Evaluation of drought representation and propagation in regional climate model simulations across Spain

Abstract

Abstract. Drought is an important climatic risk that is expected to increase in frequency, duration, and severity as a result of a warmer climate. It is complex to model due to the interactions between atmospheric and continental processes. A better understanding of these processes and how the current modelling tools represent them and characterize drought is vital. The aim of this study is to analyse how regional climate models (RCMs) represent meteorological, soil moisture, and hydrological drought as well as propagation from precipitation anomalies to soil moisture and streamflow anomalies. The analysis was carried out by means of standardized indices calculated using variables directly related to each type of drought: precipitation (SPI), soil moisture (SSMI), runoff (SRI), and streamflow (SSI). The RCMs evaluated are the CNRM-RCSM4, COSMO-CLM, and PROMES. All of the simulations were obtained from the Med-CORDEX database and were forced with ERA-Interim. The following datasets were used as references: SAFRAN (meteorological drought), offline land surface model simulations from ISBA-3L and ORCHIDEE (soil moisture drought), a SIMPA hydrological model simulation, and observed streamflow (hydrological drought). The results show that RCMs improve meteorological drought representation. However, uncertainties are identified in their characterization of soil moisture and hydrological drought, as well as in drought propagation. These are mainly explained by the model structure. For instance, model structure affects the temporal scale at which precipitation variability propagates to soil moisture and streamflow.