Impacts of vegetation dynamics on hydrological simulations under drought conditions in a humid river basin in Southern China

Vegetation plays an essential role in the atmospheric and hydrological processes, and vegetation responds differently to climate change in various regions, especially in extreme climates. Therefore, the use of static prescribed vegetation information from past years in numerical models can be a source of biases in hydrological simulations. However, previous studies have mainly focused on the effects of vegetation dynamics on hydrological processes in arid and semi-arid regions. It remains unclear how static or dynamic vegetation affects hydrological simulations in humid regions, especially under drought conditions. In this study, the Weather Research and Forecasting (WRF) model coupled with Noah-MP was used to assess the impact of vegetation dynamics on hydrological simulations in the East River basin (ERb) of China, which is a major water source for several major cities in the Pearl River Delta. The model was run with prescribed and dynamic vegetation conditions, respectively. Our model validation based on observed 2-m temperature (T2) and Leaf Area Index (LAI) showed that the model performance was improved when vegetation dynamics were considered. Our simulations with static or dynamic vegetation showed the impacts of vegetation dynamics on hydrological simulations under droughts. The model with vegetation dynamics simulated a wetter condition with higher soil moisture and runoff and lower T2, compared with the simulations of static vegetation. The results suggested that ignoring vegetation dynamics may overestimate the severity of drought in this humid basin, unlike arid and semi-arid regions. Therefore, consideration of vegetation dynamics in this humid basin will deepen our research on different types of zones and serve as a reference for other humid regions. Unrealistic LAI (vegetation information) likely causes significant biases in regional climate simulations. Dramatic changes in vegetation were observed in the East River basin. Our results showed that the performance of the WRF model was significantly improved with a more appropriate representation of vegetation, that is, dynamic vegetation. Furthermore, the drought severity was highly overestimated (higher temperature and lower LAI, drier soil moisture, less runoff, etc.) when the vegetation dynamics were ignored. Therefore, the vegetation dynamics should be considered in this humid region, especially during extreme drought conditions.