Temporal validation for landsat-based volume estimation model
This article is part of a larger document. View the larger document here.Abstract
Satellite imagery can potentially reduce the costs and time associated with ground-based forest inventories; however, for satellite imagery to provide reliable forest inventory data, it must produce consistent results from one time period to the next. The objective of this study was to temporally validate a Landsat-based volume estimation model in a four county study area for possible use in large-scale forest inventories. Data were derived from Landsat Thematic Mapper (TM) imagery of Choctaw, Clay, Oktibbeha, and Winston counties in central Mississippi and compared with field inventory data from the Mississippi Institute for Forest Inventory (MIFI) for 1999 and 2006 (Clay County 2007). Comparisons were made between Landsat-based and MIFI inventory-based (90 percent confidence level) cubic foot volume outside bark to a pulpwood top (CFVOBPW) estimates in each of 2 study years. Results showed that, with one exception for the hardwood forest type cover class, estimates from the Landsat-based volume model fell within a reasonable sampling error range (± 20 percent ) when calibrated to the MIFI field inventory. Further research will be conducted on: (1) re-parameterization of the model to correct for bias using a small sample of field plots; (2) the model’s spatial validity; and (3) its potential for providing acceptable data substitutes where field inventory data are missing, too costly, or unattainable.