E-Thesis 324 views 69 downloads
Developing a Habitat Suitability Model for Welsh Lesser Horseshoe Bats / LAWRENCE NEAL
Swansea University Author: LAWRENCE NEAL
Abstract
Lesser horseshoe bats (Rhinolophus hipposideros) are protected in the UK under the Wildlife and Countryside Act 1981, having a threatened status. Habitat Suitability Models (HSM) offer a practical way to determine species-specific predictions on potential roost sites for bats, aiding in the protecti...
| Published: |
Swansea, Wales, UK
2023
|
|---|---|
| Institution: | Swansea University |
| Degree level: | Master of Research |
| Degree name: | MRes |
| Supervisor: | Bull, J. and Harris, W. |
| URI: | https://cronfa.swan.ac.uk/Record/cronfa64627 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Abstract: |
Lesser horseshoe bats (Rhinolophus hipposideros) are protected in the UK under the Wildlife and Countryside Act 1981, having a threatened status. Habitat Suitability Models (HSM) offer a practical way to determine species-specific predictions on potential roost sites for bats, aiding in the protection and conservation of threatened species. Fuller, Shewring & Caryl (2018) present a novel HSM method for identifying roost sites for R. hipposideros in Wales, UK. This study aims to test the hypothesis that national-scale models are not appropriate for use in making accurate predictions at local levels, by recreating their HSM within Gower AONB, a region of Wales ~1% the size of the whole of Wales. The difference in environmental variables for two pseudo-absence methods (random and building) across Wales and Gower AONB were assessed, and the accuracy of both was investigated using known bat roost presences, provided by the Bat Conservation Trust. Additionally, a third ensemble model was assembled from both pseudo-absence methods and assessed. Sites within Gower AONB with high bat roost presence probability were then identified. This studies’ assessments generally supported the stated hypothesis, with Gower AONB having significantly different environmental structure to the whole of Wales, and despite predictive performance being ‘fair’ for both pseudo-absence methods (0.782 for building pseudo-absences and 0.787 for random pseudo-absences) and the ensemble model (0.700), accuracy was low throughout (known bat presences that should have probabilities of 1.00 instead had probabilities of 0.548 [building pseudo-absences, 0.57 [random pseudo-absences], and 0.571 [ensemble model]). Although Fuller, Shewring & Caryl’s HSM have practical use in determining likely roost sites of R. hipposideros across Wales, their use across Gower AONB is diminished by the variable environment of the regions being investigated. These findings suggest that applying large-scale HSM’s to smaller-scale regions is not effective for identifying potential roost sites for R. hipposideros and may potentially serve as a cautionary case study for other species-specific HSM’s. However, given that the random pseudo-absence model had ‘fair’ predictive performance, and was able to correctly predict bat roost presences more than half the time, it was used to provide three sites within Gower AONB of highest probability. These three sites would be used to feed into R. hipposideros conservation by providing them to the Bat Conservation Trust for observation. Should these sites prove to include roost sites for R. hipposideros, it would suggest the uses of such HSM’s across smaller-scale regions still have some merit for other species with large-scale HSM’s available. |
|---|---|
| Keywords: |
Rhinolophus hipposideros, Lesser Horseshoe Bat, Habitat Suitability Model, Recreation, Small-scale, Gower, Wales, UK |
| College: |
Faculty of Science and Engineering |