Crouching tiger, hidden dangers: Avian fatality rates reduced by red-blade patterning at a species-rich African wind farm

Avian collision rates are certain to rise as renewable energy industries roll out wind and solar farms to reduce fossil fuel impacts in biologically diverse areas of the world. Technological solutions are often sought to decrease mortality rates, but for developing nations automated shut downs are expensive, and alternatives required. A promising route is to increase blade visibility to birds using high contrast colours. Despite the success of the solid black-blade experiment in Norway only one other black-blade field-study in the Netherlands has explored this possibility, with no significant results. We tested the use of colour-patterned blades at a species-rich, 37-turbine, wind facility in Hopefield, South Africa. Two broad "signal red" stripes were applied to a single blade at four high-fatality turbines, in 2023 by Umoya Energy. Avian fatality rates were compared before and after painting using the Before-After-Control-Impact (BACI) approach. Seventy-five fatalities of 23 species of raptors, passerines and wetland species over 24 months were compared for the same 20 turbines after patterning with two sets of controls: (i) their four nearest neighbours (NN) and (ii) all 16 controls (AC). Over 32 months 25 fatalities were recorded, 23 occurred at the controls and only two at the patterned turbines. Testing with Bayesian Generalized Linear Models (BGLMs) revealed a median 83% reduction in fatalities at the patterned blades for both the NN turbines (credible intervals 14% - 98%) and the AC comparisons (30% - 97%). Bayes Factors (BF) revealed strong statistical support for NN (BF = 49.9) and AC comparisons (BF = 159). There was little evidence that birds avoiding patterned turbines increased fatalities at the neighbouring turbines as there was a small median 15% increase in fatality rates when NN controls were compared with other controls, and weak statistical support (BF = 0.15). Among 14 raptor species recorded on site, 10 species have suffered fatalities. Of seven individuals killed prior to treatment at the four patterned blades, only one was killed post-treatment suggesting blade patterning is equally effective at reducing raptor fatalities. Our results show that patterned blades had a high probability (83%) of reducing fatalities with strong statistical support despite the small samples. This supports the Norway experiment in a high diversity African setting, but with red patterns not a solid black design. The strong effect of red stripes may arise from both the high contrast it provides and the possible warning effect that red may elicit. We call for additional experiments to differentiate the effect of patterns and colours for the optimal design to reduce avian-turbine collisions.