Estimating Impact Likelihoods from Probabilistic Projections of Climate and Socio-Economic Change Using Impact Response Surfaces

Abstract

Estimates of future climate change impacts using numerical impact models are commonly based on a limited selection of projections of climate and other key drivers. However, the availability of large ensembles of such projections offers an opportunity to estimate impact responses probabilistically. This study demonstrates an approach that combines model-based impact response surfaces (IRSs) with probabilistic projections of climate change and population to estimate the likelihood of exceeding pre-specified thresholds of impact.The changing likelihood of exceeding impact thresholds during the 21 st century was estimated for selected indicators in three European case study regions (Iberian Peninsula, Scotland and Hungary), comparing simulations that incorporate adaptation to those without adaptation. The results showed high likelihoods of increases in heat-related mortality and of yield decreases for some crops, whereas a decrease of NPP was estimated to be exceptionally unlikely. For a water reservoir in a Portuguese catchment, increased likelihoods of severe water scarce conditions were estimated for the current rice cultivation. Switching from rice to other crops with lower irrigation demand changes production risks, allows to expand the irrigated areas, but introduces a stronger sensitivity to changes in rainfall.The IRS-based risk assessment shown in this paper is of relevance for policy making by addressing the relative sensitivity of impacts to key climate and socio-economic drivers, the urgency for action expressed as a time series of the likelihood of crossing critical impact thresholds and options to respond by incorporating alternative adaptation actions in the analysis framework, which may be useful for exploring the types, choice and timing of adaptation responses.

Keywords: impact model, sensitivity analysis, temperature, precipitation, population