Extended Abstract (292K)2A.5
Forecasting September Atlantic basin tropical cyclone activity at zero and one month lead times
Philip J. Klotzbach, Colorado State University, Fort Collins, CO
Annual tropical cyclone activity in the Atlantic basin has been forecasted by researchers at Colorado State University for the past seventeen years with considerable success. With the advent of the reanalysis dataset created by the National Center for Environmental Prediction (NCEP), a powerful new tool now exists to conduct research on shorter time periods within the hurricane season. September is the most active month during the hurricane season with over 50 percent of all tropical cyclone activity occurring during that month. Skillful prediction of September Atlantic basin tropical cyclone activity would be of value and interest to a variety of government and private interests.
Utilizing the NCEP reanalysis data, a prediction scheme to forecast tropical cyclone activity in the month of September has been developed. Based on preliminary hindcasting results from 1950-2001 (52 years), approximately 50 percent (r=0.7) of the variance in tropical cyclone activity can be forecasted by the end of July. This number increases to approximately 60-65 percent (r=0.8) by the end of August. These global parameters include zonal wind at 200 and 1000 mb, sea level pressure, geopotential heights and sea surface temperatures at widely varying global locations. These parameters also show strong correlations with United States landfalling tropical cyclones indicating that United States landfall probability forecasts can also be issued at zero and one month lead times. During the 26 most active September periods since 1950, there have been 23 landfalling hurricanes along the U.S. coast compared to 10 hurricanes during the least active 26 years or a ratio of 2.3:1. With skillful forecasts of tropical cyclone activity for the month of September, it is possible to make forecasts of the probability of a landfalling storm along the U.S. coast with appreciable skill. In addition, the July forecast of September activity considerably improves the 1 August seasonal forecast prediction of Dr. William Gray.
Session 2A, Seasonal to Interannual Prediction and Predictability II (Parallel with Sessions 2B, 2C, & 2D)
Monday, 29 April 2002, 2:00 PM-3:30 PM
Previous paper