Investigating anthropogenically transformed landscapes with remote sensing

Abstract

The assessment of the current state and spatio-temporal changes in natural landscapes is crucial, among others, for the agro-industrial and agricultural sectors of every country, including Ukraine. The main objective of this paper is to estimate the state and spatio-temporal changes in landscapes with land-cover identification using remote sensing data processing and geographic information technologies. The generalized scheme of the present study consists in the analysis of natural and anthropogenic changes of Ukraine's capital city of Kiev landscapes using satellite data. The processing involved step-by-step decipherment of Kiev's landscapes using cluster analysis, namely supervised and unsupervised classification based on a series of satellite images from Landsat missions 5, 7, 8 (1985–2020) and Sentinel-2 MSI (2015–2020). Spectral signatures were determined, and their quality was estimated based on the classification. The Normalized Difference Vegetation Index (NDVI) was calculated for vegetated landscapes. The following landscapes were identified: water bodies, vegetation, urban areas, and bare soil. Two land cover maps were compiled from Sentinel-2 MSI images for 2015 and 2020 with an overall accuracy of 89.76% and 82.05%, respectively. Two maps of anthropogenic transformations of Kiev's landscapes were also compiled. According to Landsat data (spatial resolution 30 m), the area of urban landscapes increased by about 41% from 1985 to 2020, and according to Sentinel-2 MSI data (spatial resolution 10 m), it grew around 5% from 2015 to 2020.