: Inﬂ uence of ocean-atmosphere interaction on
simulation of the global climate. Geophys. Res. Lett., 27, 3481–3484.
the Arctic Oscillation in two general circulation models. J. Clim., 14,
Schiller, A., U. Mikolajewicz, and R. Voss, 1997: The stability of the North
3240–3254.
Atlantic thermohaline circulation in a coupled ocean-atmosphere general
Robock, A., et al., 2000: The global soil moisture data bank. Bull. Am.
circulation model. Clim. Dyn., 13, 325–347.
Meteorol. Soc., 81, 1281–1299.
Schmidt, G.A., C.M. Bitz, U. Mikolajewicz, and L.B. Tremblay, 2004:
Roeckner, E., et al., 1996: The Atmospheric General Circulation Model
Ice-ocean boundary condi tions for coupled models. Ocean Modelling,
ECHAM4: Model Description and Simulation of Present-Day Climate.
7, 59–74.
MPI Report No. 218, Max-Planck-Institut für Meteorologie, Hamburg,
Schmidt, G.A., et al., 2006: Present day atmospheric simulations using
Germany, 90 pp.
GISS ModelE: Comparison to in-situ, satellite and reanalysis data. J.
Roeckner, E., et al., 2003: The Atmospheric General Circulation Model
Clim., 19, 153–192, http://www.giss.nasa.gov/tools/modelE/.
ECHAM5. Part I: Model Description. MPI Report 349, Max Planck
Schmittner, A., and T.F. Stocker, 1999: The stability of the thermohaline
Institute for Meteorology, Hamburg, Germany, 127 pp.
circulation in global warming experiments. J. Clim., 12, 1117–1133.
Roesch, A., 2006: Evaluation of surface albedo and snow cover in AR4
Schmittner, A., C. Appenzeller, and T.F. Stocker, 2000: Enhanced Atlantic
coupled climate models. J. Geophys. Res., 111, D15111, doi:10.1029/
freshwater export during El Niño. Geophys. Res. Lett., 27, 1163–1166.
2005JD006473.
Schneider, E.K., 2001: Causes of differences between the equatorial Paciﬁ c
Rooth, C., 1982: Hydrology and ocean circulation. Prog. Oceanogr., 11,
as simulated by two coupled GCM’s. J. Clim., 15, 2301–2320.
131–149.
Schneider, S.H., 2004: Abrupt non-linear climate change, irreversibility
Rosati, A., K. Miyakoda, and R. Gudgel, 1997: The impact of ocean initial
and surprise. Global Environ. Change, 14, 245–258.
conditions on ENSO forecasting with a coupled model. Mon. Weather
Rev., 125(5), 754–772.
658

Chapter 8
Climate Models and Their Evaluation
Schubert, S., et al., 1992: Monthly Means of Selected Climate Variables
Slingo, J., et al., 2003: Scale interactions on diurnal to seasonal timescales
for 1985–1989. NASA Technical Memorandum, Goddard Space
and their relevance to model systematic errors. Ann. Geophys., 46, 139–
Flight Center, Greenbelt, MD, 376 pp. Available from the NASA
155.
Technical Report Server, Accession Number: 92N29653; Document ID:
Smith, R.D., and P.R. Gent, 2002: Reference Manual for the Parallel
19920020410; Report Number: NAS 1.15104565, NASA-TM-104565,
Ocean Program (POP), Ocean Component of the Community Climate
REPT-92B00088.
System Model (CCSM2.0 and 3.0). Technical Report LA-UR-02-2484,
Scinocca, J.F., and N.A. McFarlane, 2004: The variability of modelled
Los Alamos National Laboratory, Los Alamos, NM, http://www.ccsm.
tropical precipitation. J. Atmos. Sci., 61, 1993–2015.
ucar.edu/models/ccsm3.0/pop/.
Seidel, D.J., and J.R. Lanzante, 2004: An assessment of three alternatives to
Soden, B.J., 1997: Variations in the tropical greenhouse effect during El
linear trends for characterizing global atmospheric temperature changes.
Niño. J. Clim., 10(5), 1050–1055.
J. Geophys. Res., 109, D14108, doi:10.1029/2003JD004414.
Soden, B.J., 2000: The sensitivity of the tropical hydrological cycle to
Seidov, D., E.J. Barron, and B.J. Haupt, 2001: Meltwater and the global
ENSO. J. Clim., 13, 538–549.
ocean conveyor: Northern versus southern connections. Global Planet.
Soden, B.J., 2004: The impact of tropical convection and cirrus on upper
Change, 30, 253–266.
tropospheric humidity: A Lagrangian analysis of satellite measurements.
Seidov, D., R.J. Stouffer, and B.J. Haupt, 2005: Is there a simple bi-polar
Geophys. Res. Lett., 31, L20104, doi:10.1029/2004GL020980.
ocean seesaw? Global Planet. Change, 49, 19–27.
