Download (PDF 63 KB): https://research.fs.usda.gov/download/treesearch/14429.pdf
Abstract
Stratifications developed from National Land Cover Data (NLCD) and from photointerpretation (PI) were tested for effectiveness in reducing sampling error associated with estimates of timberland area and volume from FIA plots in western Oregon. Strata were created from NLCD through the aggregation of cover classes and the creation of 'edge' strata by reclassifying pixels at class boundaries. Strata were created from aerial photography by interpretation of a sample grid for land use and cover attributes. NLCD-based stratifications are less costly than PI and sacrifice little precision on inventory estimates. Neither PI nor NLCD stratifications achieved the FIA target of 10 percent sampling error/billion cubic feet of volume, but both are near the national standard of 3 percent sampling error/million acres of timberland.
Parent Publication
Citation
Dunham, Paul; Weyermann, Dale; Azuma, Dale. 2002. A comparison of stratification effectiveness between the National Land Cover Data set and photointerpretation in western Oregon. In: McRoberts, Ronald E.; Reams, Gregory A.; Van Deusen, Paul C.; Moser, John W., eds. Proceedings of the Third Annual Forest Inventory and Analysis Symposium; Gen. Tech. Rep. NC-230. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Research Station: 63-68
