Abstract
Kalimantan poses one of the highest carbon emissions worldwide since its landscape is strongly endangered by deforestation and degradation and, thus, carbon release. The goal of this study is to conduct large-scale monitoring of above-ground biomass (AGB) from space and create more accurate biomass maps of Kalimantan than currently available. AGB was estimated for 2007, 2009, and 2016 in order to give an overview of ongoing forest loss and to estimate changes between the three time steps in a more precise manner. Extensive field inventory and LiDAR data were used as reference AGB. A multivariate linear regression model (MLR) based on backscatter values, ratios, and Haralick textures derived from Sentinel-1 (C-band), ALOS PALSAR (Advanced Land Observing Satellite’s Phased Array-type L-band Synthetic Aperture Radar), and ALOS-2 PALSAR-2 polarizations was used to estimate AGB across the country. The selection of the most suitable model parameters was accomplished considering VIF (variable inflation factor), p-value, R2, and RMSE (root mean square error). The final AGB maps were validated by calculating bias, RMSE, R2, and NSE (Nash-Sutcliffe efficiency). The results show a correlation (R2) between the reference biomass and the estimated biomass varying from 0.69 in 2016 to 0.77 in 2007, and a model performance (NSE) in a range of 0.70 in 2016 to 0.76 in 2007. Modelling three different years with a consistent method allows a more accurate estimation of the change than using available biomass maps based on different models. All final biomass products have a resolution of 100 m, which is much finer than other existing maps of this region (>500 m). These high-resolution maps enable identification of even small-scaled biomass variability and changes and can be used for more precise carbon modelling, as well as forest monitoring or risk managing systems under REDD+ (Reducing Emissions from Deforestation, forest Degradation, and the role of conservation, sustainable management of forests, and enhancement of forest carbon stocks) and other programs, protecting forests and analyzing carbon release.
Highlights
The Earth’s land surface spans approximately 149.4 million km2, of which nearly 30% is characterized by forested areas [1]
The results suggest that a MLR using backscatter values in combination with ratios and textures of Advanced Land Observing Satellite (ALOS) Phased Array Type L-band Synthetic Aperture Radar (PALSAR)-1/2 L-band and Sentinel-1 C-band data is eligible to model above-ground biomass (AGB) in tropical forests
It was shown that a multivariate linear regression model using C- and L-band synthetic aperture radar (SAR)-derived ratios and textures is able to model biomass more precisely and at a better resolution than existing models in this region
Summary
The Earth’s land surface spans approximately 149.4 million km, of which nearly 30% is characterized by forested areas [1]. At the same time the deforestation and degradation of these tropical ecosystems lead to a considerable release of carbon, making Indonesia, especially Kalimantan and Sumatra, one of the largest carbon emitters worldwide [10]. As a result of the significant emission levels and immense loss of forests and peatlands, Indonesia has become a prime target for REDD+ projects (Reducing Emissions from Deforestation, forest Degradation, and the role of conservation, sustainable management of forests, and enhancement of forest carbon stocks). One of the intentions of REDD+ includes conditional payments to developing countries for reducing their emissions [11] Within this framework, the most active projects and initiatives worldwide are taking place in Indonesia and its provinces [12]. Indonesia, especially Kalimantan, the largest isle of Indonesia, is affected by ongoing deforestation and degradation processes
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
