High time- and spatial-resolution climate scenarios of the DISTENDER project according to statistical and dynamical downscaling

Abstract

This paper discusses statistical and dynamical methods used to produce local (grid-spacing < 4 km) and Euro pean (~10 km) climate scenarios that were used as input for multi-sectoral impact models in the DevelopIng STratEgies by integrating mitigatioN, aDaptation and participation to climate changE Risks (DISTENDER) project, and shares the main results with a special focus on temperature and precipitation. The statistical downscaling consisted of three stages: (1) a parametric quantile mapping at a daily scale; (2) an analogous transference function of hourly curves for each day, and (3) a classical geostatistical downscaling. This three stage technique was applied to three representative Earth System Models according to three different climate change level (being EC-EARTH3-Veg the medium case) under four shared socioeconomic pathways (SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP5-8.5). In addition, dynamical downscaling was also considered. Particularly, the ICOsahedral Nonhydrostatic model downscaled the EC-EARTH3-Veg model to computationally costly km-scale resolution under all four pathways. Both downscaling approaches show consistent behaviour for the down scaled model under the different pathways. Results indicate historical biases in precipitation about ± 10 % in general, while temperature biases ranged from ¿ 2¿C to + 1¿C across different regions and seasons. Under SSP5- 8.5, summer precipitation in southern Europe is projected to decrease by up to 20 %, while northern Europe experiences increases of + 10 % to + 15 %. Temperature increases under the same scenario reach + 5¿C in summer across southern Europe, with smaller increases of + 2¿C to + 3¿C in northern regions. These findings on management for uncertainty levels demonstrate the utility of combined downscaling approaches for local climate risk assessment and adaptation strategies.