FORMA: Forest Monitoring for Action

FORMA: Forest Monitoring for Action FORMA uses freely-available satellite data to generate rapidly-updated online maps of tropical forest clearing. We have designed it to provide useful information for local and national forest conservation programs, as well as international efforts to curb greenhouse gas emissions by paying to keep forests intact. Although rich countries are willing to pay for forest conservation in developing countries like Indonesia, results must be monitored to ensure that effective measures are identified, and to sustain public support for payments that could reach billions of dollars annually. FORMA is designed to facilitate forest conservation by identifying where -- and when -- forest clearing has occurred on a monthly basis. FORMA identifies forest clearing by analyzing publicly-available satellite data from NASA and other public and academic institutions. In particular, it relies on changes in vegetation color and the incidence of fires that indicate forest clearing. This information makes it easier to know where to intervene to stop the spread of forest clearing, and is intended to complement national forest monitoring programs and local forest conservation efforts. We provide a brief technical introduction to FORMA in this report. A full description is available from the Center for Global Development at http://www.cgdev.org/forma. The site includes an online map implementation for Indonesia that enables users to zoom to any local area in the country. This report draws on FORMA to describe and analyze forest clearing in Indonesia from 2000 to 2010, with a particular focus on monthly changes from December, 2005 to September, 2010. During the next several months, the FORMA project will post similar reports for all tropical forest countries in Asia, Africa and the Americas. Monthly updates will follow, along with downloadable GIS datasets for local policy applications. 1

National Forest Clearing in Indonesia Since 2005 Hansen, et al. (2008) 1 provide the most recent published estimate of forest clearing in Indonesia that employs a globally consistent methodology. For 2000-2005, they report forest loss as 3.36% of Indonesia s year-2000 forested area. 2 FORMA s estimate for the same period is effectively identical: 3.29% as of February, 2006. Figure 1: Indonesian Forest Clearing Rate (%), Mar. 2006 - Sept. 2010 0.05.1.15.2 Figure 1 and Table 1 show what has happened since then. Indonesian deforestation has continued, and a surge four years ago caused total forest clearing since 2005 to be greater than clearing during the first half of the decade. However, the pace has Monthly Deforestation Rate (%) Source: FORMA www.cgdev.org/forma Trend declined during the past four years, from 1.11 million hectares cleared in 2006 to 761,000 hectares in 2009. Monthly clearing in 2010 averaged 41,336 hectares from January to September, 48.6% less than clearing for the same period in 2006 (80,483 hectares). Table 1: Forest Clearing in Indonesia, 2006-2009 Cumulative Area Deforested as % of Year-2000 Forest Area Change in % From Previous Year Jul2006 Jul2007 Jul2008 Jul2009 Jul2010 Annual Hectares Cleared Monthly Hectares Cleared (Jan-Dec) Monthly Hectares Cleared (Jan-Sept) Year 2006 4.34 1.08 1,106,572 92,214 80,483 2007 5.16 0.82 850,104 70,842 76,613 2008 5.94 0.78 812,198 67,683 68,330 2009 6.68 0.74 761,083 63,424 67,520 2010 7.04 0.36 372,020 41,336 Source: FORMA: www.cgdev.org/forma 1 Hansen, M.C., Stehman, S.V., Potapov, P.V., Loveland, T.R., Townshend, J.R.G., DeFries, R.S., Pittman, K.W., Stolle, F., Steininger, M.K., Carroll, M., Dimiceli, C. 2008. Humid tropical forest clearing from 2000 to 2005 quantified using multi-temporal and multi-resolution remotely sensed data. PNAS, 105(27), 9439-9444. www.pnas.org/cgi/doi/10.1073/pnas.0804042105 2 Hansen, Ibid, Table 1. 2

Regional Trends Indonesian provinces are divided into subprovinces (kabupatens), whose relative forest clearing rates are also tracked by FORMA. Figure 2 displays major islands, island groups and provinces in Indonesia. Figure 2: Indonesian Islands and Provinces Sumatra Kalimantan Sulawesi Maluku Irian Jaya Java Bali Figure 3 displays rates of forest clearing across Indonesia from 2000 to 2005. Subprovinces are colored from the highest rate of clearing (deep red) to the lowest (yellow). The figure shows three distinctive clusters of rapid clearing, in north-central Sumatra, central Kalimantan (on a north-south axis), and southern Irian Jaya. These patterns reflect the estimates in Hansen, et al. (2008). Figure 4 shows how clearing patterns have changed since 2005. The map assigns the deepest red to subprovinces whose forest clearing rates have strongly accelerated during 2006-2010; orange to subprovinces that have accelerated, but less strongly; and yellow to subprovinces where clearing has decelerated since 2005. The map shows striking shifts in clearing intensity: Sumatran clearing has intensified to the northwest and south of the area with the fastest clearing in 2000-2005. In Kalimantan, the most intense clearing has shifted from the previous north-south axis to the southern and western coastal areas. In eastern Indonesian, recent acceleration is also visible in Sulawesi, Maluku and, to a lesser degree, southern Irian Jaya. 3

