Discussion
We report for the first time that frequency of grazing has effects on the phylogenetic community structure that are not captured by standard diversity indexes. Plant communities grazed every other year (i.e. pulse grazing), but not communities grazed every year (i.e. press grazing) were more phylogenetically dispersed than lightly grazed communities, while there was no effect of grazing on species richness, the Shannon-Wiener index (H’) and community evenness (Evar). This does not support our predictions that grazing would act as a biotic filter, resulting in communities that are more phylogenetically clustered, but rather suggests that pulses of grazing release biological control and allow more phylogenetically diverse communities.
We observed no effect of grazing on species richness, the Shannon-Wiener index (H’) and community evenness (Evar). Since the productivity of our study site is intermediate, these results are in agreement with previous studies, which show that excluding reindeer decreases species richness in sites with low productivity, but increases species richness in productive sites (Proulx and Mazumder 1998, Bernes et al. 2015, Sundqvist et al. 2019). A neutral effect of reindeer grazing on species diversity is expected for sites that have intermediate productivity in this region (Sundqvist et al. 2019).
While we do not observe a change in the overall richness and evenness of vascular plants across grazing regimes, we do detect an effect of grazing on the species composition of communities. Plots with pulse and press grazing both differed in community structure from lightly grazed plots (Figure 2a) as there was a transition from shrubs to graminoids that occurred with grazing (Figure 3). These results are fully consistent with previous studies in Fennoscandia that report the replacement of dwarf shrubs by graminoids in grazed areas (Olofsson et al. 2001). Similar effects of grazing on vegetation composition have also been observed in other systems (McKendrick et al. 1980, Clarke et al. 1995, Rooney 2009, Begley-Miller et al. 2014, Ferreira et al. 2020). For example, grazing by sheep in Great Britain has resulted in the replacement of heather moorlands with graminoids (Clarke et al. 1995), and grazing by mammalian herbivores in northern Alaska also resulted in the replacement of tundra heaths by graminoids (McKendrick et al. 1980). Potential reasons for this shift in vegetation composition include increased soil nutrient concentrations that favour graminoids (McKendrick et al. 1980) and an increased ability of graminoids to outcompete shrubs and ferns in heavily grazed environments due to their short stature, high shoot densities, and capacity for compensatory growth (Coughenour 1985). This combination of traits allows graminoids to be the first plants to colonize following disturbances (Chapin and Shaver 1981). Interestingly, the shift from shrubs to graminoids was also observed in plots with several years of pulse grazing (Figure 2a; Figure 3). This may be explained by the decades of chronic grazing these plots were exposed to before the temporary fences were established.
Our novel analysis of the effect of reindeer grazing on the phylogenetic structure of vascular plant communities revealed that communities with pulse, but not press grazing seem more phylogenetically dispersed than lightly grazed communities (Figure 4). This is in contrast to our prediction that, assuming anti-herbivore traits are evolutionarily conserved (Loiola et al. 2012, Yessoufou et al. 2013, but see Kursar et al. 2009) and reindeer are generalist herbivores (Baskin and Danell 2003), reindeer grazing would result in communities that are more phylogenetically clustered compared to lightly-grazed communities (Cavender-Bares et al. 2009). Other studies examining the effect of grazing on phylogenetic structure have found conflicting results. Grazing by white-tailed deer (Odocoileus virginianus ) resulted in phylogenetic clumping (Begley-Miller et al. 2014), but grazing by large herbivores in the African savanna resulted in changes in phylogenetic community structure dependent on the initial community structure: communities that were initially clumped became more dispersed while communities that were initially dispersed became more clumped (Yessoufou et al. 2013). In contrast, grazing by livestock had no effect on the phylogenetic dispersion of plant communities in Chile (Salgado-Luarte et al. 2019). One potential explanation for our result is that anti-herbivore defense traits are actually evolutionarily convergent in our study region, a situation where grazing by a generalist herbivore would result in phylogenetic clumping (Cavender-Bares et al. 2009). However, it should be noted that this predicted effect of grazing on the phylogenetic structure of plant communities by Cavender-Bares et al. (2009) is based only on the direct effects of herbivory (i.e. the physical removal of species) and does not account for the indirect effects on e.g. productivity, nutrient availability, trampling disturbance and competitive interactions between species. For example, if herbivory increases soil nutrient availability and decreases competition between plant species, this might lead to phylogenetic clumping, given that both competition between distantly related taxa (Mayfield and Levine 2010) and limited nutrient availability (Hurteau et al. 2016) may drive phylogenetic clumping. In this case, the indirect effects of grazing by herbivores on phylogenetic structure may act in the opposite direction of direct effects of grazing, and the resulting phylogenetic structure may depend on the relative strength of the direct and indirect effects. Given that we found that pulse, but not press, grazing results in increased phylogenetic dispersion, it is possible that in our study system, the indirect effects of reindeer grazing are more important than the direct effects in driving community assembly for plant communities that are exposed to acute periods of grazing after decades of chronic grazing.
The results of this study may be relevant to management decisions involving reindeer in Fennoscandia. While we do not observe any richness effects of reindeer grazing, we do observe a significant effect of grazing on vegetation composition and phylogenetic structure. More specifically, several years of pulse grazing (following over forty years of chronic grazing) increased the phylogenetic dispersion of vascular plant communities compared to plots with almost no grazing and plots with continued press grazing. Changes to the phylogenetic structure of communities may have important consequences for community function if phylogenetic diversity captures genetic and functional diversity related to ecosystem productivity (see, e.g. Cadotte et al. 2009, Flynn et al. 2011). Given the numerous studies that have analyzed the impact of reindeer herbivory on vegetation composition, there is a potential to re-analyze existing datasets using the framework of phylogenetic community ecology as done here. Additionally, future studies considering variation in productivity and grazing intensity will reveal if general relationship between grazing and phylogenetic diversity exist in the same way they do for common diversity measures, or whether the responses depend on other factors such as the evolutionary history of plants and herbivores. Direct measurement of ecosystem function in such experimental communities are also needed to help understand the mechanisms driving plant community assembly in support of future management decisions.
Declarations
Ackowledgements: We thank Lauri Oksanen for knowledge of the study area and design and Oleg Dmytrenko, Jonas Gustafsson, Pirjo Isotupa, Marisano James, Maciek Koperwas, Nicolai Krichevsky, Anne Muola, Karin Nilsson, Antti Oksanen, Lotta Ström and Gunnar Öhlund for help in the field. We also thank James Bodnarchuk and Sara Durkin for help with the phylogenetic analyses.
Funding: Funding for this study was provided by the Swedish Research Council (MJM grant number 637-2013-274) and NSERC Canada (AOM).
Data accessibility: Data will be submitted to Dryad upon acceptance.
Competing interests: The authors declare no competing interests.