Meteorological and hydrological drought characteristics and their relationships in the Central Highlands of Vietnam

This study investigates meteorological and hydrological drought characteristics and their relationships in the Central Highlands of Vietnam. Monthly observed rainfall (streamflow) data from 1980–2015 were used to compute the Standardized Precipitation Index (Standardized Streamflow Index) to represent meteorological (hydrological) drought characteristics based on Run theory. The relationships between meteorological and hydrological droughts were established by Artificial Neural Network (ANN). The results show that meteorological (hydrological) drought frequencies follow one- (two-) maximum (maxima) distributions in the annual cycle. The maximum frequencies of droughts with different levels occur in different months. Four basins are chosen to describe spatial variations of drought characteristics, namely, KonTum, AnKhe, DaLat, and BaoLoc. The former two basins have more river branches and longer distances from the main upstream points to hydrological observation stations than the latter two basins. The results show that meteorological droughts must last longer with stronger magnitudes, on average, to cause hydrological droughts in the latter two basins. Although the total number of event pairs is remarkably larger, the total number of event pairs that last for more than one year is smaller in the former two basins. Moreover, it appears that much cooler weather associated with higher elevations may reduce the severity of extreme hydrological droughts in the DaLat Basin. Finally, ANN can successfully reproduce the interannual variations and the differentiations in space of the hydrological drought characteristics in the Central Highlands.