Analysis of the spatio - temporal variability of extreme droughts signatures at a global scale, over the last 50 years

The study presented assesses the variability of extreme hydrological events at a global scale over the last 50 years, both in space and time in view of assisting predictability. Hydrological extremes are observed through the Palmer Drought Severity Index (PDSI), with a monthly temporal resolution. In order to perform the analysis, the concept of compact spatial entities that we will refer to as episodes is introduced. An episode is defined as the ensemble of edge or vertex connected spatial cells with PDSI values belonging to the same severity class, constituting a compact region. With respect to other analysis which treats cells as independent, the perspective of episodes is more in agreement with the underlying catchment processes and allows accounting for additional parameters relevant for a better assessment of the temporal variability of the characteristics of extreme hydrological events. Hence, with the concept of episode, emphasis is better driven to the dynamic behavior of hydrological extreme events. In this work, the usefulness of approaching droughts through the study of the characteristics proposed at the episode temporal and spatial scale will be evaluated. Besides, a key issue for the prediction of future changes in the regional patterns of hydrological extremes is to assess to which extent the increase in the number of drought episodes can be attributed either to El Niño events and / or to the ongoing global warming signature. It is usually assumed that a close relationship exists between global climate and regional droughts, but the extent of the relationship is not fully characterized, even if the assessment of this relationship would increase predictability. Hence, the temporal variability of the global energy associated to the ENSO signal will be analyzed in conjunction to the temporal variability of the drought episodes characteristics over the last fifty years, focusing on the detection and characterization of common transients between both signals. A processing based on the Scale Dependent Correlation (SDC) method will allow the characterization of the ENSO impact on soil moisture conditions at the global scale. Finally, the drought dynamics observed through PDSI from 1950 to 2005 will be confronted to the ones observed in the first year of measures of the ESA's Soil Moisture and Ocean Salinity (SMOS) mission, based on a spaceborne radiometer. This work is funded by the MIDAS6's Spanish MICINN research project.