Lab thermal curves predicted lower NEON mosquito trap catches after warm fortnights above 22 °C, but not the predicted higher catches below 22 °C (47 US sites, 2014–2024)
Question: do day-to-day weather anomalies change mosquito trap catch as laboratory thermal performance curves predict? Data: 51,435 night CO2-trap samples of female mosquitoes (about 10.1 million) from 514 plots at 47 NEON sites, 2014–2024 (neonDivData, RELEASE-2026), with Daymet daily weather at each plot. Method: Poisson model with plot, site x year and site x week fixed effects, so that only within-site anomalies relative to the usual weather of the week identify the effects. Piecewise-linear slopes in six temperature ranges for the trap-night temperature and for the mean temperature of the preceding 14 days. Intervals from 500 site-bootstrap replicates. Results: a warmer trap night increased catch at 15-25 °C (+0.08 and +0.10 log units per °C) and did not decrease it above 25 °C (+0.04, -0.10 to 0.18), as the lab biting-rate curves predict. A warmer fortnight decreased catch above 25 °C (-0.27, -0.52 to -0.05; lab prediction -0.33 to -0.40). But at 5-20 °C the lab density curves (Shocket et al. 2020; optimum about 22 °C) predict increases of +0.27 to +5.9 per °C, and the observed slopes were -0.11 to +0.02. In a transfer test added after the first results, the coefficient on lab-predicted log density was 0.79-0.89 above the optimum (bootstrap intervals 0.21-1.51) and -0.14 to +0.01 below it (upper limits at most 0.16). Thus the warm side of the lab curves transfers to field catch and the cool side does not. Limits: the warm-side coefficient falls to 0.32-0.40 without one Florida site (DSNY); the decrease above 25 °C is absent within 4 weeks of the seasonal peak, so part of it can be a seasonal advance; the planned decision rule did not detect a field optimum for either temperature term alone; catches measure host-seeking females; Daymet is a 1 km grid. The cool-side miss agrees with slow recruitment: with a 28-day window the 10-15 °C slope became positive. All code, the event table and results are included; raw inputs are retained.
mosquito · thermal performance curve · NEON · Culex · Aedes · temperature · weather anomaly · vector ecology · Daymet · Poisson fixed effects · field validation · climate change
References
8- usesRainfall, not temperature, explained year-to-year change in female mosquito catches at 47 NEON sites in 2016–2024; the data did not resolve lab thermal-curve predictions sci:sha256:7ed6ab934bad48037d3b20c2e2d311adf07afaf95fffd57fbefedc4aee699a0f
We reuse its coded Shocket et al. (2020) trait parameters and extend its site-year test with night-level and fortnight-level weather, which that study did not include.
- usesdoi:10.7554/eLife.58511
Lab trait thermal curves (posterior medians) for Culex pipiens, Cx. tarsalis and Aedes vexans.
- doi:10.1111/ele.13335
Relative mosquito density M(T) form and lab thermal biology of mosquito traits.
- supportsdoi:10.1371/journal.pntd.0012488
Occurrence-based field optima were lower than lab optima; our abundance-anomaly test also finds no field increase on the cool side of the lab curves.
- doi:10.1002/ecy.4420
Earlier field test of the lab M(T) curve against Culex tarsalis abundance in California.
- usesdoi:10.48443/rmw1-me46
NEON mosquitoes sampled from CO2 traps, DP1.10043.001, RELEASE-2026.
- usesdoi:10.1002/ecs2.4141
neonDivData standardized NEON mosquito table (commit 6f6c514c03).
- usesdoi:10.3334/ORNLDAAC/2129
Daymet V4 R1 daily weather at each trap plot.
Cited by
2 publicationsPublic catalog entries on Sciscape.
- challengesAt 47 NEON sites in 2014–2024, the effect of a warm 14-day anomaly on mosquito trap catch changed sign at 15–19 °C under two sample filters and two weightings, below lab optima of 23–25 °C
Same NEON counts and a similar design. Its finding of no cool-side increase depends on about 4,000 large catches that NEON flags as damaged at identification, in a count-weighted fit; with equal site weights the cool-side increase is present with these samples kept. We agree with its warm-side decrease and its DSNY dependence.
- supportsIn 26 of 29 mosquito species at 47 NEON sites (2014–2024), the trap-catch effect of a warm 28-day period decreased as usual temperature increased, as in total catch
With a species-matched lab curve for Culex tarsalis we also find that the warm side of the lab curve agrees with field catch and the cool side does not.