Forest biomass estimation using optical and microwave imagery (Case study: Garazbon Series, Kheirud Forest)

Abstract

Estimation of forest aboveground biomass is important in regional carbon policies and sustainable forest management. Since forests are the largest carbon store, it is important to evaluate the forest biomass to estimate carbon storage and its impacts on climate change in global scale. Optical and active microwave remote sensing data both play important roles in forest biomass monitoring. Our aims in this research are biomass modeling and estimation using multilayer perceptron neural network in Gorazbon district, Kheyroud Forest in Mazandaran province. Estimation was performed using the Landsat and ALOS PALSAR dataset and also 201 ground sample plots in two years of 2007 and 2012. The capability of the ALOS PALSAR Global Mosaic product with 25 m resolution was also evaluated in biomass estimation. The effects of environmental factors such as slope and aspect were specifically evaluated on the accuracy of biomass estimation. Finally, the best model was presented in 2012 by ALOS PALSAR Global Mosaic product with R2= 0.83 and RMSE = 108.99 which has very little difference from other optical and radar images. According to the research results, newer sensors using up-to-date technology will deliver much better results compared to the previous generations. Of course, to ensure these results, it is necessary to conduct additional studies in this field as well.

Keywords

• biomass
• forest
• slope & aspect
• neural network
• alospalsar
• landsat