PROOF OF CONCEPT FOR AN APPROACH TO A FINER RESOLUTION INVENTORY by Cieszewski et al. 2005

摘要

This report presents a proof of conceptfor a statistical framework to develop a timely, accurate,and unbiased fiber supply assessment in the State ofGeorgia, U.S.A. The proposed approach is based onusing various data sources and modeling techniquesto calibrate satellite image-based statewide stand lists,which provide initial estimates for a State inventoryon a common timeline. The system is based on usingGeorgia ground inventory data from the forest productsindustry, enhanced by various geographic informationsystem and remote sensing data, and applied with thek-th “nearest neighbor” methods to time-series-stratifiedsatellite imagery. The initial estimates are then scaledregionally to the Forest Inventory and Analysis (FIA)summary totals to eliminate potential bias in the initialestimates. The system enhances the FIA inventorydata in four significant ways. First, it removes theneed for the specific FIA plot coordinates; although,the coordinates, if available, would probably enhancethe analysis. Second, it provides a current commontimeline of inventory estimates based on the LandsatThematic Mapper imagery for the given year and season.Third, it provides currently accurate high-resolutionarea estimates. Last, it uses various auxiliary dataavailable from private and public sources in the Stateand can easily take advantage of other data as theybecome available.