Research Projects

A Method for Direct Measurement of Impervious Cover Using Remote Sensing and Spatial Modeling Techniques: A Case Study in the Lake Tahoe Basin

Co-PIs: Tim Minor and Mary Cablk

Project Period: May 2000 - May 2001

Funded by: Tahoe Regional Planning Agency

Right: Lake Tahoe

Keywords: Lake Tahoe Basin, impervious cover, IKONOS multispectral imagery, principal components analysis, land capability

Project Description

The Tahoe Regional Planning Agency (TRPA) is responsible for drafting regulations on land development and land use and for evaluating the effects of consequential regulatory choices based on these land use models. For this reason, improved methods for calculating impervious surface cover estimates are needed for the Lake Tahoe Basin to assist with TRPA's land use modeling efforts. Accurate estimation of impervious cover in the Lake Tahoe Basin is relevant to all TRPA thresholds including wildlife and fish habitat, noise, recreation and scenic resources, air and water quality, soils, and vegetative cover.

The Desert Research Institute (DRI) conducted a pilot analysis of a small study area, (Figure 1), in the Lake Tahoe Basin to evaluate the feasibility of applying spatial and predictive modeling techniques using high resolution IKONOS^TM satellite imagery and geographic information systems (GIS) to derive more accurate estimates of impervious cover. This goal of this study was to calculate discrete measures of impervious cover using image processing techniques and spatial modeling in a GIS in a small study area in South Lake Tahoe. To accomplish this goal, we developed a protocol to model both exposed impervious cover and cover that was obscured by tree canopy and shadows. A modeling approach was proposed because the IKONOS^TM data utilized was expected to have the same difficulties that other air and space-borne sensors have exhibited with dense canopies. The results were contrary to this assumption and we found that sub-canopy and sub-shadow surfaces were not only detectable, but also discernible with respect to the underlying surface cover using image processing methods combined with spatial modeling tools. An image processing/spatial modeling method based on principal components analysis and spatial morphological operators was developed using ENVI image processing software and ArcView GIS. Based on an accuracy assessment performed using 170 ground verification points selected throughout the study area, an overall classification accuracy of 92.94% was obtained, with an even higher user or "reliability" accuracy of 95.83%. Impervious cover estimates for the entire 25 km² area were calculated at 4.15 km², or 16% of the total surface area. The resultant impervious cover image of the area is shown in Figure 2.

Return to Projects List