Examining spatio-temporal patterns of seed dispersal by a terrestrial non-obligate frugivore

Seed dispersal is a key ecological process for plant regeneration. Being sessile organisms, plants rely on vectors (e.g. wind, animals) to disperse their seeds. In the tropics and subtropics, fruit-eating animals are vital dispersers for many plants, and the same plant species is often dispersed by several animal species. However, based on differences in body size, diet, and behaviour, frugivores differ in the quantities of fruit they consume and the spatial locations to which they carry seeds within a landscape. Because seed dispersal lays the template for germination and seedling survival, variation in spatial patterns of dispersal arising from different dispersers can have profound consequences for plant regeneration. Thus, assessing the contribution of dissimilar dispersers to spatial patterns of seed dispersal would provide a clearer understanding of the ecological mechanisms underlying plant community dynamics. Our knowledge on several seed dispersers remain rudimentary. For example, large herbivores (> 2 kg) are found on nearly all continents and their diets include fruits in various amounts. Despite them being some of the largest consumers of fruit, studies on frugivory and seed dispersal by large herbivores are relatively sparse and the natural history and ecological understanding of their role in seed dispersal remains understudied. It has been suggested that large herbivores crucially assisted plant migration during past climatic changes. However, without detailed knowledge of their role as dispersers it is difficult to predict their contribution to shaping future plant communities, especially in the context of global land-use and climate change. In my thesis, I assessed seed dispersal services provided by large herbivores. Specifically, I examined how the behaviour of a forest ruminant in response to food resources and predation risk influences the spatial patterns of seed rain they generate. I used Chital (Axis axis) a deer native to the Indian sub-continent, and tree species including Terminalia bellerica and Ziziphus mauritiana as a model system. I carried out the fieldwork in Rajaji National Park, a sub-tropical dry deciduous forest in Uttarakhand, India. First, we reviewed information on large herbivores of Asia comprising of around 80 species. We compiled nearly 300 records of frugivory and seed dispersal from 27 large herbivore species for which diet information was available, but the data was rarely quantitative. The relationship between traits of fruits consumed by the herbivores and the traits of herbivores was examined. Larger-bodied herbivores consumed larger fruits on average (as measured by the maximum width of fruits), suggesting that body size affects patterns of frugivory by herbivores. Further, with the exception of ruminants (Bovids and Cervids) most herbivore families differed from one another in the type of fruits (e.g. size, colour) they consumed. Thus, large herbivores vary in their contribution to seed dispersal and may be more important to seed dispersal than previously though. Next, I assessed how movement patterns of a medium-sized ruminant, Chital (Axis axis, 50 kg), influenced spatial patterns of seed rain for four tropical dry deciduous trees. Specifically, I examined how heterogeneity in topography and habitat influenced Chital movement patterns and consequently seed dispersal. I deployed GPS collars on seven Chital individuals. Chital movement was characterized as being short-range (small movements with rapid changes in direction) or long-range (moving large distances with little change in direction). Short-range movement was greater in areas that had an even mix of habitat types (providing food and cover) but less in rugged terrain. In comparison, long-range movements occurred in habitats with less cover. Notably, Chital space use and movement was not influenced by fruit availability but by factors that likely reflect predator avoidance. Further, seeds were more likely to be dispersed to areas with more Chital activity. Therefore, prey-predator interactions, more than spatial location of fruit trees, might affect spatial patterns of seed rain generated by medium-sized ruminants. I then developed a novel analytical framework to characterize seed rain patterns of Terminalia bellerica generated by Chital. Specifically, using occupancy models I accounted for imperfect detection of seeds when sampling the forest floor. Terminalia bellerica seeds were not always detected even when present and detection varied among time periods. Seeds were less likely to be detected among taller ground vegetation. Seed rain patterns were best explained by a combination of Chital dung piles (indicating frequency of use) and the distance to the edge of closed canopy forest (proxy for flight distance). Temporal changes in seed rain patterns were predicted by a combination of grass and structural cover available in the plots. Our method is compatible with currently used field protocols for measuring seed rain, especially for terrestrial dispersers and allows more robust statistical inferences by explicitly modelling both the observation and ecological processes underlying spatial patterns of seed arrival. Using the occupancy framework, I next examined seed rain patterns generated by Chital and contrasted them with other dispersers in the system namely pigs and birds. For this, I collected seed rain data for two species of trees (T. bellerica and Ziziphus mauritiana). Chital and pigs dispersed similar quantities of Z. mauritiana seeds but to very different number of locations. Chital dispersed seeds to nearly thrice the number of locations compared to pigs or birds. Although pigs transported far more seeds than birds, both dispersed seeds to similar number of locations. Further, seed rain by Chital varied less through time than pigs and birds. While the spatial distribution of grass and structural cover best explained seed rain patterns of Chital, fruiting trees predicted patterns generated by pigs and birds. Chital generate seed rain patterns that are very different from and likely more effective than pigs and birds. Finally, I used individual-based simulation models to test whether seed rain patterns generated by Chital are driven by predator-prey interactions. Using my own data and that from other studies, I simulated four different scenarios where Chital behaved as grazers or frugivores, both in the presence or absence of perceived predation risk. I compared seed rain data from simulations against observed data. From simulations, I found that grazing behaviour in the presence of predation risk explained most variation in the observed data. Spatial concordance was also highest between observed data and simulations from grazing behaviour in the presence of predation risk. For a non-obligate frugivore such a Chital, predation risk is likely to strongly influence seed rain patterns. Overall, my results show that the movement patterns of Chital and the seed rain they generate are influenced by habitat features that mediate predator avoidance behaviour and non-fruit resources such as grass and cover. Unlike highly frugivorous birds, non-obligate frugivores like ruminants likely 'decouple' seed rain from fruit tree distribution and perceived predation risk may be a key mechanism driving this process. The ecological insights combined with the methodological advances of my study will assist in bridging the research gap in understanding ecological differences among seed dispersers, particularly non-obligate terrestrial frugivores such as ruminant. The strong possibility for prey-predator interactions to influence the outcomes of seed dispersal by Chital is a novel and timely insight. My results have implications for ongoing global declines in large herbivores and top predators which could potentially alter seed dispersal mechanisms