Inventory methods and mapping of forest extensions

Abstract

A better knowledge of forest extensions is becoming more and more urgent whether for economical, environmental or social reasons. This is shown by a demand for information, notably in the form of maps. We propose to look over the data and the processing methods available for listing and mapping spontaneous forest extensions. The dynamics of the vegetation related to the natural reforestation of abandoned agricultural land often comprises an intermediate phase corresponding to the invasion by low ligneous vegetation of the species originally cultivated. A first study performed in Cantal shows that the various classes of fallow land are generally too variable and show too much radiometric confusion to allow reliable identification using Spot satellite images taken at any given moment. However a diachronic approach by comparison of images allows identifying the areas which show radiometric changes and then characterise the changes. As for aerial photography, it seems the most suitable medium and its stereoscopic analysis is the best method for identifying the phenomena. However, the digital processing of photographs is fast developing. A second example presents the results obtained by Swiss software automatically classifying digitised aerial photographs in forest formations from isolated trees to closed forest. This type of application, made possible by recent progress in data processing, should develop further. It should be favoured by the development of the market for digital aerial photographs as well as by the arrival on the market in 1999 of new satellite data with a resolution of 1 or 2 m.