Long-term trend quantification of rainfall-induced inundation due to climate change

This study aims to develop an ANN-based evaluation model to quantify the effects of climate change on long-term trends in 2D inundation simulations (the SM_QLT_2DIS model). The proposed SM_QLT_2DIS model was developed by improving the ANN-derived two-dimensional inundation simulation model (SM_EID_2D) with Mann–Kendall (MK) long-term trend analysis, considering simulated AR5 hourly typhoon events from 1979 to 2099. Miaoli City in northern Taiwan was regarded as the study area, and the AR5 typhoons across it were treated as the study data. The model application results indicate that the average and maximum hourly inundation depths in the study area exhibit a significant, steadily increasing trend, with a high positive MK estimator (on average 790). Despite a simulated flooding area with a negative MK estimator S (around −620), the coefficient of variation shows an increasing trend (ranging from 0.23 to 0.4). Consequently, the proposed SM_QLT_2DIS model can effectively quantify and assess the impacts of climate change on 2D inundation simulations as a reference for early flood warning and mitigation.