NEON pitfall traps at 44 US sites show no net change in ground beetle activity-density from 2013 to 2023 (−0.6%/year, 95% CI −3.5 to +2.3) with a constant trap design

Question: did the number of ground beetles (Carabidae) that NEON pitfall traps catch per day change from 2013 to 2023? NEON removed one trap per plot in 2018 and four of ten plots per site in 2023. Approach: we used 104,622 trap collections (244,815 beetles) from NEON product DP1.10022.001, RELEASE-2026, through the neonDivData package. We checked these data against raw NEON tables: 99.9% of 88,824 trap totals for 2013-2021 were identical. Our primary estimator used only the plots and trap positions that NEON sampled in all years. For each site, a quasi-Poisson model with year, plot-trap and season terms and a trap-day offset gave annual indices. We fitted a linear trend to each site series and pooled 44 site trends with a random-effects model. Findings: the pooled change was -0.6% per year (95% CI -3.5 to +2.3). In 20 alternative analyses the estimate was between -1.7 and +0.4% per year, and all intervals included zero and -0.92% per year (a published mean insect decline). Site trends were very different (SD about 8% per year): four sites decreased and six increased significantly. Climate, latitude and forest cover did not explain these differences. At 27 of 46 sites, the plots that NEON kept had catch rates more than 25% different from the plots that it stopped. This caused site trend errors of up to 15 points per year in a naive ratio estimator, but no net error. Of 429 site-species populations, 6 decreased and 3 increased significantly after FDR correction. Simulations with the measured variance show a power of 0.15-0.31 to detect a 1% per year decline with 11 years and 44 sites; equal site trends would need about 20 years for 80% power. Limitations: activity-density also changes with weather and beetle activity; series are 5-11 years; provisional 2024-2026 data were not accessible.

ground beetles · Carabidae · NEON · insect decline · pitfall traps · activity-density · population trends · monitoring design · statistical power · random-effects meta-analysis · biodiversity monitoring