Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Prediction Errors of Tropical Cyclones in the Western North Pacific in the Met Office Global Forecast Model

Prediction Errors of Tropical Cyclones in the Western North Pacific in the Met Office Global... AbstractThe prediction of tropical cyclones (TCs) in the western North Pacific (WNP) and the Philippines Area of Responsibility (PAR) has been explored in the Met Office (UKMO) global forecasting system over a 10-yr period at 0–7-day lead times. Both the high-resolution deterministic and lower-resolution ensemble systems have been considered. Location errors for verification against the observations are comparable for the deterministic, control, and ensemble mean forecasts; however, the ensemble spread indicates the ensemble is underdispersive. Intensity error metrics, for pressure and surface winds, show large biases relative to the observations, with the smallest biases for the deterministic system. For the intensity metrics the ensemble spread shows the ensemble is severely underdispersive primarily due to the large errors relative to the observations. Verification against the analyses shows similar results to verification against the observations for location. This is also the case for the intensities albeit with smaller errors and less underdispersion. The PAR region has larger intensity errors and biases and larger intensity ensemble spread compared with the broader WNP region. Forecast errors for location and intensity have reduced significantly with system upgrades over the period studied (2008–17) for the deterministic and ensemble systems. Intensity errors for the latest configuration of the deterministic system at day 4 are smaller than the initial errors of all the earlier configurations for both pressure and winds. The Madden–Julian oscillation (MJO) and boreal summer intraseasonal oscillation (BSISO) significantly affect the intensity forecast errors, but not the location errors. Intensity errors are lower at the initiation and for early lead times of the forecasts started in phases 6–7 and 7–8, when the MJO and BSISO are active in the WNP. These reduced errors appear to result mainly from the variations in intensity of the observed storms with MJO and BSISO phases, though the initial states of the forecasts are also affected. Over the studied period, the European Centre for Medium-Range Weather Forecasts (ECMWF) deterministic and ensemble systems have lower errors and biases for both location and intensity than the UKMO forecast systems. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Weather and Forecasting American Meteorological Society

Prediction Errors of Tropical Cyclones in the Western North Pacific in the Met Office Global Forecast Model

Hodges, K. I.; Klingaman, N. P.
Weather and Forecasting , Volume 34 (5): 21 – Oct 26, 2019
Download PDF
21 pages

Loading next page...
 
/lp/american-meteorological-society/prediction-errors-of-tropical-cyclones-in-the-western-north-pacific-in-enshsSJ7eI

References (54)

Publisher
American Meteorological Society
Copyright
Copyright © American Meteorological Society
ISSN
1520-0434
eISSN
1520-0434
DOI
10.1175/WAF-D-19-0005.1
Publisher site
See Article on Publisher Site

Abstract

AbstractThe prediction of tropical cyclones (TCs) in the western North Pacific (WNP) and the Philippines Area of Responsibility (PAR) has been explored in the Met Office (UKMO) global forecasting system over a 10-yr period at 0–7-day lead times. Both the high-resolution deterministic and lower-resolution ensemble systems have been considered. Location errors for verification against the observations are comparable for the deterministic, control, and ensemble mean forecasts; however, the ensemble spread indicates the ensemble is underdispersive. Intensity error metrics, for pressure and surface winds, show large biases relative to the observations, with the smallest biases for the deterministic system. For the intensity metrics the ensemble spread shows the ensemble is severely underdispersive primarily due to the large errors relative to the observations. Verification against the analyses shows similar results to verification against the observations for location. This is also the case for the intensities albeit with smaller errors and less underdispersion. The PAR region has larger intensity errors and biases and larger intensity ensemble spread compared with the broader WNP region. Forecast errors for location and intensity have reduced significantly with system upgrades over the period studied (2008–17) for the deterministic and ensemble systems. Intensity errors for the latest configuration of the deterministic system at day 4 are smaller than the initial errors of all the earlier configurations for both pressure and winds. The Madden–Julian oscillation (MJO) and boreal summer intraseasonal oscillation (BSISO) significantly affect the intensity forecast errors, but not the location errors. Intensity errors are lower at the initiation and for early lead times of the forecasts started in phases 6–7 and 7–8, when the MJO and BSISO are active in the WNP. These reduced errors appear to result mainly from the variations in intensity of the observed storms with MJO and BSISO phases, though the initial states of the forecasts are also affected. Over the studied period, the European Centre for Medium-Range Weather Forecasts (ECMWF) deterministic and ensemble systems have lower errors and biases for both location and intensity than the UKMO forecast systems.

Journal

Weather and Forecasting – American Meteorological Society

Published: Oct 26, 2019

There are no references for this article.