Regional Forest Obervatories



Roadless Forest Project


Aerial view of a tropical humid forest.

Tropical forests protect us from climate change and are a haven for biodiversity. Most forest loss, degradation and fragmentation occurs in a “risk zone” around transport networks. Although communities need roads to access markets, hospitals and schools, roads also open up the forest to possible damage. In tropical areas, new road building is often followed by secondary roads being created, often unplanned or illegal, which then trigger further degradation and deforestation. The remoteness of some areas makes it especially hard to monitor and manage this change. The RoadlessForest pilot project aims to provide up-to-date data, to help plan infrastructure wisely and spot possible areas where the forest is in danger. The Joint Research Centre of the European Commission, with funding from the European Parliament, is building a road and forest atlas for tropical regions to support sustainable development. Sustainable development and sustainable forest management is a way of protecting biodiversity, fighting desertification and responding to climate change, whilst ensuring that forest ecosystems deliver goods and services.

Road in a degraded tropical forest landscape.

Tools and methods

By processing data from satellite (Landsat and Sentinel) images at fine spatial resolution with Google Earth Engine, we aim at mapping roads and assess forest cover and forest cover change (deforestation, regrowth, intact forest) on the last 32 years. With RoadlessForest, we are also piloting crowdsourced mapping to monitor roadbuilding in remote forest areas. Maps will be available on OpenStreetMap, a free-to-use open-source map built by mappers all over the world. By modelling the deforestation process with advanced statistical approaches (such as hierarchical Bayesian models), we also aim at understanding better the role of road development in driving forest cover changes.

Case-study sites

Six tropical forest sites have been selected as case-studies to test the methodology employed in the framework of the Roadless Forest project, with the aim of extended the work at the whole tropical belt later on. The case-studies are located in the three tropical continents in South-America (Brasilian Amazonia and Paraguay), Africa (Western DRC and Eastern Madagascar) and South-East Asia (Cambodia and Papua New Guinea).

Pilot study areas of the Roadless Forest project.

Available products

Accessibility map

The Global Map of Accessibility characterizes the connectedness of the human landscape by illustrating the amount of time it takes to access the nearest densely populated area. Mapping accessibility to cities is a useful proxy for the relative ease by which people in rural areas can access services and resources concentrated in more urban areas. Conversely, this map measures the relative inaccessibility and remoteness of the forest areas. The map is available at

Accessibility map.

Intact humid forest and forest cover change over the last 32 years

A methodology is currently being developed and implemented with Google Earth Engine for mapping the intact forest and the forest cover change in the humid tropical world for the last 32 years at 30m spatial resolution. This would be an unprecedent work giving globally consistent and locally relevant information on the forest state. The final version of the map will be available in the near future.

Map of intact humid forest and logging roads in North Congo
from Landsat satellite images analysis on 32 years.

Deforestation probability map

We develop a methodology to derive a map of the probability of deforestation as a function of several spatial factors including roads. We thus isolate the effect of roadbuilding in driving deforestation from the effect of other spatial factors such as population density or topography. With the deforestation probability map, we can also identify both forest areas with high risk of deforestation and potential intact forest areas in the future. A first deforestation probability map is already available for Madagascar at (hyperlink) and the methodology will be extended to the whole Africa and to the 6 case-study sites in the near future.


Uchida H. and Nelson A., 2008
Agglomeration Index: Towards a new measure of urban concentration
Urbanization and development: Multidisciplinary perspectives, 41-60. World Development Report 2009. The World Bank.
Vancutsem Ch. and F. Achard, 2016
Mapping Intact and Degraded Humid Forests over the Tropical Belt From 32 Years Of Landsat Time Series
Paper 2034 - Session title: Tropical Forest and REDD+ 1. ESA living planet symposium 2016
Vieilledent G., C. Grinand, M. Pedrono, T. Rabetrano, J.-R. Rakotoarijaona, B. Rakotoarivelo, F. A. Rakotomalala, Linjanantenaina Rakotomalala, Andriamandimbisoa Razafimpahanana and Frédéric Achard.
Global trade and bad governance are responsible for the unceasing deforestation in western Madagascar, not poverty.

Project website