To better understand possible causes of intensity overpredictions, various ocean observations were studied and employed in a modeling case study. Underwater gliders documented the rapid cooling and mixing of the warm surface layer with the cold pool below, resulting in much lower sea surface temperatures (SSTs) in the storm's wake. Infrared satellite imageryunique declouded AVHRR datacaptured the post-storm cooling of SSTs, and the data were used as bottom boundary conditions in several iterations of atmospheric hindcasts using the WRF model. Using observed spatial and temporal variations in SST reduced modeled intensity of the stormin some cases by 15 knotsto more closely match NHC best track data and available observations.
The one-dimensional ocean mixed-layer model available in the WRF package decreased modeled wind speed but overprediction and errors remained. The authors indicate the importance of a fully coupled, three-dimensional atmosphere-ocean modeling system in predicting storms similar to Hurricane Irene. Furthermore, it is critical that the modeling system correctly simulates the quick ocean mixing processes that occur during storms like Hurricane Irene, including the role of a bottom ocean boundary layer over the Mid-Atlantic continental shelf.