Reservoir outflow is an important variable for understanding hydrological processes and water resource management. Natural streamflow variation, in addition to the streamflow regulation provided by dams and reservoirs, can make streamflow difficult to understand and predict. This makes them a challenge to accurately simulate hydrologic processes at a daily scale. In this study, three Machine Learning (ML) algorithms, Random Forest (RF), Support Vector Machine (SVM), and Artificial Neural Network (ANN), were examined and compared to model reservoir outflow. Past, current, and future hydrologic and meteorological data were used as model inputs, and the outflow of next day was used as prediction. Simulation results demonstrated that all three models can reasonably simulate reservoir outflow. For Carlyle Lake, the coefficient of determination and Nash–Sutcliffe efficiency were each close to one for the three models. The coefficient of determination, relative mean bias, and root mean square error indicated that the SVM performed better than the RF and ANN, but the SVM output displayed a larger relative mean bias than that from RF and ANN. For Lake Shelbyville, the ANN model performed better than RF and SVM when considering the coefficient of determination, Nash–Sutcliffe efficiency, relative mean bias, and root mean square error. The study results demonstrate that the three ML algorithms (RF, SVM, and ANN) are all promising tools for simulating reservoir outflow. Both the accuracy and efficacy of the three ML algorithms are considered to support practitioners in planning reservoir management.

Climate change and increasing water demand are critical factors influencing water availability and the hydrologic cycle. This study examines their combined impacts on the Kaskaskia River watershed in Illinois, USA, using a comprehensive hydrologic modeling approach. The model integrates key watershed components, including existing reservoirs, multiple water withdrawal sources, and point source discharges, with necessary modifications to the governing algorithms to accurately represent system processes. A daily target release method was implemented for routing water through the watershed’s two main reservoirs, significantly improving the accuracy of storage and outflow simulations. A total of 68 hydrologic scenarios were simulated, combining 17 climate projections with four water demand patterns. Results reveal that climate change exerts a substantial influence on river flows and reservoir storages, while the effects of projected water demand are comparatively minor. Under RCP2.6, RCP4.5, and RCP8.5 scenarios, minimum storage volumes in both reservoirs are projected to decline over the next 25 to 50 years, while maximum storage volumes are expected to increase, indicating greater variability in reservoir levels. Despite these changes, future water yields from both reservoirs are projected to exceed current yields across all scenarios. These findings underscore the importance of adopting sustainable water resource management strategies to address the challenges posed by climate variability and evolving water demands, ensuring long-term water security in the region. Furthermore, as water availability is also constrained by water quality, future enhancements to the model should incorporate water quality simulations to assess the impacts of sediment and nutrient loads on overall water availability.

A nutrient loss reduction strategy is necessary to guide the efforts of improving water quality downstream of an agricultural watershed. In this study, the effectiveness of two winter cover crops, namely cereal rye and annual ryegrass, is explored as a loss reduction strategy in a watershed that ultimately drains into a water supply reservoir. Using a coupled optimization‐watershed model, optimal placements of the cover crops were identified that would result in the tradeoffs between nitrate‐N losses reduction and adoption levels. Analysis of the 10%, 25%, 50%, and 75% adoption levels extracted from the optimal tradeoffs showed that the cover crop placements would provide annual nitrate‐N loss reductions of 3.0%–3.7%, 7.8%–8.8%, 15%–17.5%, and 20.9%–24.3%, respectively. In addition, for the same adoption levels (i.e., 10%–75%), sediment (1.8%–17.7%), and total phosphorus losses (0.8%–8.6%) could be achieved. Results also indicate that implementing each cover crop on all croplands of the watershed could cause annual water yield reduction of at least 4.8%, with greater than 28% in the months of October and November. This could potentially be detrimental to the storage volume of the downstream reservoir, especially in drought years, if cover crops are adopted in most of the reservoir's drainage area. Evaluating water yield impacts, particularly in periods of low flows, is thus critical if cover crops are to be considered as best management practices in water supply watersheds.

Excess sediment and nutrients including nitrogen and phosphorus are major causes of non-point source (NPS) pollution in rivers and streams draining agricultural watersheds. Research studies indicate that a significant proportion of NPS pollutants entering the Mississippi River comes from agricultural watersheds in the Midwest. Illinois watersheds are among those that contribute the highest nutrient flux to the Gulf of Mexico. Watershed Best Management practices (BMPs) could serve as crucial control measures in reducing NPS pollutants from agricultural watersheds and have been applied in many places in Upper Mississippi River basin. This paper presents a coupled optimization-watershed model designed to identify cost-effective placement of BMPs in watersheds that could result in maximized reduction of NPS pollutants such as sediment and nutrients. The coupled model is an interface between the Soil and Water Assessment Tool (SWAT) and the Non-dominated Sorting Genetic Algorithm II (NSGA-II). The computational model was applied to Bray Creek and Frog Alley watersheds, which are tributary watersheds of the Mackinaw River in central Illinois. The BMP of choice to control sediment and nutrient loadings was constructed wetlands because there has been some practical experience in the study watersheds. Simulation results indicate the constructed wetlands appear to be more effective in reducing nutrient loads in both watersheds and the model was able to provide optimal tradeoffs between pollutant reduction and placement cost of constructed wetlands in the watersheds. The coupled model can assist as a decision support tool in prioritizing and targeting areas for BMP placement in the watersheds, resulting in an optimal reduction of pollutants.