Characteristics and dynamic evolution patterns of drought-flood abrupt alternation events based on three-dimensional identification

Drought-flood abrupt alternation (DFAA) events, as a unique hydrometeorological disaster, remain poorly understood in terms of their dynamic evolution characteristics and nonlinear mechanisms. This study introduces a daily-scale Dry-Wet Abrupt Alternation Index (DWAAI), which is constructed from daily SPEI data derived from the China Meteorological Forcing Dataset. The DWAAI integrates multi-timescale (30-, 90-, and 360-day) SPEI values through a weighted scheme that explicitly accounts for cumulative effects of antecedent drought or flood conditions. By identifying DFAA patches and applying spatial connectivity rules, temporal overlap criteria, and patch-merging procedures, we innovatively develop a three-dimensional (longitude-latitude-time) identification framework that enables continuous spatiotemporal tracking of DFAA patches. By incorporating event migration and severity transitions into the assessment framework, we systematically analyze the spatio-temporal evolution of DFAA events. The main findings are as follows: (1) Based on DWAAI, DFAA events are classified into four severity levels (mild, moderate, severe, and extreme), with mean severity, duration, affected area, and intensity all increasing progressively with event level. Moderate events occur most frequently and exhibit the most concentrated within-group distributions. (2) Spatially, event frequency shows an east-high and west-low pattern, with the South China Humid region serving as a hotspot. Over the past four decades, both drought to flood and flood to drought events have shown fluctuating increases, with the most significant increase observed in the Northwest Inland Arid Region. (3) Approximately 51.28% of events exhibited expanding affected areas during their evolution, with about half accompanied by changes in event severity, and four primary evolution patterns were identified. (4) VPD and NINO were identified as common influencing factors across all four regions, indicating that DFAA events are jointly regulated by both region-common and region-specific factors. This study incorporates the dynamic evolution dimension into DFAA assessment, deepens the understanding of DFAA evolution mechanisms, and provides a new analytical framework for risk prevention and control.