Mechanisms and simulation methods for the impact of small reservoirs on watershed hydrological processes

The widespread implementation of small-scale hydraulic engineering structures has profoundly modified the underlying surface characteristics of watersheds, leading to significant changes in rainfall-runoff processes. Understanding the mechanisms by which small reservoirs influence runoff generation and routing, as well as developing effective simulation methods, is crucial for enhancing flood forecasting accuracy at the watershed scale. This research seeks to construct aggregated reservoirs based on the topological relationships of small storage bodies and to integrate these with existing hydrological models, thereby improving the precision of flood forecasting in humid watershed regions.

This study simplifies the diverse topological structures of small reservoirs into three foundational connection units: single-reservoir, series, and parallel configurations. These units are systematically combined into mixed configurations that realistically reflect the spatial distribution of storage bodies within watersheds. Furthermore, a multi-stage weir flow discharge scheme, specifically designed for aggregated reservoirs, is proposed based on field conditions, and the corresponding reservoir outflow equations are formulated. By coupling this newly developed reservoir storage-discharge module with the traditional lumped Xin'anjiang model, an improved version of the model is created, incorporating the regulatory effects of small reservoirs. To evaluate the performance of the improved Xin'anjiang model, 13 flood events were analyzed. Results demonstrated a substantial enhancement in simulation accuracy, with the average Nash-Sutcliffe efficiency coefficient improving by 0.27 during the calibration period and 0.40 during the validation period compared to the original model. Notably, the improved model excelled in simulating floods early in the flood season or following extended dry spells. However, its ability to simulate mid-to-late-season or multi-peak floods showed comparatively modest improvements. Additionally, the model's simulation accuracy was observed to decrease as flood magnitude increased.

Compared to traditional hydrological models that exclusively consider natural watershed processes, those incorporating aggregated reservoir storage and discharge dynamics offer a more nuanced representation of watershed hydrology. By significantly enhancing flood forecasting accuracy during the critical flood season, the improved model not only mitigates the impacts of flood disasters but also bolsters local water resource management capabilities.