Using satellite data to monitor REDD+ projects: developing methodologies and error estimation for Africa

Deforestation is occurring at a rapid rate in tropical countries, accounting for 12-15 % of human emissions of greenhouse gases. This causes more problems than exacerbating climate change and the loss of species. Destruction of forests is known to reduce rainfall, increase the rate of soil erosion and increase the risk of flooding in tropical countries. A major United Nations policy process is underway to halt tropical deforestation: it is called Reducing Emissions from Deforestation and Degradation (REDD). REDD involves developed countries making direct payments to developing countries, enabling them to increase their population's standard of living without destroying their forests. This also benefits the developed world, as it will reduce the rate of climate change and preserve these biodiverse and climate-stabilising forests. One major sticking point in the negotiations for REDD is methods for measuring and monitoring forest area, and in particular 'biomass'. This is the amount of living material in an area, and corresponds directly to the quantity of carbon stored in the vegetation. It is important to be able to assess the biomass of an area at a number of different time points: in the past so that the historical rate of deforestation can be calculated; at the present day; and into the future to measure deforestation and therefore calculate (and verify) appropriate payments. The most accurate way to measure the biomass of forests is to measure the location, diameter and height of every tree. This is expensive and time-consuming, and thus most projects are constrained to a maximum of about a hundred plots over a project site. Therefore satellite data is used to produce biomass maps for the past and present, using field data to test and improve the accuracy of estimated biomass. There is little independent work being done to assess the accuracy of the many different satellite monitoring systems available. There are three major types to consider. First the various optical satellite sensors, which are essentially similar to digital cameras, taking images of the Earth from space; secondly radar satellite sensors, which send pings of microwave radiation at the surface, and 'listen' for their return; and thirdly LiDAR data, which is similar to radar but uses pulses of laser light. All of the systems have different advantages and disadvantages: optical data cannot see through cloud, and can only determine forest area rather than biomass and so may miss degradation; radar data is a new technology and the consistency of biomass maps in different areas and at different times is unknown; LiDAR produces the most accurate biomass maps, but there is currently no satellite system that collects it, so it has to be collected from an aircraft, greatly increasing costs. We aim to objectively estimate the errors involved in these different methods using REDD forest projects in three African countries with differing vegetation types: Mozambique, Uganda and Gabon, as well as looking at new ways to use and combine their data to increase accuracies. We have already established collaborations with the management of these REDD projects, which are in various stages of development, and they have agreed to give us access to all their data, and to allow us to influence ground collection regimes and analyse satellite data over their sites. The project will quantify the accuracy of different satellite methods for assessing changes in biomass in REDD projects. The data and analyses emerging from this study will be widely used, by project implementers trying to find optimal methodologies, and by investors and independent monitoring agencies wishing to estimate the accuracy of the monitoring regime in a project or country. The project has the potential for a huge impact, potentially contributing to a reduction in the rate of global deforestation by reducing the cost and increasing the accuracy of all forest monitoring systems.