Abstract
Convergence of model projections is often considered by climate scientists to be an important objective in so far as it may indicate the robustness of the models’ core hypotheses. Consequently, the range of climate projections from a multi-model ensemble, called “model spread”, is often expected to reduce as climate research moves forward. However, the successive Assessment Reports of the Intergovernmental Panel on Climate Change indicate no reduction in model spread, whereas it is indisputable that climate science has made improvements in its modelling. In this paper, after providing a detailed explanation of the situation, we describe an epistemological setting in which a steady (and even slightly increased) model spread is not doomed to be seen as negative, and is indeed compatible with a desirable evolution of climate models taken individually. We further argue that, from the perspective of collective progress, as far as the improvement of the products of a multi-model ensemble (e.g. means) is concerned, reduction of model spread is of lower priority than model independence.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
State-of-the-art GCMs minimally include components of the physical climate, i.e., atmosphere, ocean and their interactions, as well as external forcing induced by the Sun, volcanoes and human activities. They can be Coupled Atmosphere-Ocean General Circulation Models (AOGCMs) or Earth System Models (ESMs). Unlike the former, the latter include ice sheets and biogeochemical processes.
GCMs differ in the discretisation of the physical differential equations, in the grid resolution, in the parameterisations, in the possible inclusion of stochastic components, etc. Importantly, variations between models depend on their respective parameterisations, which are approximate descriptions of subgrid-scale processes (e.g. cloud processes) used when explicit representation of these processes requires very high computer power, or when understanding of these processes is simply lacking.
Nonetheless, the conclusions in this paper should extend to model spreads from ensembles of regional models, e.g. EURO-CORDEX ensembles.
As an illustration, CMIP5 puts forward four scenarios of anthropogenic forcing called “Representative Concentration Pathways” based on different assumptions about future global greenhouse-gas emissions. Twenty-three models contributed to CMIP5, and eighty-eight models are currently running for CMIP6.
In addition to the methodological problem we point out, this also raises concerns from the perspective of public communication. A steady model spread and the even more baffling slight increase may engender public doubt about the progress being made in climate modelling, in that model spread is usually interpreted as quantifying uncertainty.
We should note that an additional estimation of the uncertainty is in practice applied in the model spread based on expert judgments (using e.g. a Bayesian probabilistic methodology) since the multi-model ensemble is an “ensemble of opportunity” that imperfectly spans model uncertainty (e.g. Thompson et al., 2016).
Note that the evolution of climate models is intrinsically limited by the internal variability of the climate and constrained by the choice of emissions scenarios.
Interestingly, in Pichelli et al. (2021), the estimated ensemble means are more accurate, which, as we argue in the rest of the paper, is an important progress in climate modelling.
These authors discuss model fit, variety of evidence or consistency with background knowledge as criteria of confirmation, whereas we take the respective evolutions of model fit, variety of evidence and consistency with background knowledge as indicators of improvement.
Another lesson can be learned from the comparison between model spread and other indicators of improvement in climate modelling. When used without taking model spread into consideration, the other indicators are insufficient for two reasons. First, no climate model can be said to perform better that others in respect of every purpose. Second, the fact that a particular model improves its performance is not a sufficient indication that our general understanding of the climate system has improved.
As Parker (2011, 2013, 2018) argues, a challenge in climate science is to produce better ensembles by properly sampling the space of models. We therefore believe that, in this respect, the priority is to build a statistical sample of models that are independent of each other, in order to get notably better means.
Hence the model spread is sometimes interpreted as the “distance from the truth”.
There are at least three interpretations which can be assigned to model spread as a quantification of uncertainty. As Parker (2013) remarks, it can be seen as 1. “a lower bound on response uncertainty, indicating changes in climate that cannot yet be ruled out” (218), 2. “precise probabilities” (218) assigned to model projections, or 3. “interval probability specifications” (218), i.e. as specifications of the range of imperfect yet plausible projections.
As emphasised notably by Carrier and Lenhard (2019, 4) and supported by the scientific literature, the tails in those functions are currently not well estimated.
References
Abramowitz, G. (2017). Calibrating ensembles for model independence. Talk available online, last checked the 18. Sept. 2020. https://www.agci.org/lib/17s2/calibrating-ensembles-model-independencehttps://www.agci.org/lib/17s2/calibrating-ensembles-model-independence.
Abramowitz, G., Herger, N., Gutmann, E., Hammerling, D., Knutti, R., Leduc, M., Lorenz, R., Pincus, R., & Schmidt, G.A. (2018). Model dependence in multi-model climate ensembles: Weighting, sub-selection and out-of-sample testing. Earth System Dynamics Discussions, 1–20.
