Does enhanced horizontal resolution improve hurricane intensity forecasts?

As part of the High-resolution Hurricane (HRH) test for NOAA's Hurricane Forecast Improvement Project (HFIP), the Advanced Hurricane-research WRF (AHW) was used to perform 69 pairs of simulations for a total of 10 Atlantic tropical cyclones, varying only the horizontal grid spacing. The model was initialized using an ensemble Kalman filter consisting of 96 members at 36-km grid spacing. Conventional observations, plus vortex central pressure and location were assimilated for each case in a 6-h cycling procedure beginning about 2 days prior to formation of a tropical depression. Pairs of forecasts, one integrated on a single 12-km grid, the other with storm-centered, moving nests of 4-km and 1.33-km grid spacing, were integrated to 126 h or until the time the observed storm dissipated. The high-resolution simulations marginally improved upon the intensity prediction at coarse resolution. Both versions exhibited measurable skill for rapid intensification, but not for rapid weakening. By comparison, statistical and operational forecasts for a homogeneous dataset offered no skill for rapid intensity change. Track errors were nearly identical in the two sets of simulations, and slightly larger than official forecast tracks. Overall, we find that using high resolution marginally improves intensity prediction but does not improve track prediction.