Title: Comparison of U.S. Forest Land AreaEstimates From Forest Inventory and Analysis, National Resources Inventory, and Four Satellite Image-Derived Land Cover Data Sets

Author: Nelson, Mark D.; McRoberts, Ronald E.; Lessard, Veronica C.

Year: 2005

Publication: In: Proceedings of the fifth annual forest inventory and analysis symposium; 2003 November 18-20; New Orleans, LA. Gen. Tech. Rep. WO-69. Washington, DC: U.S. Department of Agriculture Forest Service. 222p.

Abstract: Our objective was to test one application of remote sensing technology for complementing forest resource assessments by comparing a variety of existing satellite image-derived land cover maps with national inventory-derived estimates of United States forest land area. National Resources Inventory (NRI) 1997 estimates of non-Federal forest land area differed by 7.5 percent from estimates based primarily on Forest Inventory and Analysis data reported in the Forest and Rangeland Renewable Resources Planning Act of 1974 (RPA) draft 2002 forest resource assessment. The NRI estimates differed only 2.2 percent from non-Federal land area, with the NRI estimate slightly smaller than the RPA estimate. Comparisons of statewide forest land area estimates derived from these two inventories with four satellite image-derived maps reveal area-weighted root mean square deviations ranging from 2.5 to 41.0 percent across the conterminous United States. In general, estimates of non-Federal forest land area from RPA and NRI were more closely related to each other than to image-derived estimates. The Forest Cover Types map and the National Land Cover Data set produced image-derived estimates that were most similar to the RPA estimate of forest land area across all land ownerships.