![]() |
ИСТИНА |
Войти в систему Регистрация |
ИПМех РАН |
||
A major part of subarctic Eurasian landcover is composed of boreal forest, and understanding interactions between this forest and climate is essential to the unravelling of ecosystem dynamics. Satellite remote sensing is well established as a method of assessing the strength of vegetation, and time series of satellite images are routinely analysed to monitor vegetation dynamics. However, most such approaches depend on the use of vegetation indices, which are only indirectly related to the amount of vegetation. Above-ground biomass (AGB) or Growing Stock Volume (GSV) are more difficult to estimate from satellite data, although recently some global-coverage, courseresolution data products have become available. In this presentation we describe a new approach to estimate spatial variation in GSV at a resolution of 20 m, using summer imagery from Sentinel-2 MSI and a satellite-derived land cover product, trained using field measurements of individual tree geometries from field plots in two contrasting areas of the Russian boreal forest. The study areas were located on the Kola Peninsula, where the forest is mainly composed of pine, spruce and birch, and in the Sakha Republic, where the forest is mainly composed of larch, pine and birch. Preliminary results suggest that the method is capable of estimating GSV to within 50% of its true value.