Modelling flight heights of marine birds to more accurately assess collision risk with offshore wind turbines

The number of offshore wind farms is rapidly increasing as they are a critical part of many countries' renewable energy strategies. Quantifying the likely impacts of these developments on wildlife is a fundamental part of the impact assessments required in many regions before permission for developments is granted. A key concern related to wind turbines is the risk of birds colliding with turbine blades. We present a novel method to generate species‐specific flight height distributions which can be used to improve the assessment of collision risk by better reflecting the proportion of in‐flight populations at risk of collision. Data describing the flight heights of birds from surveys of 32 potential offshore wind farm development sites were combined to estimate continuous distributions for 25 marine bird species. Observations of flying birds assigned to discrete height categories were treated as observations from independent multinomial distributions with a shared underlying continuous distribution. This analysis enables calculation of the uncertainty around the estimates of the proportion of the in‐flight population at risk and consideration of different turbine designs. The mean r² for model fit across species was 0·85, and for seven of the species, good independent model validation (80% of independent observations within 95% confidence intervals) provides some confidence for use of the results at alternative sites. All species exhibited positively skewed flight height distributions. These results demonstrate that under the conditions in which the data were collected, raising hub height and using fewer, larger turbines are effective measures for reducing collision risk. Synthesis and applications. The methods presented here for modelling continuous flight height distributions provide measures of uncertainty and enable comparison of collision risk between different turbine designs. This approach will improve the accuracy of impact assessments and provide estimates of uncertainty, allowing better evidence to inform decision‐making.