"When developing a ground water model, the quality of the dataset should first be evaluated. Spatial outliers can lead to predictions which are not representative of actual conditions. In order to isolate misrepresentative points, a method is presented which examines the experimental variogram of a ground water elevation dataset. To define a threshold variance between pairs of ground water elevation measures, ground elevation values from a digital elevation model (DEM) are used to determine a maximum reasonable variance expected to occur on the experimental variogram. To determine appropriate DEM parameters, a separate study was also done which observed characteristic behavior of gradient calculations for a DEM with fluctuating resolution and extent. This method is applied first to a synthetic dataset and then to a monitoring well network at Fort Leonard Wood, Missouri. Results of the analysis show that all points targeted as spatial outliers in the case study are justified for removal. This approach can readily be incorporated into the development of a regional groundwater model by kriging. The strengths of this method are that it incorporates supplemental DEM building of the concept that the groundwater surface is a smoothed version of the topographic surface. This method also takes advantage of every point pair relationship in that both neighboring points and distant pairs are compared"--Abstract, page iv.
Guggenberger, Joe D.
Guggenberger, Joe D.
Elmore, Andrew C.
Geosciences and Geological and Petroleum Engineering
M.S. in Geological Engineering
United States. Army. Corps of Engineers.
Missouri University of Science and Technology
Journal article titles appearing in thesis/dissertation
- Development of a variogram procedure to identify spatial outliers using a supplemental digital elevation model
x, 98 pages
© 2017 Zane Daniel Helwig, All rights reserved.
Thesis - Open Access
Electronic OCLC #
Helwig, Zane Daniel, "Development of a variogram approach to spatial outlier detection using a supplemental digital elevation model dataset" (2017). Masters Theses. 7685.