Earth Research from Space

The work is devoted to the study of the vegetation response in the tundra, forest-tundra, and taiga zones of Western Siberia to climate change (temperature and precipitation) according to MODIS/Terra+Aqua satellite data for the period (2000–2022). The study is based on the maps of land cover (The MODIS Land Cover Type). Based on maps from 2001, 2010, and 2020, the areas of vegetation classes in tundra, forest-tundra, and taiga were determined. Multidirectional dynamics of the areas of the vegetation classes were noted. Positive trends in the long-term series of average summer (June, July, August) temperature and NDVI values of tundra, forest-tundra, and taiga vegetation for the period 2000–2022 were identified. A slight positive trend was noted in the long-term series of average annual precipitation in the territory of the grass tundra, and a negative trend in the territory of the shrub tundra. Positive trends in the long-term series of average summer NDVI values of vegetation and temperature were identified in the territory of permanent wetlands and forested swamps. Precipitation trends in wetland areas remained unchanged. For the periods (2001–2010) and (2010–2020), the following phenomena were revealed: the movement of the shrub tundra to the north, into the territory of the grass tundra, and the movement of the shrub tundra and northern woodland into the territory of the forest-tundra. The movement of the forest to the north, into the forest-tundra zone, was also noted.

Currently, northern ecosystems are changing under the influence of climate warming, which makes it necessary to study their dynamics and predict their further development. The Nadym geocryological observation site landscapes was classified using the Support Vector Machine method with high accuracy (86–97 percent) based on Landsat satellite images from 2001, 2013, and 2023. Temporal sampling with selected landscapes (surface types) made it possible to determine changes in their area and to simulate the development of the area up to 2033 in the MOLUSCE 4.0 QGIS module based on the algorithm of artificial neural networks (multi-layer perceptron), taking into account geophysical influencing factors. It was revealed that the dynamics of landscapes is largely due to the periodic occurrence of fires. Analysis of the transition matrix showed the presence of local substitutions of flat shrub-grass-moss bogs with grass-shrub-moss-lichen peatlands. Forecasting for a 10-year period has demonstrated that, with the current trend, vegetation will recover on forest and peat burns with an increase in lichen abundance and a further smooth transition of a small proportion of bogs to peatlands, due to the processes of drainage and phytomass growth against the background of an increase in the surface atmospheric temperature.