Estimation of Biomass for Pinus brutia Ten.Using SPOT-6 Satellite Data and Machine Learning Algorithms

 Hassan Ali*(1)

(1). General commission for the administration and development of Al-Ghab, Al-Ghab, Syria.

(*Corresponding author: Dr.Hassan Ali, E-mail: hso414516@gmail.com)

Received: 11/09/2023         Accepted: 6/11/2023

Abstract: 

Today, artificial forests are one of the most important sources of carbon storage in forests, saving timber and reducing the degradation of natural forests. Forest biomass plays a major role in reducing global warming, an indicator of environmental sustainability, and an important source of information at the national and international levels. In recent years, remote sensing techniques using machine learning algorithms such as random forest and multiple linear regression have been widely used to estimate forest tree biomass. This research was conducted in Arab Dagh Forest in Golestan Province, Iran, and field data were collected by systematic cluster sampling method. 180 samples of Pinus brutia Ten. were inventoried with an area of ​​400 square metres. At the level of each plot, the diameter and tree height were measured. 135 samples were selected for modeling (training data) and 45 samples for modeling validation (test data). The aim of this research is to estimate the biomass of Pinus brutia using Random Forest and the multiple linear regression algorithms. And compare the results obtained from using these algorithms to estimate biomass. The results of modeling using the multiple linear regression algorithm showed that the coefficient of determination (R2) was equal to 0.55, and the percentage root mean square error (%RMSE) was equal to 32.14%. While the results of modeling using the Random Forest algorithm showed that the coefficient of determination was equal to 0.92, and the percentage of root mean square error was equal to 18.19%. Estimating the biomass of Pinus brutia using the Random Forest algorithm gave encouraging results compared with multiple linear regression.

Keywords: Biomass, Random Forest, Multiple Linear Regression, Pinus brutia.

Full paper in Arabic: pdf