Journal Article
Quantification of the above ground carbon stock (AGC) is important in sus-
tainable forest management and policy advice on climate change mitigation.
Traditional ground vegetation survey methods have been used to provide data
for estimation of AGC stock but constrained by inadequate time and often too
costly. Remote sensing when combined with few ground collected data has the
potential of improving forest resource assessment even though this opportu-
nity has not well been utilised. In this study, we mapped AGC through com-
bination of ground survey data collected from 51 permanent sapling plots
with Normalized Difference Vegetation Index (NDVI) derived from Landsat 5
Thematic Mapper image. Linkage of the two data sources was made during a
training stage of supervised classification. The overall classification accuracy
was 98%, suggesting that reliable estimate of AGC for a large area can be
made through combination of medium resolution satellite imagery and few
samples from the ground.