Multiscale patterns of mammal diversity and occurrence in the Karoo

Woodgate, Zoë Anne

Multiscale patterns of mammal diversity and occurrence in the Karoo

Doctoral Thesis

2022

Abstract

The transformation of natural habitat for urban, industrial, and agricultural activities is the leading driver of terrestrial biodiversity loss. Yet how such land–use changes will impact the global drylands is poorly understood, despite the vulnerability of these once isolated regions. One such region is South Africa's Karoo, which is characterised by low human density, high levels of endemism and extensive pastoralism. Whilst centuries of small–livestock farming have severely impacted the Karoo's biophysical environment it remains relatively intact from an ecological perspective. However, novel threats have emerged, including a proposal for the extraction of natural gas by hydraulic fracturing extensive shale reservoirs. A major impediment to understanding how this, and other land–use changes, may impact indigenous wildlife is the lack of updated, multiscale foundational biodiversity data. To address this knowledge gap, my thesis evaluates the distribution and ecology of the mammalian (>0.5kg) community across the Karoo, with a focus on developing potential methods that can effectively record both common and rare species. My primary aim was to understand how biophysical drivers and intraguild interactions have shaped the distribution of mammals at different spatial scales, relative to existing and proposed land–uses. I had three main objectives: 1) to provide a foundational understanding of the mammalian community present throughout a 171 811km2 area under consideration for shale gas extraction, 2) to compare species occurrence and diversity across different land–use types (i.e., farmland, protected area [PA] and private protected areas [PPA]), and 3) to understand the drivers of the Karoo's rarest mammal species, namely the critically endangered riverine rabbit (Bunolagus monticularis), presence. To accomplish these objectives, I deployed three camera trap surveys at different geographic extents, with differing array designs. At the broadest geographical extent, I utilised an array consisting of 25 sites (125 camera traps total) that were selected using the Latin hypercube method. At the intermediate extent I compared three different land–uses, each sampled using a standardised 2km2 grid of sites (451 camera traps total). Lastly, at the finest scale (223.24km2 ), I deployed a stratified random design (30 clusters of 5 camera traps deployed within 400m) with the explicit goal of detecting riverine rabbits. I found that at the broadest scale wildlife diversity and occupancy was largely influenced by landscape–level abiotic processes, namely longitude and precipitation. None of the wide–ranging large predator (e.g., leopard or brown hyena) or endangered species (i.e., riverine rabbit) known to occur in the region were detected, suggesting the survey design was too coarse, and that protected areas were underrepresented. At the intermediate extent I used a hierarchical multi–species single–season occupancy model that incorporates species–specific responses to management treatments to show that land–use has a significant impact on species richness and occupancy, both of which were highest in the PPA and lowest in commercial small livestock farms. It was only at the finest sampling scale, with numerous camera traps arranged in discrete clusters at independent sites, that I detected all of the rare species present in the region, including the targeted riverine rabbit. Using a multispecies occupancy model that accounts for ≥2 interacting species, I showed that riverine rabbit occurrence was conditional on the absence of close competitors– namely scrub and cape hares– and smooth terrain. These results will better align ongoing conservation efforts for the species, which currently focus exclusively on environmental restoration of assumed preferred habitat. Overall, my findings reveal a diverse community of mostly medium sized generalist species that have persisted throughout the Karoo despite hundreds of years of sustained human impact. Larger predators and herbivores were almost entirely restricted to protected areas, and the PPA provided the only refuge for the critically endangered riverine rabbit. Whilst the predicted impacts of climate change and extractive mining on the mammals of the Karoo remain poorly understood, this thesis provides an important baseline of the extant mammal communities across the three dominant land–uses. Long term monitoring of select sites would provide a better understanding of how local and global anthropogenic impacts may affect the future of mammals in the drylands of South Africa.

Keywords

Biological Sciences

Reference:

Collections

PhD / Doctoral

Full item page