Assessing High-Resolution Weather Research and Forecasting (WRF) Forecasts Using an Object-Based Diagnostic Evaluation.

摘要

The Model Assessment Project conducted an investigation into the applicability of the Model Evaluation Tools (MET), Method for Object-Based Diagnostic Evaluation (MODE) tool, which was designed to perform spatial verification of numerical weather prediction (NWP) model forecasts. The NWP model used during the investigation was a version of the Weather Research and Forecasting-Advanced Research Weather Research and Forecasting (WRF-ARW) model, which is tailored to address Armyscale horizontal spatial resolutions of 1 3 km. This model is called the Weather Running Estimate Nowcast (WRE-N). The WRE- N was run over three nested grids with the 1-km inner nest grid spacing being the study focus. The observations were surface meteorological variables from independent gridded analyses. MODE compared meteorological features or objects defined from the forecast and observed fields for the same valid time on the basis of measureable attributes. It then quantified the differences between corresponding objects as a measure of forecast error. MODE was designed to evaluate errors in precipitation forecasts, but little work had been done with continuous variable fields. The focus of this study was to assess the application of MODE to NWP models and high-resolution meteorological variables over small, Army-relevant domains.