Baumberger, C., Knutti, R., & Hadorn, G.H. (2017). Building confidence in climate model projections: An analysis of inferences from fit. WIREs Climate Change, 8, e454.
Betz, G. (2015). Are climate models credible worlds? prospects and limitations of possibilistic climate prediction. European Journal for Philosophy of Science, 5(2), 5.
Brunner, L., Pendergrass, A.G., Lehner, F., Merrifield, A.L., Lorenz, R., & Knutti, R. (2020). Reduced global warming from cmip6 projections when weighting models by performance and independence. Earth System Dynamics, 11 (4), 995–1012.
Carrier, M., & Lenhard, J. (2019), Climate models: How to assess their reliability. International Studies in the Philosophy of Science.
Charney, J.G., et al. (1979). Carbon dioxide and climate: A scientific assessment : Report of an ad hoc study group on carbon dioxide and climate, woods hole, massachusetts, july 23-27, 1979 to the climate research board, assembly of mathematical and physical sciences, national research council. National Academy of Sciences: Available from Climate Research Board, http://books.google.com/books?id=cj0rAAAAYAAJ.
Clarke, L. (2008). Possibilistic thinking: A new conceptual tool for thinking about extreme events. Social Research, 75, 669–690.
Collins, M., Knutti, R., Arblaster, J., Dufresne, J.-L., & et al. (2013). Long-term climate change: Projections, commitments and irreversibility. In In Stocker et al. (2013), chapter 12.
Dahan, A. (2010). Putting the earth system in a numerical box ? the evolution from climate modeling toward climate change. Studies in History and Philosophy of Modern Physics, 41, 282–292.
Fosser, G., Kendon, E.J., Stephenson, D., & Tucker, S. (2020). Convection-permitting models offer promise of more certain extreme rainfall projections, (Vol. 47.
Hausfather, Z. (2019). Cmip6: The next generation of climate models explained. Carbon Brief, https://www.carbonbrief.org/cmip6-the-next-generation-of-climate-models-explained, last checked the 05/04/2020.
Herger, N., Abramowitz, G., Knutti, R., Angélil, O., Lehmann, K., & Sanderson, B.M. (2018). Selecting a climate model subset to optimise key ensemble properties. Earth System Dynamic, 9, 135–151.
IPCCFAQ 1.1 (2013). If understanding of the climate system has increased, why hasn’t the range of temperature projections been reduced?. In Stocker et al. (2013), chapter 1 (pp. 140–141).
IPCCFAQ 12.1 (2013). Why are so many models and scenarios used to project climate change?. In Stocker et al. (2013), chapter 12(pp. 1036–1037).
IPCCFAQ 9.1 (2013). Are climate models getting better, and how would we know?. In Stocker et al. (2013), chapter 9 (pp. 824–825).
Katzav, J. (2014). The epistemology of climate models and some of its implications for climate science and the philosophy of science. Studies in History and Philosophy of Modern Physics, 46, 228–238.
Katzav, J., & Parker, W.S. (2015). The future of climate modeling. Climatic Change, 132, 475–487.
Klein, S.A., & Hall, A. (2015). Emergent constraints for cloud feedbacks. Current Climate Change Reports, 4(1), 276–287.
Knutti, R. (2010). The end of model democracy? Climatic Change, 102, 395–404.
Knutti, R., & Sedláček, J. (2012). Robustness and uncertainties in the new cmip5 climate model projections. Nature Climate Change, 3, 369–373.
Le Treut, H. (2009). Nouveau climat sur la Terre : Comprendre, prédire, réagir. Flammarion.
Leduc, M., Laprise, R., Elía, R., & Šeparović, L. (2016). Is institutional democracy a good proxy for model independence?. American Meteorological Society, 29.
Lloyd, E.A. (2009). I–varieties of support and confirmation of climate models. Aristotelian Society Supplementary Volume, 83, 213–232.
Lloyd, E.A. (2010). Confirmation and robustness of climate models. Philosophy of Science, 77(5), 971–984.
Lloyd, E.A. (2015). Model robustness as a confirmatory virtue: The case of climate science. Studies in History and Philosophy of Science Part A, 49, 58–68.
Murphy, J.M., Sexton, D.M., Barnett, H., & et al. (2004). Quantification of modelling uncertainties in a large ensemble of climate change simulations. Nature, 430, 768–772.
Parker, W.S. (2011). When climate models agree: The significance of robust model predictions. Philosophy of Science, 78(4), 579–600.
