Introduction
Herbivory is an important factor influencing plant community assemblages directly through the physical removal of plant species (Lubchenco 1978), and indirectly by altering patterns of ecosystem productivity (McNaughton et al. 1988) and nutrient availability (Mazumder et al. 1988). As a result, herbivores may alter species composition (Augustine and McNaughton 1998), measures of richness (Olff and Ritchie 1998), and the phylogenetic structure (Cavender-Bares et al. 2009) of plant communities. Plant species richness and evenness (Tilman et al. 1997) and phylogenetic diversity (Faith 1992) are both important for ecosystem function and productivity (Liu et al. 2018). Thus, understanding the impact of reindeer grazing on plant communities is essential for directing conservation efforts in regions where reindeer are managed (Olofsson et al. 2004).
Herbivore activity can both increase and decrease the diversity of plant species (Olff and Ritchie 1998, Proulx and Mazumder 1998). According to the Intermediate Disturbance Hypothesis plant communities will increase or decrease their biodiversity as a function of the level of disturbance (Grime 1973, Connell 1978). Low disturbance levels (e.g., grazing) may increase plant species richness by removing dominant and competitive species, increasing light exposure to soil and increasing nutrient availability (Olff and Ritchie 1998, Proulx and Mazumder 1998, Bakker et al. 2003). High levels of disturbance/grazing, however, may decrease species richness due to insufficient recovery periods, trampling and erosion (Olff and Ritchie 1998). The nutrient availability of an ecosystem may play a role in determining the grazing intensity that results in the greatest biodiversity overall (Proulx and Mazumder 1998). If nutrient-rich ecosystems are more likely to have a few dominant species that respond quickly to disturbance (Rosenzweig 1971, Huston 1979), nutrient-rich communities may experience peak levels of biodiversity at greater grazing intensities. Nutrient-poor ecosystems, however, which are common in the North, are limited by their regrowth ability and thus are expected to achieve the greatest species richness at lower grazing intensities (Proulx and Mazumder 1998, Sundqvist et al 2019). Thus, experimentally increased grazing is more likely to result in increased species richness in nutrient-rich ecosystems but have no effect or result in decreased diversity in nutrient-poor ecosystems.
Herbivores may also alter the phylogenetic structure of communities, though predictions are complex (Cavender-Bares et al. 2009). If anti-herbivore defense traits have a significant phylogenetic signal (Loiola et al. 2012, Yessoufou et al. 2013, but see Kursar et al. 2009), then generalist herbivores may act as a biological filter (Begley-Miller et al. 2014), resulting in a community that is more phylogenetically clumped i.e. communities including more closely-related species than drawn from the regional pool. However, if anti-herbivore defense traits are evolutionarily convergent, a generalist herbivore may increase the phylogenetic dispersion (communities including more distantly related species) of a community. In contrast, when multiple specialist herbivores are present and defense traits are highly conserved within plant clades, the community may become phylogenetically clumped. However, if defense traits are evolutionarily convergent then seemingly random patterns of species relatedness may result (Cavender-Bares et al. 2009). Alternatively, if herbivores both (1) decrease competition via removal of dominant plant species from a community and (2) increase nutrient availability via changes in nutrient cycling, herbivory may result in phylogenetic dispersion as intense competition between distantly related taxa (Mayfield and Levine 2010) and limited nutrient availability (Hurteau et al. 2016) both drive phylogenetic clumping.
The most extensive form of human land use in the northern Fennoscandian tundra is grazing by reindeer (Rangifer tarandus ), and the major populations of reindeer across different regions have increased, decreased or remained stable over recent decades (Uboni et al. 2016). As the tundra is considered a low productivity environment with low nutrient availability, we expect low levels of reindeer grazing to have no effect or a negative effect on plant species diversity. The effects of reindeer grazing on tundra plant communities are significant (Suominen and Olofsson 2000, Austrheim and Eriksson 2001), as demonstrated by the pronounced replacement of dwarf shrubs by graminoids in grazed areas (Olofsson et al. 2001, Sundqvist et al 2019). However, the direction and size of the effect on other measures of biodiversity varies (Bernes et al. 2015, Sundqvist et al 2019): plant species richness generally increases with reindeer grazing in areas dominated by moss-like lichens, i.e. Cladonia species (Suominen and Olofsson 2000, Austrheim and Eriksson 2001), while communities lacking lichen mats experience no effects (Olofsson and Oksanen 2005), weak effects (Suominen and Olofsson 2000), or mixed effects that vary among regions (Olofsson et al. 2001). Additionally, a Scandinavia-wide study found that reindeer grazing decreased species richness in sites with low productivity, but increased species richness in productive sites (Sundqvist et al. 2019). Although many of the effects of reindeer grazing on biodiversity are well-studied, they are still not well-understood, and we have no data on the impact of reindeer grazing on phylogenetic community structure. Understanding the impact of reindeer on phylogenetic structure may provide insight into mechanisms driving community assembly (Webb et al. 2002) in this region and help inform good management practise.
Here we study the effects of reindeer grazing on plant community structure using data from a multi-year experiment including varying intensities of grazing: light (almost never grazed), acute i.e. pulse (grazed every other year), and chronic i.e. press (grazed every year), in the region of Raisduoddar in northern Norway. First, we analyzed the effect of grazing on the diversity of vascular plant species. Based on previous work in Fennoscandia, we predicted that in the nutrient-poor tundra, both pulse and press grazing would have no effect or negative effects on species richness and evenness. Second, we analyzed the effect of grazing on plant species composition with the prediction that the pulse and press grazed areas would be composed of different species from the lightly grazed areas, specifically with the replacement of dwarf shrubs by graminoids (Olofsson et al. 2001). Finally, we analyzed the effect of grazing on the phylogenetic structure of the vascular plant communities. As reindeer are generalist herbivores (Baskin and Danell 2003) and anti-herbivore traits are generally evolutionarily conserved (Loiola et al. 2012, Yessoufou et al. 2013, but see Kursar et al. 2009), we predicted that grazing would result in appreciable phylogenetic clustering such that species in areas with pulse and press grazing would be more closely related to one another than those in lightly grazed areas.