Changes in the form of forests in the Central Russian upland in the submeridional direction (quantitative description based on landscape metrics)

The article is dedicated to a quantitative description of the geometric shape of forests in the territory of the Central Russian Upland. This is a new direction in the research of forest insularity in the forest-steppe zone. It complements the existing works in the field of landscape ecology, which are dedicated to the area, location and fragmentation of forests in the forest-steppe zone. This study was carried out at four key sites with a size of 30 by 30 km each. The sites are located along the north-south line, and each site is located in one of the natural subzones of the Central Russian Upland (subzones of the northern, typical, southern forest-steppe and the steppe zone). For each site, forests were mapped using images from the Sentinel-2 satellite in the ArcGIS 10.5 program. Then, in the “R” free software environment for statistical computing, landscape metrics were calculated that characterize the complexity of the shape of forest areas, using the additional landscapemetrics package. Six landscape metrics were used, namely: average related circumscribing circle (CIRCLE), average contiguity index (CONTIG), average fractal dimension index (FRAC), perimeter-area fractal dimension (PAFRAC), average perimeter-area ratio (PARA) and average shape index (SHAPE). A comparison of landscape metrics for different key sites showed that the complexity of the forest form increases in the direction from north to south. This is due to a change in the landscape localization of forests. From north to south, the proportion of forests growing on watersheds decreases, and the proportion of ravine forests increases. The maximum increase in the complexity of the form is observed during the transition from the forest-steppe zone to the steppe zone. Of all the landscape metrics used, the PAFRAC and SHAPE metrics showed the greatest sensitivity to changes in the shape of forest areas. These metrics are most useful when comparing the forests shapes in different areas.