Índices SPI e RDI via produto terraclimate aplicado às capitais do nordeste brasileiro

Abstract

This study evaluated drought and extreme rainfall events in the capitals of Northeast Brazil (NEB) based on the Standardized Precipitation Index (SPI) and Reconnaissance Drought Index (RDI), using precipitation and potential evapotranspiration (PET) data from TerraClimate through the DrinC software. The research used rainfall data from conventional meteorological stations (CMS) provided by INMET for the nine NEB capitals, spanning from 1980 to 2021. The INMET data were used to validate the TerraClimate data through statistical indicators. Cluster analysis (CA) with the complete linkage method was applied to identify homogeneous regions for rainfall and evapotranspiration (PET). The dendrogram generated using the complete linkage method was validated using the cophenetic correlation coefficient (CCC), resulting in CCC = 0.79 for precipitation data and CCC = 0.65 for evapotranspiration data. These climatic data were used as input for the SPI and RDI indices. The CA technique identified three homogeneous PET regions and two precipitation regions, with Salvador remaining ungrouped (Not Aggregated - NA) among the NEB capitals. The obtained results indicated that monthly precipitation values from TerraClimate showed strong correlations, ranging from very high (0.7 - 0.9) to nearly perfect (0.9 - 1.0), and the error statistics (SEE and RSME, mm) were satisfactory. The DrinC software demonstrated satisfactory performance in SPI and RDI at all adopted scales, successfully identifying the intensity and frequency of drought and rainfall events in the NEB capitals. The SPI was useful in identifying rainfall deficits and extreme events, while the RDI efficiently identified agricultural and hydrological droughts at quarterly and semi-annual scales. The application of both indices is recommended for the entire Northeast region.