Take-home message:

  • Driving a car for 1 km kills an estimated average of: ~90 insects (by the estimate that I find most rigorous), ~0.5 - 2.5 insects (by another reasonable estimate, but an underestimate), and ~0.005 insects (I think that estimate is problematic, but including for honesty’s sake).
  • But there’s lots of variance, and different papers have wildly different numbers. The variation arises due to both methodology (attaching sticky traps to cars vs counting dead insects on the side of the road) and physical location.
  • I think the best guess is an average of ~90 insects, with large error bars, and a reasonable possibility that the true number is ~1-10% that esimate.

Background to this blog post

  • The first half of this blog post was written by Rebecca Eddington in 2022, as part of a research contract with me.
  • The second half of this blog post was written by me, just now. Like with some of my other blog posts, the second half of this post summarizes key passages from journal publications, and doesn’t contain much original addition from me.

Question: If a car drives for X km, how many insects could the driver expect to kill (assuming a suburban area, like the suburbs of Adelaide where I live)?

In a 2015 review paper, ‘Effects of roads on insects: a review’ by Tamayo et al., the authors discuss the variation across studies calculating the amount of insects killed per km despite similar methods of using transects along the road to collect dead insects. Some of the studies removed the insects from the road before counting, some did not, and some only provided the total number of insects found, with no specification of the number per km. While a few of these studies measured the amount of insects killed in relation to traffic density, they concentrated either more widely on roadkill of all animals or particular species of insects. The authors of the review also discuss the difficulties in comparing these studies due to the extent of the studies, the number and types of road assessed and transect particularities. Factors such as the speed of vehicles, traffic volume, road width, time of year/day and habitat along the road all influence the collision rate per km too.

But, let’s try and answer the above question with some assumptions.

Estimate 1: Martin et al 2018

In a paper written by Martin et al. 2018 ‘Flying insect abundance declines with increasing road traffic density’ the authors used sticky traps, attached to a vehicle, to measure insect collisions on high and low density traffic roads. Both the high density and low density roads were in rural locations, with two traffic lanes and with paving alongside, so while neither road location is representative of a suburban area, the road structure (two lanes/paved) is somewhat similar to a suburban area. The sampling took place between June - September 2014. In the study the authors used 20 road segments of ~1-3km, which were sampled twice and in each sample the road segment was driven down twice (2022=80). The total number of insects that hit the sticky traps on the vehicle was 7225. To standardise the varying road distances, the authors sampled less of the surface area of the sticky traps on longer roads. So, 7225/80 gives us an estimate of ~90 insects killed per vehicle per km of road.

Estimate 2: Gilbert et al 2015

In a paper authored by Gilbert et al., 2015 ‘Road mortality potentially responsible for billions of pollinating insect deaths annually’ the authors used a 2km portion of the Canadian Highway 69/400 and conducted 235 transect surveys, during the summer months of May - August in 2012 and 2013, collecting insects found on the road and the adjacent shoulders. They calculated that Lepidopterans are struck and killed by traffic at a rate of 10.1 individuals/km/day, Hymenopterans at a rate of 26.8/km/day and Dipterans at a rate of 10.4/km/day (excluding one abnormally large dataset) totalling a rate of 47.3 individuals/km/day. The paper includes data from the Ministry of Transportation Ontario that on average 9700 vehicles per day drive on the highway which measures 388km.

Ren’s addition:

  • (47.3 insects/km/day) / (9700 vehicles/day)
  • = ~0.005 insects/car/km
  • But the numbers from the paper are a little messy, we might be conflating “km of road” with “km driven”
  • In any case, it’s insects collected from the side of the road. Insects killed = insects found on side of road + insects found on windscreen
  • Gilbert et al 2015 would be only “insects found on side of road”, so an underestimate of “insects killed”
  • Whereas Martin et al 2018 used sticky traps so would probably be the closest paper to “insects killed”
  • Seems difficult to get any kind of comparable numbers from differences in sampling methods. Gilbert et al = transect, might miss smaller insects - Martin et al results = 6996 of 7225 insects under 5 mm long. Another example Martin et al, only when not raining, Gilbert et al non weather dependent sampling. Too many different factors for any kind of comparison - why I’ve found this difficult! I think Martin et al methodology best I’ve found to answer question but can’t find anything with similar methodology to back this up (yet!)

Everything from here onwards is also from me.

Estimate 3: Wilson et al 2024 x Russo 2025

Wilson et al 2024

  • Utah, USA
  • Sticky traps, focused on bees
  • “One particular stretch of road, from Stansbury Park, UT to St. George UT was travelled 10 times between April and July of 2021.”
  • “Accounting for the different distances driven on each trip, the average number of bees hit per kilometer by any car is 0.26, and ranges from 0 to 1.15”

But what contribution do bees make to insects as a whole, at least among insects who are killed by cars?

Enter Russo (2025).

Russo 2025

  • Knoxville, Tennessee, 2023-24
  • Roadside collection
  • Some quick maths on Table S2 in the Supporting Information reveals that Russo’s survey found 4,916 arthropods, of which 2,307 were bees (i.e., belonged to one of the six families of bees listed on Wikipedia)
  • Notably, Russo also wrote this passage, which has implications for insect suffering: “Many specimens were badly damaged. […] It was very common for the insects to be headless, potentially an indication of the trauma of impact with a vehicle…”

Russo’s results imply that ~47% of arthropods killed by vehicles (and then fall to the side of the road, and aren’t damaged too badly to be identified…) are bees. This indicates that Wilson et al’s results of 0.26 bees/car/km is equivalent to 0.26/0.47 = ~0.55 arthropods.

For another estimate, Gilbert et al 2015 report that ~11% of insects killed by vehicles (with the same massive caveats) are Hymenoptera (roughly speaking, bees, wasps, and ants). This indicates that Wilson et al’s results of 0.26 bees/car/km is equivalent to 0.26/0.11 = ~2.4 arthropods.

So, Wilson et al 2024’s results imply roughly ~0.5 - 2.5 insects killed per vehicle per km of road (rounding because my maths here is very very dodgy).