Calibration of Global Above Ground Biomass Estimate Using Multi-Source Remote Sensing Data
Wijaya, Arief; Sasmito, Sigit Deni; Purbopuspito, Joko; Murdiyarso, Daniel
Center for International Forestry Research (CIFOR), INDONESIA
Mangrove ecosystem is among the most richest carbon in the tropics . The study by Donato, et.al (2011) estimated the carbon density of about 1.028 Tg (1 Tg = 1,000 Mg) per hectare. It was also estimated that mangrove deforestation resulted in emissions of 0.02 and 0.12 Pg carbon per year, which contributes to 10% of global deforestation. Study on mangrove mapping and estimation of carbon stocks and emissions over this particular ecosystem especially in the tropical regions is essential, as this ecosystem is under serious threats, highly degraded and deforested to other land uses. Over the past 50 years the area of mangrove forests has sharply declined by 30-50% due to coastal development, aquaculture and over-harvesting (Murdiyarso et al. 2012). Mapping of above ground biomass (AGB) on mangrove ecosystem in the tropics is important, since it may reduce uncertainty in carbon accounting over this ecosystem and to support climate change initiatives to decide on benefits distribution following carbon emissions reduction .
This study employs integration of Lidar and Landsat data combined with in-situ measurement to assess and calibrate global above ground biomass estimate (Baccini et al. 2012). The study objectives are to: 1) estimate the extent of carbon stocks over mangrove ecosystem in the tropics combining spatial and in-situ data, and 2) to find possibilities for more accurate biomass mapping over this particular ecosystem.
2. Data and Methods
This study focuses on mangrove forests in Eastern Indonesia, at the island of Papua (Figure 1). Field data measuring biophysical parameters and above ground biomass was conducted in Timika (TMK, Site 1), Bintuni (BTN, Site 2) and Teminabuan (TMB, Site 3) late 2011 (Table 1). We have measured vegetation properties and above ground biomass in 12 transects, which represent 72 measurement plots. We collected Landsat TM/ETM data for each project site with the acquisition date between 2007 and 2011 to fit in with the ICESAT Glas Lidar acquisition date (i.e. from year 2007) and the date of field survey (i.e. from late 2011). We took a subset of global Lidar acquired between 2003 and 2009. In total, more than 6000 Lidar shots covering three study sites in the Island of Papua were acquired. Attempt to map the above ground biomass follows the approach proposed by Baccini (Baccini et al. 2012). Firstly, field biomass data were directly modeled to Landsat images (i.e. 30m and 60m resolution) and the estimate was validated using field data. Secondly, Lidar data calibrated to above ground biomass model for Southeast Asian region was used to generate biomass equations applying inputs of original Landsat data (30m resolution) and resampled ones (60m) to reduce systematic bias (Walker et al. 2010), thus the models were validated with the test data. Another possibility was to calibrate Lidar biomass estimate using a subset of field data, through Landsat TM/ETM data which co-locate with Lidar and field measurement. The modeling results and error propagation were analyzed and discussed.
Figure 1 Distribution of sampling plots, ICESAT Glas Lidar data, extent of biomass density (Ton/ha) from 500 m MODIS data and distribution of mangrove forest (Ministry of Forestry 2009, Giri et al. 2011).
3. Results and discussion
We observed that 60 meter resampled Landsat data provides higher accuracy than the 30 meter Landsat scene in estimating AGB. Also, calibration and refinement of 500 meter global biomass map product provided by previous study (Baccini, 2012) is possible using higher resolution Landsat data (60 meter) and the availability of field measurement data. The main challenge to calibrate global biomass dataset into a nation-wide estimate is to collect additional field data which overlap with Lidar data. Unfortunately, since ICESAT Glas has stopped its operations, the launching of next satellite generation that is designed to map global biomass is urgently needed.
Table 1 Summary of field measurement data
Baccini, A., S. J. Goetz, W. S. Walker, N. T. Laporte, M. Sun, D. Sulla-Menashe, J. Hackler, P. S. A. Beck, R. Dubayah, M. A. Friedl, S. Samanta, and R. A. Houghton. 2012. Estimated carbon dioxide emissions from tropical deforestation improved by carbon-density maps. Nature Clim. Change 2:182-185.
Donato, D. C., J. B. Kauffman, D. Murdiyarso, S. Kurnianto, M. Stidham, and M. Kanninen. 2011. Mangroves among the most carbon-rich forests in the tropics. Nature Geosci 4:293-297.
Giri, C., E. Ochieng, L. L. Tieszen, Z. Zhu, A. Singh, T. Loveland, J. Masek, and N. Duke. 2011. Status and distribution of mangrove forests of the world using earth observation satellite data. Global Ecology and Biogeography 20:154-159.
Ministry of Forestry. 2009. Digital Land Cover and Land Use Map of Indonesia for Years 2000, 2003, 2006 and 2009. Spatial Planning Agency, Ministry of Forestry of Indonesia, Jakarta.
Murdiyarso, D., J. B. Kauffman, M. Warren, E. Pramova, and K. Hergoualch. 2012. Tropical wetlands for climate change adaptation and mitigation: Science and policy imperatives with special reference to Indonesia. Center for International Forestry Research (CIFOR), Bogor, Indonesia.
Walker, W. S., C. M. Stickler, J. M. Kellndorfer, K. M. Kirsch, and D. C. Nepstad. 2010. Large-Area Classification and Mapping of Forest and Land Cover in the Brazilian Amazon: A Comparative Analysis of ALOS/PALSAR and Landsat Data Sources. Selected Topics in Applied Earth Observations and Remote Sensing, IEEE Journal of 3:594-604.