Title: Can a Forest/Nonforest Change Map Improve the Precision of Forest Area, Volume, Growth, Removals, and Mortality Estimates?

Author: Gormanson, Dale D.; Hansen, Mark H.; McRoberts, Ronald E.

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: In an extensive forest inventory, stratifications that use dual-date forest/nonforest classifications of Landsat Thematic Mapper data approximately 10 years apart are tested against similar classifications that use data from only one date. Alternative stratifications that further define edge strata as pixels adjacent to a forest/nonforest boundary are included in the test. The variance of stratified estimates of total forest area, volume, growth, removals, and mortality are used to compare results. The two classifications tested are the 1992 and 2001 National Land Cover Data (NLCD) maps, referred to as NLCD-92 and NLCD-01, respectively. The study area is the Minnesota portion of Mapping Zone 41, an area that covers 60 percent of Minnesota and contains 90 percent of the State?s forest. Permanent plot data from nearly 6,500 samples measured over the period 1999 to 2002 are used to compare the alternatives that include four different edge stratifications at the forest/nonforest boundary: one 2-pixel wide stratum, one 4-pixel wide stratum, two 1-pixel wide strata, and two 2-pixel wide strata. Estimates that use stratifications based only on the NLCD-92 were found to be more precise than estimates that used only the more recent NLCD-01 for stratification. Change-based stratifications (those that incorporated the NLCD-92 and NLCD-01) produced estimates with lower variances than estimates based on either single-date stratification alone, with the largest differences observed for the forest area estimates and smaller differences observed for estimates of volume, mortality, growth, and removals.