Estimation of above ground forest biomass using unmanned aerial vehicle image

Abstract

Forests provide the biggest carbon pool used to counter-balance the concentration of Green House Gases (GHG) in the atmosphere. Threats of excessive concentration of GHG to climate patterns has drawn much attention around the world that different measures have been taken. Knowledge of the amount of Carbon sequestered by forests is important for appropriate mitigation measures, to this end, various methods are employed for quantification of above ground biomass (AGB). Field techniques yields accurate results but tedious, time consuming and sometimes unsafe for workers. Light Detection and Ranging (LiDAR) and Radio Detection and Ranging (RADAR) are appropriate techniques but involves high costs and complexity in data processing. AGB estimation based on Unmanned Aerial Vehicle (UAV) images is the simple and cost-effective technique suited for small and medium size forests. In the current study, AGB were estimated using UAV images and compared to AGB estimated based on field observations. The mean AGB estimated from field data was 0.576 t/ha, 0.622 t/ha, and 0.309 t/ha compared to 0.613 t/ha, 0.546 t/ha and 0.245 t/ha estimated using UAV images in sample plot1, sample plot2 and sample plot3 respectively. Likewise, the Root Mean Square Error (RMSE) computed for sample plot1 was 0.087 t/ha while for sample plot2 the RMSE was 0.015 t/ha and for sample plot3 RMSE was 0.516 t/ha. The results suggest the application of the method for small and medium size forests and is recommended down to local government authorities and individual companies in Tanzania to collect forest information that helps combat excessive GHG by taking appropriate measures to prevent much threats.