Diagnostic verification of COAMPS-TC performance for the HFIP retrospective forecasts

COAMPS-TC is a new version of the Coupled Ocean/Atmosphere Mesoscale Prediction System designed specifically for prediction of tropical cyclones (TCs). Recently, a large sample of retrospective COAMPS-TC forecasts of storms in the Atlantic and Eastern Pacific basins was produced in support of the Hurricane Forecast Improvement Project (HFIP). This large sample of cases (over 500 forecasts) provides an excellent testbed for application of new approaches to the verification of TC forecasts. Here, we show an array of verification tools designed to diagnose model deficiencies in predictions of TC position, TC intensity, TC surface wind field, and the near-storm atmospheric environment. Our verification approach incorporates graphical techniques to effectively display the verification data sample, diagnostic techniques to characterize the joint distribution of forecasts and observations, and summary measures to encapsulate different attributes of forecast quality (accuracy, bias, etc.). We will demonstrate that an in-depth characterization of the quality of COAMPS-TC forecasts can lead to useful suggestions for improvements to the model.