Authors: Dina Rasquinha*, UGA, Deepak Mishra, University of Georgia
Topics: Remote Sensing, Coastal and Marine, Global Change
Keywords: mangroves, hyperspectral, biomass, carbon, remote sensing
Session Type: Paper
Presentation File: No File Uploaded
Mangrove forests play a crucial role in the global carbon cycle, mitigating climate change impacts and serving as carbon sinks, that reduce greenhouse emissions. However, these forests are cleared across the tropics to accommodate aquaculture, agriculture and other coastal development infrastructure. The loss leading to substantial changes in carbon stored within these forests. In the last few years, several studies have looked at carbon storage and sequestration potential of mangrove ecosystems across the globe. The efforts have translated into assessments at the global, regional and local scales. However, many aspects of these ecosystems remain poorly characterized especially in India due to the challenges in quantifying and monitoring both aboveground and belowground mangrove forest biomass and carbon stocks. In this context, remote sensing provides a strategic avenue to monitor mangrove species biomass and carbon stocks. In this study, we use a combination of multispectral (Worldview-2 & 3) and hyperspectral images (AVIRIS) to understand the spatial heterogeneity of biomass and carbon stocks of the mangrove forests of Bhitarkanika, India. Specifically, we used an object-based classification scheme on high resolution Worldview images in combination with machine learning algorithms such as random forest to delineate mangrove species classes and estimate their biomass and carbon stocks. Our results corroborate with hyperspectral airborne imagery and field data on forest species composition, structure and biomass values. In conclusion, we propose a methodology to monitor mangrove biomass by building a biomass model which can help policy makers design monitoring mechanisms for the region.