Nihonbashi 1-chome Mitsui Building, 15th floor, 1-4-1 Nihonbashi, Chuo-ku, Tokyo 103-0027, Japan
Dr. Tien Dat Pham (University of Tsukuba)
Monitoring Biomass of Mangrove Species Using Remote Sensing Data and Machine Learning Techniques for Implementation of REDD+ Policies in Vietnam
Climate change is one of the main concerns and the biggest challenges for all scientists to deal with. The “reducing emissions from deforestation and forest degradation” (REDD+) activities and “blue carbon” programs under the United Nations Framework Convention on Climate Change (UNFCCC) are expected to offer the reliable methods for monitoring, reporting, and verification (MRV) for providing reference levels (or baselines), and protect biodiversity and ecosystem services. Mangroves play an important role in the global carbon cycle by reducing greenhouse gas (GHG) emissions, and mitigating climate change impacts. However, these forests have been lost worldwide, resulting in the systematic loss in carbon stocks. Additionally, the roles of mangrove forests remain poorly quantitatively characterized as compared to other forest ecosystems, due to the practical difficulties and the cost-effectiveness in measuring and monitoring mangrove forests biomass and their carbon stocks. Without a quantitative method for effectively monitoring the carbon stocks in the mangrove ecosystems, sensible policies and actions for conserving mangroves in the context of climate change can be hard to be made.
My current work presents a novel technology to retrieve biomass of mangrove species and monitor their changes using optical and SAR remote sensing data combined with different machine learning techniques and to promote the implementation of the REDD+ policies by introducing the willingness to pay (WTP) concept to mangrove ecosystem services. The present work selected Hai Phong City located on the northern coast of Vietnam, where the mangroves are distributed within zones I and II of the four mangrove zones in Vietnam as a case study. This city is vulnerable to a sea level rise associated with climate change impacts, which are forecasted to become more prevalent and stronger as climate change intensifies.
This work first determines the relationship between biophysical parameters of specific mangrove species and remote sensing data and then attempt to estimate aboveground biomass (AGB) of these species using multiple linear regression models delivered from dual-polarization HH and HV backscatters of ALOS-2 PALSAR-2 imagery. In the second part, the present work investigates the applicability of machine learning techniques and remotely sensed data for estimating AGB and carbon stocks of mangrove species. For the improvement of model performance, the current work test the usability of selected machine learning techniques with an integration of optical and SAR data for the AGB estimation of mangrove forests. In the last part, the study shows a case study at Cat Ba Biosphere Reserve for monitoring mangrove forests change between 2010 and 2015, and attempt to evaluate the economic values of mangrove ecosystem services and the social benefits of mangrove restoration in the context of climate change by estimating WTP using contingent valuation method (CVM).
This work will support provincial decision making on mangrove conservation and management. The results of this study may promote the implementation of mangrove conservation and restoration strategies in climate change mitigation approaches such as REDD+ and blue carbon programs.
Keywords: Blue carbon, climate change, contingent valuation method, mangrove species, mangrove restoration, machine learning, ALOS PALSAR, Sentinel-2, REDD+, Vietnam, willingness to pay.
Public events of RIKEN Center for Advanced Intelligence Project (AIP)Join community