How well must climate models agree with observations?

The usefulness of a climate-model simulation cannot be inferred solely from its degree of agreement with observations. Instead, one has to consider additional factors such as internal variability, the tuning of the model, observational uncertainty, the temporal change in dominant processes or the uncertainty in the forcing. In any model-evaluation study, the impact of these limiting factors on the suitability of specific metrics must hence be examined. This can only meaningfully be done relative to a given purpose for using a model. I here generally discuss these points and substantiate their impact on model evaluation using the example of sea ice. For this example, I find that many standard metrics such as sea-ice area or volume only permit limited inferences about the shortcomings of individual models.

Download Full-text

Influence of Ocean and Atmosphere Components on Simulated Climate Sensitivities

Journal of Climate ◽

10.1175/jcli-d-12-00121.1 ◽

2013 ◽

Vol 26 (1) ◽

pp. 231-245 ◽

Cited By ~ 27

Author(s):

Michael Winton ◽

Alistair Adcroft ◽

Stephen M. Griffies ◽

Robert W. Hallberg ◽

Larry W. Horowitz ◽

...

Keyword(s):

Sea Ice ◽

Climate Models ◽

Climate Model ◽

Model Simulation ◽

Coupled Model ◽

Ocean Model ◽

Historical Simulation ◽

Model Version ◽

Salinity Response ◽

The Impact

Abstract The influence of alternative ocean and atmosphere subcomponents on climate model simulation of transient sensitivities is examined by comparing three GFDL climate models used for phase 5 of the Coupled Model Intercomparison Project (CMIP5). The base model ESM2M is closely related to GFDL’s CMIP3 climate model version 2.1 (CM2.1), and makes use of a depth coordinate ocean component. The second model, ESM2G, is identical to ESM2M but makes use of an isopycnal coordinate ocean model. The authors compare the impact of this “ocean swap” with an “atmosphere swap” that produces the GFDL Climate Model version 3 (CM3) by replacing the AM2 atmospheric component with AM3 while retaining a depth coordinate ocean model. The atmosphere swap is found to have much larger influence on sensitivities of global surface temperature and Northern Hemisphere sea ice cover. The atmosphere swap also introduces a multidecadal response time scale through its indirect influence on heat uptake. Despite significant differences in their interior ocean mean states, the ESM2M and ESM2G simulations of these metrics of climate change are very similar, except for an enhanced high-latitude salinity response accompanied by temporarily advancing sea ice in ESM2G. In the ESM2G historical simulation this behavior results in the establishment of a strong halocline in the subpolar North Atlantic during the early twentieth century and an associated cooling, which are counter to observations in that region. The Atlantic meridional overturning declines comparably in all three models.

Download Full-text

Marginal ice zone fraction benchmarks sea ice and climate model skill

Nature Communications ◽

10.1038/s41467-021-22004-7 ◽

2021 ◽

Vol 12 (1) ◽

Author(s):

Christopher Horvat

Keyword(s):

Sea Ice ◽

Climate Models ◽

Climate Model ◽

Global Climate ◽

Internal Variability ◽

Arctic Sea Ice ◽

Global Climate Models ◽

Global Mean Temperature ◽

Marginal Ice Zone ◽

The Impact

AbstractGlobal climate models (GCMs) consistently underestimate the response of September Arctic sea-ice area (SIA) to warming. Modeled SIA losses are highly correlated to global mean temperature increases, making it challenging to gauge if improvements in modeled sea ice derive from improved sea-ice models or from improvements in forcing driven by other GCM components. I use a set of five large GCM ensembles, and CMIP6 simulations, to quantify GCM internal variability and variability between GCMs from 1979–2014, showing modern GCMs do not plausibly estimate the response of SIA to warming in all months. I identify the marginal ice zone fraction (MIZF) as a metric that is less correlated to warming, has a response plausibly simulated from January–September (but not October–December), and has highly variable future projections across GCMs. These qualities make MIZF useful for evaluating the impact of sea-ice model changes on past, present, and projected sea-ice state.