Soden, B.J., and I.M. Held, 2006: An assessment of climate feedbacks in
Sellers, P.J., Y. Mintz, Y.C. Sud, and A. Dalcher, 1986: A simple biosphere
coupled ocean-atmosphere models. J. Clim., 19, 3354–3360.
model (SiB) for use within general circulation models. J. Atmos. Sci.,
Soden, B.J., A.J. Broccoli, and R.S. Hemler, 2004: On the use of cloud
43, 505–531.
forcing to estimate cloud feedback. J. Clim., 17, 3661–3665.
Selten, F.M., and G. Branstator, 2004: Preferred regime transition routes
Soden, B.J., R.T. Wetherald, G.L. Stenchikov, and A. Robock, 2002: Global
and evidence for an unstable periodic orbit in a baroclinic model. J.
cooling after the eruption of Mount Pinatubo: A test of climate feedback
Atmos. Sci., 61, 2267–2268.
by water vapour. Science, 296, 727–730.
Semtner, A.J., 1976: A model for the thermodynamic growth of sea ice in
Soden, B.J., et al., 2005: The radiative signature of upper tropospheric
numerical investigations of climate. J. Phys. Oceanogr., 6, 379–389.
moistening. Science, 310(5749), 841–844.
Seneviratne, S.I., J.S. Pal, E.A.B. Eltahir, and C. Schär, 2002: Summer
Sohn, B.-J., and J. Schmetz, 2004: Water vapor-induced OLR variations
dryness in a warmer climate: A process study with a regional climate
associated with high cloud changes over the tropics: a study from
model. Clim. Dyn., 20, 69–85.
Meteosat-5 observations. J. Clim., 17, 1987–1996.
Senior, C.A., and J.F.B. Mitchell, 1993: Carbon dioxide and climate: The
Sokolov, A., and P. Stone, 1998: A ﬂ exible climate model for use in
impact of cloud parameterization. J. Clim., 6, 393–418.
integrated assessments. Clim. Dyn., 14, 291–303.
Senior, C.A., and J.F.B. Mitchell, 2000: The time dependence of climate
Sokolov, A.P., et al., 2005: The MIT Integrated Global System Model
sensitivity. Geophys. Res. Lett., 27, 2685–2688.
(IGSM), Version 2: Model Description And Baseline Evaluation. Report
Severijns, C.A., and W. Hazeleger, 2005: Optimising parameters in an
No. 124, Joint Program on the Science and Policy of Global Change,
atmospheric general circulation model. J. Clim., 18, 3527–3535.
Massachusetts Institute of Technology, Cambridge, MA, http://web.mit.
Shaffrey, L., and R. Sutton, 2004: The interannual variability of energy
edu/globalchange/www/MITJPSPGC_Rpt124.pdf.
transports within and over the Atlantic Ocean in a coupled climate model.
Spelman, M.J., and S. Manabe, 1984: Inﬂ uence of oceanic heat transport
J. Clim., 17, 1433–1448.
upon the sensitivity of a model climate. J. Geophys. Res., 89, 571–586.
Shibata, K., et al., 1999: A simulation of troposphere, stratosphere and
Sperber, K.R., S. Gualdi, S. Legutke, and V. Gayler, 2005: The Madden-
mesosphere with an MRI/JMA98 GCM. Papers in Meteorology and
Julian Oscillation in ECHAM4 coupled and uncoupled GCMs. Clim.
Geophysics, 50, 15–53.
Dyn., 25, doi:10.1007/s00382-005-0026-3.
Shindell, D.T., R.L. Miller, G.A. Schmidt, and L. Pandolfo, 1999:
Stainforth, D.A., et al., 2005: Uncertainty in predictions of the climate
Simulation of recent northern winter climate trends by greenhouse-gas
response to rising levels of greenhouse gases. Nature, 433, 403–406.
forcing. Nature, 399, 452–455.
Stein, O., 2000: The variability of Atlantic-European blocking as derived
Shiogama, H., M. Watanabe, M. Kimoto, and T. Nozawa, 2005:
from long SLP time series. Tellus, 52A, 225–236.
Anthropogenic and natural forcing impacts on the Paciﬁ c Decadal
Stenchikov, G., et al., 2002: Arctic Oscillation response to the 1991 Mount
Oscillation during the second half of the 20th century. Geophys. Res.
Pinatubo eruption: Effects of volcanic aerosols and ozone depletion. J.
Lett., 32, L21714, doi:10.1029/2005GL023871.
Geophys. Res., 107(D24), 4803.
Shukla, J., et al., 2006: Climate model ﬁ delity and projections of climate
Stephens, G.L., 2005: Cloud feedbacks in the climate system: a critical
change. Geophys. Res. Lett., 33, L07702, doi:10.1029/2005GL025579.
review. J. Clim., 18, 237–273.
