Darstellung maskierter Nutzungseffekte auf naturraumspezifische Artengemeinschaften grasiger Feldraine mithilfe von Restvarianzmustern

Background and aim Grassy field margins have a high relevance in agricultural landscapes regarding the preservation of typical arthropod communities, their biodiversity and the ecological system functions linked with it. The structure of terrestrial communities is affected by anthropogenic impairment, which can lead to the replacement of sensitive by more tolerant species. The negative influence of pesticides on fauna and flora as well as the associated functional aspects (ecological system functions) and the reduction of biodiversity are undisputed since longer for agrarian systems and can also be assumed for grassy field margins due to spray-drift. The case study presented here examined the effect of influences related to utilization on the plant and arthropod communities of grassy field margins. Reference sites, on which due to missing farming in the direct surrounding countryside no utilization influences on the communities were present, were compared to non-target sites, on which these influences could not be excluded. Sites in three German macrochores were examined: in the Jülicher Börde, at the northeastern edge of the Leipziger low lands and in the area Mainfranken near Würzburg, all of them intensely used agricultural landscapes. Beside the vegetation, the epigeic arthropod communities of carabid beetles, spiders, springtails, hymenopterans, hover flies and ladybirds as well as abiotic parameters were included in the analysis. The aim of the study was the development of a statistical exclusion procedure which is capable to quantify the amount of variation in field community data which can be attributed to isolated factors. Special attention was paid to non-observable utilization impacts like undocumented pesticide application. The extraction of patterns of residual variance allowed for the uncovering of masked effects on a scale below the obvious abundance pattern. Materials and methods The variability in the species composition was visualized with the help of non-metric multi-dimensional scaling (NMDS). Indicator species analysis revealed those species which could be recognized as statistically significant indicators for local conditions. The relationship between the species composition and the environmental factors was statistically modeled by canonical correspondence analysis. By variance partitioning it was possible to extract the variance portion which could be bound to a set of covariables. For the remaining residual variance it could not be excluded that this resulted from a pesticide influence. Results The analyses based on the ecological distance (Bray-Curtis) showed that the communities in references sites and non-target sites could be clearly distinguished in all three landscapes. Based on the portions of variance which could be explained statistically by a utilization related influence, two different directions of reaction to the utilization related variables could be stated. On the one hand there were sensitive species, showing reduced abundances in the non-target sites, on the other hand there were species increasing in abundance in the non-target sites, recruiting from the group of strong competitors. Discussion By the use of residual variances a clear influence of utilization related parameters on the community of soil arthropods and vegetation could be shown. The observed abundances shifts between reference sites and non-target sites in the raw data were attributed to a multi-dimensional factor complex which could be split up by the use of partial ordination methods, quantifying the relationships to the utilization related parameters. It was demonstrated that special utilization correlated patterns could be found in the raw data after eliminating the influence of the covariables. Conclusions Utilization related influences in the non-target sites evened out the characteristic communities for the three landscapes towards a comprehensive, ubiquitous species composition. The hypothesis of decreasing abundance of sensitive species and the augmentation of tolerant species due to a potential pesticide influence could be underpinned and quantified. Sensitive species which showed a significant reduction in abundance between reference and non-target sites showed a much higher sensitivity to the influences and thus were assigned a higher indicator potential than tolerant species. From the group of the sensitive species Pardosa palustris and Poecilus cupreus could be isolated as two promising macrochore-specific indicator species for the masked effects. Recommendations and perspectives A macrochore-specific assessment of utilization effects was requested for future studies due to the different sensitivities of the species in the three landscapes. The elaboration and specifying of reference conditions for terrestrial agrarian systems is an important task for the future. The presented approach deduces a macrochore-specific, complex effect pattern of anthropogenic impairment on terrestrial species communities. It can be used to extract masked effects and by this facilitates a more sustainable use of e.g. plant protection products. Furthermore it provides an opportunity to validate evaluation systems for the effects of utilization impacts on terrestrial vegetation and arthropods. Additionally the species sensitivities with respect to the macrochores allow a regionalized assessment of ecotoxicological effects and the integration into spatially explicit effect assessment models.