115 Ensemble Hurricane forecasts from a suite of Mesoscale models

Tuesday, 25 January 2011
Mrinal K. Biswas, Florida State University, Tallahassee, FL; and T. N. Krishnamurti

One of the main goals of the Hurricane Forecast Improvement Project (HFIP) is to reduce the hurricane track and intensity by 50% in the next decade. A multi-agency effort was initiated starting 2009 hurricane season and considerable progress was made in issuing realtime forecasts from a suite of state of the art mesoscale models with twice a day forecast going out to 120 hours. A consensus forecasts was also made by FSU and found to have great potential in reducing the track and intensity errors compared the member models as well as the HFIP baseline errors. Since all NWP models have inherent biases and systematic errors a bias corrected ensemble forecasts is tested from a cross validation approach and was found to have considerable skill compared to a simple ensemble mean during the 2009 season. This methodology will be applied to the 2010 season storms and a detailed report will be presented on its performance for the whole season. An in-house version of the ARW will be part of the ensemble and the performance of the model will also be presented. The physics options for this ARW were arrived at from a large experimentation, this model provided the least errors among all models for the hurricane forecasts.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner