Effects of Storm-Related Parameters on the Accuracy of the Nested Tropical Cyclone Model.
NAVAL POSTGRADUATE SCHOOL MONTEREY CA
Pagination or Media Count:
The performance of the Nested Tropical Cyclone Model NTCM for 542 track forecasts in the western North Pacific is evaluated with respect to five storm-related parameters intensity, 12-h change in intensity, latitude, longitude and size. This study is intended to aid the operational forecaster in deciding when to use the NTCM based on storm-related parameters at the forecast time. The storm-related parameters are divided into three subsamples about 180 in each and the forecasts are evaluated in terms of the mean forecast error, median forecast error and systematic zonal and meridional error. Cross-track CT and along-track AT components are computed relative to a Climatory and Persistence CLIPER track. A scoring system M that assesses penalty points for forecasts in incorrect terciles is used to compare the accuracy of the NTCM and CLIPER forecasts within the subsamples. For the entire sample, the NTCM has a slow bias, especially at the 12- through 36-h forecast periods. It also performs better for storms with initial latitudes south of 13 N and initial longitudes west of 129 E. For very large storms, the NTCM forecasts have both left-of-track and westward biases which indicate problems of the NTCM in predicting recurvature of such systems. The NTCM which has a 60-kt bogus forecasts for storms with initial intensities between 50 and 75 kt have much lower CTAT M scores and smaller forecast errors than the subsamples with initial intensities less than 50 kt or greater than 75 kt. Keywords Tropical cyclone forecasting Forecast error statistics Typhoons Recurvature Tropical cyclones Cross track error Along track error Theses.
- Statistics and Probability