David A. Stainforth

Quantification of modelling uncertainties in a large ensemble of climate change simulations.

Comprehensive global climate models are the only tools that account for the complex set of processes which will determine future climate change at both a global and regional level. Planners are typically faced with a wide range of predicted changes from different models of unknown relative quality, owing to large but unquantified uncertainties in the modelling process. Here we report a systematic attempt to determine the range of climate changes consistent with these uncertainties, based on a 53-member ensemble of model versions constructed by varying model parameters. We estimate a probability density function for the sensitivity of climate to a doubling of atmospheric carbon dioxide levels, and obtain a 5–95 per cent probability range of 2.4–5.4 °C. Our probability density function is constrained by objective estimates of the relative reliability of different model versions, the choice of model parameters that are varied and their uncertainty ranges, specified on the basis of expert advice. Our ensemble produces a range of regional changes much wider than indicated by traditional methods based on scaling the response patterns of an individual simulation.

Uncertainty in predictions of the climate response to rising levels of greenhouse gases.

The range of possibilities for future climate evolution needs to be taken into account when planning climate change mitigation and adaptation strategies. This requires ensembles of multi-decadal simulations to assess both chaotic climate variability and model response uncertainty. Statistical estimates of model response uncertainty, based on observations of recent climate change, admit climate sensitivities—defined as the equilibrium response of global mean temperature to doubling levels of atmospheric carbon dioxide—substantially greater than 5 K. But such strong responses are not used in ranges for future climate change because they have not been seen in general circulation models. Here we present results from the ‘climateprediction.net’ experiment, the first multi-thousand-member grand ensemble of simulations using a general circulation model and thereby explicitly resolving regional details. We find model versions as realistic as other state-of-the-art climate models but with climate sensitivities ranging from less than 2 K to more than 11 K. Models with such extreme sensitivities are critical for the study of the full range of possible responses of the climate system to rising greenhouse gas levels, and for assessing the risks associated with specific targets for stabilizing these levels.

Confidence, uncertainty and decision-support relevance in climate predictions

Over the last 20 years, climate models have been developed to an impressive level of complexity. They are core tools in the study of the interactions of many climatic processes and justifiably provide an additional strand in the argument that anthropogenic climate change is a critical global problem. Over a similar period, there has been growing interest in the interpretation and probabilistic analysis of the output of computer models; particularly, models of natural systems. The results of these areas of research are being sought and utilized in the development of policy, in other academic disciplines, and more generally in societal decision making. Here, our focus is solely on complex climate models as predictive tools on decadal and longer time scales. We argue for a reassessment of the role of such models when used for this purpose and a reconsideration of strategies for model development and experimental design. Building on more generic work, we categorize sources of uncertainty as they relate to this specific problem and discuss experimental strategies available for their quantification. Complex climate models, as predictive tools for many variables and scales, cannot be meaningfully calibrated because they are simulating a never before experienced state of the system; the problem is one of extrapolation. It is therefore inappropriate to apply any of the currently available generic techniques which utilize observations to calibrate or weight models to produce forecast probabilities for the real world. To do so is misleading to the users of climate science in wider society. In this context, we discuss where we derive confidence in climate forecasts and present some concepts to aid discussion and communicate the state-of-the-art. Effective communication of the underlying assumptions and sources of forecast uncertainty is critical in the interaction between climate science, the impacts communities and society in general.

Realistic quasi-biennial oscillations in a simulation of the global climate

The tropical quasi-biennial oscillation is one of the most spectacular examples of low frequency variability observed in the Earths atmosphere, yet the oscillation is noted for its absence from numerical simulations of the global climate. Recent studies suggest that much of the required wave forcing for the oscillation is likely to come from buoyancy (gravity) waves [Sato and Dunkerton, 1997; Dunkerton, 1997] that are not well resolved in the numerical models currently used for climate prediction and global weather forecasting. Here we show that when the effects of these missing waves are parametrized in a comprehensive numerical model of the atmosphere, the simulation of the climate is improved by the generation of a realistic quasi-biennial oscillation in the stratosphere.

Constraining Climate Sensitivity from the Seasonal Cycle in Surface Temperature

Abstract The estimated range of climate sensitivity has remained unchanged for decades, resulting in large uncertainties in long-term projections of future climate under increased greenhouse gas concentrations. Here the multi-thousand-member ensemble of climate model simulations from the climateprediction.net project and a neural network are used to establish a relation between climate sensitivity and the amplitude of the seasonal cycle in regional temperature. Most models with high sensitivities are found to overestimate the seasonal cycle compared to observations. A probability density function for climate sensitivity is then calculated from the present-day seasonal cycle in reanalysis and instrumental datasets. Subject to a number of assumptions on the models and datasets used, it is found that climate sensitivity is very unlikely (5% probability) to be either below 1.5–2 K or above about 5–6.5 K, with the best agreement found for sensitivities between 3 and 3.5 K. This range is narrower than most prob...