Sinclair, M.R., 1996: A climatology of anticyclones and blocking for the
Stephenson, D.B., and V. Pavan, 2003: The North Atlantic Oscillation in
Southern Hemisphere. Mon. Weather Rev., 124, 245–263.
coupled climate models: a CMIP1 evaluation. Clim. Dyn., 20, 381–399.
Sitch, S., et al., 2003: Evaluation of ecosystem dynamics, plant geography
Stephenson, D.B., A. Hannachi, and A. O’Neill, 2004: On the existence of
and terrestrial carbon cycling in the LPJ dynamic global vegetation
multiple climate regimes. Q. J. R. Meteorol. Soc., 130, 583–605.
model. Global Change Biol., 9, 161–185.
Stocker, T.F., D.G. Wright, and L.A. Mysak, 1992: A zonally averaged,
Six, K.D., and E. Maier-Reimer, 1996: Effects of plankton dynamics on
coupled atmosphere-ocean model for paleoclimate studies. J. Clim., 5,
seasonal carbon ﬂ uxes in an ocean general circulation model. Global
773–797.
Biogeochem. Cycles, 10, 559–583.
Stocker, T.F., et al., 2001: Physical climate processes and feedbacks. In:
Slater, A.G., et al., 2001: The representation of snow in land-surface
Climate Change 2001: The Scientiﬁ c Basis. Contribution of Working
schemes: Results from PILPS 2(d). J. Hydrometeorol., 2, 7–25.
Group I to the Third Assessment Report of the Intergovernmental Panel
Slingo, J.M., P.M. Inness, and K.R. Sperber, 2005: Modelling the Madden
on Climate Change [Houghton, J.T., et al. (eds.)]. Cambridge University
Julian Oscillation. In: Intraseasonal Variability of the Atmosphere-
Press, Cambridge, United Kingdom and New York, NY, USA, pp. 419–
Ocean Climate System [Lau, W.K.-M., and D.E. Waliser (eds.)]. Praxis
470.
Publishing.
Stommel, H., 1961: Thermohaline convection with two stable regimes of
Slingo, J.M., et al., 1996: Intraseasonal oscillations in 15 atmospheric
ﬂ ow. Tellus, 13, 224–230.
general circulation models: Results from an AMIP Diagnostic Subproject.
Stouffer, R.J., 2004: Time scales of climate response. J. Clim., 17(1),
Clim. Dyn., 12, 325–357.
209–217.
659

Climate Models and Their Evaluation
Chapter 8
Stouffer, R.J., and K.W. Dixon, 1998: Initialization of Coupled Models for
Thompson, D.W.J., and J.M. Wallace, 2000: Annular modes in the
Use in Climate Studies: A Review. Research Activities in Atmospheric
extratropical circulation. Part I: Month-to-month variability. J. Clim., 13,
and Oceanic Modelling, Report No. 27, WMO/TD-No. 865, World
1000–1016.
Meteorological Organization, Geneva, Switzerland, I.1–I.8.
Thompson, D.W.J., and S. Solomon, 2002: Interpretation of recent Southern
Stouffer, R.J., and S. Manabe, 2003: Equilibrium response of thermohaline
Hemisphere climate change. Science, 296, 895–899.
circulation to large changes in atmospheric CO2 concentration. Clim.
Thorndike, A.S., D.A. Rothrock, G.A. Maykut, and R. Colony, 1975: The
Dyn., 20(7/8), 759–773.
thickness distribution of sea ice. J. Geophys. Res., 80, 4501–4513.
Stouffer, R.J., A.J. Weaver, and M. Eby, 2004: A method for obtaining pre-
Thorpe, R.B., R.A. Wood, and J.F.B. Mitchell, 2004: The sensitivity of
twentieth century initial conditions for use in climate change studies.
the thermohaline circulation response to preindustrial and anthropogenic
Clim. Dyn., 23, 327–339.
greenhouse gas forcing to the parameterisation of mixing across the
Stouffer, R.J., et al., 2006: Investigating the causes of the response of the
Greenland-Scotland ridge. Ocean Modelling, 7, 259–268.
thermohaline circulation to past and future climate changes. J. Clim., 19,
Thorpe, R.B., et al., 2001: Mechanisms determining the Atlantic
1365–1387.
thermohaline circulation response to greenhouse gas forcing in a non-
Stowasser, M., and K. Hamilton, 2006: Relationship between shortwave
ﬂ ux-adjusted coupled climate model. J. Clim., 14, 3102–3116.
cloud radiative forcing and local meteorological variables compared
Tiedtke, M., 1993: Representation of clouds in large-scale models. Mon.
in observations and several global climate models. J. Clim., 19, 4344–
Weather Rev., 121, 3040–3061
4359.
