Date of Award:
Master of Science (MS)
Mathematics and Statistics
Creating maps of continuous variables involves estimating values between measurement locations scattered throughout a geographic region. These maps often leverage observed similarities between geographically close measurements, but may also make predictions using other geographic information such as elevation. The relationship between the available geographic information and the variable of interest can vary with location, especially when mapping large areas like a continent. A simple way to account for the changing relationship is to divide the space into different sub-regions and model the relationship at each region. The naive implementation of this approach has the side effect of making sudden changes in predictions at the borders of each region. This thesis describes a novel regional border smoothing method that allows for the formation of a continuous map built with regional models. The method is implemented and available to the public through the open source R package remap. Improvements in model accuracy are demonstrated using a national scale and a state scale dataset.
Wagstaff, Jadon S., "Regionalized Models with Spatially Continuous Predictions at the Borders" (2021). All Graduate Theses and Dissertations. 8065.
Copyright for this work is retained by the student. If you have any questions regarding the inclusion of this work in the Digital Commons, please email us at .