Assessing multi-temporal Landsat 7 ETM+ images for estimating above-ground biomass in subtropical dry forests of Argentina

Abstract

Above-ground biomass (AGB) is important to estimate total carbon pools in forests, where it has a key role in the global carbon cycle. We assessed correlations between spectral information and ground data to estimate AGB in the Semiarid Chaco, Argentina. Ground data (DBH, height and species of trees) were obtained from 15 samples (0.8 ha each) and AGB was estimated. Multi-temporal Landsat images were used to obtain spectral data (single bands/vegetation indexes) of the samples. Correlation tests between AGB and spectral bands and between AGB and vegetation indexes were performed for all dates. A strong correlation was found between spectral indexes and AGB in the early dry season (fall – May 12, 2002) while poorer results were obtained for summer and winter. This would result from a differential phenological response of trees, shrubs and grasses to environmental conditions. A biomass predictive model was fitted using the NDVI of May 12, 2002 and a biomass map was obtained applying this regression. There was a rain-related regional pattern of AGB decrease in an east–west direction, and a land-use related local pattern. Our results offer a great potential for increasing the understanding of dry Chaco forest structure and for improving carbon pools estimates.