For the purpose of the periodic global forest resources assessment, FAO has developed a simplified classification system of the forest vegetation based essentially on defined physiognomic parameters, such as crown cover and height, and on agricultural impacts. Vegetation with crown coverage less than 10% is not classified as forest. Trees outside forests are also not recognised by the classification. Crown cover classes are very wide (cc = 10-40% and 40-100%) to enable detecting changes when degradation or recovery are concerned. The processes of degradation and/or recovery of natural vegetation are very slow to be detected with the required accuracy over a period of 10 years. They require many decades to be easily measurable.
There is therefore clear need for an improved classification that fits the GFS requirements, which will rely, to a large extent, on field measurements of forest characteristics and socio-economic factors. The component trees outside forest should be incorporated within the classification. This component includes wood-lots with less than 0.5 ha, line plantations, gallery vegetation, isolated trees, etc. These may be planted for forest purposes or for agricultural aims. The forest categories should be revised and improved to enable measuring in the field every change in forest canopy and detecting the degradation or recovery processes at their earliest stage.
The classification should all the time be easy to implement on remote sensing material (aerial photographs, satellite images, radar images). But what is feasible on air photos is not fully achievable on satellite imagery as both media have different picture proprieties as well as spatial, spectral and radiometric resolutions. This requires a standardisation of the classification scheme for better use of available remote sensing media but also to encompass the various eco-types, tree physiognomies and stand structures.
Co-ordination with other countries and specialised institutions will be needed for the standardisation of the forest classification system at global level while ecological, social and economic specificities are reflected in country classifications.
Use of aerial photographs is the most appropriate tool for production of accurate thematic maps and generation of various other statistical informations. It is however a costly material in forest inventories and can only be supported when the investment is socially and economically justified. Use of aerial photography in the GFS will remain depending on their availability and on the extent of co-ordination with potential users in other sectors, when new coverage is to be decided in a given country.
Nowadays, satellite imagery has been improved tremendously in terms of spatial, spectral, radiometric and temporal resolutions. The quality of these spatial data for natural resources monitoring allows, nowadays, extracting better quality information such as thematic maps, statistics and other qualitative data. It will be more cost-effective for the GFS and easier2 to rely on satellite imagery.
2 Satellite images are more available than air photos in most countries and particularly in the developing world.
