Implications of surveyor accuracy in bird flight height estimation for wind farm collision risk assessment

The increasing anthropogenic use and modification of the airspace,
particularly through the expansion of wind energy, has led to an increase
in collision-related mortality of aerial wildlife. Environmental impact
assessments rely on accurate bird flight height estimates to predict
collision risk, yet these estimates are typically made by surveyors
visually estimating flight height, with limited validation of their
accuracy. This study evaluates the accuracy of surveyor flight height
estimates against measurements from an unmanned aerial vehicle (UAV) and
investigates whether accuracy can be improved by the use of annotated maps
(spatial cues) and/or prior survey experience. Results indicate that there
was generally a discrepancy between surveyor flight height estimates and
the UAV measurements, particularly for inexperienced surveyors. Annotated
maps reduced variability in estimates and improved accuracy, with the
greatest benefit observed for inexperienced surveyors. There was a
discrepancy between the number of model predicted collisions based on
surveyor estimates and UAV-based predictions, but this was reduced with
the use of annotated maps and prior experience. Discrepancies in flight
height estimation and subsequent discrepancies in collision predictions
resulted in underestimating and overestimating mortality and translated
into differences in impact assessment classifications, which could
influence wind farm planning and mitigation decisions. Practical
implication. Overall, our results suggest that while annotated maps are
beneficial, particularly for inexperienced surveyors, they are not a
substitute for experience. Standardising survey methods through tools like
annotated maps can improve data reliability, particularly for less
experienced surveyors, ultimately strengthening impact assessments and
wind farm management strategies. Beyond the value of spatial cues and
surveyor experience, our work underlines the importance of quantifying
sources of error in data collection for environmental impact assessments
and consequences for subsequent management decisions.