Modeling the Forest Cover of the Vengeri River Basin (Sakhalin Island) with Remote-Sensing Data

Abstract—The study concerns mapping of the actual vegetation cover of the Vengeri River basin (eastern coast of the central Sakhalin) with an area of more than 33 000 ha. A cartographic model has been created with the use of the algorithm of stepwise discriminant analysis based on the original geobotanical data, a digital elevation model, and the spectral data of Landsat 5 images. Nine physiognomical vegetation types were recognized: three of them are nonforest (open rock vegetation, alpine tundra, and alpine meadows), while six feature different kinds of forests: fir–spruce forests (Abies sachalinensis, Picea ajanensis), larch forests and sparse stands (Larix cajanderi), shrub thickets (Pinus pumila, Dushekia fruticosa), Erman’s birch forests (Betula ermanii), and broad-leaved and riparian deciduous forests (Alnus hirsuta, Betula platyphylla, Chosenia arbutifolia, Populus suaveolens, Salix udensis). The dependence between the physiognomical types of the forest and the syntaxa of traditional Russian dominant (ecophytocoenotic) and floristic classification systems was determined. Forest communities evolved on 89.1% of the river basin; the largest area was covered by larch forests and open woodlands (natural sparse forests) (33.32%). The landscape and climatic conditions of the region allow the existence of fir–spruce forests, which are typical of the zone, but they account for as little as 11.23% of the entire territory. The widespread distribution of larch forests results from the natural pyrogenic transformation of the dark coniferous forests. Under the conditions of the strongly rugged mountain relief, it was impossible to decode and model the vegetation areas situated on the shadowed northern slopes automatically. The paper proposes that the reconstruction should be performed with visual decoding of high-resolution images and field-study data.