Offshore wind developments assessment - seabird collision risk, displacement and barrier effects: study

3 Background: Displacement risk modelling

Displacement effects can be defined to be:

Displacement mortality = Baseline exposure * Displacement rate * Mortality rate for displaced birds

(Equation 1)

"Baseline exposure" represents the number of birds that are estimated to use the wind farm footprint (plus any appropriate buffer to represent the distance to which birds may be affected outside of the windfarm footprint) using data on the baseline spatial distribution of birds - which may either be at-sea survey data or GPS tracking data. The "displacement rate" represents the proportion of birds that are susceptible to displacement – i.e. the proportion of birds that will undertake displacement behaviour if they encounter a wind farm. The "mortality rate for displaced birds" represents the proportion of displaced birds that die as a result of being displaced – note that this represents the mortality rate for birds that are both exposed and susceptible to displacement. Some methods also include an impact of displacement effects on chicks via changes to the productivity rate or breeding success of adults (e.g. individual based models such as SeabORD), whilst other methods, such as the displacement matrix only consider impacts on survival.

Two approaches are currently used for the estimation of displacement effects within Scottish waters. The 'Displacement Matrix' approach involves calculating baseline exposure from site-based density estimates, and then calculating effects for pre-specified values of the "displacement rate" and "mortality rate for displaced birds". The Displacement Matrix often involves calculating displacement risk for a range of values of the latter two inputs, but we focus here upon running it for a single, 'best estimate', of each rate.

SeabORD (Searle et al. 2014, 2017) is a mechanistic individual-based model of seabird foraging, energetics, demographics and OWF interactions, providing an alternative to the Displacement Matrix approach. SeabORD takes a map of baseline spatial distribution of birds from the breeding colonies of interest, and the footprint(s) for the OWFs of interest and produces an estimate of displacement and barrier effects - the increase in mortality (as a percentage of population size) associated with displacement or barrier effects caused by the OWFs, for both breeding adults and chicks. Displacement and barrier rates are entered in to the model by the user, specified as a proportion of the breeding adult population that are displacement or barrier susceptible, and therefore impacted by OWFs should they interact with footprints during the course of the chick-rearing period. These rates are set in consultation with statutory agencies. This creates sets of birds in three categories: 1. Non-susceptible birds that are unaffected by OWFs; 2. Displacement susceptible birds that will not forage within an OWF, but will fly directly through it; and 3. Displacement and barrier susceptible birds that are both barriered by an OWF and will not forage within it. In practice, SeabORD is most often set up such that only Categories 1 (unaffected) and 3 (displacement & barrier susceptible) are used. SeabORD predicts the time/energy budgets of breeding seabirds during the chick-rearing period and translates these into projections of adult annual survival and productivity. The model simulates foraging decisions of individual seabirds under the assumption that they are acting in accordance with optimal foraging theory. Each individual selects a suitable location for feeding during each foraging trip from the colony based on bird density maps derived from a range of methods, and the subsequent behaviour of birds is then simulated, incorporating realistic assumptions and constraints derived from observed behaviour. Fundamentally, the model assumes that the foraging behaviour of individual seabirds is driven by prey availability, travel costs, provisioning requirements for offspring, and behaviour of conspecifics. Barrier effects are implemented within the model by susceptible birds flying around OWFs to reach foraging grounds, and displacement effects are incorporated by causing susceptible birds to re-select foraging locations out-with the OWF should an initial foraging location within the OWF be selected. The resulting outputs for adult mass at the end of the breeding season are then translated into an estimate of population level adult survival for each colony, with and without one or multiple OWFs present. Impacts on chick mortality are also included within the model, with provisioning and attendance by both parents affecting chick growth and probability of mortality within simulations. The model provides individual and population level estimates for the change in adult mortality and breeding success for individual colonies affected by one or more OWFs, providing a direct link from observed or estimated spatial foraging patterns of breeding birds through to population demographics. The model also enables the behaviour and fate of individual birds to be tracked and summarised in a range of different ways. This permits a direct quantification of the demographic consequences of displacement for individual birds (for full details of SeabORD model specification see Appendix A).

SeabORD does not directly use Equation 1 in calculating displacement risk, but because it does require users to specify both baseline exposure (via a bird utilisation distribution map) and the displacement rate, it is straightforward to express SeabORD outputs using Equation 1. Specifically, Equation 1 can be rearranged to give:

Mortality rate for displaced birds = Displacement mortality / (Baseline exposure * Displacement rate).

which is directly comparable to the 'mortality rate for displaced birds' used within the Displacement Matrix approach.

Expressing it in this way illustrates that the fundamental difference between SeabORD and the Displacement Matrix lies in the way that the mortality rate for displaced birds is calculated. SeabORD calculates this using a mechanistic model, whereas the Displacement Matrix approach currently derives these rates from expert judgement. The Displacement Matrix approach also uses a particular way of visualising and summarising uncertainty (via the 'Matrix'), which is not currently used when applying SeabORD, but the connection between the approaches described here shows that this could also potentially be used in conjunction with SeabORD. Presentation of uncertainty in SeabORD is currently limited to that resulting from variation in prey availability across good, moderate and poor environmental conditions and its impact on adult and chick mortality. However, uncertainty in the displacement rate used within SeabORD could also be expressed using something similar to the visualisation used in the matrix method – running simulations with different displacement rates and tabulating or graphing the resulting uncertainty in model outputs.

Equation 1 can be used to calculate mortality rates for either adults (i.e. survival) or chicks (i.e. productivity). Note that SeabORD considers both of these, but impacts on chicks are usually ignored within the Displacement Matrix approach.