Joris P. G. M. Cromsigt

RESOURCE PARTITIONING AMONG SAVANNA GRAZERS MEDIATED BY LOCAL HETEROGENEITY: AN EXPERIMENTAL APPROACH

Recent theoretical studies predict that body size-related interspecific differences in spatial scale of perception and resource use may contribute to coexistence of species that compete for the same class of resources. These studies provide a new theoretical framework for explaining resource partitioning patterns among African ungulates that coexist in spatially heterogeneous savanna grasslands. According to these studies, different-sized ungulates can coexist because larger species forage at a coarser scale but can tolerate lower quality food, whereas smaller species need higher quality food but forage at a finer scale. To test this hypothesis in an African savanna, we created an experimental mosaic with variation in grain (spatial detail) and quality of short-grass patches and directly observed the visitation of naturally occurring grazers to this mosaic over a two-year period (total of 903 observation hours). Of the seven species that visited our experiment, warthog, impala, zebra, and white rhino visited long enough to allow data analysis. We showed that warthog and impala avoided plots with a finer grain of short grass and that warthog preferred fertilized plots to unfertilized plots. Zebra and white rhino did not avoid the finer grain plots. Our results suggest that differences in grain and quality of a resource might indeed contribute to partitioning of this resource by savanna ungulates. Although four focal species is unusually high for an experimental study on resource partitioning among naturally occurring savanna ungulates, this number is too low to evaluate the allometric basis of our hypothesis. Our results, however, encourage wider experimental testing of the role of spatial heterogeneity in facilitating the coexistence of potentially competing savanna herbivores.

Using models in the management of Black Rhino populations

A study was undertaken, modeling Black rhino, Diceros bicornis, populations from Hluhluwe-Umfolozi Game Reserve and Mkuzi Game Reserve, Republic of South Africa. This study had two specific goals. Firstly, we tried to find a simple model to predict the population number in the following year and give insight in the number of animals that can be removed from the population. Secondly, the use of different models gave us the opportunity to evaluate the effect of small differences in model structure on the outcome of the models. Five different models were tested, mainly focusing on the structure of density dependence. The density dependent model, that incorporated stronger density dependence at high densities, fitted best to the observed census values. The study, furthermore, indicated that the population in Hluhluwe-Umfolozi is stabilizing. The situation in Mkuzi was not that clear. Based on these results, managers should therefore be more careful with removing animals from this population. Finally the results clearly showed that different model structures can result in very different model outcomes. Besides in the values of the model parameters, this was also shown by different values for the population number where the yearly increase is maximal. Based on these results it does not seem wise to base management decisions solely on the outcome of one model. Therefore, we strongly encourage managers to compare the outcome from different model structures, based on different biological assumptions.