Timmermann, A., and H. Goosse, 2004: Is the wind stress forcing essential
Stowasser, M., K. Hamilton, and G.J. Boer, 2006: Local and global climate
for the meridional overturning circulation? Geophys. Res. Lett., 31(4),
feedbacks in models with differing climate sensitivity. J. Clim., 19, 193–
L04303, doi:10.1029/2003GL018777.
209.
Tomé, A., and P.M.A. Miranda, 2004: Piecewise linear ﬁ tting and trend
Stratton, R.A., and V.D. Pope, 2004: Modelling the climatology of storm
changing points of climate parameters. Geophys. Res. Lett., 31, L02207,
tracks - Sensitivity to resolution. In: The Second Phase of the Atmospheric
doi:10.1029/2003GL019100.
Model Intercomparison Project (AMIP2) [Gleckler, P. (ed.)]. Proceedings
Tompkins, A., 2002: A prognostic parameterization for the subgrid-scale
of the WCRP/WGNE Workshop, Toulouse, pp. 207-210.
variability of water vapor and clouds in large-scale models and its use to
Stuber, N., M. Ponater, and R. Sausen, 2001: Is the climate sensitivity to
diagnose cloud cover. J. Atmos. Sci., 59, 1917–1942.
ozone perturbations enhanced by stratospheric water vapor feedback?
Tompkins, A.M., and G.C. Craig, 1999: Sensitivity of tropical convection
Geophys. Res. Lett., 28, doi:10.1029/2001GL013000.
to sea surface temperature in the absence of large-scale ﬂ ow. J. Clim.,
Stuber, N., M. Ponater, and R. Sausen, 2005: Why radiative forcing might
12, 462–476.
fail as a predictor of climate change. Clim. Dyn., 24, 497–510.
Toyota, T., et al., 2004: Thickness dis tribution, texture and stratigraphy, and
Sud, Y.C., and G.K. Walker, 1999: Microphysics of clouds with the relaxed
a simple probabilistic model for dynamical thick ening of sea ice in the
Arakawa-Schubert Cumulus Scheme (McRAS). Part I: Design and
southern Sea of Okhotsk. J. Geophys. Res., 109, C06001, doi:10.1029/
evaluation with GATE Phase III data. J. Atmos. Sci., 56, 3196–3220.
2003JC002090.
Sugi, M., A. Noda, and N. Sato, 2002: Inﬂ uence of the global warming on
Trenberth, K.E., and J.M. Caron, 2001: Estimates of meridional atmosphere
tropical cyclone climatology: An experiment with the JMA global model.
and ocean heat transports. J. Clim., 14, 3433–3443.
J. Meteorol. Soc. Japan, 80, 249–272.
Trenberth, K.E., J. Fasullo, and L. Smith, 2005: Trends and variability in
Sun, D.-Z., and I.M. Held, 1996: A comparison of modeled and observed
column-integrated atmospheric water vapour. Clim. Dyn., 24, 741–758.
relationships between interannual variations of water vapor and
Trenberth, K.E., D.P. Stepaniak, J.W. Hurrel, and M. Fiorino, 2001: Quality
temperature. J. Clim., 9, 665–675.
of re-analyses in the tropics. J. Clim., 14, 1499–1510.
Sun, D.-Z., C. Covey, and R.S. Lindzen, 2001: Vertical correlations of
Trenberth, K.E., et al., 1998: Progress during TOGA in understanding
water vapor in GCMs. Geophys. Res. Lett., 28, 259–262.
and modeling global teleconnection associated with tropical sea surface
Sun, Y., S. Solomon, A. Dai, and R. Portmann, 2006: How often does it
temperatures. J. Geophys. Res., 103, 14291–14324.
rain? J. Clim., 19, 916–934.
Trigo, R.M., I.F. Trigo, C.C. DaCamra, and T.J. Osborn, 2004: Climate
Suzuki, T., et al., 2005: Projection of future sea level and its variability
impact of the European winter blocking episodes from the NCEP/NCAR
in a high-resolution climate model: Ocean processes and Greenland
reanalyses. Clim. Dyn., 23, 17–28.
and Antarctic ice-melt contributions. Geophys. Res. Lett., 32, L19706,
Tselioudis, G., and W.B. Rossow, 1994: Global, multiyear variations of
doi:10.1029/2005GL023677.
optical-thickness with temperature in low and cirrus clouds. Geophys.
Takahashi, M., 1996: Simulation of the stratospheric quasi-biennial
Res. Lett., 21, 2211–2214
oscillation using a general circulation model. Geophys. Res. Lett., 23,
Tselioudis, G., and W.B. Rossow, 2006: Climate feedback implied by
661–664.
observed radiation and precipitation changes with midlatitude storm
Takahashi, M., 1999: The ﬁ rst realistic simulation of the stratospheric
strength and frequency. Geophys. Res. Lett., 33, L02704, doi:10.1029/
quasi-biennial oscillation in a general circulation model. Geophys. Res.