Parker, W.S (2013). Ensemble modeling, uncertainty and robust predictions. Climate Change, 4, 213–223.
Parker, W.S. (2018). Climate science, the stanford encyclopedia of philosophy (summer 2018 edition). Zalta, E.N. (ed.), https://plato.stanford.edu/archives/sum2018/entries/climate-science/.
Parker, W.S., & Lusk, G. (2019). Incorporating user values into climate services. American Meteorological Society, (September, 2019, 1643–1650.
Pichelli, E., Coppola, E., Sobolowski, S., Ban, N., Giorgi, F., Stocchi, P., Alias, A., Belušić, D., Berthou, S., Caillaud, C., Cardoso, R.M., Chan, S., Christensen, O.B., Dobler, A., de Vries, H., Goergen, K., Kendon, E.J., Keuler, K., Lenderink, G., ..., Vergara-Temprado, J. (2021). The first multi-model ensemble of regional climate simulations at kilometer-scale resolution part 2: Historical and future simulations of precipitation. Climate Dynamics.
Räisänen, J. (2007). How reliable are climate models? Tellus A, 59, 2–29.
Reichler, T., & Kim, J. (2008). How well do coupled models simulate today’s climate? Bulletin of the American Meteorological Society, 89, 303–312.
Ribes, A., Qasmi, S., & Gillett, N.P. (2021). Making climate projections conditional on historical observations. Science Advances, 7(4).
Sanderson, B.M., Knutti, R., & Caldwell, P. (2015a). Addressing interdependence in a multimodel ensemble by interpolation of model properties. Journal of Climate, 28, 5150–5170.
Sanderson, B.M., Knutti, R., & Caldwell, P. (2015b). A representative democracy to reduce interdependence in a multimodel ensemble. Journal of Climate, 28(13), 5171–5194.
Schmidt, G.A. (2018). Model independence day. http://www.realclimate.org/index.php/archives/2018/07/model-independence-day/, last checked March 5, 2021.
Schupbach, J.N. (2018). Robustness analysis as explanatory reasoning. The British Journal for the Philosophy of Science, 69(1), 275–300.
In T. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S. Allen, & et al. (Eds.) (2013). IPCC climate change 2013: The physical science basis. Contribution of working group I to the fifth assessment report of the intergovernmental panel on climate change. Cambridge, United Kingdom and New York: Cambridge University Press.
Thompson, E., Frigg, R., & Helgeson, C. (2016). Expert judgment for climate change adaptation. Philosophy of Science, 83(5), 1110–1121.
Vezér, M. A. (2016). Computer models and the evidence of anthropogenic climate change: An epistemology of variety-of-evidence inferences and robustness analysis. Studies in History and Philosophy of Science, 56, 95–102.
Vezér, M. A. (2017). Variety-of-evidence reasoning about the distant past: A case study in paleoclimate reconstruction. European Journal in Philosophy of Science, 7, 257–265.
Webb, M.J., Andrews, T., Bodas-Salcedo, A., Bony, S., Bretherton, C.S., Chadwick, R., Chepfer, H., Douville, H., Good, P., Kay, J.E., Klein, S.A., Marchand, R., Medeiros, B., Siebesma, A.P., Skinner, C.B., Stevens, B., Tselioudis, G., Tsushima, Y., & Watanabe, M. (2017). The cloud feedback model intercomparison project (cfmip) contribution to cmip6. Geoscientific Model Development, 10(1), 359–384.
Winsberg, E. (2018). What does robustness teach us in climate science: A re-appraisal. Synthese. https://doi.org/10.1007/s11229-018-01997-7.
Acknowledgements
This work was supported by “MOVE-IN Louvain” Incoming Post-doctoral Fellowship and by the Swiss National Science Foundation project PP00P1_170460 “The Epistemology of Climate Change”. We thank the anonymous reviewers for their helpful comments. We also thank the participants, Gab Abramowitz, Mathias Frisch and Eric Winsberg, as well as the audience of the symposium “Diversity, Uncertainty, and Action: Coping with a Plurality of Climate Models” held in Seattle at the Philosophy of Science Association (PSA) 2018 conference, for their feedback on the oral presentation of the paper. JJ is in debt to Michel Crucifix and Samuel Somot for insightful discussions on issues related to the addressed topic.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Jebeile, J., Barberousse, A. Model spread and progress in climate modelling. Euro Jnl Phil Sci 11, 66 (2021). https://doi.org/10.1007/s13194-021-00387-0
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s13194-021-00387-0