Estimation of the aboveground biomass and carbon sequestration in an urban forest remnant using aerial photogrammetry from a low-cost Unmanned Aerial Vehicle

Abstract

For low-income nations, a low-cost Unmanned Aerial Vehicle (UAV) offers an alternative to the traditional time-intensive forest survey that requires many resources. This study was conducted in the remnant forest with the dominant tree species, Dipterocarpus alatus Roxb., forming the upper canopy. Photogrammetry techniques with UAV images were used to obtain the Canopy Height Model (CHM). The results of tree height, individual tree detection, biomass, and carbon sequestration were compared between ground truthing and photogrammetry estimation. The large percentages of trees were automatically recognized. However, due to a closed forest canopy, some trees might have been left out from the actual count, leading to an undercounting of trees and an underestimation of the aboveground biomass (AGB). The photogrammetric dataset demonstrated a good tree height extraction accuracy and did not differ significantly from that determined by ground truthing (RMSE=2.59 m and 8.24%). The mean predicted height AGB from direct measurement was 24.76 tons ha-1, higher than those obtained from single-and multi-set photogrammetry, 5.41 and 17.99 tons ha-1, respectively. AGB and carbon sequestration estimated from photogrammetry were 72.66% of the ground truthing value. The accuracy of the photogrammetry results was acceptable and feasi ble for detecting individual tree heights, biomass, and carbon sequestration in the remnant forest. Overall, low-cost UAV could create a cost, time-efficient, and reasonably accurate local-scale forest inventory.