Drivers of belowground β-diversity
In grasslands, extensive grazing generally leads to spatial heterogeneity in aboveground vegetation (Adler et al. 2001) where different patches represent different phases on a successional gradient (Olff et al. 1999). Cessation of grazing thereby becomes a homogenizing factor that pushes patches towards a climax stage of generally lower aboveground β-diversity (Olff & Ritchie 1998). As a result, one might expect belowground β-diversity to exhibit a similar decrease in response to cessation of grazing (i.e. greater homogenization in community composition). However, in contrast to the consistent negative responses for ɑ-diversity, our results show remarkably mixed responses for β-diversity of different groups of belowground biota. We observed a strong decline in β-diversity for eukaryotic soil microbes (fungi, protists), no change in β-diversity of prokaryotic soil microbes, and an in increase in β-diversity for springtails, mites and nematodes. These differences in response between larger bodied and smaller bodied soil organisms may result from differences in the sensitivity of these groups of organisms to shifts in plant species composition and/or changes in soil physical parameters that happen as a result of grazer removal. A plethora of studies has shown that soil microbial community composition is strongly related to the composition of the plant species community (Grayston et al.1998; Kowalchuk et al. 2002; Berg & Smalla 2009; Leff et al. 2018), as many microbial taxa are directly dependent on carbon sources (e.g., exudates and litter) from plants. Soil animals are also ultimately dependent on plant-derived carbon, but, due to their greater mobility and size, may also be affected by other environmental factors that change in response to the cessation of grazing. We therefore propose that the changes in vegetation community and the accompanying changes in root exudation patterns, litter quality, local changes in pH-gradients (Fierer & Jackson 2006; Rousk et al. 2010) as well as variation in litter recalcitrance (Freschet et al. 2012) may explain the observed differences in β-diversity for soil microbial groups, whereas the observed physical differences in soil properties, soil organic matter and soil structure, and the spatial variation therein might be more important for the spatial distribution larger bodied species (Ettema & Wardle 2002; Quist et al. 2019).
A last remaining question is why β-diversity of soil fauna actuallyincreases in response to cessation of grazing. Here, we speculate that this might be due to increased cover and patch size of mid-late successional (clonal) plant species, and the observed coinciding increase in air filled porosity. Indeed, clonal plant species characteristic of mid to late successional stages (e.g., C. vulgaris , Pteridium aquilium. Molinia caerulea , D. cespitosa ) were more abundant in the abandoned plots. These species are generally associated with more complex food webs and a greater abundance of higher trophic levels as a result of larger belowground carbon inputs (Morriën et al. 2017). Patches that consist of different functional groups (heather species, legumes, grasses, other shrubs, ferns, mosses) can differentially affect the diversity of organisms through changes in resource supply and other abiotic properties (Wikberg & Mucina 2002; Ward et al. 2015). Evidence for the idea that such patches are associated with different soil communities comes from vegetation removal experiments, which show that removal of entire functional groups has major effects on belowground species composition and functioning (Ward et al. 2009). In our study sites, this increasing patch size is exemplified by the replacement of species rich grasslands dominated by short grasses and herbs by clonal growth of patch forming species such as ferns (e.g., P. aquilinum ), dwarf shrubs (e.g., C. vulgaris, V. myrtillus, E. tetralix ) and tall grasses (e.g. M. caerulea and D. cespitosa ). This change in patchiness in the vegetation after cessation of grazing may thus result in increased medium-large scale spatial heterogeneity. We hypothesize that the resulting divergence in belowground communities may start once the lack of grazing permits these clonal structures to become locally dominant. Although more rigorous experiments are needed to test the relative importance of these different possible mechanisms, we conclude that β-diversity of soil fauna and soil microbial taxa respond markedly differently to cessation of grazing and that changes in vegetation properties likely underlie these patterns.