Using Statistical Process Control Charts (SPCC) to Determine Optimum Resolution for Geomorphometric Analyses.

Abstract

Studying landscapes can give great insight into what geological processes lead to their current appearance. The use of digital elevation models (DEMs) helps scientists better understand the processes that created or modified landscapes, especially with study areas that are geographically extensive. As resolution has increased, scientists have started to notice more subtle details in features that have not been previously reported in coarser resolution studies. This study’s focus was on assessing the impact of DEM resolution on the delineation of landforms. There were two main objectives of this thesis: (1) Determine the optimum resolution for measuring drumlins and (2) automating the analysis by customizing ArcGIS 9.3 for the morphometric measurements for use with Statistical Process Control Charts (SPCCs) to determine the optimum resolution of drumlins. SPCCs are primarily used for process and quality control in manufacturing and industrial applications. The principles of this method can be applied similarly to the problem of coarsening of DEM resolution. The problem was treated like an industrial process where the finest resolutions was treated as the control set, then as resolutions get coarser, a point is reached where the values exceed acceptable values and are deemed unreliable. The last resolution before the unreliable value is then deemed the optimum resolution. The study area was in Palmyra, NY, chosen due to the abundance of drumlins, which are streamlined glacial landforms that are easily recognizable on contour maps or DEMs. Results indicate that (1) 10 m DEMs were consistently within the control limits. (2) Morphometric drumlin analysis can be automated once drumlins are delineated using a bounding container script. The only task not automated was delineating the drumlins themselves. This study’s focus was on assessing the impact of DEM resolution on the delineation of landforms. SPCC helped give statistical significance to optimum resolution rather than more simplified methods (e.g., inflection point or relative error). All of this combined could lead to discoveries in landform delineations, patterns, or genesis of not only drumlins, but possibly other landforms or landscapes