Solar's Bird Bill: New Research Links Panel Land Use to a 0.66% Drop in Bird Abundance

Most of the argument about solar farms and birds has been about collisions — birds striking panels after mistaking their polarised reflection for water. A 2026 paper in Ecological Economics moves the question one step earlier in the project, to where the panels sit and what they replace, and the finding is uncomfortable for anyone who treats solar as ecologically neutral by default.

What the study found

The paper, by Xi Ji, Jie Lin, Pengfei Liu and Xiaoxue Zong, uses high-resolution county-month panel data from China with an instrumental-variable strategy and rich fixed effects. Its central estimates: a 1% increase in photovoltaic panel area is associated with a 0.659% decline in bird abundance and a 0.473% decline in species richness. The mechanism is not collision but habitat — the loss of surface vegetation that panels and their footprint remove, reducing the vegetation birds depend on.

Three of the study's refinements matter for how the finding should be used:

• Land type dominates. The negative effects are most pronounced on forest land, followed by cropland, then grassland — so the same installed capacity has very different consequences depending on what it covers.
• Configuration matters. Low-density, large-scale centralised installations show larger adverse effects than more spatially distributed ones.
• Season matters. Impacts are most pronounced in autumn — the period that overlaps with intensive bird migration.

The authors also report that Key Ecological Function Zones did not mitigate the effects, and they frame the result as an under-appreciated trade-off: decarbonisation and biodiversity are both stated goals, and land-intensive solar can pull them against each other unless siting and vegetation are planned deliberately.

The collision story, next door

The habitat finding sits alongside the collision concern rather than replacing it. In New Zealand, the 300-hectare Glorit Solar Farm proposed by Contact Energy and Lightsource bp — 290,000 panels — was sited under regular flight paths of the tara iti fairy tern, of which fewer than 40 remain. Conservation group Forest & Bird argued the panels' polarised reflections would be mistaken for water, leading to fatal or injurious collisions, and took the fast-tracked consent to the High Court. The appeal was discontinued after a settlement that added fairy-tern collision-monitoring measures and contingency actions. Further developers are queuing for fast-track consents in the Mackenzie Basin, several near internationally important bird areas around Lake Benmore.

Together the two strands describe a project's full bird exposure: what the land was before the panels (habitat, the paper's subject) and what the panels do to birds that fly over them (collision, the New Zealand case).

What it means for developers and suppliers

• Ecological baseline before siting, not after. The study's land-type result is the planning lesson: forest and wetland conversion carries a different burden from low-value grassland, and the cheapest land is often the most ecologically sensitive.
• Treat autumn as the critical monitoring window. Seasonal effects peaking during migration argue for survey and mitigation effort in that window specifically, not a single year-round average.
• Manage vegetation and water, not just collisions. If habitat loss is the mechanism, the mitigation is siting choice, vegetation management and avoiding designs that turn a site into a green desert — plus attention to how panels read as water to birds in flight.
• Keep the categories straight. Habitat loss and collision are different problems needing different responses. A deterrence device is not an answer to habitat loss, and a habitat study is not evidence that a deterrent works.

Our position

We make solar-powered and variable-frequency bird repellers, and we are precise about where they sit in this picture. They are site-scale deterrents — well suited to the edges of a solar array, substations, storage yards and other unpowered, unmanned locations where a compact acoustic device can make a specific spot less attractive. They are not a collision-monitoring system and they do not restore habitat, and we will not market them as either. If a project needs a deterrence layer as part of a broader ecological response, we can supply and document the equipment; the monitoring and siting decisions belong with the project's ecologists. Our note on the measured evidence for painted blades and acoustic deterrence covers what deterrence can and cannot claim, and our bird repeller range is here.