ESTIMATION OF ABOVE GROUND FOREST BIOMASS IN A TIGER HABITAT OF THE WESTERN NEPAL USING ALOS DATA AND FIELD INVENTORY

摘要

This study aims to apply remote sensing approach for estimating above ground forest biomass of the Western Tiger Landscape of Nepal. Multi-temporal radar imageries of L-band ALOS/PALSAR were used to provide data of radar backscatter coefficients. In-situ data of above ground forest biomass was calculated from the field inventory data using an allometric equation. In total, there were 14 ALOS/PALSAR scenes and 358 forest plots which were analyzed to meet the objectives. The relationships between radar backscatter coefficient and in-situ above ground forest biomass were analyzed over a grid with cell size of 9 by 9 pixels (approximately 1 hectare). Then, a non-linier regression approach was used to develop a biomass model. Finally, the best-fit biomass model was applied to produce biomass map of the study area. In comparison between the radar backscatter of mono- and multi-temporal L-band ALOS/PALSAR imageries, the multi-temporal data generally shows a better correlation with the above ground forest biomass calculated from field data. The multi-temporal approach compensates variable climatic conditions of a single radar image. The biomass model which developed using a non-linier regression analysis successfully produced biomass map. However, the accuracy should be improve and like other similar research that utilize radar backscatter coefficient to estimate biomass, use of L-band also has a constraint in terms of biomass saturation (around 100-150 t/ha). Although, the saturation level for L-band is quite high (＞100 t/ha) compared to the reported values for X- and C-band (30-50 t/ha).