Figure 3: Forest Clearing Rates in Indonesian Subprovinces, 2000-2005 Sumatra Kalimantan Sulawesi Irian Jaya Java Maluku Forest-Clearing Rate Figure 4: Acceleration and Deceleration of Forest Clearing, 2006-2010 Sumatra Kalimantan Sulawesi Irian Jaya Much Faster Faster Slower Java Maluku How FORMA Estimates Indonesian Forest Clearing FORMA utilizes data recorded daily by the Moderate Resolution Imaging Spectrometer (MODIS), which operates on NASA's Terra and Aqua (EOS PM) satellite platforms. Although its signal-processing algorithms are relatively complex, FORMA is based on a common-sense observation: Tropical forest clearing typically involves the burning of biomass and pronounced temporary or long-term changes in vegetation color, as the original forest is cleared and replaced by pastures, croplands or plantations. Accordingly, FORMA constructs forest-clearing indicators from MODIS-derived data on the incidence of fires and changes in vegetation color as identified by the Normalized Difference Vegetation Index (NDVI). It then calibrates to local forest clearing by fitting a statistical model that relates the MODIS-based indicator values to the best available information on actual forest clearing in each area. FORMA incorporates local characteristics of the forest by dividing Indonesia into WWF ecoregions 3 and separately fitting the model to 3 http://wwf.panda.org/about_our_earth/ecoregions/ecoregion_list/ 4

data for each ecoregion. The dependent variable for each pixel is coded 1 if it has experienced forest clearing within the relevant time period, and 0 otherwise. The MODIS-based indicator values are the independent variables. For all tropical countries except Brazil, the best available data on recent forest clearing have been published in the Proceedings of the National Academy of Sciences by Hansen, et al. (2008) 4, who estimate the incidence of forest clearing for 500m parcels in the humid tropics. We calibrate FORMA using the map of forest cover loss hotspots (henceforth referred to as the FCLH dataset) published by Hansen, et al. for the period 2000-2005. 5 Using the FCLH pan-tropical dataset for 2000-2005, FORMA fits the calibration model to observations on forest clearing for 1 km 2 (100-hectare) cells in each Indonesian ecoregion. As we document in our technical paper on FORMA, 6 the model s predicted spatial probability distribution provides a very close match to the spatial incidence of FCLH forest clearing. FORMA then applies the fitted model to monthly MODIS indicator data for the period December, 2005 to September, 2010. The output for each month is a predicted forest clearing probability for each 1 km 2 cell outside of previously deforested areas, as identified in the FCLH map. Since FORMA estimates a probability of forest clearing for each cell, it is straightforward to add probabilities across cells to calculate the expected area cleared in a province, subprovince or other geographic area. Since even small areas can include thousands of 1 km 2 cells, error averaging ensures relatively small deviations from FCLH forest clearing estimates for those areas. 7 We use this approach to estimate forest clearing in each area of interest, by month. For the Indonesia report, we develop forest clearing estimates at the national, provincial and subprovincial levels. FORMA s online mapping implementation can be used to view the spatial and temporal pattern of forest clearing within each geographic area. 4 See footnote 1 for the citation. 5 In Brazil, higher resolution estimates are also available annually from the INPE PRODES program. We have used these estimates to test the accuracy of our FCLH-based calibration methodology. For more information on PRODES, see Projeto PRODES: Monitoramento da Floresta Amazonica Brasileira por Satelite. http://www.obt.inpe.br/prodes/ 6 Dan Hammer, Robin Kraft and David Wheeler. 2009. FORMA: Forest Monitoring for Action--Rapid Identification of Pan-tropical Deforestation Using Moderate-Resolution Remotely Sensed Data. Center for Global Development Working Paper No. 192. http://www.cgdev.org/content/article/detail/1423248/ 7 For example, a square area 50 km on a side contains 2,500 1 km 2 cells. 5