Estimation of above ground biomass in tropical heterogeneous forests in India using GEDI

Abstract

Quantifying above ground biomass (AGB) and its spatial distribution can significantly contribute to monitor carbon stocks as well as the carbon storage dynamics in forests. For effective forest monitoring and management in the case of complex tropical Indian forests, there is a need to obtain reliable estimates of the amount of carbon sequestration at regional as well as national levels, but the estimation of biomass is quite challenging. The main objective of the study is to validate the usefulness of the gridded above ground biomass density (AGBD) estimates (ton/ha) of the spaceborne LiDAR Global Ecosystem Dynamics Investigation data (GEDI L4B, Version 2) across two tropical heterogeneous forests in India, Betul and Mudumalai forests. Methodology includes, for each forest area, a linear regression model which predicts AGB from Sentinel-2 MSI data was developed using ground reference data and comparing it with the GEDI AGBD values. The AGB model for the Betul forests in Central India had a RMSE of 13.9 ton/ha, relative RMSE = 8.7% and a R2 of 0.88, with a bias of-0.28 ton/ha, and a comparison between modelled AGB and GEDI gridded AGBD at 1 km resolution show a relatively strong correlation (0.66) and no or little bias. It also found that GEDI AGBD footprint value is underestimated compared to the AGB according to the Sentinel-2 model. For the Mudumalai forest in southern India, the AGB model had an RMSE of 29.1 ton/ha, relative RMSE = 10.8%, and an R2 of 0.79 and a bias of-0.022. The correlation between modelled AGB and gridded GEDI AGBD was 0.84, and GEDI AGBD is underestimated compared to AGB from the Sentinel-2 model. The field values of AGB of Betul lies between 42.2 ton/ha and 238.8 ton/ha and for the Mudumalai forests, the AGB lies between 75.9 ton/ha and 353.6 ton/ha. The results indicates that the GEDI gridded AGBD underestimates AGB, and that the model used to produce the gridded AGBD product needs to be adjusted to provide reliable information on the carbon balance and its changes over time for the type of Indian forests that exists in the two test areas.

Keywords

GEDI L4A; GEDI L4B; Sentinel-2; GEDI footprint biomass

Published in

Ecological Informatics
2024, volume: 82, article number: 102712
Publisher: ELSEVIER

SLU Authors

• Indirabai, Indu

  • Department of Forest Resource Management, Swedish University of Agricultural Sciences
    

• Nilsson, Mats

  • Department of Forest Resource Management, Swedish University of Agricultural Sciences
    

UKÄ Subject classification

Forest Science

Publication identifier

• DOI: https://doi.org/10.1016/j.ecoinf.2024.102712

Permanent link to this page (URI)

https://res.slu.se/id/publ/139302