Determination of Optimal Locations of Shelterbelts in Southern Ukraine Using Remote Sensing Data

The principal component of southern Ukrainian agrolandscapes that significantly influences their sustainability is shelterbelts. Determining their optimal locations in order to maximize water-holding and snow-retention effects requires the usage of novel techniques, in particular decision support tools. We propose to build them on the base of the dependencies built using remote sensing data for different agrolandscapes’ organization. We used Landsat-8 satellite images acquired in 2015–2019 to obtain the relevant regional-level dependencies. For this purpose each agricultural field in the Kherson region of Ukraine was classified by the criteria of irrigation presence and surrounding shelterbelts’ number. In the case of the water-holding function of shelterbelts, we construct exponential dependencies between NDVI (Normalized Difference Vegetation Index) and VTCI (Vegetation Temperature Condition Index) indices’ values, which describe the moisture state of a specific type of fields for a fixed biomass. These dependencies describe both the phenomenon of moisture extraction by shelterbelts and their water-holding effect which is more evident on rain-fed fields. In the case of snow retention function, appropriate dependencies were build using the statistical characteristics of snow distribution on the fields accessed using NDWI (Normalized Difference Water Index). Having the appropriate dependencies, we pose the problem of shelterbelts creation planning in order to increase the moisture content and snow retention on surrounding fields. The genetic algorithm for the solution of this problem along with algorithms of automated input data preprocessing are proposed and tested.