Methods for assessing the temporal dynamics of landscape cover based on procrustean analysis of spectral indices v1
Year: 2025
Type: preprint
Abstract: Comparisons in different time periods using individual spectral indicesare often used to assess the dynamics of vegetation cover. The NDVI is a verypopular index for indicating disturbances in the structure of vegetation cover.Various vegetation indices are highly correlated for homogeneous vegetationcover, so assessing dynamics based on one of them is quite appropriate.Similarly, other single spectral indices can be used for other homogeneoussurface types, such as agricultural land, rocks, or water. However, in the caseof complicated landscape complexes, there is a need to assess the dynamics ofheterogeneous land cover types using indices that are sensitive to thecharacteristics of vegetation, soil and water surface. Anthropogenic pressurealso strongly affects the physiological state of plants, so it is alsoappropriate to use spectral indices that are sensitive to plant health. Themultidimensionality of the spectral index space raises the problem of reducingthe dimensionality of this space, as well as comparing different landscapestates over time. The method uses principal component analysis to reduce thedimensionality of the spectral index space and Procrustes analysis to comparethe structures established by the results of principal component analysis. TheR code uses as an example an image of the surface of the southern part ofKhortytsia Island, where floodplain ecosystems are located, which werenegatively affected by the environmental disaster caused by the undermining ofthe Kakhovka Dam by the Russian occupiers in 2023. Accordingly, the case studycompares the state of floodplain ecosystems in 2022 and 2024.
Language: English
Subfield: Ecology
Field: Environmental Science
Domain: Physical Sciences
SDG: Life on land
Open Access status: green
OpenAlex ID: W4409410659