Assessment of species composition, productivity and functionality of grassland in the Greater Caucasus (Georgia, Kazbegi Region) by means of remote sensing
High mountain ecosystems and their services are of global importance. They take 20 percent of the terrestrial surface, offer habitat to humans, a diverse fauna and flora, and play a key role for fresh water supply. But for all that, they are prone to global change. Remote, high mountain regions, especially from former Soviet Union States experience a strong change in the agricultural sector, mainly the abandonment of agricultural land use, which impacts socio economy, ecosystem functioning and biodiversity in the end. The unique species composition of high mountain grassland and the related ecosystem services, such as biomass or erosion mitigation are most likely to change with the ongoing abandonment in high mountain regions, mapping and monitoring of high mountain grassland properties and key ecosystem processes is therefore highly needed. Thus, this study focuses on modelling the vegetation composition, biomass and plant functional groups (content of graminoid, herbaceous and legume plants) of the subalpine grassland, as well as the recognition of shrub encroachment, by means of remote sensing. Shrub encroachment has already been observed in the study region and plays a key role in many arctic and alpine environments. The identification of species identity is of great importance to assess potential consequences. The species-rich, extensive grassland in the isolated Kazbegi region, Greater Caucasus, Georgia is the main winter fodder source and thus of importance for the local subsistence agriculture, based on pastured cattle for dairy production. The results of the vegetation analysis show that the subalpine grassland vegetation is closely related to the physical site conditions (elevation a.s.l. and exposition to the east), characterized by broad transitions between grassland types with distinct attributes, such as aboveground biomass, vegetation cover, species richness and plant functional groups. The shrub stands of Betula litwinowii, Rhododendron caucasicum and Hippophae rhamnoides are of distinct growth forms and species composition. However, Betula litwinowii, Rhododendron caucasicum and the tall herb Verartrum lobelianum co-occur in the tree line ecotone, challenging a clear delineation. The results of the gradual modelling approaches show that species composition, biomass and to a lesser degree plant functional groups can be modelled by multispectral imagery, vegetation indices and topographical parameters. Species composition is furthermore a good predictor for biomass, even though the saturation effect is present at high yields in both approaches. Moreover, hyperspectral and multispectral data result in similar model fits. The separation of shrub reflectance indicated that Betula litwinowii and Hipopphae rhamnoides can be delineated in imagery, whereas Betula litwinowii and Rhododendron caucasicum, as well as Betula litwinowii and Veratrum lobelianum show a more similar reflectance pattern. However, overall separability was high, especially in simulated RapidEye and WorldView-2 reflectance. The satellite image taken in September further showed a better separability than the one acquired in July. Identification of species-rich grassland types with a high conservational value, from the resulting multi scale maps helps to apply site specific and case sensitive grassland management. At the same time the yield map can be used to identify the few highly productive sites in the landscape. The distribution of plant functional groups provides important information about the usability, such as hay meadow, and the erodibility. The encroaching shrub species and their distinct growth forms offered an excellent opportunity to study distinctiveness and the applicability of different sensors. A multi scale prognosis of grassland properties, as well as the delineation of shrub encroachment on species level, are important factors for sustainable land use planning in the high mountain landscape. They further allow an integrative analysis of the high mountain landscape for modified, site specific agricultural land use measures, with the help of normative scenarios.
Verknüpfung zu Publikationen oder weiteren Datensätzen
