Estimation of above-ground biomass of oil palm trees using phased array L-band synthetic aperture radar data

• Main Author: Shashikant, Veena
• Format: Thesis
• Language: English
• Published: 2014
• Subjects: Plant biomass; Oil palm
• Online Access: http://psasir.upm.edu.my/id/eprint/70253/1/FK%202014%20174%20IR.pdf; http://psasir.upm.edu.my/id/eprint/70253/
• Institution: Universiti Putra Malaysia

Quantification of Above Ground Biomass (AGB) over oil palm plantation is essential for a wide range of modern day research and management demands. Spatial distribution information of oil palm plantation AGB is therefore important for the palm oil industry. In this study, the oil palm plantation in Perak, Malaysia was assessed to obtain the AGB information. Lack of biomass information due to the uncertainties and cost has limited its determination and changes over the years. Hence, this research studies the quantification of AGB of oil palm plantation. In a tropical country such as Malaysia, cloud covers are hindrance to visible light sensing. For that reason, all types of Synthetic Aperture Radar (such as Phased Array type L-band Synthetic Aperture Radar (PALSAR)) sensor is an added advantage to overcome the cloud problem to analyze the biomass of the oil palm plantation. Ground data of oil palm biomass for age of 6, 8, 10, and 12 years old are trees compared to the estimated AGB using PALSAR data with all polarizations. Four filters were applied on the PALSAR images and compared among the four window sizes which were 3x3, 5x5, 7x7 and 9x9. The filters used in this study were Gaussian low, Gaussian high, Laplacian and Median filter implemented in PALSAR data with HH, HV, VH and VV polarimetry bands. Speckle suppression index was applied to check the filters’ efficiency and thus selection of model building. A valid model was constructed to show the VV polarization degree of relationship between the field data results and filtered PALSAR data of biomass. The model developed had an R2 value of 0.90, between the VV backscattering against the AGB values. This model was subsequently validated and found to have an accuracy of 85%.This study can be useful to the management of the logging cycle and the volume of the harvested biomass in the felling season in the oil palm plantations.