Download Full-text

Climate Model Evaluation in the Presence of Observational Uncertainty: Precipitation Indices over the Contiguous United States

Journal of Hydrometeorology ◽

10.1175/jhm-d-18-0230.1 ◽

2019 ◽

Vol 20 (7) ◽

pp. 1339-1357 ◽

Cited By ~ 9

Author(s):

Peter B. Gibson ◽

Duane E. Waliser ◽

Huikyo Lee ◽

Baijun Tian ◽

Elias Massoud

Keyword(s):

United States ◽

Model Evaluation ◽

Climate Models ◽

Climate Model ◽

Dynamical Downscaling ◽

Daily Precipitation ◽

Regional Climate ◽

Climate Model Evaluation ◽

Observational Uncertainty ◽

Precipitation Indices

Abstract Climate model evaluation is complicated by the presence of observational uncertainty. In this study we analyze daily precipitation indices and compare multiple gridded observational and reanalysis products with regional climate models (RCMs) from the North American component of the Coordinated Regional Climate Downscaling Experiment (NA-CORDEX) multimodel ensemble. In the context of model evaluation, observational product differences across the contiguous United States (CONUS) are also deemed nontrivial for some indices, especially for annual counts of consecutive wet days and for heavy precipitation indices. Multidimensional scaling (MDS) is used to directly include this observational spread into the model evaluation procedure, enabling visualization and interpretation of model differences relative to a “cloud” of observational uncertainty. Applying MDS to the evaluation of NA-CORDEX RCMs reveals situations of added value from dynamical downscaling, situations of degraded performance from dynamical downscaling, and the sensitivity of model performance to model resolution. On precipitation days, higher-resolution RCMs typically simulate higher mean and extreme precipitation rates than their lower-resolution pairs, sometimes improving model fidelity with observations. These results document the model spread and biases in daily precipitation extremes across the full NA-CORDEX model ensemble. The often-large divergence between in situ observations, satellite data, and reanalysis, shown here for CONUS, is especially relevant for data-sparse regions of the globe where satellite and reanalysis products are extensively relied upon. This highlights the need to carefully consider multiple observational products when evaluating climate models.

Download Full-text

The Impact of Poleward Moisture and Sensible Heat Flux on Arctic Winter Sea Ice Variability*

Journal of Climate ◽

10.1175/jcli-d-15-0074.1 ◽

2015 ◽

Vol 28 (13) ◽

pp. 5030-5040 ◽

Cited By ~ 69

Author(s):

Hyo-Seok Park ◽

Sukyoung Lee ◽

Seok-Woo Son ◽

Steven B. Feldstein ◽

Yu Kosaka

Keyword(s):

Sea Ice ◽

Climate Model ◽

Model Simulation ◽

Early Stage ◽

Sensible Heat Flux ◽

The Arctic ◽

Sea Ice Concentration ◽

Sea Ice Thickness ◽

Surface Warming ◽

The Impact

Abstract The surface warming in recent decades has been most rapid in the Arctic, especially during the winter. Here, by utilizing global reanalysis and satellite datasets, it is shown that the northward flux of moisture into the Arctic during the winter strengthens the downward infrared radiation (IR) by 30–40 W m−2 over 1–2 weeks. This is followed by a decline of up to 10% in sea ice concentration over the Greenland, Barents, and Kara Seas. A climate model simulation indicates that the wind-induced sea ice drift leads the decline of sea ice thickness during the early stage of the strong downward IR events, but that within one week the cumulative downward IR effect appears to be dominant. Further analysis indicates that strong downward IR events are preceded several days earlier by enhanced convection over the tropical Indian and western Pacific Oceans. This finding suggests that sea ice predictions can benefit from an improved understanding of tropical convection and ensuing planetary wave dynamics.

Download Full-text

How does internal variability influence the ability of CMIP5 models to reproduce the recent trend in Southern Ocean sea ice extent?

The Cryosphere Discussions ◽

10.5194/tcd-6-3539-2012 ◽

2012 ◽

Vol 6 (5) ◽

pp. 3539-3573 ◽

Cited By ~ 5

Author(s):

V. Zunz ◽

H. Goosse ◽

F. Massonnet

Keyword(s):

Southern Ocean ◽

Sea Ice ◽

Climate Models ◽

Internal Variability ◽

Cmip5 Models ◽

Positive Trend ◽

External Forcing ◽

Sea Ice Extent ◽

The One ◽

The Impact

Abstract. Observations over the last 30 yr have shown that the sea ice extent in the Southern Ocean has slightly increased since 1979. Mechanisms responsible for this positive trend have not been well established yet and climate models are generally unable to simulate correctly this expansion. In this study, we focus on two related hypotheses that could explain the misrepresentation of the positive trend in sea ice extent by climate models: an unrealistic internal variability and an inadequate initialization of the system. For that purpose, we analyze the evolution of sea ice around the Antarctic simulated by 24 different general circulation models involved in the 5th Coupled Model Intercomparison Project (CMIP5). On the one hand, historical simulations, driven by external forcing and initialized without observations, are examined. They provide information about the mean state, the variability and the trend in sea ice extent simulated by each model. On the other hand, decadal prediction experiments, driven by external forcing and initialized with some observed fields, allow us to assess the impact of the representation of the observed initial state on the quality of model predictions. Our analyses show that CMIP5 models respond to the forcing, including the one induced by stratospheric ozone depletion, by reducing the sea ice cover in the Southern Ocean. Some simulations display an increase in sea ice extent. However, models strongly overestimate the variability of sea ice extent and the initialization methods currently used in models do not improve systematically the simulated trends in sea ice extent. On the basis of those results, a critical role of the internal variability in the observed increase in the sea ice extent in the Southern Ocean could not be ruled out but current models results appear inadequate to test more precisely this hypothesis.

