A revised methodology for quantifying ‘Acceptable Level of Impact’ from offshore wind farms on seabird populations

The Acceptable Level of Impact (ALI) methodology defines acceptable limits for the predicted population effect of mortality imposed by offshore wind farms (OWFs) on marine bird populations and was first developed in 2021 as a replacement for the Potential Biological Removal (PBR) method. The ALI is a probabilistic framework and defined as: ‘The probability of a population decline of X% or more, 30 years after the impact, cannot exceed Y’. The methodology is based on a comparison of future population abundance between two scenarios: one unimpacted scenario without impact from OWFs, and one impacted scenario which includes additional mortality resulting from bird collisions with offshore wind turbines and/or habitat loss from avoiding OWFs. In the ALI definition, X represents the threshold value above which the population effects of OWFs are considered undesirable and Y is the threshold value for the probability that a population decline larger than X would still occur. Following its development and first use, several reviews and an in-depth analysis revealed a number of methodological issues. These were related to the use of a causality measure. This quantified the probability that an unacceptable decline in population abundance was caused by OWFs, instead of by uncertainty or biological variability (e.g. environmental stochasticity). Discounting the effect of uncertainty and variation was considered contrary to the precautionary principle. In addition, this may lead to the counterintuitive result that a more strict threshold for the acceptable level of population decline (X) results in a lower probability that such a decline is judged as unacceptable. Thus, only choosing a more strict X threshold would permit larger impacts of OWFs. Although these dependencies can be accounted for through the choice of threshold values, this requires expert judgement and complicates the methodology. It was therefore recently recommended to revise the ALI methodology and abandon the use of a causality measure. The current report proposes a revision for the ALI methodology based on a comparison between many (100,000) impacted and many unimpacted simulations using stochastic population models. Here, impacted simulations include additional OWF-induced mortality. The adopted approach is similar to the original methodology, but the crucial difference is that this comparison is made per replicate simulation. Essentially, a comparison between scenarios is made while keeping constant all other processes that affect the predicted development of the population. Therefore, variation between the simulations do not contribute directly to the differences between scenarios. This approach makes the ALI methodology more user-friendly and it no longer uses a measure of causality. Furthermore, the X and Y threshold values of the probabilistic ALI framework are no longer interrelated, which obscured the use and applicability of the original ALI. We present an updated set of recommendations for choosing these threshold values.