Flattening the curve: approaching complete sampling for diverse beetle communities

Insects are a hyper diverse and ecologically important group. Their high diversity, however, presents challenges in sampling methodology, because rare species are unreliably detected with low sampling effort. However, the relationship between effort and species detections, critical for effective monitoring and evaluation of population trends, is too seldom quantified. We sampled forest beetles for 3 months in a 4-ha stand of mixed deciduous forest in southeastern Norway using 110 flight intercept (four types) and Malaise traps, the highest trap density (29 traps ha(-1)) that we have seen reported. We examined species accumulation curves to quantify the benefits of each additional trap, compared capture rates among several trap designs and trap emptying frequencies, and tested for spatial autocorrelation. In total, we captured 566 beetle taxa (19 854 individuals) from 52 families, yet our species accumulation curve was only beginning to flatten. Trap types differed considerably in their effectiveness. Nevertheless, 20 of our most effective window traps detected 75% of all taxa in our dataset. We found no evidence of spatial correlation within the scale of the study (100 m radius), nor did trap-level forest covariates (5 m radius) explain much variation. This implies that low-to-moderate sampling effort dramatically underestimates species richness, but that a limited number of effective traps can nonetheless achieve relatively thorough sampling for some applications. Immediate trap surroundings and spacing appeared unimportant. Insect ecologists should take particular care in selecting trap types, and be cautious comparing studies that employed different trap types.