Attribution analysis of non-stationary hydrological drought using the GAMLSS framework and an improved SWAT model

Climate change and human activities are two key factors that modify the hydrological cycle and drought characteristics of a watershed. Quantifying the effects of climate change and human activities on hydrological drought is of great significance for water resources management. The traditional method of hydrological drought assessment is based on the assumption of stationarity, which is no longer applicable. In this study, using the Hanjiang River basin (HRB) as a study area, the Pettitt test and heuristic segmentation algorithm are first used to identify the change points in the runoff time series. Then, a distributed hydrological model (SWAT) is used to reconstruct the natural runoff series during the impacted period. Finally, the impacts of climate change and human activities on hydrological drought are quantitatively assessed using the non-stationary Standardized Runoff Index (NSRI), developed based on the GAMLSS model. Results show that runoff processes in the basin are non-stationary, and the year 1991 is identified as the change point of the runoff series. Both climate change and human activities exacerbate the hydrological drought by increasing its duration, severity, and peak, with climate change playing a dominant role in the 1990s. Human activities, however, have gradually surpassed the impact of climate change on hydrological drought since the 21st century. In addition, reservoir regulation extends the drought duration and severity but mitigates the peak. These findings provide valuable information for drought management and water resource regulation under changing environments.