Next-Generation Water Resources Modeling System: A Community Research, Development, Testing and Operations Software Framework

Abstract:

Legacy water resources models use dissimilar inputs and setup workflows, run-time environments, discretizations, solvers, and require different forcing data. The sheer variety of approaches impedes model comparison and interoperability. However, all water resources model applications seek improved understanding or predictions at some scale and area of interest on an interconnected set of hydrographic features. We suggest that large-scale vector-based hydrographic datasets provide a common starting point. The use of a standard to describe hydrologic network topology and geometry suggest eliminating the need for stream/catchment delineation, and represents an essential starting point for developing a flexible and extensible software framework for community-accessible water resources modeling on a common “hydrofabric”. The WaterML Hy_Features standard provides a stable meta-model for water resources modeling. This standard provides three fundamental topological elements: “catchment”, “water body”, and “nexus” points that link them. Open-source software frameworks that emphasize contributions by domain scientists promote community use for research to operations. For example, the lack of a complete and verifiable runoff generation theory hinders progress in continental scale hydrologic and water-resources predictions. However, models that emphasize different processes are sensitive to different sets of parameters, which aids in formulation selection. This suggests that formulations that are known to perform well in specific situations will do so using a smaller number of model parameters. The proposed next-generation water resources modeling software framework promotes interoperability, inter-comparison, model-based testing of research hypotheses, and ultimately improved agency-specific operational predictions while incorporating rapid adoption of advancements by research communities.