Download Full-text

Seasonal transition dates can reveal biases in Arctic sea ice simulations

10.5194/tc-2020-81 ◽

2020 ◽

Author(s):

Abigail Smith ◽

Alexandra Jahn ◽

Muyin Wang

Keyword(s):

Sea Ice ◽

Annual Cycle ◽

Model Evaluation ◽

Climate Model ◽

Internal Variability ◽

Arctic Sea Ice ◽

Seasonal Transition ◽

Climate Model Evaluation ◽

Arctic Sea ◽

Break Up

Abstract. Arctic sea ice experiences a dramatic annual cycle, and seasonal ice loss and growth can be characterized by various metrics: melt onset, break-up, opening, freeze onset, freeze-up and closing. By evaluating a range of seasonal sea ice metrics, CMIP6 sea ice simulations can be evaluated in more detail than by using traditional metrics alone, such as sea ice area. We show that models capture the observed asymmetry in seasonal sea ice transitions, with spring ice loss taking about 1.5–2 months longer than fall ice growth. The largest impacts of internal variability are seen in the inflow regions of melt and freeze onset dates, but all metrics show pan-Arctic model spreads exceeding the internal variability. Through climate model evaluation in the context of both observations and internal variability, we show that biases in seasonal transition dates can compensate for other unrealistic aspects of simulated sea ice. In some models, this leads to September sea ice areas in agreement with observations for the wrong reasons.

Download Full-text

Seasonal transition dates can reveal biases in Arctic sea ice simulations

The Cryosphere ◽

10.5194/tc-14-2977-2020 ◽

2020 ◽

Vol 14 (9) ◽

pp. 2977-2997

Author(s):

Abigail Smith ◽

Alexandra Jahn ◽

Muyin Wang

Keyword(s):

Sea Ice ◽

Annual Cycle ◽

Model Evaluation ◽

Climate Model ◽

Internal Variability ◽

Arctic Sea Ice ◽

Seasonal Transition ◽

Ice Growth ◽

Climate Model Evaluation ◽

Arctic Sea

Abstract. Arctic sea ice experiences a dramatic annual cycle, and seasonal ice loss and growth can be characterized by various metrics: melt onset, breakup, opening, freeze onset, freeze-up, and closing. By evaluating a range of seasonal sea ice metrics, CMIP6 sea ice simulations can be evaluated in more detail than by using traditional metrics alone, such as sea ice area. We show that models capture the observed asymmetry in seasonal sea ice transitions, with spring ice loss taking about 1–2 months longer than fall ice growth. The largest impacts of internal variability are seen in the inflow regions for melt and freeze onset dates, but all metrics show pan-Arctic model spreads exceeding the internal variability range, indicating the contribution of model differences. Through climate model evaluation in the context of both observations and internal variability, we show that biases in seasonal transition dates can compensate for other unrealistic aspects of simulated sea ice. In some models, this leads to September sea ice areas in agreement with observations for the wrong reasons.

Download Full-text

A Cloudier Picture of Ice-Albedo Feedback in CMIP6 Models

Frontiers in Earth Science ◽

10.3389/feart.2021.769844 ◽

2021 ◽

Vol 9 ◽

Author(s):

Anne Sledd ◽

Tristan S. L’Ecuyer

Keyword(s):

