The “Windshield Phenomenon” — Are We Measuring Insect Decline, or Car Design Evolution?
A new citizen science initiative out of France is making headlines this week, asking drivers to count dead insects on their license plates after road trips and submit the data via a smartphone app called Bugs Matter. The effort, run by France’s National Museum of Natural History along with several environmental groups, aims to quantify what researchers are calling the “windshield phenomenon” — the widely noticed decline in bug splats on vehicles over recent decades.
The numbers being cited are dramatic. A two-decade Danish study reportedly found declines of 80–97% on two monitored road stretches. A UK study using the same Bugs Matter app found nearly a 63% drop between 2021 and 2024. A 2017 German study documented a greater than 75% decline in flying insect biomass even in protected nature reserves. Scientists are attributing these trends to habitat loss, pesticide use, pollution, and of course, drum roll please…..”climate change.” There may be other factors, like competition from non-flying insects, or some other sort of natural suppression.
I don’t doubt that insect populations have changed over time and that’s worth studying carefully. But before we accept the windshield/license plate data as a clean signal of population collapse, there’s a significant confounding variable that I haven’t seen addressed in any of the coverage: modern cars are dramatically more aerodynamic than they were 30–50 years ago.
The Aerodynamics Problem
Here’s the physics. A vehicle moving at highway speed creates a bow wave of compressed air in front of it. As aerodynamic efficiency improves — lower drag coefficients, more raked windshields, smoother underbodies, active grille shutters — that airflow increasingly wraps around and over the vehicle rather than hitting it head-on. Insects flying at bumper or windshield height get deflected upward and around the car rather than impacting the surface.
This isn’t speculation. It’s the same aerodynamic principle that makes modern EVs so range-efficient. The average sedan drag coefficient (Cd) has dropped from roughly 0.30–0.35 in the early 2000s to 0.20–0.25 today, with some models pushing below 0.20. That’s a 30–40% improvement in aerodynamic slipperiness over roughly the same timeframe as the insect decline studies.
The French study made a deliberate choice to use license plates rather than windshields for counting, specifically because plates are standardized across vehicles — same size, same perpendicular-to-road orientation. That’s smart methodology. But license plates are mounted at bumper height on the front fascia, a surface that on modern vehicles is also increasingly swept back and aerodynamically optimized compared to the bluff, vertical grilles of older cars. The aerodynamic deflection argument applies there too, just perhaps somewhat less than to windshields.
What the Data Would Need to Show
For the windshield/plate studies to be genuinely robust, researchers would need to control for at least the following:
1. Fleet composition over time. The mix of vehicles on European and American roads in 2024 is fundamentally different aerodynamically from the fleet in 2004. As older, blunter vehicles were retired and replaced with more aerodynamic designs, the average “bug-catching efficiency” of the fleet declined. Has anyone modeled this?
2. Vehicle type stratification. Full-size pickup trucks and older box-shaped vans have changed little aerodynamically. Do those vehicle classes show a smaller decline in splat counts compared to modern sedans and crossovers? If the phenomenon is real insect decline, you’d expect similar rates across all vehicle classes. If aerodynamics is a confounding factor, you’d expect lower declines on blunter vehicles.
3. Speed and road type. Higher speeds create more intense bow waves and deflect insects more efficiently. If average highway speeds have changed, or if the monitored road types have shifted, that introduces another variable.
4. The protected reserve data as a partial control. To their credit, researchers can point to the 2017 German study showing comparable declines in flying insect biomass in nature reserves — measured without vehicles at all. That’s important, and it does suggest something real is happening with insect populations independent of measurement artifacts.
The Bottom Line
None of this means insect populations are healthy or that the “windshield phenomenon” is entirely an artifact of car design. The German reserve data in particular is hard to explain away. But the citizen science studies using vehicle-based counts should not be presented as unambiguous population proxies without explicitly accounting for fleet aerodynamic evolution over the study period.
This is the kind of methodological detail that tends to get glossed over when a narrative — in this case, an alarming environmental decline story — is already well established in the press. The decline may well be real and serious. But the magnitude being reported from vehicle-based studies could be meaningfully inflated if no one has corrected for the fact that today’s cars are fundamentally better at not hitting things than the cars of 20, 30, or 50 years ago.
It would be a relatively straightforward analysis: pull historical vehicle registration data, apply known Cd values by model year and class, estimate fleet-average aerodynamic interception efficiency over time, and see how much of the reported decline could be explained by the vehicle fleet itself. If someone has already done this work, I’d genuinely like to see it. If they haven’t, it seems like a glaring gap.
As always, the data may be right — but the methodology deserves scrutiny before the conclusions are carved in stone.
