Measured Deterrence: Painted Blades Cut Bird Collisions 83%, Audio Alarms Double Avoidance

Two studies published in October 2026 are worth reading together, because one shows how cheap visual deterrence can be and the other shows when active deterrence actually changes behaviour. Both speak directly to anyone specifying bird control on a solar farm, wind farm or industrial site.

The paint result

At a 37-turbine wind farm in South Africa, two broad red stripes were painted on one blade of four machines. Carcass monitors worked the site for 32 months after the painting. The result: two carcasses beneath the painted turbines against 23 beneath unpainted controls, with Bayesian modelling putting the median reduction at about 83%. Raptors were not left out — 14 raptor species use the airspace over that farm, and at the four treated machines seven raptors had been killed before the stripes went on against one afterwards. Red was partly a regulatory choice: aviation rules in South Africa restrict black on tall structures.

Two earlier trials point the same way. A Norwegian island farm painted a single blade black and reported overall collisions down around 70%, with the clearest gains among birds of prey. At a farm in Cadiz, one blade in three was painted black on four of 20 turbines and fatality rates fell 74%, with the largest gains among passerines and large waterbirds and less effect on raptors.

Why it works, and where it does not

The mechanism proposed is motion smear. A rotor does not read as three separate blades to a bird; at speed the pale arcs blur into one pale disc with no edge and no obvious motion to steer around. Breaking the repetition restores contrast with every rotation. A competing explanation is simpler still: many birds never register a pale blade against a pale sky, because their sensitivity to light and dark contrast is far weaker than ours. Both explanations lead to the same prescription — mark one blade, and the rotor stops reading as empty air.

Keep the caveats attached. The South African work is a preprint and has not cleared peer review; four treated turbines buys you wide credible intervals, quoted at 14 to 98% against the nearest neighbours. A recent review rates blade painting as promising but unproven, with successes and failures in roughly equal numbers. And the cost is as small as it sounds — a tin of exterior paint and a crew already scheduled for blade inspection, no sensor and no curtailment — which is precisely why it matters most at smaller and older sites that cannot fund radar-based detection.

The acoustic result

The second study looked at active deterrence. Researchers from H.T. Harvey & Associates and the Renewable Energy Wildlife Institute tested automated detection and audio deterrent units at two commercial wind farms, California's Manzana Wind Power Project and Washington's Goodnoe Hills. Audio warnings triggered evasive behaviour in 79% of probable golden eagles in California and 61% in Washington. Active audio alarms doubled deterrence effectiveness compared with operational turbines alone, and only 27% of raptors showed no response at all. Birds already on moderate to high collision courses reacted more strongly than those at low risk, golden eagles and turkey vultures responded better at higher wind speeds, and smaller buteos showed a slight negative relationship. The work appears in the Journal of Raptor Research.

What this means if you are buying deterrence

• Passive and active do different jobs. Contrast marking reduces the chance a bird fails to see the hazard at all. Active deterrence works at the moment of approach. A site that uses only one of the two is relying on a single mechanism.
• Habituation is the real enemy. Devices that broadcast continuously train birds to ignore them, which is why the studies above trigger a response at the moment a bird is on a collision course rather than sounding all day. Look for detection-triggered or randomised output, and for the ability to change pattern, angle, speed or frequency.
• Ask for evidence that behaviour changed, not that the device switched on. The useful number is the percentage of approaches that resulted in avoidance, measured on site. A unit that logs detections is worth more than one that only logs uptime.
• Design for an unmanned site. Solar and wind sites are rarely staffed. IP-rated housings, UV-stable materials, wide temperature range, salt-air tolerance, solar or battery autonomy through overcast weeks and remote configuration all belong in the specification.
• Never at the nest. Deterrence belongs at the array, the boundary and the structures — not at active nests of protected species. Have the site ecologist sign off on placement.

Solar-specific note

On PV sites the measurable cost of birds is not only collision. Droppings and debris soil modules, which reduces yield and adds cleaning labour, and nesting under and around arrays creates both maintenance and grazing conflicts. That is an operations line item, and it is easiest to fund when deterrence is written into the O&M budget rather than arriving later as an unbudgeted complaint.

Where we fit

We build solar-powered bird repellers — see our bird repeller range, the selection criteria in our solar bird repeller buyer guide, and the planning context in our note on the October UK solar consents. Tell us the site size, climate, species mix and whether the requirement is construction-phase or operational, and we will say which units suit it — including when a unit is the wrong tool.