2005GL024513.
Lett., 26, 1307–1310.
Tselioudis, G., Y.-C. Zhang, and W.R. Rossow, 2000: Cloud and radiation
Takemura, T., et al., 2002: Single scattering albedo and radiative forcing of
variations associated with northern midlatitude low and high sea level
various aerosol species with a global three-dimensional model. J. Clim.,
pressure regimes. J. Clim., 13, 312–327.
15, 333–352.
Tsushima, Y., A. Abe-Ouchi, and S. Manabe, 2005: Radiative damping
Takemura, T., et al., 2005: Simulation of climate response to aerosol direct
of annual variation in global mean surface temperature: Comparison
and indirect effects with aerosol transport-radiation model. J. Geophys.
between observed and simulated feedback. Clim. Dyn., 24, 591–597,
Res., 110, D02202, doi:10.1029/2004JD005029.
doi:10.1007/s00382-005-0002-y.
Tang, Y.M., and M.J. Roberts, 2005: The impact of a bottom boundary
Tsushima, Y., et al., 2006: Importance of the mixed-phase cloud
layer scheme on the North Atlantic Ocean in a global coupled climate
distribution in the control climate for assessing the response of clouds to
model. J. Phys. Oceanogr., 35(2), 202–217.
carbon dioxide increase: a multi-model study. Clim. Dyn., 27, 113–126,
Terray, L., S. Valcke, and A. Piacentini, 1998: OASIS 2.2 Guide and
doi:10.1007/s00382-006-0127-7.
Reference Manual. Technical Report TR/CMGC/98-05, Centre Europeen
Turner, A.G., P.M. Inness and J.M. Slingo, 2005: The role of the basic state
de Recherche et de Formation Avancée en Calcul Scientiﬁ que, Toulouse,
in monsoon prediction. Q. J. R. Meteorol. Soc., 131, 781–804.
France.
Uppala, S.M., et al., 2005: The ERA-40 Reanalysis. Q. J. R. Meteorol.
Thompson, C.J., and D.S. Battisti, 2001: A linear stochastic dynamical
Soc., 131, 2961–3012, doi:10.1256/qj.04.176.
model of ENSO. Part II: Analysis. J. Clim., 14, 445–466.
660

Chapter 8
Climate Models and Their Evaluation
Valcke, S., E. Guilyardi, and C. Larsson, 2006: PRISM and ENES:
Warrach, K., H.T. Mengelkamp, and E. Raschke, 2001: Treatment of
A European approach to Earth system modelling. Concurrency and
frozen soil and snow cover in the land surface model SEWAB. Theor.
Computation: Practice and Experience, 18(2), 247–262.
Appl. Climatol., 69(1–2), 23–37.
Van Oldenborgh, G.J., S.Y. Philip, and M. Collins, 2005: El Nino in a
Washington, W.M., et al., 2000: Parallel Climate Model (PCM) control and
changing climate: a multi-model study. Ocean Sci., 1, 81–95.
transient simulations. Clim. Dyn., 16, 755–774.
Vallis, G.K., E.P. Gerber, P.J. Kushner, and B.A. Cash, 2004: A mechanism
Watterson, I.G., 2001: Zonal wind vacillation and its interaction with
and simple dynamical model of the North Atlantic Oscillation and
the ocean: Implications for interannual variability and predictability. J.
Annular Modes. J. Atmos. Sci., 61, 264 –280.
Geophys. Res., 106, 23965–23975.
Vavrus, S., 2004: The impact of cloud feedbacks on Arctic climate under
Watterson, I.G., 2006: The intensity of precipitation during extratropical
greenhouse forcing. J. Clim., 17, 603–615.
cyclones in global warming simulations: a link to cyclone intensity? Tellus,
Vavrus, S., and S.P. Harrison, 2003: The impact of sea-ice dynamics on the
58A, 82–97.
Arctic climate system. Clim. Dyn., 20, 741–757.
Weare, B.C., 2004: A comparison of AMIP II model cloud layer properties
Vavrus, S., J.E. Walsh, W.L. Chapman, and D. Portis, 2006: The behavior of
with ISCCP D2 estimates. Clim. Dyn., 22, 281–292.
extreme cold air outbreaks under greenhouse warming. Int. J. Climatol.,
Weaver, A.J., O.A. Saenko, P.U. Clark, and J.X. Mitrovica, 2003:
26, 1133–1147.
Meltwater pulse 1A from Antarctica as a trigger of the Bølling-Allerød
Vellinga, M., and R.A. Wood, 2002: Global climate impacts of a collapse
warm interval. Science, 299, 1709–1713.
of the Atlantic thermohaline circulation. Clim. Change, 54, 251–267.
Weaver, A.J., et al., 2001: The UVic Earth System Climate Model: Model
Vellinga, M., R.A.Wood, and J.M. Gregory, 2002: Processes governing the
description, climatology and application to past, present and future
recovery of a perturbed thermohaline circulation in HadCM3. J. Clim.,
climates. Atmos.-Ocean, 39, 361–428.
