Assessing Carbon Stocks over Tropical Peat Swamp Forests Using Multi-Temporal L-Band SAR Data
Wijaya, Arief1; Liesenberg, Veraldo2; Susanti, Ari3; Karyanto, Oka3
1Center for International Forestry Research (CIFOR), INDONESIA; 2TU-Bergakademie Freiberg, GERMANY; 3Faculty of Forestry, Gadjah Mada Univ., INDONESIA



1. INTRODUCTION

In the context of the recent mechanism pact by United Nations â€oeReducing Emissions from Deforestation and Forest Degradation†(REDD), provision of reliable forest carbon stocks information is urgently needed to help reducing anthropogenic greenhouse gas emissions [1]. Tropical rain forests covers approximately 15% of the Earth’s land surface [1] and contain up to 40% of the terrestrial carbon [2]. One of them is the peat swamp forest that covers 3% of the of the Earth’s land surface and contain as much carbon as all terrestrial biomass. Peatland ecosystem has therefore an important role as carbon sink that stores huge amounts of carbon. Within the context of global climate change, there has been an increased interest in tropical peatlands because of their importance as carbon store and their role in carbon and water fluxes [2-4]. Peatlands store carbon for thousands of years and play critical role in biodiversity conservation and hydrological regulation [5].
Application of microwave remote sensing for biomass estimation allows the provision of additional information measured from the ground objects, since these measurements are almost insensitive to the cloud-free daylight conditions for image acquisition. The application of SAR data for mapping of tropical forest properties has been widely applied [6-10]. However, few studies are found applying this data in tropical peatlands. Besides, empirical models of microwave instrument data are known to be very sensitive to the density, shape, length, dielectric properties, and orientation of the scatterers [11]. L-band SAR data (‰15 â€'' 30 cm) may show good ability to model forest parameters under dense vegetation [12, 13]. The capability of L-band radar backscatter to penetrate through the forest canopy makes this data an useful tool for mapping the forest structure, including above ground biomass (AGB) estimation [13]. Previous investigations using dual polarimetric SAR data found out that the backscatter signal is more sensitive to forest biomass and forest structure because of tree trunk scattering, enhancing the discriminations of different forest successional stages [14-16]. This study aims to assess forest carbon stocks and certain stand biophysical properties using combination of both multi-temporal and multi-polarizations (quad-polarimetric) of the Advanced Land-Observing Satellite (ALOS) Phased Array L-band Synthetic Aperture Radar (PALSAR) data.
This work focuses on tropical peat swamp forest site located over the Kampar Peninsula at Riau Province, Sumatra, Indonesia (Fig. 1). The Kampar Peninsular forest represents the second largest tropical peatland in Indonesia after Central Kalimantan Province. Peatland conversion into Acacia mangium plantations for pulp and paper industries and excessive expansion of oil palm plantations are major threats for this particular ecosystem. During the last decade, the deforestation rate in the Riau Province is among of the highest in the whole country, as a trade off of accelerating local economic growth. In this study, 83 sampling plots data (e.g. red circle in Fig. 1) have been collected between July 2009 to August 2010. Various biophysical parameters, namely diameter at breast height (dbh), basal area, tree height, canopy cover percentage and leaf area index (LAI) have been measured. Furthermore, allometric equations were applied to estimate tree volume and above ground biomass. Measured of peat depth by peat drilling over different land cover types, i.e. secondary peat swamp forest, primary swamp forest and regenerating plantation forests at different age were also performed.



Fig 1. Study area location

2. DATA AND METHODS

We used quad four polarimetric SAR (PLR) scenes from ALOS satellite for the purpose of the study. The were acquired on April 04, 2007; May 20, 2007; April 09, 2009 and April 12, 2010. The pre-processing steps include image coregistration, topographic and radiometric normalization. Imagegeocoding and multi-temporal speckle filtering were performed afterwards. Attempts to estimate carbon stocks were applied to both mono-temporal and multi-temporal approaches (e.g HHt1 versus HHt2, etc). Additionaly, we also combine different polarizations (e.g HHt1+HVt1+VHt1+VVt1 versus HHt2+HVt2+VHt2+VVt2).

3. PRELIMINARY RESULTS AND ONGOING WORK

In-situ measurements showing the properties of peat swamp forests (PSF) and successional plantation forests (PF) are presented in Fig. 2. In this figure, the stand volume, biomass, carbon stocks and LAI increased following forest regenerations stages (e.g. from PF1 to PF3, and PSF respectively; Fig. 2). Co-polarizations channels at HH showed a higher correlation coefficient than cross-polarization at HV and VH in estimating biomass. This behavior may be related to the general differences in both horizontal and vertical structure of the canopy strata that enhanced therefore volume scattering.

On the other hand, forest age was better explained with the cross-polarization channels than co-polarization bands. Since these results may have been influenced by local seasonal variations, multi-temporal approach is suggested to solve this limitation.



Fig 2. Field measurement data over Kampar Peninsular

We intend to demonstrate the spatial patterns of the estimated biomass and carbon stocks and to compare them. Land cover/land use maps of the area and validate them with independent test dataset. The method proposed by Saatchi [17] will be used as a comparison to assess the quality of global carbon estimate over project level estimate. Using multi-temporal SAR dataset, we can compare seasonal aspects of the forest (changes in random distribution of twigs, leaves and branches) and forest changes due to growing nature, human exploration or tree-dieback.