Landscape structure of the Middle Lena valley

Abstract

In this article we present the results of landscape structure study in the Middle Lena valley (Erkeni, Ensieli and Tuimaada), made using GIS mapping based on Sentinel 2 satellite images and ASTER GDEM datasets. Alas complexes and croplands were identified using K-means image segmentation with different initial numbers. Three machine learning algorithms were tested for pixel-based land cover classification. The best result was achieved with the Support Vector Machine classifier. The mapping revealed the landscape structure of 20 spatial units, which were organized into groups of upland, alas, slope, valley and floodplain landscapes. The contrasting landscape structure of the Middle Lena valley was revealed. The northern Ensieli valley is characterized by the predominance of forest larch-pine landscapes, alas complexes mostly dry with steppe meadows are widespread in the watershed. The slopes are dominated by larch forests with patches of birch. The central Tuimaada valley is the most anthropogenically transformed, here steppe meadows and anthropogenic steppes dominate. The plateau part is mainly forested with croplands. The southern valley of the Erkeni is characterized by the smallest area of the valley part dominated by steppe meadows, the slopes of the main bank of the southern and south-eastern exposition are covered with xerophytic meadows. The landscape structure of valleys, floodplain and plakor parts largely determines the specifics of agricultural land use of the local population. The results obtained can be used for further research into the response of permafrost landscapes of the Middle Lena valleys to increased anthropogenic pressure in the context of climate change.