Design and implementation of a spatio-temporal interpolation model

This dissertation presents and evaluates the viability of a spatio-temporal interpolation model (STIM). The first of three research questions that are addressed relates to defining a generalized framework for spatio-temporal data interpolation. This was addressed by developing a generalized framework for spatio-temporal data interpolation to create a comprehensive spatio-temporal data set from high-resolution spatial data lacking temporal information and data containing detailed temporal information but without the spatial resolution.; The second question addressed the possibility of developing a spatio-temporal interpolation model that incorporates the resolution of the initial data to analyze geographic processes. A STIM, the Diad model, was developed to address the feasibility and functionality of this approach using a case study of urban change in Phoenix, Arizona. The implementation of the Diad STIM focused on the spatio-temporal data needs of the case study. The data used in the case study were the spatially rich remotely sensed Landsat TM satellite imagery and the temporally rich meteorological temperature data. Diad uses interpolation, statistical, and cellular modeling procedures to integrate point temporal data with continuous spatial data to create spatio-temporal datasets having the resolutions of the original data sources.; The last question relates to the ability of the resulting data sets to provide the information needed to detect and analyze geographic variations over space and time. This analysis was undertaken by assessing the data sets accuracy against independent information on data already used by other researchers to detect and analyze geographic variations over space and time in Phoenix.; This research contributes to Geographic Information Science by explicitly defining STIM and the theoretical framework needed to converge the empirical studies of geographic processes and spatio-temporal interpolation. This project provides the STIM framework with a “proof of concept” case study that encourages additional research both in the development of other STIM and in the use of the resultant data to analyze and understand geographic variation and change.