Variants of the Palmer Drought Severity Index for Portugal using fine resolution gridded datasets

Monthly time series of the Self-Calibrating Palmer Drought Severity Index (SC-PDSI) and SC-MedPDSI, a modified PDSI for the Mediterranean environment, are computed for Portugal using high-resolution gridded datasets of monthly precipitation totals retrieved from the PT02 daily gridded dataset, and monthly averages of maximum and minimum temperature purposefully developed. The datasets have 0.2˚ × 0.2˚ spatial resolution and cover the period 1950–2003. The effects of the weighing and self-calibrating (SC) methods are studied by comparing the SC-PDSI and SC-MedPDSI with the corresponding indices computed using the original calibration for Portugal. Particular attention is given to the frequency of wet and dry events. The two indices are compared with the SPI on 9-month time scale (SPI-9) computed using the PT02 dataset. Results show that extreme events are less frequent when using the SC indices, and that there is a significant agreement between the SPI-9 and the SC-PDSI or the SC-MedPDSI. Moreover, the space and time variability of the indices are analysed using the Principal Component Analysis (PCA). The spatial patterns identify two sub-regions consistently with previous studies. The Modified Mann-Kendal test is applied to the PC score time series