Accuracy of tropical cyclone induced winds using TYDET at Kadena AB

When a tropical cyclone (TC) is within 360 nautical miles of Kadena AB, the Air Force's Typhoon Determination (TYDET) program is used to estimate TC-induced winds expected at the base. Best-track data and Joint Typhoon Warning Center (JTWC) forecasts are used to evaluate systematic errors in TYDET. The largest contributors to errors in TYDET are a systematic error by which wind speeds are too large and the lack of size and symmetry parameters. To examine these parameters, best-track and forecasts are used to classify TCs as small or large and symmetric or asymmetric. A linear regression technique is then used to adjust TYDET forecasts based on the best-track and forecast position, size, and symmetry categories. Using independent data, over 65 percent of the overall cross-wind forecasts were improved and more than 60 percent of the cross-wind forecasts were improved when verifying conditions noted a cross-wind of 20 knots or greater. The effectiveness of the corrections and implications for TYDET forecasts are examined in relation to errors in forecast data used to initialize TYDET. A similar approach as developed here for the TYDET model at Kadena AB is proposed for other bases within the Pacific theater.