IDA@SMU Banner

PIIH: Learning Prediction Intensity Interval model for Hurricanes

Hurricane intensity (maximum sustained wind speed) is one of the most important indicators for a hurricane's destructive power. Unfortunately, hurricane intensity is notoriously difficult to predict. At the IDA@SMU research lab, we apply a novel data stream modeling technique called TRACDS to improve the prediction of hurricane intensity. This new approach named PIIH dynamically models hurricane life cycle behavior using historic data and then applies these models to predict hurricane intensities up to 5 days in advance. What is completely new with this approach is the fact that it models the hurricane life cycle and, in addition to single value predictions, it also provides ranges (high to low) for the expected intensity.

The new approach was used for the fist time for live real-time predictions for the 2011 hurricane season. Results can be found on the 2011, 2012 and the live 2013 prediction pages.

Results

Media Coverage

Publications

Acknowledgements

We would like to thank James Franklin (Hurricane Specialist Unit, NHC, NOAA) and Dr. Mark Demaria (NESDIS Regional and Mesoscale Meteorology Branch, CIRA) for their help.

NSF This work is based on TRACDS sponsored by the National Science Foundation under Grant No. IIS-0948893.

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.