"Automating the Use of Geostatistical Tools for Lake Tahoe Area Study" describes how powerful analysis tools in the ArcGIS Geostatistical Analyst 9.2 extension were applied in a study of air quality degradation around Lake Tahoe. A model to optimize the monitoring network by locating additional monitoring stations was built using ModelBuilder.
[Click the start button below to view a series of maps showing the process employed in optimizing the monitoring network.]
New monitoring point locations were chosen to improve the overall reliability of the geostatistical interpolation by sampling at the locations within the study area where reliability was the lowest. In addition, all supplemental points had to be located within the forested portion of the study area. As expected, increasing the density of the monitoring network decreased the standard error of prediction for the entire study area. To better measure the increase in reliability caused by adding supplemental stations, the output geostatistical surfaces of prediction standard error generated by the model were converted into rasters with the same color schema.