Non-stationarity analysis of extreme water level in response to climate change and urbanization in the Taihu Basin, China

Regional climate change as well as urbanization are likely to introduce characteristics of non-stationarity in the extreme water levels (EWL), which in turn will affect the possibility of flooding in the plain river network region. For this study, China’s Taihu Basin as a typical case for analyzing the impacts that climate change and urbanization will have on changes in the EWL. Additionally, changes in the response that EWL will have toward future climate change and urbanization are also studied. A generalized additive model for location, scale and shape (GAMLSS) model with time (t), precipitation (pre) and population (pop) as major predictors was constructed to empirically investigate. Results indicate that the time dependent variables of the GAMLSS clearly describe the temporal variation of the EWL. The GAMLSS with physically-based covariates can capture additional details regarding the changing properties of EWL. According to changes exhibited in the EWL Akaike information criterion values, it is found that the impacts of climate change covariates are larger than that of the urbanization covariates. Additionally, for the study’s sample period of 2015–2099, an increasing trend is found for EWL under the RCP2.6 and RCP4.5 scenarios. It is concluded that flooding risk is likely to increase in China’s Taihu Basin with changing environment conditions, and that some necessary measures should be taken by local water resource management authorities to address the potential risks posed to the region.