Hurricane Forecast Modeling

Ramsay (2017, DOI: 10.1093/acrefore/9780199389407.013.79)

Project description: Analyze historical hurricane data to build predictive models for the following purposes:

• determine whether or not a hurricane will go through “rapid intensification” in the next 24 hours
• predict hurricane intensity in the next 24 hours
• predict the total economic loss that will be created by a given hurricane

This project was in collaboration with NASA (JPL) to provide them a new predictive solution.

1. Data Sets

The following data sets were utilized:

1. Statistical Hurricane Intensity Prediction Scheme (SHIPS)
2. Second-Order Statallite IR Data: Hurricane Precipitation
3. Economic Loss Caused By Each Hurricane
4. GDP Level by each Latitude and Longitude

2. Hurricane Rapid Intensification (RI) Forecast Model (Classifier)

Hurricane rapid intensification Forecast is the key to preventing damage caused by a hurricane. The most damage is caused when a hurricane with low intensity suddenly intensifies as it lands on major cities. This model utilizes various machine learning and neural network algorithms to create an ensemble model to accurately predict whether or not a hurricane will go through RI in next 24 hours.

3. Hurricane Intensity Forecast Model (Regressor)

This model is alternative form of the RI Forecast model described above. However, this model predicts a numerical value of hurricane intensity. The model was able to achieve RMS Error less than 13 knots.

Additional information regarding the model cannot be shared due to NDA.