Simulating how rivers flow is important for managing water supplies, especially in areas where people rely on reservoirs, canals, and legal water rights. However, many large-scale models used to simulate river flows don’t include these human-made systems, which can lead to inaccurate results.

In this study, we used a newly developed “hydrofabric”—a detailed map-based dataset that brings together information from different sources and links it to real-world locations. This hydrofabric made it easier to build models that represent how water is stored, moved, and used across a watershed.

We combined a hydrologic model that simulates natural water inflows with a water management model that simulates how people operate reservoirs and meet water demands. We tested this approach in a part of the Upper Colorado River Basin, running daily simulations from 2003 to 2013.

By adding water management features one at a time—such as reservoirs, water demands, and water rights—we were able to see how each one improved the model’s ability to match observed river flows. Our results show that including these features leads to more accurate simulations and can help improve national water models used for planning and decision-making.