Application of the Method for Object-based Evaluation (MODE) for Evaluation of High- resolution Precipitation Forecasts
Abstract
In recent years, the verification and weather forecasting communities have invested a great deal of effort toward the development of new diagnostic approaches for evaluating spatial forecasts of precipitation. These methods are needed in order to provide more informative evaluations of forecast performance than can be obtained from traditional metrics such as the Critical Success Index or Mean-Squared Error. The Method for Object-based Evaluation (MODE) is one of the new approaches, and can be categorized as a "features-based" approach. MODE identifies and compares attributes of precipitation objects that are relevant for particular applications; object attributes can include such features as storm location, size, and intensity, as well as other aspects of forecast and observed precipitation areas that are relevant for a particular application. During the summer of 2008, MODE and other verification approaches were applied to precipitation forecasts from several high-resolution numerical weather prediction models that were included in the "Spring Experiment" sponsored by the NOAA Storm Prediction Center. MODE provided evaluations and comparisons of the heavy precipitation events that were of interest for this project. One of the goals of the 2008 project was to determine what information could be obtained from the MODE and other evaluations that would be relevant for forecasters and other users of the model forecasts. Results from the application of MODE and the other verification methods will be compared and summarized to provide guidance for methods to be used in future evaluations of high-resolution forecast models.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2008
- Bibcode:
- 2008AGUFM.H31A0827B
- Keywords:
-
- 1854 Precipitation (3354)