Masters Thesis

Spatial interpolation: a simulated analysis of the effects of sampling strategy on interpolation method

Spatial interpolation is a procedure for estimating data at unsampled locations using known, measured locations within the same area. Choice of sampling strategy and sample size play an important role in determining an interpolator's effectiveness in deriving a spatially accurate data surface. Given the widespread application of interpolated datasets as input to geospatial analyses and modeling procedures, it is critical to understand what effect sampling strategy, sample size, and interpolation method have on the quality and accuracy of these data. The objective of this study was to evaluate the influence of sampling strategy and sample size on the ability of interpolation methods to minimize error in a data surface of estimates for a range of possible conditions. Script tools were developed to automate data processing and evaluation of the various model scenarios. Results obtained from this study suggest that a uniform sampling strategy, when compared to a random sampling strategy, consistently minimizes interpolation error for a variety of surface conditions. The results also suggest that regardless of sampling strategy or interpolation method, as sample size increases interpolation error decreases. In all instances, as the complexity of the data increases, in terms of variability, the ability of all interpolation methods to provide spatially accurate estimates diminishes. The methodologies and tools developed for this study provide a baseline approach for evaluating the role that sampling strategy and sample size have on an interpolator's ability to create accurate data surfaces for a variety of surface conditions and provide further insight into additional areas for future exploration.

Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.