AUTOMATED DETECTION OF DISTURBED LAND SURFACE RELATED TO OIL AND GAS DEVELOPMENT

Comprehensive land management requires the quantitative monitoring of surface activities and impacts. The Bureau of Land Management (BLM) is responsible for regulating the development of Federal oil and gas resources to avoid significant disturbance of the land surface (Federal Land Policy and Management Act of 1976) and ensuring that operators reclaim the disturbed surface upon the conclusion of operations (43 CFR 3162.5-1(b)). Currently, surface disturbance data is provided by operators in Applications for Permit to Drill (APD) and is not typically captured in a geospatial format. The precise location and intensity of surface disturbance is instead monitored through boots-on-the-ground inspections and sporadic manual digitization, neither of which enforce a high degree of precision. This inhibits the ability of the BLM to verify if the extent of surface disturbance falls within permissible limits and prevents thorough numerical analyses needed for regional management plans.

Here we present a method for automating the identification of disturbed surfaces from various imagery products to gauge the accuracy of previous surface disturbance estimates and to generate a geospatial product that can provide the basis for further analysis and reference. We perform both morphometric and spectral analyses of the imagery in order to take advantage of the unique geometries and building materials of oil and gas infrastructure. This method is applied to several imagery products of varying spatial resolutions from oil fields in New Mexico, Colorado, and Wyoming. The results are compared to surface disturbance estimates presented in official land management planning documents and manually digitized surface disturbance products. Our method displays a number of advantages over current estimation techniques, and the limitations of the method are addressed.