15, 764–780.
Webb, M., C. Senior, S. Bony, and J.-J. Morcrette, 2001: Combining ERBE
Verseghy, D.L., N.A. McFarlane, and M. Lazare, 1993: A Canadian land
and ISCCP data to assess clouds in the Hadley Centre ECMWF and LMD
surface scheme for GCMs: II. Vegetation model and coupled runs. Int. J.
atmospheric climate models. Clim. Dyn., 17, 905–922.
Climatol., 13, 347–370.
Webb, M.J., et al., 2006: On the contribution of local feedback mechanisms
Visbeck, M., J. Marshall, T. Haine, and M. Spall, 1997: Speciﬁ cation of
to the range of climate sensitivity in two GCM ensembles. Clim. Dyn.,
eddy transfer coefﬁ cients in coarse-resolution ocean circulation models.
27, 17–38.
J. Phys. Oceanogr., 27, 381–402.
Wentz, F.J., and M. Schabel, 2000: Precise climate monitoring using
Volodin, E.M., 2004: Relation between the global-warming parameter and
complementary satellite data sets. Nature, 403, 414–416.
the heat balance on the Earth’s surface at increased contents of carbon
Wigley, T.M.L., and S.C.B. Raper, 1992: Implications for climate and sea
dioxide. Izv. Atmos. Ocean. Phys., 40, 269–275.
level of revised IPCC emissions scenarios. Nature, 357, 293–300.
Volodin, E.M., and V.N. Lykossov, 1998: Parameterization of heat and
Wigley, T.M.L., and S.C.B. Raper, 2001: Interpretation of high projections
moisture processes in the soil-vegetation system: 1. Formulation and
for global-mean warming. Science, 293, 451–454.
simulations based on local observational data. Izv. Atmos. Ocean. Phys.,
Wild, M., 2005: Solar radiation budgets in atmospheric model
34(4), 453–465.
intercomparisons from a surface perspective. Geophys. Res. Lett., 32,
Volodin, E.M., and N.A. Diansky, 2004: El-Niño reproduction in a coupled
doi:10.1029/2005GL022421.
general circulation model of atmosphere and ocean. Russ. Meteorol.
Wild, M., C.N. Long, and A. Ohmura, 2006: Evaluation of clear-sky
Hydrol., 12, 5–14.
solar ﬂ uxes in GCMs participating in AMIP and IPCC-AR4 from
Waliser, D.E., K.M. Lau, and J.H. Lim, 1999: The inﬂ uence of coupled
a surface perspective. J. Geophys. Res., 111, D01104, doi:10.1029/
sea surface temperatures on the Madden–Julian oscillation: A model
2005JD006118.
perturbation experiment. J. Atmos. Sci., 56, 333–358.
Wild, M., et al., 2001: Downward longwave radiation in General Circulation
Wallace, J.M., Y. Zhang, and L. Bajuk, 1996: Interpretation of interdecadal
Models. J. Clim., 14, 3227–3239.
trends in Northern Hemisphere surface air temperature. J. Clim., 9, 249–
Williams, K.D., M.A. Ringer, and C.A. Senior, 2003: Evaluating the cloud
259.
response to climate change and current climate variability. Clim. Dyn.,
Walsh, J.E., et al., 2002: Comparison of Arctic climate simulations by
20(7–8), 705–721.
uncoupled and coupled global models. J. Clim., 15, 1429–1446.
Williams, K.D., et al., 2006: Evaluation of a component of the cloud
Walsh, K.J.E., K.C. Nguyen and J.L. McGregor, 2004: Fine-resolution
response to climate change in an intercomparison of climate models.
regional climate model simulations of the impact of climate change on
Clim. Dyn., 26, 145–165.
tropical cyclones near Australia. Clim. Dyn., 22, 47–56.
Williamson, D.L., et al., 2005: Moisture and temperature balances at the
Wang, B., et al., 2004: Design of a new dynamical core for global
Atmospheric Radiation Measurement Southern Great Plains Site in
atmospheric models based on some efﬁ cient numerical methods. Science
forecasts with the Community Atmosphere Model (CAM2). J. Geophys.
in China, Ser. A, 47 Suppl., 4–21.
Res., 110, D15S16, doi:10.1029/2004JD00510.
Wang, G.L., and E.A.B. Eltahir, 2000: Ecosystem dynamics and the Sahel
Winton, M., 2000: A reformulated three-layer sea ice model. J. Atmos.
drought. Geophys. Res. Lett., 27, 795–798.
Ocean. Technol., 17(4), 525–531.
