Comparing Geospatial Interpolation Methods for Modeling Snow Depths

Kryzsko Commons Ballroom

Predicting values of unknown points at geographic locations using known points at other locations seems simple, but many methods offer different solutions to try to predict such points. The goal of this research was to investigate how different geospatial interpolation methods work by comparing their accuracy of predicting snow depths in western Colorado. This data is publicly available and was collected by the Remote Sensing Center at the University of Alabama, who used an FMCW radar system to measure the snow depths. The interpolation methods explored in this project included ordinary kriging, inverse distance weighting, k nearest neighbor, thin plate splines, and universal kriging. Out of all these methods, inverse distance weighting performed the best on this subset of snow depths data.