Issues in the interpretation of climate model ensembles to inform decisions

There is a scientific consensus regarding the reality of anthropogenic climate change. This has led to substantial efforts to reduce atmospheric greenhouse gas emissions and thereby mitigate the impacts of climate change on a global scale. Despite these efforts, we are committed to substantial further changes over at least the next few decades. Societies will therefore have to adapt to changes in climate. Both adaptation and mitigation require action on scales ranging from local to global, but adaptation could directly benefit from climate predictions on regional scales while mitigation could be driven solely by awareness of the global problem; regional projections being principally of motivational value. We discuss how recent developments of large ensembles of climate model simulations can be interpreted to provide information on these scales and to inform societal decisions. Adaptation is most relevant as an influence on decisions which exist irrespective of climate change, but which have consequences on decadal time-scales. Even in such situations, climate change is often only a minor influence; perhaps helping to restrict the choice of ‘no regrets’ strategies. Nevertheless, if climate models are to provide inputs to societal decisions, it is important to interpret them appropriately. We take climate ensembles exploring model uncertainty as potentially providing a lower bound on the maximum range of uncertainty and thus a non-discountable climate change envelope. An analysis pathway is presented, describing how this information may provide an input to decisions, sometimes via a number of other analysis procedures and thus a cascade of uncertainty. An initial screening is seen as a valuable component of this process, potentially avoiding unnecessary effort while guiding decision makers through issues of confidence and robustness in climate modelling information. Our focus is the usage of decadal to centennial time-scale climate change simulations as inputs to decision making, but we acknowledge that robust adaptation to the variability of present day climate encourages the development of less vulnerable systems as well as building critical experience in how to respond to climatic uncertainty.

Analyzing the climate sensitivity of the HadSM3 climate model using ensembles from different but related experiments

Global climate models (GCMs) contain imprecisely defined parameters that account, approximately, for subgrid-scale physical processes. The response of a GCM to perturbations in its parameters, which is crucial for quantifying uncertainties in simulations of climate change, can—in principle—be assessed by simulating the GCM many times. In practice, however, such “perturbed physics” ensembles are small because GCMs are so expensive to simulate. Statistical tools can help in two ways. First, they can be used to combine ensembles from different but related experiments, increasing the effective number of simulations. Second, they can be used to describe the GCM’s response in ways that cannot be extracted directly from the ensemble(s). The authors combine two experiments to learn about the response of the Hadley Centre Slab Climate Model version 3 (HadSM3) climate sensitivity to 31 model parameters. A Bayesian statistical framework is used in which expert judgments are required to quantify the relationship between the two experiments; these judgments are validated by detailed diagnostics. The authors identify the entrainment rate coefficient of the convection scheme as the most important single parameter and find that this interacts strongly with three of the large-scale-cloud parameters.

Association of parameter, software, and hardware variation with large-scale behavior across 57,000 climate models

In complex spatial models, as used to predict the climate response to greenhouse gas emissions, parameter variation within plausible bounds has major effects on model behavior of interest. Here, we present an unprecedentedly large ensemble of >57,000 climate model runs in which 10 parameters, initial conditions, hardware, and software used to run the model all have been varied. We relate information about the model runs to large-scale model behavior (equilibrium sensitivity of global mean temperature to a doubling of carbon dioxide). We demonstrate that effects of parameter, hardware, and software variation are detectable, complex, and interacting. However, we find most of the effects of parameter variation are caused by a small subset of parameters. Notably, the entrainment coefficient in clouds is associated with 30% of the variation seen in climate sensitivity, although both low and high values can give high climate sensitivity. We demonstrate that the effect of hardware and software is small relative to the effect of parameter variation and, over the wide range of systems tested, may be treated as equivalent to that caused by changes in initial conditions. We discuss the significance of these results in relation to the design and interpretation of climate modeling experiments and large-scale modeling more generally.

Adaptation to Global Warming: Do Climate Models Tell Us What We Need to Know?

Scientific experts have confirmed that anthropogenic warming is underway, and some degree of adaptation is now unavoidable. However, the details of impacts on the scale of climate change at which humans would have to prepare for and adjust to them are still the subject of considerable research, inquiry, and debate. Planning for adaptation requires information on the scale over which human organizations and institutions have authority and capacity, yet the general circulation models lack forecasting skill at these scales, and attempts to “downscale” climate models are still in the early stages of development. Because we do not know what adaptations will be required, we cannot say whether they will be harder or easier—more expensive or less—than emissions control. Whatever improvements in regional predictive capacity may come about in the future, the lack of current predictive capacity on the relevant scale is a strong argument for why we must both control greenhouse gas emissions and prepare to adapt.

Constraints on Model Response to Greenhouse Gas Forcing and the Role of Subgrid-Scale Processes

A climate model emulator is developed using neural network techniques and trained with the data from the multithousand-member climateprediction.net perturbed physics GCM ensemble. The method recreates nonlinear interactions between model parameters, allowing a simulation of a much larger ensemble that explores model parameter space more fully. The emulated ensemble is used to search for models closest to observations over a wide range of equilibrium response to greenhouse gas forcing. The relative discrepancies of these models from observations could be used to provide a constraint on climate sensitivity. The use of annual mean or seasonal differences on top-of-atmosphere radiative fluxes as an observational error metric results in the most clearly defined minimum in error as a function of sensitivity, with consistent but less well-defined results when using the seasonal cycles of surface temperature or total precipitation. The model parameter changes necessary to achieve different values of climate sensitivity while minimizing discrepancy from observation are also considered and compared with previous studies. This information is used to propose more efficient parameter sampling strategies for future ensembles.