Modelling Carbon Sequestration in Drylands of Kazakhstan Using Remote Sensing Data and Field Measurements

Abstract

Landsat ETM+ data were related with field measurements of carbon stocks residing in the vegetation and in the soil along a 230 km west-east transect within Kazakhstan’s grassland zone. The biomass carbon was correlated to the mid-infrared corrected Normalized Difference Vegetation Index (NDVIc). Soil carbon responded most strongly to the multi-spectral features of the Kauth-Thomas transformation (KT) and was modelled using the non-linear multiple regression. Retrieved models were applied to the Landsat ETM+ image and carbon stocks were mapped over Shetsky district of Karaganda province at the pixel-by-pixel scale. Total terrestrial carbon stocks estimated for the area of Shetsky district account to 97.2 million tons, whereas most part of it (97.6%) is residing in soil.