Faculty of Science;
School of Animal, Plant and Environmental Sciences;
MSC Research Report / Three arachnid species, the rock scorpion Hadogenes gunningi, the burrowing scorpion
Opistophthalmus pugnax and the baboon spider Harpactira hamiltoni have been identified as species of conservation interest for inclusion in a bioregional systematic conservation planning project by the provincial conservation authority in Gauteng province, South Africa. The systematic conservation planning procedure requires information on the spatial distribution and an estimate of the minimum
viable area (MVA) required to support a population for species of conservation interest. The purpose of this report is to provide this information for these three arachnid species. 47 sites were sampled on
a regular grid across Gauteng province where data were collected for habitat distribution modeling and density estimation for MVA calculation. Sites were sampled by two field workers. Distance
sampling methodology was used for the estimation of density and the genetic algorithm for rule set production (GARP) was used for habitat distribution modeling. Analysis of distance data comprised
fitting several alternative models to both continuous and interval data, and data for each field worker were analyzed both separately and pooled. To calculate MVA from estimates of population density a
minimum viable population size of 2000 adult individuals was assumed. Based on composite models fitted to continuous data collected by both field workers mean MVA for Opistophthalmus pugnax was 431.57 ha (279.44 ha to 666 ha, 95% confidence interval), while that for Harpactira hamiltoni was 909.09 ha (518.00 ha to 1594.90 ha, 95% confidence interval). Insufficient data were collected for the
estimation of population density for Hadogenes gunningi, but based on encounter rate relative to the other two species a subjective estimate of MVA between 380 ha and 570 ha is presented. Habitat distribution modeling was conducted at two grains of predictor variable data. As GARP produces highly variable results models were selected according to the criteria of having less than 5% omission and less than 10% non-prediction. Selected models were stacked and predictions of presence and
absence summed for each map pixel across all models. The resultant maps of proportion of positive predictions per pixel were multiplied to obtain a final composite map of probability of occurrence.
Accuracy of the coarse, fine and composite maps was assessed using receiver operating
characteristic analysis. Mean AUC for models for Hadogenes gunningi were 0.893, 0.857 and 0.886.
For Opistophthalmus pugnax and Harpactira hamiltoni mean AUC values were 0.831, 0.790, 0.856 and 0.783, 0.765, 0.805 respectively. Probabilities of occurrence were converted to presence absence
at the threshold where false positive and false negative prediction rates were equivalent. Hadogenes gunningi is predicted to occur on most ridges within the province, while Opistophthalmus pugnax and
H. hamiltoni are predicted to have a patchy distribution in the southern two thirds of the province. The results presented are a significant improvement on the data previously available for these species and
it is recommended that their conservation status be revised in light of the results. Concerns regarding the utility of GARP in conservation planning and suggestions for further research are outlined.
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:wits/oai:wiredspace.wits.ac.za:10539/1498 |
| Date | 27 October 2006 |
| Creators | Engelbrecht, Ian |
| Source Sets | South African National ETD Portal |
| Language | English |
| Detected Language | English |
| Type | Thesis |
| Format | 1084603 bytes, application/pdf, application/pdf |
Page generated in 0.0022 seconds