Biodiesel is one of the attractive renewable fuel sources in Thailand. The Alternative Energy Development Plan: AEDP 2012-2021 by Department of Alternative Energy Development and Efficiency, Ministry of Energy, has set the target of oil palm at 5.97 million liters per day by 2021. Such that, 880 thousands hectares of oil palm plantation will be required by 2021. In a process of land transformation to oil palm, rubber plantation is chosen as a priority replacing cropland. A trend of rubber market prices in Southeast Asia industry during 2012-2015 has significantly declined due mainly to the economic crisis in the EU and US. In response to the crisis, the Thai government has taken an opportunity to induce the parallel plans in replacing older rubber plantation aged more than 20 years old or lands with lower production by a palm-oil energy crop.
Today, a combination of GIS and remote sensing techniques can be used to determine crop types, crop yield, and, crop age by its growth cycle. The aim of this study was (1) to investigate potential land availability and land suitability for oil palm expansion focusing on the AEDP policy implementation for rubber land conversion using GIS, and (2) to demonstrate an example of rubber plantation age classification using remote sensing data from THailand Earth Observation System, THEOS or a so-called Thaichote. The crop age determination will be made explicit advantages for land conversion restriction acknowledged in the national renewable energy plan.
We proposed a method for sampling various groups of stand ages for rubber plantation using a stratified cluster-window sampling method. This method was described in Keson et al. (2015). In short, there is the two-stage sampling method. The first stage is the stratified sampling using the ISODATA algorithm for preliminary grouping of multispectral data. The second stage is the cluster-window sampling procedure using a human visual process to cognitively learning and discriminating texture, pattern, tone, and color in the spatial image. A process of moving window throughout the image was then carried out by visual interpretation for data filtering to best select a representative sample that comprised the proportion of each stratified stand age group evenly distributed. After that a classification method based on the support vector machine was applied to group rubber plantation with respect to their stand ages.
A map of para rubber age classification in Krabi province: a prototype.
|1.||Keson J., Ratchaniphont A., Wongsai S. and Wongsai N. 2015. Policy assessment of potential biodiesel feedstock supply in Thailand. Energy Procedia. 79. Pp. 710 – 718.|