Tropical cyclone intensity change predictability estimates using a statistical-dynamical model

The NOAA Hurricane Forecast Improvement Project (HFIP) has set 5- and 10-year goals to improve tropical cyclone intensity forecasts by 20% and 50%, respectively, relative to a baseline from 2006-2008. An underlying assumption in these goals is that they do not exceed the limits of predictability for intensity change forecasting. At the present time, the most accurate objective intensity forecasts are from statistical-dynamical models, although there is a considerable effort within HFIP to improve those from three-dimensional coupled ocean-atmosphere prediction systems. In this paper, the limits of predictability of intensity prediction will be assessed using a methodology developed by Charles Neumann in the mid-1980s for track models. At that time, statistical-dynamical track models were generally better than those from dynamical track models, similar to the situation today for intensity prediction. In the Neumann study, a statistical-dynamical track model was run with “perfect-prog” input where the initial motion input was from the National Hurricane Center best track and the environmental predictors were from analysis fields rather than forecast fields. Using this method, Neumann concluded that Atlantic track error reductions of about 50% were possible out to 72 hr, relative to the 1976-1985 baseline, and that the leveling off of the track error improvements seen at that time was only temporary. As it turned out, both of these predictions were true. The track forecast error reduction accelerated beginning in the 1990s, and the 50% improvement was achieved by the early 2000s.

Given that Neumann's method provided reliable estimates of track forecast improvements in the following two decades, the same methodology will be applied to intensity prediction using the statistical-dynamical Logistic Growth Equation Model (LGEM). Atlantic forecasts from 2004-2008 will be considered. The replacement of the forecast track with the best track and the model fields with analysis fields will be performed separately and together in LGEM. This procedure will allow the assessment of the impact of track errors on intensity errors and the impact of large-scale environmental forecast errors on intensity errors.

The views, opinions, and findings contained in this report are those of the authors and should not be construed as an official National Oceanic and Atmospheric Administration or U.S. Government position, policy, or decision.