Performance of Experimental HAFS-B During the 2024 Hurricane Season

Abstract The Hurricane Analysis and Forecast System (HAFS) was implemented operationally in 2023, and work is ongoing to develop and improve future versions of HAFS. During the 2024 hurricane season, an experimental version of HAFS-B (HAFSV2.0.1B or HFXB) was tested in near-real-time in the Atlantic and eastern Pacific basins. This experimental version features several upgrades to the model, including data assimilation improvements as well as testing the scale-aware Tiedtke convective scheme in HAFS for the first time. The track bias characteristics were very different from other versions of HAFS run in 2024, with much less across-track bias. This difference is promising for maintaining diversity between HAFS-A and HAFS-B moving forward, and also for continuing to improve the overall track forecast skill. HFXB also showed comparable skill to operational models for rapid intensification, though some lingering negative intensity bias is something that will need to be examined further. Detailed analysis of several key storms from the 2024 Atlantic season, including Hurricane Beryl, Hurricane Helene, and Hurricane Milton, highlights some of the strengths of these experimental forecasts, including the ability to accurately predict the fine-scale structure changes associated with rapid intensification, which helps forecasters improve intensity predictions and warnings ahead of the storm. These case studies also highlight areas for improvement in future upgrades to HAFS, motivating further research for both operational and experimental testing.