Patch configuration affects alpine plant distribution

Abstract

The relative importance of niche requirements and dispersal limitation in controlling the landscape‐scale distribution of plants is still contentious. Local occurrence and abundance of alpine plants are commonly thought to be driven by abiotic site conditions due to pronounced environmental gradients over short distances. However, explicit tests of the additional role of dispersal‐related processes for alpine plant distribution patterns are lacking. Here, we combine niche‐based species distribution models with variables describing patch size and connectivity to evaluate if, besides abiotic limitations, spatial habitat configuration affects the occurrence and abundance of six plant species inhabiting patchy snowbed mosaics of the northeastern Calcareous Alps in Austria. Moreover, we assess if eventual effects of spatial patch configuration are more clearly detectable when calculating connectivity based on parameterized mechanistic dispersal kernels for both wind and animal vectors instead of using nearest neighbour metrics. We show that patch size and connectivity are significantly correlated to the occurrence of all and to the abundance of four out of six study species, although the relative importance of these variables, as compared to niche constraints, varies among species. In addition, connectivity measures derived from parameterized dispersal kernels were more closely related to occupancy, and in particular to abundance patterns than a simple nearest neighbour metric. The fitted kernels also suggest that dispersal by alpine chamois plays an important role for inter‐patch seed exchange. We conclude that, despite evident abiotic limitations, recurrent local extinctions and delayed re‐colonizations indeed play a role for the distribution of our study species, and that alpine plants may hence be less in equilibrium with their abiotic environment than commonly thought. Moreover, the relatively high long‐distance dispersal probabilities of animal kernels indicate that the ability of alpine plants to adapt their ranges to a rapidly warming climate may, among other factors, depend on the availability of dispersal services by large mammals.

Experimental Evaluation of Seed Limitation in Alpine Snowbed Plants

“…The relative roles of site and seed limitation in driving the distribution of these four species hence obviously differ with seed limitation being evident in case of A. clusiana and limitation by site conditions clearly important for A. caerulea . This interpretation is consistent with a recent analysis of patch occupancy patterns of six snowbed plant species (the four included in this study and two additional ones) in the same study area: A. clusiana is the most common and A. caerulea the rarest species, most probably because A. clusiana is least, and A. caerulea most sensitive to variation in abiotic conditions among the individual patches [15]. Morevover, measurements of snow melt dates by means of temperature loggers buried directly at 30 of our 55 plots (loggers at the remaining plots unfortunately failed during the experiment) also suggested that the incidence of A. clusiana (and G. hoppeanum ) at or around these plots is insensitive to melt-out dates, whereas A. caerulea (and A. atrata ) prefer latest melting sites.…”

A simple post-hoc method to add spatial context to predictive species distribution models

“…There is reduced seed rain in small fragments (Hobbs and Yates, 2003), but it is worth noting that the ability for species to persist in small patches is determined by its mortality rate and colonisation capabilities, which is more than simply dispersal. Factors such as fecundity, dormancy, seedling establishment characteristics, species interactions, and habitat quality also influence colonisation ability (Fahrig and Merriam, 1994;Levin et al, 2003;Levine and Murrell, 2003;Franken and Hik, 2004;Guisan and Thuiller, 2005;Dullinger et al, 2011). We therefore predict that species that are better modelled using larger radii will have traits such as low fecundity, poor dispersal, or low probability of establishment, which is assessed below for the three categories of species we studied.…”

“…Landscape ecology and island biogeography models were once based on the binary model of favourable or unfavourable habitat, but researchers have now realised that variations in the habitat quality of patches can have a profound influence on species' persistence as well as extinction and colonisation rates (Franken and Hik, 2004;Schooley and Branch, 2007;McAlpine et al, 2008). While these studies now consider differences in habitat quality between different patches, they are still limited to situations where the habitats are distinct, well-defined habitats such as snowbeds, wetlands or talus patches (Franken and Hik, 2004;Schooley and Branch, 2007;Dullinger et al, 2011), and they are based on the premise that these patches are relatively homogenous in habitat quality.…”

Structural uncertainty in models projecting the consequences of habitat loss and fragmentation on biodiversity

“…For example, to date several studies have incorporated patch and landscape characteristics into SDM. Dullinger et al (), Betts et al (), and McCune () combine patch (e.g. area) and landscape (e.g.…”