Extended Abstract (668K)15D.7
Objective tropical cyclone warning guidance using Monte Carlo wind speed probabilities
Andrea B. Schumacher, CIRA/Colorado State Univ., Fort Collins, CO; and M. DeMaria, J. Knaff, C. R. Sampson, and D. P. Brown
Tropical Cyclone Wind Speed Probability products became operational at the National Hurricane Center (NHC) in 2006, replacing its Strike Probability Program. These products are derived from a Monte Carlo method that estimates the probability of 34, 50 and 64-kt winds out to 5 days. Error distributions for this method are based on the previous five years of NHCs track and intensity errors, and wind structure errors from a climatology and persistence model. This Monte Carlo Wind Probability Model (MCWPM) is also run for cases in the central and western North Pacific using error distributions from the Central Pacific Hurricane Center and the Joint Typhoon Warning Center. Since 2006, several applications of these probabilistic products have been explored. One such application is the use of product-generated probabilities to provide objective guidance for NHC's hurricane watch and warning issuance and Department of Defense (DOD) Tropical Cyclone Conditions Of Readiness (TC-COR) issued by individual DOD installations. To provide a guide for applying wind speed probabilities to the deterministic problem of NHC watches and warnings, a statistical analysis comparing past official warnings and wind speed probabilities was performed. Tropical storm and hurricane watch/warning data and wind observations were available from 2004-2008 in the Atlantic. The MCWPM was re-run for all cases in the developmental dataset. Initial results for hurricane warnings suggest that there is a strong statistical relationship between previous warning locations and wind speed probabilities that can be used to define a deterministic warning guidance scheme. Similar results have been found for hurricane watches and tropical storm watches and warnings.
An experimental objective TC-COR guidance product, where probability thresholds were based on the 2004-2008 NHC's Atlantic watch/warning data, was tested in 2009 and well received by DOD operators in the West Pacific. Encouraged by this result, data collection of historical TC-COR information is currently under way that will allow the extension of similar analyses to the N.W. Pacific basin. A preliminary analysis indicates that this more systematic methodology using the historical TC-COR data should result in even further improvements in MCWM-based guidance for the issuance of TC-COR levels. The deterministic watch, warning and TC-COR models will be presented with initial validation results. In addition to providing objective guidance, these models make it possible to relate a set reduction in forecast errors to reductions in coastal length over-warned. This relationship and related socioeconomic benefits will also be discussed.
DISCLAIMER: The views, opinions, and findings in this report are those of the authors and should not be construed as an official NOAA and/or U.S. Government position, policy, or decision.
Session 15D, Probabilistic Forecasting
Friday, 14 May 2010, 8:00 AM-9:45 AM, Tucson Salon A-C
Previous paper
Recorded presentation