Sea Ice ◽

Snow Cover ◽

Climate Models ◽

Climate Model ◽

Internal Variability ◽

Climate Forcing ◽

The Arctic ◽

Cmip5 Models ◽

Solar Absorption ◽

Albedo Feedback

Increased solar absorption is an important driver of Arctic Amplification, the interconnected set of processes and feedbacks by which Arctic temperatures respond more rapidly than global temperatures to climate forcing. The amount of sunlight absorbed in the Arctic is strongly modulated by seasonal ice and snow cover. Sea ice declines and shorter periods of seasonal snow cover in recent decades have increased solar absorption, amplifying local warming relative to the planet as a whole. However, this Arctic albedo feedback would be substantially larger in the absence of the ubiquitous cloud cover that exists throughout the region. Clouds have been observed to mask the effects of reduced surface albedo and slow the emergence of secular trends in net solar absorption. Applying analogous metrics to several models from the 6th Climate Model Intercomparison Project (CMIP6), we find that ambiguity in the influence of clouds on predicted Arctic solar absorption trends has increased relative to the previous generation of climate models despite better agreement with the observed albedo sensitivity to sea ice variations. Arctic albedo responses to sea ice loss are stronger in CMIP6 than in CMIP5 in all summer months. This agrees better with observations, but models still slightly underestimate albedo sensitivity to sea ice changes relative to observations. Never-the-less, nearly all CMIP6 models predict that the Arctic is now absorbing more solar radiation than at the start of the century, consistent with recent observations. In fact, many CMIP6 models simulate trends that are too strong relative to internal variability, and spread in predicted Arctic albedo changes has increased since CMIP5. This increased uncertainty can be traced to increased ambiguity in how clouds influence natural and forced variations in Arctic solar absorption. While nearly all CMIP5 models agreed with observations that clouds delay the emergence of forced trends, about half of CMIP6 models suggest that clouds accelerate their emergence from natural variability. Isolating atmospheric contributions to total Arctic reflection suggests that this diverging behavior may be linked to stronger Arctic cloud feedbacks in the latest generation of climate models.

Download Full-text

Do Climate Models Underestimate the Sensitivity of Northern Hemisphere Sea Ice Cover?

Journal of Climate ◽

10.1175/2011jcli4146.1 ◽

2011 ◽

Vol 24 (15) ◽

pp. 3924-3934 ◽

Cited By ~ 65

Author(s):

Michael Winton

Keyword(s):

Sea Ice ◽

Northern Hemisphere ◽

Climate Models ◽

Climate Model ◽

Ice Cover ◽

Natural Variability ◽

Global Temperature ◽

Temperature Trends ◽

Gfdl Model ◽

The Impact

Abstract The sensitivity of Northern Hemisphere sea ice cover to global temperature change is examined in a group of climate models and in the satellite-era observations. The models are found to have well-defined, distinguishable sensitivities in climate change experiments. The satellite-era observations show a larger sensitivity—a larger decline per degree of warming—than any of the models. To evaluate the role of natural variability in this discrepancy, the sensitivity probability density function is constructed based upon the observed trends and natural variability of multidecadal ice cover and global temperature trends in a long control run of the GFDL Climate Model, version 2.1 (CM2.1). This comparison shows that the model sensitivities range from about 1 to more than 2 pseudostandard deviations of the variability smaller than observations indicate. The impact of natural Atlantic multidecadal temperature trends (as simulated by the GFDL model) on the sensitivity distribution is examined and found to be minimal.

Download Full-text

Observations and Simulations of Meteorological Conditions over Arctic Thick Sea Ice in Late Winter during the Transarktika 2019 Expedition

Atmosphere ◽

10.3390/atmos12020174 ◽

2021 ◽

Vol 12 (2) ◽

pp. 174

Author(s):

Günther Heinemann ◽

Sascha Willmes ◽

Lukas Schefczyk ◽

Alexander Makshtas ◽

Vasilii Kustov ◽

...

Keyword(s):

Sea Ice ◽

Climate Models ◽

Climate Model ◽

Regional Climate ◽

Open Water ◽

The Arctic ◽

Late Winter ◽

Good Representation ◽

Hourly Temperature ◽

Ice Model

The parameterization of ocean/sea-ice/atmosphere interaction processes is a challenge for regional climate models (RCMs) of the Arctic, particularly for wintertime conditions, when small fractions of thin ice or open water cause strong modifications of the boundary layer. Thus, the treatment of sea ice and sub-grid flux parameterizations in RCMs is of crucial importance. However, verification data sets over sea ice for wintertime conditions are rare. In the present paper, data of the ship-based experiment Transarktika 2019 during the end of the Arctic winter for thick one-year ice conditions are presented. The data are used for the verification of the regional climate model COSMO-CLM (CCLM). In addition, Moderate Resolution Imaging Spectroradiometer (MODIS) data are used for the comparison of ice surface temperature (IST) simulations of the CCLM sea ice model. CCLM is used in a forecast mode (nested in ERA5) for the Norwegian and Barents Seas with 5 km resolution and is run with different configurations of the sea ice model and sub-grid flux parameterizations. The use of a new set of parameterizations yields improved results for the comparisons with in-situ data. Comparisons with MODIS IST allow for a verification over large areas and show also a good performance of CCLM. The comparison with twice-daily radiosonde ascents during Transarktika 2019, hourly microwave water vapor measurements of first 5 km in the atmosphere and hourly temperature profiler data show a very good representation of the temperature, humidity and wind structure of the whole troposphere for CCLM.

Download Full-text