The NWM is a hydrologic model that simulates streamflow over the entire continental United States. It provides real-time estimates and forecasts out to 30 days of streamflow across 2.7 million stream reaches as well as distributed snowpack, soil moisture, and evapotranspiration at 1-km resolution. The NWM provides complementary hydrologic guidance at current National Weather Service river forecast locations and significantly expands guidance coverage and type in underserved locations.
Recently Tolson, whose research focuses on advanced methods for environmental simulation model development and use in environmental decision-making, was invited to the University Corporation’s National Center for Atmospheric Research (NCAR) laboratory in Boulder, Colorado to present his research and deliver a workshop about model parameter estimation. The workshop demonstrated the open source optimization software package called Ostrich which houses various single and multi-objective optimization algorithms including Tolson’s single-objective, multi-objective, and parallel optimization algorithms.
“During my visit I learned that the National Water Model is calibrated operationally across the continental United States with my algorithm, Dynamically Dimensioned Search (DDS), and this calibration is repeated annually as the model is revised and updated,” said Tolson. “The use of the DDS algorithm allows the NWM to automatically optimize model parameters to best fit observed streamflow data.”
“Model parameter calibration has contributed significantly to version-over-version improvements in NWM streamflow prediction,” said Tolson. “The DDS algorithm has been shown to narrow in on high-performing parameter sets in fewer number of iterations than other algorithms, and therefore provides an efficient methodology for traversing the NWM parameter space.”
To learn more about Bryan Tolson’s research, visit his research page.