Researchers at The University of Melbourne in Australia, recognizing that traditional flood models are very time-consuming even though very accurate, have created a new model for flood prediction that maintains high accuracy but takes significantly less time to run. With increasing flooding around the world, particularly as the climate continues to change and rainfall intensity increases, being able to model where water will flow is imperative before the floods occur.

This model does not forecast locations of the rainfall itself, but once intense rain occurs, it can be run for large river basins in less than a minute instead of hours to very accurately simulate where water inundation will occur. Called the Low-fidelity, Spatial analysis, and Gaussian Process learning (LSG) model, this can simulate both the extent and depth of the flooding. With a changing and always chaotic atmosphere where locations of intense precipitation can be extremely difficult to pinpoint in advance, this new model should facilitate faster evacuation and mitigation once the intense rainfall does occur.