Wang, J., H.L. Cole, and D.J. Carlson, 2001: Water vapor variability in the
Winton, M., 2006a: Surface albedo feedback estimates for the AR4 climate
tropical western Paciﬁ c from 20-year radiosonde data. Adv. Atmos. Sci.,
models. J. Clim., 19, 359–365.
18(5), 752–766.
Winton, M., 2006b: Ampliﬁ ed Arctic climate change: what does surface
Wang, L.R., and M. Ikeda, 2004: A Lagrangian description of sea ice
albedo feedback have to do with it? Geophys. Res. Lett., 33, L03701,
dynamics using the ﬁ nite element method. Ocean Modelling, 7, 21–38.
doi:10.1029/2005GL025244.
Wang, S., R.F. Grant, D.L. Verseghy, and T.A. Black, 2002: Modelling
Winton, M., R. Hallberg, and A. Gnanadesikan, 1998: Simulation of
carbon dynamics of boreal forest ecosystems using the Canadian land
density-driven frictional downslope ﬂ ow in z-coordinate ocean models.
surface scheme. Clim. Change, 55, 451–477.
J. Phys. Oceanogr., 28, 2163–2174.
Wang, W., and M. Schlesinger, 1999: The dependence on convection
Wittenberg, A.T., A. Rosati, N.-C. Lau, and J.J. Ploshay, 2006: GFDL’s
parameterization of the tropical intraseasonal oscillation simulated by the
CM2 global coupled climate models, Part 3: Tropical Paciﬁ c climate and
UIUC 11-layer atmospheric GCM. J. Clim., 12, 1423–1457.
ENSO. J. Clim., 19, 698–722.
Wang, X.L.L., V.R. Swai, and F.W. Zwiers, 2006: Climatology and changes
Wolff, J.-O., E. Maier-Reimer, and S. Lebutke, 1997: The Hamburg Ocean
of extratropical cyclone activity: Comparison of ERA-40 with NCEP-
Primitive Equation Model. DKRZ Technical Report No. 13, Deutsches
NCAR reanalysis for 1958-2001. J. Clim., 19, 3145–3166.
KlimaRechenZentrum, Hamburg, Germany, 100 pp., http://www.mad.
zmaw.de/Pingo/reports/ReportNo.13.pdf.
661

Climate Models and Their Evaluation
Chapter 8
Wood, R.A., A.B. Keen, J.F.B. Mitchell, and J.M. Gregory, 1999:
Yu, Y., Z. Zhang, and Y. Guo, 2004: Global coupled ocean-atmosphere
Changing spatial structure of the thermohaline circulation in response to
general circulation models in LASG/IAP. Adv. Atmos. Sci., 21, 444–455.
atmospheric CO2 forcing in a climate model. Nature, 399, 572–575.
Yu, Y., R. Yu, X. Zhang, and H. Liu, 2002: A ﬂ exible global coupled climate
Wright, D.G., and T.F. Stocker, 1992: Sensitivities of a zonally averaged
model. Adv. Atmos. Sci., 19(1), 169–190.
global ocean circulation model. J. Geophys. Res., 97, 12707–12730.
Yukimoto, S., and A. Noda, 2003: Improvements of the Meteorological
Wright, D.G., and T.F. Stocker, 1993: Younger Dryas experiments. In: Ice
Research Institute Global Ocean-Atmosphere Coupled GCM (MRI-
in the Climate System, NATO ASI Series, I12 [Peltier, R. (ed.)]. Springer-
GCM2) and its Climate Sensitivity. CGER’s Supercomputing Activity
Verlag, London, pp. 395–416.
Report, National Institute for Environmental Studies, Ibaraki, Japan.
Wu, P., R.A. Wood, and P. Stott, 2005: Human inﬂ uence on increasing
Yukimoto, S., et al., 2001: The new Meteorological Research Institute
Arctic river discharges. Geophys. Res. Lett., 32, L02703, doi:10.1029/
global ocean-atmosphere coupled GCM (MRI-CGCM2)--Model climate
2004GL021570.
and variability. Papers in Meteorology and Geophysics, 51, 47–88.
Wu, Q., and D.M. Straus, 2004a: On the existence of hemisphere-wide
Zhang, C., 2005: Madden-Julian Oscillation. Rev. Geophys., 43, RG2003,
climate variations. J. Geophys. Res., 109, D06118, doi:10.1029/
doi:10.1029/2004RG000158.
2003JD004230.
Zhang, C., B. Mapes, and B.J. Soden, 2003: Bimodality of water vapour.
Wu, Q., and D.M. Straus, 2004b: AO, COWL, and observed climate trends.
Q. J. R. Meteorol. Soc., 129, 2847–2866.
J. Clim., 17, 2139–2156.
Zhang, J., and D. Rothrock, 2001: A thickness and enthalpy distribution
Wunsch, C., 2002: What is the thermohaline circulation? Science, 298,
sea-ice model. J. Phys. Oceanogr., 31, 2986–3001.
1179–1180.
