Flood Prediction Based on Multidimensional Analysis of Precipitation and Inundation in the Mekong River Delta

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Sunday, 4 January 2015
Andrew Fitzgerald, NOAA, New York, NY

Cross decomposition algorithms such as, partial least squares regression and canonical correlation analysis, are crucial techniques for understanding higher dimensional data sets. They reduce data to lower dimensions that represent the strongest representative factors. In this project, we applied techniques to precipitation and inundation data of the Mekong river delta from the CCNY Crossroads Initiative and the Dartmouth Flood Observatory. By reducing the dimensions of this data we aim to predict anomalous inundation by observing precipitation.