-
- A33E-04: Land-Ocean Comparison of Forecast Errors in Track and Intensity of North American Landfalling Tropical Cyclones
-
Board 2231‚ Hall EFG (Poster Hall)NOLA CC
Author(s):Generic 'disconnected' Message
Vishal Juneja, Louisiana State University (First Author, Presenting Author)
Huanping Huang, Louisiana State University
To improve hurricane forecasts when they matter most—as storms approach land—we studied how forecasting models behave differently over the ocean versus over land. By analyzing over 30 years of forecast data, we acted like detectives to pinpoint specific patterns in their mistakes. We made several key discoveries. First, we found that most models struggle more with predicting a storm's strength (its wind speed) while it's still at sea, likely due to intensification. For predicting a storm's path, we uncovered a fascinating split: models that simulate the atmosphere's physics were more accurate over land (Dynamical), while models that rely on historical data (Statistical) performed better over the ocean. The good news is that forecasts are improving, and we saw that the modern, physics-based models are getting better at predicting storms over land at a faster rate. Our research provides a diagnostic report for forecasters, showing them the new regime of improvements. This allows them to target specific weaknesses, build better statistical/dynamical models, and ultimately create more reliable forecasts that help communities better prepare for the critical moment a storm makes landfall.
Scientific DisciplineSuggested ItinerariesNeighborhoodType
Enter Note
Go to previous page in this tab
Session