Zhang, J., and D. Rothrock, 2003: Modeling global sea ice with a thickness
Wyant, M.C., et al., 2006: A comparison of low-latitude cloud properties
and enthalpy distribution model in generalized curvilinear coordinates.
and their response to climate change in three US AGCMs sorted into
Mon. Weather Rev., 131, 845–861.
regimes using mid-tropospheric vertical velocity. Clim. Dyn., 27, 261–
Zhang, M., 2004: Cloud-climate feedback: how much do we know? In:
279.
Observation, Theory, and Modeling of Atmospheric Variability, World
Xie, P., and P.A. Arkin, 1997: Global precipitation: A 17-year monthly
Scientiﬁ c Series on Meteorology of East Asia, Vol. 3 [Zhu et al. (eds.)].
analysis based on gauge observations, satellite estimates, and numerical
World Scientiﬁ c Publishing Co., Singapore, 632 pp.
model outputs. Bull. Am. Meteorol. Soc., 78, 2539–2558.
Zhang, M.H., R.D. Cess, J.J. Hack, and J.T. Kiehl, 1994: Diagnostic study
Xie, S.-P., W.T. Liu, Q. Liu and M. Nonaka, 2001: Far-reaching effects of
of climate feedback processed in atmospheric general circulation models.
the Hawaiian Islands on the Paciﬁ c ocean-atmosphere system. Science,
J. Geophys. Res., 99, 5525–5537.
292, 2057–2060.
Zhang, M.H., et al., 2005: Comparing clouds and their seasonal variations
Xu, Y., et al., 2005: Detection of climate change in the 20th century by
in 10 atmospheric general circulation models with satellite measurements.
the NCC T63. Acta Meteorol. Sin., Special Report on Climate Change,
J. Geophys. Res., 110, D15S02, doi:10.1029/2004JD005021.
4, 1–15.
Zhang, X., and J.E. Walsh, 2006: Toward a seasonally ice-covered Arctic
Yang, G.Y., and J. Slingo, 2001: The diurnal cycle in the tropics. Mon.
Ocean: scenarios from the IPCC AR4 model simulations. J. Clim., 19,
Weather Rev., 129, 784–801.
1730–1747.
Yang, G.Y., B. Hoskins, and J. Slingo, 2003: Convectively coupled
Zhang, Y., W. Maslowski, and A.J. Semtner, 1999: Impacts of mesoscale
equatorial waves: A new methodology for identifying wave structures in
ocean currents on sea ice in high-resolution Arctic ice and ocean
observational data. J. Atmos. Sci., 60, 1637–1654.
simulations. J. Geophys. Res., 104(C8), 18409–18429.
Yao, M.-S., and A. Del Genio, 2002: Effects of cloud parameterization on
Zhu, Y., R.E. Newell, and W.G. Read, 2000: Factors controlling upper-
the simulation of climate changes in the GISS GCM. Part II: Sea surface
troposphere water vapour. J. Clim., 13, 836–848.
temperature and cloud feedbacks. J. Clim., 15, 2491–2503.
Yeh, P. J.-F., and E.A.B. Eltahir, 2005: Representation of water table
dynamics in a land surface scheme. Part 1. Model development. J. Clim.,
18, 1861–1880.
Yeh, S.-W., and B.P. Kirtman, 2004: Decadal North Paciﬁ c sea surface
temperature variability and the associated global climate anomalies
in a coupled GCM. J. Geophys. Res., 109, D20113, doi:10.1029/
2004JD004785.
Yin, H., 2005: A consistent poleward shift of the storm tracks in simulations
of 21st century climate. Geophys. Res. Lett., 32, L18701, doi:10.1029/
2005GL023684.
Yiou, P., and M. Nogaj, 2004: Extreme climatic events and weather
regimes over the North Atlantic: When and where? Geophys. Res. Lett.,
31, doi:10.1029/2003GL019119.
Yokohata, T., et al., 2005: Climate response to volcanic forcing: Validation
of climate sensitivity of a coupled atmosphere-ocean general circulation
model. Geophys. Res. Lett., 32, L21710, doi:10.1029/2005GL023542.
Yoshimura, J., M. Sugi, and A. Noda, 2006: Inﬂ uence of greenhouse
warming on tropical cyclone frequency. J. Meteorol. Soc. Japan, 84,
405–428.
Yoshizaki, M., et al., 2005: Changes of Baui (Mei-yu) frontal activity in the
global warming climate simulated by a non-hydrostatic regional model.
Scientiﬁ c Online Letters on the Atmosphere, 1, 25–28.
Yu, Y., and X. Zhang, 2000: Coupled schemes of ﬂ ux adjustments of the
air and sea. In: Investigations on the Model System of the Short-Term
Climate Predictions [Ding, Y., et al. (eds.)]. China Meteorological Press,
Beijing, China, pp. 201–207 (in Chinese).
662