Bounding probabilistic sea-level rise projections under the IPCC constraints
Résumé
Despite large progresses in climate change science, projections of future sea-level rise remain highly uncertain, especially due to large unknowns in the melting processes affecting the ice-sheets in Greenland and Antarctica. Based on climate-models outcomes and the expertise of scientists concerned with these issues, the IPCC provided constraints to the quantiles of sea-level projections. In addition, some physical limits to future sea-level rise by 2100 have been established, although approximately. These constraints to future sea-level change projections are essentially associated to epistemic uncertainties, so that the theory of probabilities appears too constraining in terms of quality and quantity of necessary data to account for this type of uncertainty. In this contribution, we propose to go beyond the classical use of probabilities by relying on extra-probabilistic theories (namely the possibility theory): this prevents from selecting a single cumulative probability distribution, which is hardly achievable given the level of imprecision and lack of knowledge, but rather enables to identify all the possible cumulative probability distributions that are consistent with the available objective pieces of information, namely the constraints provided by the IPCC as well as by recent publications regarding the future melting of ice sheets. We focus on the case of sea-level rise by 2100 under the RCP 8.5 scenario. Importantly, we make no assumptions regarding the actual shape of the distributions. Today, coastal impact studies are increasingly moving away from deterministic sea-level projections. However, we show that the cumulative probability distributions used so-far have only explored a subset of rather conservative sea-level probabilistic projections. Hence, coastal impacts studies of sea-level rise could be larger than previously estimated while complying with the constraints provided by the IPCC.