Changes in 21st century scenarios for temperature extreme indices over South America derived from an ensemble of regional climate models

CARRIL A. F.; LÓPEZ DE LA FRANCA AREMA, N; MENENDEZ C.G.; SANCHEZ E.

Bruselas

Conferencia; International Conference on Regional Climate - CORDEX 2013; 2013

CORDEX/WCRP

This study examines a set of temperature extreme indices derived from regional climate models (RCMs) for present and future climate conditions over South America. Our aim is to provide an analysis on the performance of RCMs to reproduce such indices and to explore in their projected changes, with a focus on annual and local scales. The RCMs (ETA, LMDZ, MM5, PROMES, RCA, RegCM3, REMO) were integrated at 0.5 degree in the framework of CLARIS-LPB EU-project, and forced by different CMIP3 global models. Eight selected annual indicators derived from daily temperature time series are considered: WSDI/CSDI (warm/cold spell duration index), CSU/CFD (maximum number of consecutive summer/frost days), TX90/TN90 (warm day-times/nights) and TX10/TN10 (cold day-times/nights). Maps of observed (HadEX2) and simulated indices for present (1991-2010) and future (2079-2098) climates have been compared. The ensemble can be considered as ?satisfactory? for simulating annual indexes of extremes, nevertheless the analysis is partially limited by the lack of observations. Most errors possibly result from the fact that some processes are still poorly represented (e.g., land-atmosphere interactions in La Plata Basin) or not included in RCMs (e.g., effects of aerosol forcing or regional scale changes in land use). This is especially the case for climate extremes and associated indices if they are involved in feedback effects. Extremes may also be impacted by mesoscale and regional circulations that RCMs (even with 50 km resolution) cannot wholly resolve, such as low-level jets or mesoscale convective systems impacting precipitation and surface temperature through surface?atmosphere feedbacks. These difficulties limit quantitative assessments of the magnitude, variability and regional details of extreme indices. All the indices analyzed display sensitivity to climate change projections and the strength of that sensitivity depends on the index and on the sub regions considered.