|
Skill Assessment The first step of implementing this ELCIRC model application at NOAA/NOS was to evaluate the model performance over a one-year period. The year 2002 was selected for a hindcast simulation, as there was a significant amount of data available for evaluating the model during this time frame. For this 2002 hindcast simulation, external model forcings included tides from a regional tide model, oceanic temperature and salinity profiles (for nudging along the open ocean boundary) from the Navy Coastal Ocean Model, atmospheric forcing variables (for winds, pressure and heat flux) from NCEP reanalysis fields, Columbia and Willamette riverflows from USGS gauges (14128870 and 14211720, respectively), and Fraser riverflows from the Hope gauge of Environment Canada. These inputs were compiled by OHSU and transferred for implementation on NOS computers and evaluation of results. NOS has developed a set of skill assessment criteria and software for evaluating the performance of hydrodynamic models in development for operational implementation. Some of the key statistics used in this assessment are summarized in the table below, and target accuracies are given in its rightmost column. The acceptable error, X, is 15 cm for water levels (based on underkeel clearance needs) and 26 cm/sec for velocities (based on maneuvering estimates in ports/channels).
More information is available at http://www.csc.noaa.gov/csp |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||