Next Generation Water Resources Modeling Framework: Opportunities for Community Involvement
The current operational NOAA-NWS National Water Model applies a uniform formulation to make continental scale flow predictions on the NHD+ drainage network. However, the literature demonstrates that given the spatial variability in dominant runoff generation mechanism and associated uncertainties in processes and parameters, skillful predictions require scientific evaluation of different model formulations in different hydrologic regions. Providing timely inland and coastal continental-scale predictions requires operations in an HPC environment. Legacy water resources models have dissimilar inputs and setup workflows, run-time environments, discretizations, solvers, and required forcing data. The sheer variety of approaches impedes model comparison and interoperability. The WaterML 2.0 Hy_Features standard provides a stable meta-model to describe the hydrologic landscape, and includes four fundamental topological elements: “catchment”, “flowpath”, “water body”, and their “nexus” linkages, which represent internal boundary conditions and provide natural breakpoints between models. The Hy_Features data model standard helps to unify model setup workflows. The Next Generation Water Resources Modeling Framework that is currently under development promotes interoperability, inter-comparison, model-based testing of research hypotheses, and ultimately improved agency-specific operational predictions while incorporating rapid adoption of advancements from the academic and federal research communities. This is achieved by using the drainage network as a graph to organize parallelization, and by extending the CSDMS Basic Model Interface (BMI) to include state-serialization functionality and to accommodate models with parallel formulations. This work in progress uses the open source development paradigm and participation by the research community is welcomed.