Theory 17/Insects & Entomology /Insect ecology/3 min
Declines Track Thermal Variance
Warming is the headline. The swings may be doing the killing.
Insect population declines are driven by the variance of temperature rather than by its mean, so a stable warm climate should support more insects than a variable one at the same average, and decline rates should correlate with variance metrics.
01The evidence
- Continental-scale radar monitoring showed fewer insects above warming landscapes, with strong year-to-year variation in density in response to climatic variables, which points to variability as well as trend.NSF, radar insect monitoring
02The tension
Insect decline is framed as a response to mean warming and to land use, and the two explain a large part of the trend. But the residual is large, and it tracks variability rather than mean in several datasets. Variance is rarely used as a predictor because it is harder to model.
03The theory
Insects have thermal limits and short life cycles, so they experience temperature as a sequence of extremes rather than as an average. A population can be killed by a single extreme event even if the mean is favourable.
If extremes dominate, then the variance of temperature is a better predictor than the mean, and the relation should be nonlinear because damage depends on exceeding a threshold rather than on the average exposure. That makes variance a first-order variable.
The prediction is that decline should correlate with the frequency of extremes better than with the mean, and that two regions with the same mean but different variance should differ in insect density. It also predicts that climate change should increase decline even where the mean changes little, which is observed.
The practical implication is that monitoring should report variance metrics alongside means, and that projections based on mean warming will underestimate loss in high-variance regions.
04What this is not
Those models use mean temperature and land-use intensity. This adds variance as a first-order driver, and predicts declines where the mean has barely moved.
05Predictions
- P1Decline should correlate with the frequency of temperature extremes better than with the mean.
- P2Regions matched on mean but differing in variance should differ in insect density.
- P3Projections including variance should predict larger losses in high-variance regions.
06What would refute it
Decline tracking the mean at fixed variance would leave variability as a secondary factor.
07What it would change
It would change how insect monitoring and climate projections are built, and where conservation effort goes.
08Evidence base
- Continental-scale radar monitoring shows fewer insects above warming landscapes — NSFhttps://par.nsf.gov/biblio/10653983/media/xml