Remote sensing in support of French Guiana

A European team headed by CIRAD has shown that spatial remote sensing could be of use in analysing these types of areas, in such a context. Remote sensing, which has not previously been used to any extent in the region, can be used to observe wide and often inaccessible areas, repeatedly and objectively. The eventual aim is to inventory the situation across the whole of French Guiana, so as to enable decision-makers to make informed choices about development.

The team used two tools: a set of daily low-resolution data covering a full year, obtained using the VEGETATION sensor carried by the SPOT-4 satellite. The preliminary information derived from those data was used to define the main vegetation types and identify the landscape structure. The researchers then superimposed maps produced using high-resolution data obtained by the Landsat ETM+ and SPOT-HRVIR satellites, characterizing human impact.

By combining those data, the researchers were able to produce a map of vegetation type distribution. Three main soil cover families were identified: swamps, including mangroves and varzeas, humid riverside zones, and forests and savannahs. Three types of activity were detected within those areas. Forestry operations (felling and skidding) were identified from gaps in the forest. Gold panning, and more specifically sludge decantation ponds, was pinpointed from deforested areas. Lastly, agricultural activity in western French Guiana was assessed: infrastructures such as roads, tracks and forest trails were detected, as were farmed areas. Only regenerated forest areas escaped detection, depending on the degree of regeneration.

This technique has laid the technical foundations for planning and managing vast areas as efficiently as possible.