Separating the forest from the trees
|Posted by Andreas Persbo (andreas.persbo) on May 19 2011|
|VERTIC Blog >> Environment|
Joseph Burke, London
If a global agreement on climate change is to succeed, parties need to be able to trust that the others are keeping to their part of the bargain. Indispensable for this is the timely provision of reliable data. Recently, however, some have highlighted that a lack of resources and know-how in developing countries on forest carbon monitoring is leading to great errors in their reporting. In the long run, such errors could mean that the credibility of the monitoring regime could be undermined. However, how serious is that risk?
The UN REDD programme
Data quality is very important for the UN REDD+ programme, which has been lauded as a flagship undertaking for achieving significant cuts in greenhouse gas emissions. REDD was launched in 2008 as a collaboration to reduce emissions from deforestation and forest degradation in developing countries. Later, its remit was widened to include conservation and improvement of forest carbon stocks and sustainable manage¬ment of forests—hence REDD+. Several initiatives are carried out under the broad REDD banner, such as the Forest Investment Program and the World Bank’s Forest Carbon Partnership. Presently, there are 29 partner countries across Africa, Asia and Latin America. Some estimates predict that the programme could eventually mobilise some US$30bn in North-South transfers.
The complexity of measuring and monitoring forest carbon has long been recognized—it was one reason for the failure to credit the reduction of greenhouse gas emissions from actions on tropical deforestation in the Kyoto Protocol’s first commitment period. Now, though, policy makers are working on a final outline of the mechanism to coordinate the many multilateral and bilateral projects. The issue of data is a persistent concern. Those designing the REDD mechanism realize that for it to perform effectively it is essential that emission cuts are ‘measurable, reportable and verifiable’ (a phrase often shortened to MRV).
Monitoring Forest Carbon
A combination of remote sensing and ground measurement is required to give a precise picture of a forested area.
There are several remote sensing methods. Optical remote sensors establish spectral indices and correlates to ground measurements utilizing both visible and infrared wavelengths; Fine resolution air-borne optical remote sensors measure size and shape of the forested area; Radar remote sensors—for instance JERS-1, ERS-1, ALOS PALSAR or Envisat—use microwaves or radar signal to estimate the forest’s vertical structure. Laser remote sensors, such as LiDAR, are also used.
Monitoring deforestation and degradation, and estimating carbon stocks often involves a trade-off between cost and accuracy. Costs can be high because of the need to invest in: (i) research and development of tools; (ii) equipment and other capital costs of undertaking research; and (iii) training in instrument use and data interpretation. Comprehensive imaging needs expensive fine resolution equipment that can detail uneven surfaces and distinguish between multiple forest ecotypes.
At the national level remote sensing techniques are indispensable. However, ground measurement can provide crucial information and benefits. A common approach is to use forest inventories that relate tree diameters to forest carbon stocks using allometric relationships. This can be relatively cheap and less technical allowing participation by local stakeholders. Such methods also pick up on the collection of deadwood and under canopy vegetation. Ultimately, a pragmatic use of both sets of measurement techniques gives the best outcome.
A study by the UNFCCC in 2009 on the costs of setting up forest monitoring systems found that most developing countries fell short of the wanted capacities to deliver satisfactory data on forest-related greenhouse gas emission reductions.
Important new research by Pelletier, Ramankutty and Potvin—published in 2011—focuses on Panama. This study has symbolic value as Panama was one of the first countries to be supported by UNREDD. The researchers undertook sensitivity analyses using a land-cover change emissions model to assess uncertainty, and its causes, around the quantification of emissions reductions. The findings were dramatic. In one instance, the team found a staggering 54.5% error when compared with the reference emissions level established by the team.
Pelletier, Ramankutty and Potvin come up with some suggestions on how to reduce this uncertainty. First, they argue, the overall aim must be to arrive at a more precise national baseline from which to measure a cut in emissions. One improvement would be to expel the unreliable practice of ‘mosaicking’ (piecing together images and data from different years). Another is to undertake multi-temporal land-cover assessments at shorter intervals than eight to ten years. Third, the researchers also recommended harmonizing land-cover classification definitions to improve comparability.
Addressing capacity: the case of Brazil
Nevertheless, such criticisms, though warranted, should not veil the significant progress that has been made in recent years in forest monitoring. Brazil, for instance, has emerged as a world leader in the analysis of forest depletion. Under Brazil’s National Institute for Space Research (INPE) two Earth monitoring systems have been developed: the DETER and PRODES programmes. The former is produced from MODIS and AWIFS-ResourceSat data every 15 days. The latter provides annual estimates of clear-cut data using LANDSAT, CBERS, and DMC systems. Brazil has joint programmes with Argentina, France, the United States, the Ukraine and China. In fact, the launching of the China Brazil Earth Resources Satellite Programme 3 (1 and 2 are already operational) is expected sometime in 2011. This will provide ‘full support to the MRV requirements for REDD’ says the institute’s head, Mr Gilberto Camara.
Brazil has also taken an enthusiastic position towards the Data Democracy Initiative, which was launched by South Africa as chair of the Committee on Earth Observation Satellites (CEOS) in 2008. In December 2010 Brazil reiterated its commitment to share its expertise and knowledge with others: ‘Brazil is fully committed to the Data Democracy initiative led by South Africa. We are willing to share our data, software, expertise and technology in earth observation with other developing nations.’
Such initiatives will be key to improving capacity across developing countries and thereby helping to reduce error in forest related greenhouse gas emissions reduction.
However, even Brazil has its own difficulties. Recent satellite images from the country’s space research institute show deforestation increased from 103 to 593 square kilometres from March and April 2010 to the same period of 2011, mainly in the farming state of Mato Grosso. The jump is linked to the controversial proposed easing of land-use rules, which is seen by some as akin to offering an amnesty to farmers who have illegally cleared land in the past. Now, the government have been forced to spring into action, announcing a crisis cabinet meeting to address the issue. This shows the undeniable value of accurate and reliable data.
There are several recent projects that show exciting innovation in forest monitoring. In March 2011, the Carnegie Institution for Science and the USDA Forest Service published an article detailing how they developed a 30-meter-resolution map of above ground carbon density for Hawaii. Using airborne LiDAR-based observations, satellite-based imagery alongside field measurements the research team estimated the total above ground carbon for the island to be 28.3 teragrams. This result is 56% lower than the value estimated by IPCC Tier 1 methods. The researchers recognise that their innovative approach ‘could support end-users from many sectors … to monitor emissions over time, providing transparent information flow to support [carbon] accounting and REDD-type projects and programs.”
A further recent study of forest depletion in the Indonesian islands of Sumatera and Kalimantan used data from the Landsat 7 enhanced thematic mapper plus (ETM+) and MODIS sensors and digital elevation data from the Shuttle Topography Radar Mission (SRTM). The researchers determined the total forest cover loss for the two islands between 2000 and 2008 to 539 square kilometers. They also found out that at least 6.5% of all mapped forest cover loss had occurred in land allocation zones where clearing was banned. Their study infers that ‘accurate, annual forest cover loss maps will be integral to many REDD+ objectives, including policy formulation, definition of baselines, detection of displacement and the evaluation of the permanence of emission reduction.’
While the significant results and recommendations of the Pelletier, Ramankutty and Potvin study do indeed require urgent attention and the increase in Brazilian deforestation is worrying, there is reason to be hopeful. Existing data sharing initiatives, if supported, and continuing technical advances, if multiplied, have the potential to address these concerns. There is still time for progress in 2011, the International Year of Forests.
Last changed: May 19 2011 at 4:31 PMBack