Exploring multi-scale forest above ground biomass estimation with optical remote sensing imageries

International Symposium on Earth Observation for One Belt and One Road

地球科学;政治学;社会学;经济学

Koju, U.^1 ; Zhang, J.^1 ; Gilani, H.^2

Institute of Remote Sensing and Digital Earth, CAS, Beijing

100094, China^1

International Centre for Integrated Mountain Development, Kathmandu

44600, Nepal^2

关键词: Classification technique;    Crown projection areas;    Multi-scale approaches;    Object based image analysis (OBIA);    Optical remote-sensing imagery;    Remote sensing imagery;    Terrestrial ecosystems;    Tree crown delineation;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/57/1/012011/pdf
DOI  :  10.1088/1755-1315/57/1/012011

来源: IOP

【 摘 要 】

Forest shares 80% of total exchange of carbon between the atmosphere and the terrestrial ecosystem. Due to this monitoring of forest above ground biomass (as carbon can be calculated as 0.47 part of total biomass) has become very important. Forest above ground biomass as being the major portion of total forest biomass should be given a very careful consideration in its estimation. It is hoped to be useful in addressing the ongoing problems of deforestation and degradation and to gain carbon mitigation benefits through mechanisms like Reducing Emissions from Deforestation and Forest Degradation (REDD+). Many methods of above ground biomass estimation are in used ranging from use of optical remote sensing imageries of very high to very low resolution to SAR data and LIDAR. This paper describes a multi-scale approach for assessing forest above ground biomass, and ultimately carbon stocks, using very high imageries, open source medium resolution and medium resolution satellite datasets with a very limited number of field plots. We found this method is one of the most promising method for forest above ground biomass estimation with higher accuracy and low cost budget. Pilot study was conducted in Chitwan district of Nepal on the estimation of biomass using this technique. The GeoEye-1 (0.5m), Landsat (30m) and Google Earth (GE) images were used remote sensing imageries. Object-based image analysis (OBIA) classification technique was done on Geo-eye imagery for the tree crown delineation at the watershed level. After then, crown projection area (CPA) vs. biomass model was developed and validated at the watershed level. Open source GE imageries were used to calculate the CPA and biomass from virtual plots at district level. Using data mining technique, different parameters from Landsat imageries along with the virtual sample biomass were used for upscaling biomass estimation at district level. We found, this approach can considerably reduce field data requirements for estimation of biomass and carbon in comparison with inventory methods based on enumeration of all trees in a plot. The proposed methodology is very cost effective and can be replicated with limited resources and time.