Using ALOS PALSAR L-Band SAR to Detect Changes in Aboveground Carbon Stocks in Tropical Forests
Mitchard, Edward
University of Edinburgh, UNITED KINGDOM


Mapping forest carbon stocks, and changes in these stocks (i.e. deforestation and degradation) is becoming increasingly essential. This is because there are an increasing number of national and sub-national projects and mechanisms emerging where financial payments are contingent on forest carbon. For example there are approximately 200 voluntary sector avoided deforestation projects already in operation throughout the tropics [1], negotiations in the United National Framework Convention on Climate Change (UNFCCC) will probably soon lead to an international agreement on Reducing Emissions from Deforestation and forest Degradation (REDD+), and many other conservation and bilateral deals outside REDD+ rely on a maintenance or enhancement of forest carbon stocks.

Traditional estimation of the carbon stocks of a project or country rely on mapping its landcover, typically using optical satellite data, and then assigning carbon stock values to each class using standard or locally-derived carbon density values [2]. This approach however has its limitations, including failing to account for changes in stocks when landcover type remains the same (i.e. forest degradation), missing the natural heterogeneity of carbon stocks within classes (deforestation often preferentially occurs in highest biomass areas, all else being equal), and cloud cover. For this reason there has been interest in using Synthetic Aperture Radar (SAR) sensors to map AGB and AGB change in the tropics. It has been shown previously that L-band SAR can map AGB and AGB change in tropical landscapes [3-5]. However such a methodology has not yet been adopted more widely due to problems of data availability, scaling up algorithms, and issues of topography.


We use SAR data from the Advanced Land Observing Satellite (ALOS) Phased Array L-band SAR (PALSAR) sensor collected from 2007-2011 over sites in Cameroon (Mbam Djerem Reserve, Dja Reserve), Sierra Leone (in and around the Gola Forest Reserve) and Uganda (Budongo Forest Reserve). Repeat scenes for 3 or 4 years of coverage were available for all sites.

We processed the SAR data to correct for topography, and produced seamless, co-registered mosaics. These were then compared to field data on both biomass and the location of disturbance.

Results and Discussion

We found, as expected, a strong relationship between SAR backscatter and aboveground biomass (AGB), especially using the HV polarisation. Saturation occurred between 150 and 200 Mg ha-1. This adds to the evidence that there is a relatively consistent signal between increasing AGB and backscatter, up to some limit [5].

We also found that a sudden drop in SAR backscatter could be related to deforestation or degradation events, and that the extent of reduction in backscatter was directly related to the quantity of AGB removed. However, the influence of moisture and radar scatter can both result in false positive detections of change, and so ground data is necessary to calibrate the sensitivity of any detection method.


[1] D. Diaz, et al., "State of the Forest Carbon Markets 2011," Ecosystem Marketplace; Forest Trends. 2011.

[2] GOFC-GOLD, "A sourcebook of methods and procedures for monitoring and reporting anthropogenic greenhouse has emissions and removals caused by deforestation, gains and losses of carbon stocks in forests, remaining forests, and forestation.," Alberta, Canada. 2009.

[3] E. T. A. Mitchard, et al., "A novel application of satellite radar data: measuring carbon sequestration and detecting degradation in a community forestry project in Mozambique," Plant Ecology & Diversity, 2012.

[4] E. T. A. Mitchard, et al., "Measuring biomass changes due to woody encroachment and deforestation/degradation in a forest–savanna boundary region of central Africa using multi-temporal L-band radar backscatter," Remote Sensing of Environment, vol. 115, pp. 2861-2873, 2011.

[5] E. T. A. Mitchard, et al., "Using satellite radar backscatter to predict above-ground woody biomass: A consistent relationship across four different African landscapes," Geophysical Research Letters, vol. 36, p. L23401, Dec 2, 2009.