Many soil and water conservation (SWC) measures have been applied in the Jinghe River Basin to decrease soil erosion and restore degraded vegetation cover. Analysis of historical streamflow records suggests that SWC measures may have led to declines in streamflow, although climate and human water use may have contributed to observed changes. This paper presents an application of a watershed-scale, physically-based eco-hydrological model—the Regional Hydro-Ecological Simulation System (RHESSys)—in the Jinghe River Basin to study the impacts of SWC measures on streamflow. Several extensions to the watershed-scale RHESSys model were made in this paper to support the model application at larger scales (>10,000 km2) of the Loess Plateau. The extensions include the implementation of in-stream routing, reservoir sub-models and representation of soil and water construction engineering (SWCE). Field observation data, literature values and remote sensing data were used to calibrate and verify the model parameters. Three scenarios were simulated and the results were compared to quantify both vegetation recovery and SWCE impacts on streamflow. Three scenarios respectively represent no SWC, vegetation recovery only and both vegetation recovery and SWCE. The model results demonstrate that the SWC decreased annual streamflow by 8% (0.1 billion m3), with the largest decrease occurring in the 2000s. Model estimates also suggest that SWCE has greater impacts than vegetation recovery. Our study provides a useful tool for SWC planning and management in this region.

As the Regional Hydro-Ecological Simulation System (RHESSys) is a tool to simulate the interactions between ecological and hydrological processes, many RHESSys-based studies have been implemented for sustainable watershed management. However, it is crucial to review a RHESSys updating history, pros, and cons for further improving the RHESSys and promoting ecohydrological studies. This paper reviewed the progress of ecohydrological studies employing RHESSys by a bibliometric analysis that quantitatively analyzed the characteristics of relevant studies. In addition, we addressed the main application progress, parameter calibration and validation methods, and uncertainty analysis. We found that since its release in 1993, RHESSys has been widely applied for basins (<100 km²) within mainly seven biomes. The RHESSys model has been applied for evaluating the ecohydrological responses to climate change, land management, urbanization, and disturbances, as well as water quality and biogeochemical cycle. While most studies have paid their attention on climate change, the focus has shifted to the application for land management in recent years. This study also identified many challenges in RHESSys such as the inaccessible data and parameters, oversimplified calibration approach, few applications for large-scale watersheds, and limited application fields. Therefore, this study proposed a set of suggestions to overcome the limitations and challenges: (1) Developing a new approach for parameter acquisition and calibration from multi-source data, (2) improving the applicability for a large-scale basin, and (3) extending the scope of application fields. We believe RHESSys can improve the understandings of human–environment relationships and the promotion of sustainable watersheds development.

The Yellow River is the largest inflow into the Bohai Sea, and its inflow changes directly affect the ecological environment and marine health of the Bohai Sea. Therefore, accurate prediction of the inflow of the Yellow River is crucial for maintaining the ecological balance of the Bohai Sea and protecting marine resources. Time decomposition algorithms, combined with machine learning, are effective tools to enhance the capabilities of inflow prediction models. However future data leakage from decomposition items was ignored in many studies. It is necessary to develop the right method to operate time decomposition to avoid future data leakage. In this study, the inflow from the Yellow River into the sea was predicted based on a machine learning model (light gradient boosting machine, LightGBM) and a time decomposition algorithm (seasonal and trend decomposition using loess, STL), and the future data leakage in different ways of using STL were evaluated. The results showed that the overall performance of the STL–LightGBM model was better than that of the LightGBM model. The STL–LightGBM took the historical inflow for 8 days as the input, and predicted that the average NSE of the next 1–7 days would reach 0.720. Even when the forecast period was 7 days, the STL–LightGBM (NSE: 0.549 for 7-day lead time) was 0.105 higher than the LightGBM (NSE: 0.444 for 7-day lead time). We found that STL pretreatment of the entire test set overestimated the true performance of STL–LightGBM. It is recommended that the STL preprocesses each sample of the test set to avoid future data leakage. The study can provide help for water resources management and offshore environmental management.