A scalable and modular reservoir implementation for large-scale integrated hydrologic simulations

Recent advancements in integrated hydrologic modeling have enabled increasingly high-fidelity models of the complete terrestrial hydrologic cycle. These advances are critical for our ability to understand and predict watershed dynamics, especially in a changing climate. However, many of the most physically rigorous models have been designed to focus on natural processes and do not incorporate the effect of human-built structures such as dams. By not accounting for these impacts, our models are limited both in their accuracy and in the scope of the topics they are able to investigate. Here, we present the first implementation of dams and reservoirs in ParFlow, an integrated hydrologic model. Through a series of idealized and real-world test cases, we demonstrate that our implementation (1) functions as intended, (2) maintains important qualities such as mass conservation, (3) works in a real domain, and (4) is computationally efficient and can be scaled to large domains with thousands of reservoirs. Our results have the potential to improve the accuracy of current ParFlow models and enable us to ask new questions regarding conjunctive management of ground and surface water in systems with reservoirs.