Science based remote sensing services for REDD & tropical forest management (ReCover)

Science based remote sensing services for REDD & tropical forest management (ReCover)





Project coordinator: VTT Technical Research Centre of Finlandn FinlandOther Partners:Albert-Ludwigs-Universität Freiburg GermanyArbonaut Oy Ltd., (SME) FinlandColegio de Postgraduados MexicoEl Colegio de la Frontera Sur MexicoGMV Aerospace and Defence SA Unipersonal SpainNorut - Northern Research Institute Tromsø AS NorwayWageningen Universiteit The NetherlandsThe Hydrological, Meteorological and Environmental Studies Institute (IDEAM) ColombiaUsers: Conafor, National Forestry Commission Mexico Secretary of Environment and Natural History of Chiapas State Government MexicoComision Nacional para el Conocimiento y Uso de la Biodiversidad, MexicoComisión Nacional de Áreas Naturales Protegidas, Región Frontera Sur, Itsmo y Pacífico Sur MexicoPMC Mexican Carbon Program Mexico The Hydrological, Meteorological and Environmental Studies Institute (IDEAM) ColombiaGuyana Forestry Commission GFC GuyanaObservatoire Satellital des Forêts d'Afrique Centrale The Democratic Republic of CongoFiji Forestry Department Fiji

Improving tropical forest monitoring

Final Report Summary - RECOVER (Science based remote sensing services to support REDD and sustainable forest management in tropical region)

Executive Summary:
The ReCover project was executed in 2011-2013 under the Framework Programme 7 of the European Union to develop science based remote sensing services to support REDD and sustainable forest management in the tropical region. More precisely, ReCover focused on the Measurement, Reporting and Verification (MRV) process of REDD+.
Nine research partners developed remote sensing services for the REDD MRV for users in Mexico, Guyana, Colombia, DRC and Fiji. The interaction with the users was close through the service level agreements, user workshops, and bilateral communications. The users expressed their content of the services.
Two-stage sampling approaches were developed for forest and land cover assessment. A statistical sampling ensures provision of objective statistical data of the variables with confidence intervals, relevant to REDD. The sampled data are also used to assess the accuracy of wall-to-wall mapping and they can be used to correct the bias in the map estimates. Very fine resolution satellite data were used to collect the statistical sample.
In General, (semi-)automated classification approaches based on high resolution (10 – 30 m) optical satellite imagery could distinguish forest land from other land with overall accuracy levels close to or above 90%. The classification accuracy exceeded 90 % in areas with high forest cover and limited anthropogenic influence. It could be somewhat below 90 % when a large proportion of forest was degraded and the changes between forest and non-forest were very frequent due to shifting cultivation, for instance. Classification using radar data of the lower frequency L-band with similar resolution led to similar or slightly lower accuracy levels.
The overall accuracy was somewhat lower when considering all six IPCC classes (between 60-70%) than in forest and non-forest classification. For example, the separation of grassland from cropland remained problematic.
Forest cover was monitored between 1990 and 2012 by moving backwards from near-present images for which ground reference data were available. Also experimental biomass and degradation (biomass reduction within the forest class 1990-2010) were computed.
Using the common concept that was developed and tested in ReCover, forest and non-forest mapping for REDD+ with optical data could be implemented operationally without major additional research efforts. The mapping should include a statistical sampling framework to produce reliable information on forest area and its confidence intervals.
The Sentinel-2 satellites with 10 m spatial resolution, 290 km image size, high radiometric resolution, and 5 days imaging cycle at the equator will dramatically improve possibilities for obtaining cloud-free and high quality data from forests globally. Radar data preferably from L-band sensors are important sources to provide timely information on changes in forest cover. Also the C-band instrument of Sentinel-1 will be applicable for monitoring. The Copernicus program will enable application of earth observation to support the MRV operationally.
Access to data should be made as easy as possible also for the developing countries. The development of the REDD services should be combined with national forest inventories. Capacity building is crucial if REDD+ countries are to monitor and account for the changes in greenhouse gases. Common standards for the MRV would make its development easier.

News: Hva gjør Norut i Kongo? (Norwegian language only)