Spatial Conservation Status Modelling of the Great Crested Newt in South Wales
Authors: Fletcher, D.H., Arnell, A., French, G.C.A., Wilkinson, J.W. and Jones, G.
Pages: 1-35
Publisher: Natural Resources Wales
Place of Publication: Bangor
Abstract:This study builds on previous work, in North Wales and Powys, on the spatial modelling of the great crested newt. This research utilises high resolution (25m) modelling to develop maps and metrics for great crested newts in 12 South Wales Unitary Authorities (Blaenau Gwent, Bridgend, Caerphilly, Cardiff, Merthyr Tydfil, Monmouthshire, Neath Port Talbot, Newport, Rhondda Cynon Taff, Swansea, The Vale of Glamorgan and Torfaen). Project aims include assessing: i) known distribution, ii) suitable habitat, iii) survey target areas iv) important populations for metapopulation connectivity, v) pond/habitat creation areas and vi) Local Population Targets (LPTs) for each unitary authority.
We compiled high resolution, recent (1990-2011) great crested newt records and, after cleaning, used 149 to represent known populations. We used the distribution modelling software, MaxEnt, along with 24 environmental layers and a subset of 108 spatially filtered records, to produce an estimated probability of presence output.
From this output, and known barriers, we derived a resistance surface for calculating Least Cost Path distances and corridors between known populations. We ran graph theory software, Conefor, with these distances to produce probability of connectivity indices (dPC) to highlight important populations for maintaining metapopulation connectivity. To delimit pond creation target areas, we selected low pond density areas within modelled suitable habitat and within an appropriate Unitary Development Plan (UDP) designation. We set local population target figures to mitigate putative historical pond losses in each unitary authority, using a reference date of 1843, local pond loss rates and current pond data. We apportioned these targets to appropriate pond creation areas within cost weighted buffers surrounding known populations, or within corridors between them.
MaxEnt model testing produced acceptably high values (>0.7) for the area under the receiver operating characteristic (ROC) curve. Due to limited record numbers in individual Unitary Authorities, we built and tested the MaxEnt model with records from all 12 Unitary Authorities The recommended Local Population Targets for pond creation produced realistic and achievable figures (<10 ponds per square km), when apportioned to appropriate target areas, for most Local Authorities. However, some LAs have only negligible areas available indicating the need for regional spatial planning in order to enhance and maintain great crested newt status in South Wales.
Source: Manual