Stock Volume Loss Estimation in Poplars using Regression Models and ALOS-2/PALSAR-2 backscatter

Abstract

Stock volume is an important forest inventory parameter. In case of agro-forests and plantation forests, stock volume estimates are important as they provide reliable indicator of the productivity of these species. In this study stock volume loss due to harvest of polar plantations between 2017 and 2018 are estimated using ALOS-2/PALSAR-2 backscatter data. Using simple linear regression models the AGB of the plantations before and after harvest are estimated. These are converted to stock volume loss per hectare. From field inventory, the actual stock volume during harvest are measured. These are validated against the estimations using two models – M1 and M2. Model M1, utilizes only HV-pol backscatter data and provides a lower accuracy with r2 = 0.46. Model M2 utilizes HH- and HV-pol backscatter and provides stock volume loss estimation with r2 = 0.51.