Estimation of Aboveground Carbon using KNN Algorithm: A Case Study in Danyang, Korea

摘要

Forest biomass stores a large amount of terrestrial carbon and the need of periodic monitoring forest biomass has increased. kNN algorithm is one of the common method to estimate forest biomass and has been widely used for a variety of biomass mapping and applications. The objective of this study is to estimate aboveground carbon stock of Danyang area in South Korea. Field data from NFI and Landsat ETM+ satellite image were used, and carbon stock were estimated at k = 1 to 10. As a result, the lowest RMSE was found at k = 5, and the mean and the total carbon stock of Danyang area were estimated to be 28.33 ton C/ha and 2209748.911 respectively.