The use of coherence information from ERS tandem pairs for determining forest stock volume in SIBERIA

SAR Interferometry yields coherence maps that have been found useful for forest classification. In this paper the possibility of using coherence information derived from ERS Tandem pairs for estimating growing stock volume of boreal forest is investigated. The problem is that the coherence is not only dependent on land cover, but also on imaging characteristics (baseline), environmental effects (rainfall, wind, frost), and topography. An example is shown in the form of a mosaic of nine ERS coherence images from a region around the Yenisey river in Central Siberia. It is clear that the coherence may change drastically from track to track. Also, the strong influence of topography is evident in the lower part of the image. To study the relationship between forest and radar parameters an extensive data set was built up within the framework of the SIBERIA project, which has the aim to generate large-scale forest maps for Russian foresters using ERS and JERS imagery. Forestry data from 36 test sites, each covering from 20,000 to 100,000 hectares of forested land, were made available. The sites were chosen to represent different forest and vegetation zones. This extensive database provides a good basis for understanding the potential and limitations of using ERS coherence in forestry applications.