Above-ground Carbon Stock Assessment of Khuan Khaeng Swamp Forest after severe Burning in 2012 using Satellite Imagery

Abstract

This study aimed to assess the above-ground carbon stock of woody plants in Khuan Khaeng swamp forest after severe burning in 2012. Firstly, the swamp area was classified into different land use and land cover types affected by forest fire using the supervised classification of Thaichote image. Ground truth and accuracy assessment have been done. Consequently, the classified image was improved quality through the post-classification technique. The classified image data was then transformed into GIS layers. A number of 30x30 meter sample plots was randomly established in each forest class. The total height and DBH of trees in the plots were measured as variables for calculating the above-ground biomass based on appropriate allometric equations. Some dead trees found in the study area were also sampled to measure dry weight in order to perform specific allometric equations as well as for finding the percentage of carbon concentration. The linear relationship between the data of carbon stock per unit area derived from allometric calculation and Landsat 8 (OLI) image data were analyzed. Finally, the amount of above-ground carbon sequestration for the whole area of Khuan Khaeng swamp forest were spatially estimated using the derived regression equation.

The results showed that the overall accuracy of satellite image classification was 72.58% (Kappa coefficient = 0.63). The best-suited allometric equation for estimating the above-ground biomass of the dead swamp tea-tree (Melaleuca cajuputi) dominant species in this area, is W = 0.0381(DBH²H)^0.8952 with a coefficient of determination (R²) of 0.93 (W = Biomass). The concentration of carbon by dry weight is about 48.21%. The total above-ground biomass of the Khuan Khaeng swamp forest is 37.92 ton per hectare and 17.83 ton-carbon per hectare, averagely. Based on statistical test, the carbon stock per unit area of the healthy swamp forest area, of the degraded forest area, and of the burned forest area were significantly different with 95% confidential level. Meanwhile, there was no statistically different for the other type of forest classes. Based on the conventional method that relies on field data measurement and allometric analysis, the total amount of above-ground carbon stock of the whole study area is 174,464.27 ton-carbon. The best regression equation for estimating the spatial above-ground carbon stock per unit area from Landsat 8 (OLI) image data for this site is 0.0323 (G-R) -6.5495 with R² = 0.2126. Using this model, the total above-ground carbon stock for the study area is 167,570.46 ton-carbon that is slightly lower than the conventional method.