A novel method to test non-exclusive hypotheses applied to Arctic ice projections from dependent models

R Olson,W Chang,J.-Y Lee,J P Evans,S.-I An,Y Fan

doi:10.1038/s41467-019-10561-x

Abstract

A major conundrum in climate science is how to account for dependence between climate models. This complicates interpretation of probabilistic projections derived from such models. Here we show that this problem can be addressed using a novel method to test multiple non-exclusive hypotheses, and to make predictions under such hypotheses. We apply the method to probabilistically estimate the level of global warming needed for a September ice-free Arctic, using an ensemble of historical and representative concentration pathway 8.5 emissions scenario climate model runs. We show that not accounting for model dependence can lead to biased projections. Incorporating more constraints on models may minimize the impact of neglecting model non-exclusivity. Most likely, September Arctic sea ice will effectively disappear at between approximately 2 and 2.5 K of global warming. Yet, limiting the warming to 1.5 K under the Paris agreement may not be sufficient to prevent the ice-free Arctic.

Highlights

A major conundrum in climate science is how to account for dependence between climate models
To provide pdf of the global mean surface temperature (GMST) change to melt, we randomly sample all GMST changes that fall within our sample space: GMST changes that are within some distance of at least one model whose present-day mean September sea ice extent (SIE) is within some distance of the observed mean, and whose present-day SIE sensitivity is within some distance of the observed value
We account for the uncertainty in present-day modeled and observed SIE, as well as for the uncertainty in distances themselves

Summary

Introduction

A major conundrum in climate science is how to account for dependence between climate models. In the worst-case scenario, the future projections may be centered on output from a large group of models which rely on the same erroneous code, yet little weight may be assigned to an independent correct model. This is one of the reasons Intergovernmental Panel on Climate Change (IPCC) has outright removed any probabilistic global multimodel climate projections from its Fifth Assessment Report[12]. We apply the new method to provide the first multi-model probabilistic projections of the global mean surface temperature (GMST) change from the preindustrial period at which September Arctic sea ice will effectively disappear (thereafter called GMST change to melt). We calibrate the method in a suite of observation system simulation experiments to provide approximately correct coverage of the 90% posterior credible intervals

Methods

Results

Conclusion