Kravtsova V.I.  

Methodical approaches to investigations of the nothern forests boundaries dynamics using aerospace images

The problem of the northern boundary of forests and its dynamics is study in connection of global warming. Fuzzy character of this boundary (forest line) transforms this problem to dynamic of spatial structure of tundra-taiga ecotone. Remote sensing investigations on this problem are carry out by scientists of Aerospace Methods Laboratory, Department of Cartography and Geoinformatics, Faculty of Geography Moscow State University in the frame of PPS Arctic international project at test sites in plane and mountain regions on Kola Peninsula and Taimyr.
Forest-tundra zone at the northern edge of Russia is prolonged to 6000 km. Its structure depends on type of forest and its species, which changes from rare birch forests on Kola Peninsula and birch–fir-tree (picea) forests at European part of Russia to very rare larch forests and redina at Central and Eastern Siberia.
These regional differences and presence of aerospace images for the period of warming lead to different methodological approaches to processing of aerospace images for forest line and ecotone tundra-taiga structure studies. It is clear seen at the example of our investigations in two contrast regions – Kola Peninsula and Taimyr. In this regions we had tried to use various materials and the next approaches:
1. Comparison of multitemporal topographic maps, compiled in 1960’s (1:50 000) and 1980’s (1:25 000), have been created at Kola test site, but it was unsatisfactory due to not clear representation of tundra–taiga ecotone structure at the maps.
2. Comparison of multitemporal high resolution images – airphotos of 1950–1960’s with modern satellite images of very high resolution (VHR). This approach had allowed to study forest boundary claiming up at 30 m in Khibiny Mountains and advancing of dwarf shrub tundra to lichen tundra in Kanentiavr test site at plain part of Kola Peninsula. To regret, this successful approach may be apply only at local sites because of too small width of surveying pass, which lead to practically useless for wide territories.
3. Moving  from VHR satellite images to Landsat. Landsat images are the best remote sensing materials for our purposes due to long period of surveying from 1970’s–1980’s. But their resolution (30 m) is not enough for representation of forest line and tundra–taiga ecotone structure changes. Different components of this ecotone structure are mixed in Landsat pixel. So the approach of unmixing is developing with using VHR images as etalons. Analyzing of VHR images and special spectrometric experiments have allowed to create the methodic of compiling maps, which characterize quantity relations of ecotone components at Kola test sites and maps with quantity characteristic of forests density – at Taimyr (Ary-Mas) test site. But compiling of such useful maps is possible only for modern period, when VHR satellite images have been appeared, and for the future.
4. Simple and universal approach based on comparison of multitemporal NDVI images taken by Landsat pictures, allows to investigate changes in tundra-taiga ecotone vegetation in essential period at the both test site regions, but it requires the execution of some conditions and additional actions during images processing.

Abstracts file: Кравцова-Интеркарто17-краткий текст.doc
Full text file: Интеркарто-севгранлеса.doc

To reports list