EXPERIMENTS IN NUMERICAL FORECASTING OF TROPICAL STORM MOVEMENT.

• Main Author: Kerr, James Earl
• Other Authors: NAVAL POSTGRADUATE SCHOOL MONTEREY CA
• Format: Text
• Language: English
• Published: 1966
• Subjects: Meteorology; Statistics and Probability; (*TROPICAL CYCLONES; WEATHER FORECASTING); NUMERICAL METHODS AND PROCEDURES; THESES; TRACKING; CORIOLIS EFFECT; BAROMETRIC PRESSURE; PERIODIC VARIATIONS; ATMOSPHERIC MOTION; ATMOSPHERIC PRECIPITATION; OPTIMIZATION; ATLANTIC OCEAN; PACIFIC OCEAN; ACCURACY; STATISTICAL ANALYSIS; AIR MASS ANALYSIS; NORTH ATLANTIC OCEAN; NORTH PACIFIC OCEAN; Pacific; North Atlantic
• Online Access: http://www.dtic.mil/docs/citations/AD0807381; http://oai.dtic.mil/oai/oai?&verb=getRecord&metadataPrefix=html&identifier=AD0807381

Numerical prediction is rapidly becoming the most accurate approach to forecasting tropical cyclones. The numerical model in this study utilizes the U.S. Navy Fleet Numerical Weather Facility's (FNWF) so-called decomposition (SR) fields to produce geostrophic steering currents for tropical storms and/or hurricanes and typhoons. Fifteen of the 1965 hurricanes and typhoons were used to obtain over 200 twelve and twenty-four hour storm predictions, in one-hour time steps, for each of the following FNWF analyses: 1000, 500, 200, 1000/500, 1000/200, and 500/200 -mb SR fields. Results indicate that the 500-mb SR fields, including a statistical correction for latitude and/or longitude, yielded the least forecast error for Pacific typhoons with average 12 and 24-hour storm errors of 69 nm and 154 nm, respectively. Forecasts for Atlantic hurricanes verified best for 12-hour forecasts at 1000 mb, using a similar geographical modification, with an average error of 97 nm. For 24 hours the hurricanes were steered best at 500 mb, with modification, resulting in an average forecast error of 177 nm. (Author)