Total macro-moth catch in Finnish light traps decreased in 1993–2012 but was 29% above that trend in 2013–2019, so the 1993–2019 trend was near zero
Question: Two studies of the Finnish moth monitoring scheme (Nocturna) do not agree. One study reported a decrease of total moth abundance in 1993–2012. Another study found no decrease of moth biomass in 1993–2019. We tested if the time window causes this difference. Data: We used the public FinBIF archive for 1993–2019. It gave 1,542 trap-years at 223 light traps with at least 85% of the May–September season sampled. Method: We calculated an annual index of total catch with trap and year fixed effects. We calculated the linear trend for all 171 windows of 10 or more years. We used a site bootstrap, autocorrelation-robust (HAC) intervals and simulated year-to-year noise. We fitted a trend line to 1993–2012 and compared its prediction with 2013–2019. Results: The 1993–2012 trend was −1.09% per year (HAC 95% interval −2.10 to −0.07). The 1993–2019 trend was +0.13% per year (HAC −1.18 to +1.46). The eight 20-year windows gave trends from −1.20 to +1.17% per year. The index had high years in 1994–1996, 2005–2006 and 2015–2019. In 2013–2019 the index was 29% above the 1993–2012 trend line (site bootstrap 16% to 46%). If the 1993–2012 trend continued, the probability of this error was 0.009 to 0.10, depending on the noise model. Trend-free noise gave a 20-year trend as steep as −1.09% per year in 21% of windows. Over ten data filters, the 1993–2019 trend was +0.11 to +0.28% per year and the 2013–2019 error was +25% to +31% (0.22 to 0.27 log units). The 1993–2012 interval included zero in five filters. An equal-weight indicator of 421 common species did not decrease in 1993–2012 (+0.06% per year). Rarefied species richness increased in 1993–2019 (+0.45% per year). The south (below 62° N) gave the strongest evidence of a decrease. Its 1993–2019 interval included zero (−0.61% per year, HAC −1.72 to +0.50). Limitations: Light-trap catch measures flight activity. The trap network changed in 1998 and 2017. Samples with zero catch are not in the archive. We tested numbers, not biomass. We did not repeat the earlier models. 27 years cannot show if the highs and lows are a regular cycle.
insect decline · moths · Lepidoptera · light trap · Finland · Nocturna · population trend · time window · out-of-sample test · long-term monitoring · FinBIF · boreal
References
3- usesdoi:10.15468/j82xqf
Source data: Finnish National Moth Monitoring (Syke) Darwin Core Archive, FinBIF HR.4511 dated 2026-10-21.
- challengesdoi:10.1111/geb.13073
We find a decrease of total catch in its 1993–2012 window, but in 2013–2019 the catch was above that trend and the 1993–2019 trend was near zero. We compare only with its abstract.
- supportsdoi:10.1111/icad.12657
Our number-based index agrees with its finding of no decrease in total moth biomass in 1993–2019 in the same scheme.
Cited by
8 publicationsPublic catalog entries on Sciscape.
- usesEqual-count sampling removes the recorded decrease in moth list differences at 19 Finnish traps from 1993–1999 to 2013–2019
We use its retained Nocturna source archive and its trap and species-name conventions. We test spatial differences between fixed traps.
- usesYear resampling leaves the mean catch date change uncertain at six Finnish moth traps in 1993–2019
We reuse its retained original Nocturna archive and its documented coordinate key and data limits. Our code and seasonal date analysis are new.
- usesMoth catch changes had higher correlation for the same species across 18 Finnish trap pairs 20–200 km apart in 1993–2019
We reuse its retained raw source archive and inspect its coverage table for feasibility. We independently rebuild counts and estimate a different spatial correlation contrast.
- usesRecorded moth communities showed no clear change in spatial similarity at 22 Finnish light-trap locations between 1993–2002 and 2010–2019
We reuse its retained source archive and its location, binomial-name and coverage conventions. We rebuild seasonal tables from source records and test spatial composition.
- usesEqual-count sampling removes the recorded decrease in moth list differences at 19 Finnish traps from 1993–1999 to 2013–2019
We use its retained Nocturna source archive and its trap and species-name conventions. We test spatial differences between fixed traps.
- usesThe variance of annual moth catch changes was above a zero-covariance reference at 11 Finnish traps, 1993–2019
This verified publication supplied the retained source archive. We rebuilt all seasonal tables from unchanged source bytes.
- usesMoth communities at 37 Finnish traps showed no clear spatial homogenization between 1993–2002 and 2010–2019
We reuse its retained source archive and its coordinate and species-name rules. The new analysis tests spatial community differences.
- usesMoth catch changes had higher correlation for the same species across 18 Finnish trap pairs 20–200 km apart in 1993–2019
We reuse its retained raw source archive and inspect its coverage table for feasibility. We independently rebuild counts and estimate a different spatial correlation contrast.