Changes in hydrological behaviours triggered by earthquake disturbance in a mountainous watershed

Landslides induced by strong earthquakes often destroy large amounts of landscape vegetation which can trigger significant changes in runoff potential and flood flow. Little is known about hydrological behaviours imposed by co-seismic landslides and their post-earthquake evolution. Therefore, we collected time-series datasets (2007-2018) of underlying surface conditions (USC) changes including landslide expansion and recovery in a watershed affected by the Wenchuan earthquake to further quantify how the large physical disturbance affected the flood hydrological behaviours. The hydrological model HEC-HMS was calibrated and validated to predict the historical hydrological behaviours based on 5 min time-series data in rainfalls and streamflow (2018-2019), showing a good model performance with a mean Nash-Sutcliffe efficiency of 0.76. It was found that, shortly after the earthquake, the sharp expansion with 11% of landslide areas elevated the magnitudes of runoff potential, peak discharge, and runoff volume by >10%, and the peak to time for the high-magnitude flood was advanced by 25 min compared to the pre-earthquake levels. The tipping point along the hydrological disturbance-recovery trajectory was detected within 2011 with higher flood peaks and volumes, and the periods of 2011-2013 (i.e. 3-5 years post-earthquake) were deemed to be a rapid recovery period, revealing an unstable hydrological function. These findings are significant for clearly understanding the magnitude and timing, as well as greater risks of post-earthquake catastrophic flooding in earthquake-stricken regions. Additionally, the post-earthquake accompanied rainstorm-induced geohazards, which limited the recovery of landscape vegetation, triggering an undulant but clear recovery process (1-7 years post-earthquake) of hydrological behaviours. These findings promoted our understanding of the spatiotemporal evolution of hydrological behaviours triggered by the earthquake, and further contribute to the development of adaptation and mitigation strategies for the unpredictable flash floods triggered by future abrupt natural hazards in earthquake-affected regions.