Importance of Model Resolution on Discriminating Rapidly and Non-rapidly Intensifying Atlantic Basin Tropical Cyclones

详细信息    查看全文

• 作者：Andrew Mercer ; Alexandria Grimes
• 关键词：Tropical cyclones ; principal component analysis ; support vector machines
• 刊名：Procedia Computer Science
• 出版年：2016
• 出版时间：2016
• 年：2016
• 卷：95
• 期：Complete
• 页码：223-228
• 全文大小：216 K

文摘

The ability to discriminate rapidly intensifying tropical cyclones (TCs) from their non-rapidly intensifying counterparts remains a major forecasting challenge in operational meteorology. Primarily, approaches to this forecast problem utilize dynamic model data as input into either numerical weather prediction models or statistical algorithms. Recent work suggested higher spatial resolution dynamic simulations will have greater success in distinguishing rapid intensification (RI) of TCs from those that do not, owing to the dynamic model's ability to depict smaller scale features explicitly within the simulation. Despite these preliminary findings, this approach has not been tested with a statistical modeling approach. As such, the scope of this work was to identify the importance of spatial resolution on the ability to forecast the onset of RI and non-RI TCs at 24 hour lead times. To accomplish this, 8 storms of each type were simulated using the Weather Research and Forecasting (WRF) model at varying spatial resolutions (54km, 18km, and 6km). Meteorological fields from the WRF were used as input into a support vector machine classification algorithm trained to discriminate RI and non-RI TC environments.