All analysis was undertaken in Great Britain and associated islands over 20 km2. All prioritisations were undertaken at a 10 × 10 km landscape-scale on cells with greater than half land coverage. We considered designations ‘protected for biodiversity’ to be Sites of Special Scientific Interest (SSSI) and National Nature Reserves (NNR); and landscape protection designations to include National Parks (NP), Areas of Outstanding Natural Beauty (AONB), and Scottish National Scenic Areas (NSA). Different cell protection ‘cutoffs’ were tested at 30, 40, 50, 60, and 70% (Supplementary Table 1). Hence cells were considered to be ‘protected for biodiversity’ at the landscape-scale if SSSI/NNR coverage was above the percentage land cutoff, e.g. at least 40% IUCN IV protection (Fig. 1: black cells). ‘Protected landscapes’ were 10 × 10 km cells with total coverage from all of the designations above the cutoff, e.g. at least 40% IUCN V (or greater) protection, but under 40% level IV protection (Fig. 1b: grey cells). Results were qualitatively similar for all cutoffs (Supplementary Tables 2 and 3). The joint proportion of cells protected for biodiversity and protected landscapes were most similar to the actual coverage at the 40% ‘cutoff’ (27.80% of 10 × 10 km cells ‘protected’ compared to 26.71% actual area coverage), and this is presented in the main text. All designation data used is publicly available from the respective national spatial data repositories for England21 (SSSI/NNR/NP/AONB), Scotland22 (SSSI/NNR/NP/NSA), and Wales23 (SSSI/NNR/NP/AONB).
We used the recorded distributions of 445 priority species listed under the Section 41 (Natural Environment and Rural Communities Act, 2006), provided by Butterfly Conservation (BC), Biological Records Centre (BRC); and breeding bird atlas data from British Trust for Ornithology (BTO)24. BTO bird atlas data are only available at the 10 × 10 km scale, which limited the spatial resolution of the analysis. We used all priority species that we were able to acquire from the above recording bodies between 2000 and 2014 (Supplementary Data 1). We used the raw distribution records for 156 species that were very localised (10 or fewer presence records) and for a further 77 species which could not be modelled (most of which were also very rare, and for which models did not converge). For the remaining 212 species with over 10 presence records, we interpolated their range using Integrated Nested Laplace Approximations (INLA) in the inlabru R package25. We used a joint model predicting distribution while accounting for recording effort, including biologically relevant covariates: seasonality, growing degree days, water availability, winter cold26, and soil pH from the Countryside Survey 2007 dataset27. These covariates were calculated from monthly means of weather data (mean temperature, sunshine and rainfall) for the decade to 2014 provided by the Met Office28. We also included soil moisture in the calculation of water availability29. We used raw data records from all 445 species, along with broad habitat layers extracted from the Land Cover Map 201530, in a Frescalo analysis31 to estimate recorder effort. See Supplementary Methods for further details of modelling.
We carried out a spatial prioritisation using Core Area Zonation32, whereby cells are removed iteratively, first removing those that contribute the smallest cell value: the maximum proportion of species distributions within the remaining cells. In this way cells remaining longer within the solution complement species representation of other cells to a greater extent, and hence contribute most to underrepresented species’ distributions. However, priorities were constrained by masking or ‘locking in’ different relevant areas to each scenario such that all other cells must be removed first; reducing overall solution optimality but ensuring complementarity to masked areas. Scenario 1 only masked cells protected for biodiversity and didn’t consider other designations beyond that. Scenario 2 also masked cells protected for biodiversity but, corresponding to the 30by30 pledge, additionally masked protected landscapes.
We undertook a parallel analysis additionally incorporating opportunity costs calculated from agricultural land classification and urban areas33,34,35 (Supplementary Fig. 2, Supplementary Table 4). Although urban areas are often excluded from SCP analyses, it is important to consider species complementarity of all landscapes (the government 30% target applies to the entire land surface). Since some urban/near-urban areas contain nationally rare species, we include urban areas, albeit imposing the maximum opportunity cost in these cells. In this analysis, cell value was calculated as the maximum proportion of species distributions within the remaining cells divided by the mean opportunity cost of the cell (Supplementary Fig. 3, Supplementary Tables 2 and 3).
Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
Source: Ecology - nature.com
