Use of the Triangular Greenness Index (TGI) and high-resolution UAV data to analyse the vegetation development (2014-2022) in a calanchi badland in southern Tuscany (Val d'Orcia)

Kurzfassung/Abstract

Badlands are of great significance for humans and the environment due to the rapid geomorphic development. The establishment of vegetation is a significant factor in the stabilisation of these geomorphic highly dynamic landscapes, and thus represents a crucial step in the restoration of these ecosystems. In Italy, these landscapes are pervasive throughout the entire Apennine region. This study examines the use of the Triangular Greenness Index (TGI) in combination with high-resolution UAS data to analyse the vegetation development in a calanchi badland in southern Tuscany (Val d'Orcia) between 2014 and 2022. Furthermore, the study defines limitations and potentials of the RGB-based index in such complex landscapes. In order to ascertain the vegetation, it was first necessary to establish the requisite threshold values on the basis of both the prevailing lightning conditions and the corresponding colourisation of the respective orthophotos. Subsequently, binary grids (vegetation/no vegetation) were calculated in order to facilitate the automated mapping of the vegetation cover. To validate the automated mapping, a series of plots were delineated, and the vegetation within them was mapped manually with high precision. In the case of vital vegetation, high correspondences of 83-95% were observed. The results show an increase in vegetation cover of 10% over the entire study period. However, the development of the vegetation is very dynamic. While the period from 2014 - 2016 shows a decline of -21% in vegetation development, the following years are characterised by a phase of strong vegetation growth (36% between 2016 and 2018) and stabilisation (2018 – 2022: +1%). From a spatial perspective, this results in the extensive growth of vegetation in the areas of the valley floor of the main valley and the northern exposed slopes. In particular, areas affected by earlier landslides and mass movements demonstrate a notable increase in vegetation growth within a few years. In light of the aforementioned findings, it can be posited that the TGI represents an efficacious and cost-effective methodology for the semi-automatic mapping of vegetated regions with high precision over a large spatial extent (badlands). Nevertheless, the TGI is not free from limitations. In particular, the quality of the results is influenced by the lighting conditions (shading) and the phenological status (chlorophyll contect), which is expressed in the green value of a pixel. These limiting factors are particularly pronounced in topographically complex study areas, such as calanchi, and delimit the vegetation analysis in these areas. In these areas, the validation plots demonstrate a low level of agreement, with a value of only 23%. However, this is limited to severely damaged or dead vegetation, which represents a relatively small proportion of the total vegetation. In conclusion, the TGI is a suitable index for the automated mapping of vegetation in calanchi badlands on basis of high-resolution UAS-data, particularly when data collection is conducted under optimal lightening conditions.