Abstract
The Apportioning in the Non-Breeding Season (ANBS) tool is designed to apportion breeding adult common guillemot (Uria aalge) and razorbill (Alca torda) in the non-breeding season within the UK Exclusive Economic Zone. The approach uses geolocator tag information collected from both species over three field seasons between 2017 and 2020 from up to 12 colonies. Utilisation distributions were derived from the geolocators and simulations were generated based on locational uncertainty, producing 100 simulations for each species, colony, and month. The tool propagates locational uncertainty through the analysis to give mean estimates of apportioning along with estimates of uncertainty. The tool uses populations defined in BDMPS, including UK SPAs, UK non-SPAs, and non-UK non-SPAs (Furness, 2015). The tool can be used in map mode, in which colonies of interest are selected and mean estimates of apportioning are mapped along with 95% CIs. Alternatively, footprint mode allows users to upload an offshore renewable energy footprint and mean estimates of apportioning with 95% CIs are calculated based on the spatial extent of the footprint. User-defined inputs allow flexibility in how the ANBS tool is run including the ability to choose the species, months, and season, although default values are provided to aid users in their choices. Populations from the BDMPS are provided along with the number of breeding pairs of adults, which can be changed by users. A hierarchical approach to calculating apportioning is used, based on whether colonies are data rich (using utilisation distributions) or data poor (where utilisation distributions are unavailable). A user-defined distance threshold (with a default value of 270 km) determines how colonies with geolocator data are linked to all other UK SPA colonies. Colonies within the distance threshold are assigned a weighted proportion of the utilisation distributions from colonies with geolocator data, whereas colonies outside of the distance threshold use apportioning from BDMPS.
When in footprint mode, usage is aggregated over the footprint by colony and simulation. Mean, lower, and upper 95% CIs are calculated for each colony and scaled to population level using the number of breeding pairs of adults multiplied by 2. For colonies relying on BDMPS, usage is calculated as the proportion of time adult birds spend in the specified BDMPS region multiplied by the footprint area divided by the BDMPS region area. To scale to population level, each colony is multiplied by the number of pairs of adult breeding birds and multiplied again by 2 to get to the total mean number of adult birds. Standard deviation is estimated by multiplying the standard deviation across all UK SPA colonies by the user-specified BDMPS uncertainty scalar. An uncertainty probability distribution is constructed from colony usage and the BDMPS standard deviation. Lower and upper 95% CIs are calculated from the probability distribution and scaled to population level by is multiplying by the number of pairs of adult breeding birds and multiplied again by 2 to get to the total mean number of adult birds.
An output report is produced when running the tool, which presents a summary, user inputs, results in the form of tables and maps, caveats and limitations, and appendices showing system information.