Global map of a comprehensive drought/flood index and analysis of controlling environmental factors.

Developing an index that can assess the marked changes in water resources is necessary for drought/flood hazard prevention and water resource planning, particularly in the context of global climate change and intensified human activities. In this study, spatiotemporal variations of water resources variables, namely precipitation (P), evapotranspiration (ET), and runoff (R), regarded as water input, output, and storage, respectively, were systematically statistically analyzed. Following this, a comprehensive Copula-based Drought/Flood Index (CDFI) was constructed to identify the drought/flood hazards, and machine learning and game theory were then successively applied to quantify the influence of the Large-scale Climate Indices (LCIs) on drought/flood intensity. The results reveal that temporally, P, ET, and R present similar turning-point years in the 1990s, with values changing from significantly decreasing to insignificantly decreasing, significantly increasing, and insignificantly decreasing, respectively. Spatially, the water resources become drier in the arid area and wetter in the humid area. The CDFI results demonstrate an increased flood probability in the mountains and near-polar rivers, and drought in extremely arid areas. Moreover, the influencing factors of CDFI exhibited notable spatial heterogeneity. CDFI is controlled by the Arctic Oscillation in most regions of the world, and the influence of sea surface temperature on CDFI increases gradually with the decrease of latitude, and extreme events are generally linked to extreme LCIs. Notably, vegetation change (area, type, and growth) affects ET and R, thus enhancing the climate's impact on drought/flood hazards. The conclusions provide an essential basis for the identification and policy planning of drought/flood hazards caused by the spatiotemporal variations of water resource variables. [ABSTRACT FROM AUTHOR]