Spatial and temporal modelling of aboveground carbon stocks using Landsat TM and ETM plus for a subboreal forest

Forest carbon (C) stocks and sequestration have become an important management consideration for countries with large forested regions. A series of Landsat Thematic Mapper (TM) images was obtained for the Aleza Lake Research Forest (ALRF) in subboreal British Columbia. Plot-based aboveground biomass and woody debris C stocks measured in 2003 and 2004 were related through regression analysis to TM and spatially explicit forest cover information. Two empirical models were developed, namely a biomass C regression model (BCRM) and a woody debris C regression model (WDCRM), with r(2) values of 0.67 and 0.64, respectively. Uncertainties in C stock estimates were determined using a Monte Carlo uncertainty analysis. In 2003, the total C stocks in biomass and woody debris over the 6034 ha area were 588 +/- 7 and 77 +/- 2 kt, respectively. During the time period from 1992 to 2003, forest harvesting operations accounted for a loss of 39.9 +/- 2.6 kt C, and remaining areas accounted for a gain of 31.3 +/- 10.3 kt C. Therefore, there was a small and nonsignificant loss of 8.7 +/- 10.6 kt C over this 11 year interval for the ALRF, implying that aboveground carbon stock losses associated with clearcuts were offset by carbon stock gains achieved through forest growth.