Armston (CMS 2020): Savanna-Bio: Biomass estimation with new spaceborne missions for MRV in Dry Forests and Savannas
John Armston, University of Maryland, armston@umd.edu (Presenter)
Konrad Johan Wessels, George Mason University, kwessel4@gmu.edu
Paul Robert Siqueira, University of Massachusetts, siqueira@umass.edu
Laura Duncanson, University of Maryland, lduncans@umd.edu
Mikhail Urbazaev, University of Maryland, urbazaev@umd.edu
Narayanarao Bhogapurapu, University of Massachusetts Amherst, nbhogapurapu@umass.edu
Xiaoxuan Li, George Mason University, xli50@gmu.edu
Rajashekar Gopalakrishnan, National Remote Sensing Centre, rajashekar_g@nrsc.gov.in
Sean P Healey, USDA Forest Service, sean.healey@usda.gov
Russell Main, Scion, russ.main@gmail.com
Laven Naidoo, Council for Scientific and Industrial Research (CSIR), lnaidoo@csir.co.za
Stuart Ross Phinn, University of Queensland, s.phinn@uq.edu.au
Peter F Scarth, The University of Queensland, p.scarth@uq.edu.au
This proposal addresses the key objective of NASA’s CMS to develop accurate and precise Monitoring Reporting and Verification (MRV) systems required by stakeholders, by reducing the uncertainty in biomass estimates and derived change products. While tropical Dry Forests and Shrublands (or “Savannas”) have lower biomass densities (below 60 Mg/ha) than Tropical Rainforests and Moist Forests, they are expansive, covering more than 20% of the earth and representing the third highest total carbon stock by ecosystem. A very large portion of the world’s population, most of which live in less economically developed countries, are directly dependent on local savanna ecosystem services (e.g. fuelwood and grazing). Despite the obvious global importance of savannas, remote sensing efforts have been skewed towards moist tropical forested environments. Due the structural heterogeneity of savannas and the very sparse field data available for the parameterization of current lidar and SAR-based models, the error of biomass estimation in these ecosystems is very high (relative RMSE 50-100%), precluding the accurate monitoring for MRV of biomass changes due forest degradation, regrowth, debushing, bush encroachment and mitigating activities, e.g. restoration and afforestation. The overall goal of this research is to use the billions of measurements made by the new spaceborne lidar data from NASA’s GEDI (launched December, 2018) and ICESat-2 (launched September, 2018) to improve the accuracy of SAR-based models, with a view to the upcoming NASA/ISRO NISAR mission. To achieve this goal our research objectives are to: (i) develop GEDI and ICESat-2 biomass models representative of savanna ecosystems; (ii) produce prototype multi-date aboveground biomass maps for international pilot sites (India, South Africa and Australia) using coincident GEDI, ICESat-2, and spaceborne SAR data (ALOS PALSAR2, Sentinel1); (iii) use independent airborne lidar and field data to validate the products and evaluate the uncertainty of prototype aboveground biomass estimates from 1 ha to 1 km2 scales; and (iv) work with existing in-country partners to prototype products for biomass gains and losses that do not constitute land cover changes but are desired inputs to national carbon accounting systems, international reporting obligations, emissions trading and mitigating activities. By focusing on quantifying change at the lower end of the biomass range where SAR is most sensitive (20-80 Mg/ha), we will improve the accuracy of lower magnitude biomass change estimation, such as regrowth and degradation, not only in savannas, but for other biomes as well. The project will leverage the international experience of the PI and Co-PI’s in savannas and their existing networks of collaborators/users, as well as existing airborne lidar and biomass field data. By developing prototype products for new pilot areas in India, South Africa and Australia, the project will expand the capability, impact and societal relevance of the CMS program towards its global aspirations.
Associated Project(s):
Poster Location ID: 1
Presentation Type: Poster
Session: Poster Session 1
Session Date: Wednesday (9/27) 1:15 PM