Changes in insect biomass, abundance, and diversity are challenging to track at sufficient spatial, temporal, and taxonomic resolution. Camera traps can capture habitus images of ground‐dwelling insects. However, currently sampling involves manually detecting and identifying specimens. Here, we test whether a convolutional neural network (CNN) can classify habitus images of ground beetles to species level, and estimate how correct classification relates to body size, number of species inside genera, and species identity. We created an image database of 65,841 museum specimens comprising 361 carabid beetle species from the British Isles and fine‐tuned the parameters of a pretrained CNN from a training dataset. By summing up class confidence values within genus, tribe, and subfamily and setting a confidence threshold, we trade‐off between classification accuracy, precision, and recall and taxonomic resolution. The CNN classified 51.9% of 19,164 test images correctly to species level and 74.9% to genus level. Average classification recall on species level was 50.7%. Applying a threshold of 0.5 increased the average classification recall to 74.6% at the expense of taxonomic resolution. Higher top value from the output layer and larger sized species were more often classified correctly, as were images of species in genera with few species. Fine‐tuning enabled us to classify images with a high mean recall for the whole test dataset to species or higher taxonomic levels, however, with high variability. This indicates that some species are more difficult to identify because of properties such as their body size or the number of related species. Together, species‐level image classification of arthropods from museum collections and ecological monitoring can substantially increase the amount of occurrence data that can feasibly be collected. These tools thus provide new opportunities in understanding and predicting ecological responses to environmental change.

Environmental DNA (eDNA) metabarcoding is increasingly being implemented as a non‐invasive and efficient approach for biodiversity research and monitoring across ecosystems. However, accurate detection of species with eDNA requires robust experimental designs as eDNA analysis carries a risk of contamination at every step of the fieldwork and laboratory processes. Several studies focus on rigorous laboratory procedures and processing of sequencing data, but surprisingly, little research investigates the process of background input of DNA in the field. For example, airborne DNA from localities outside the study area could potentially contaminate eDNA samples. Here, we use an experimental setup and eDNA metabarcoding to study the diversity and accumulation of airborne eukaryotic eDNA on exposed surfaces in the field. At two different natural locations, a coastal marine site and a terrestrial grassland site, we placed open containers each filled with 0.5 liters of water, which was then sampled at eight successive time points after exposure to the surroundings. We found an accumulation of detected species richness in the samples, which reached its maximum at the end of the experiment, 24 h after exposure. This result was consistent across both sites and across two markers (COI for eukaryotes and 12S for vertebrates). While many of the detected species were contaminants commonly found in eDNA studies, we also detected several other eukaryotic taxa. Most notable were metazoan species such as birds, fish, and insects, likely originating from airborne transport of eDNA. We also found that increasing the number of PCR cycles tended to have a positive impact on richness for the unfiltered reads but a negative impact on the richness after bioinformatic filtering. Our results add to the sparse evidence that metazoan eDNA can be transported by air, which have wide implications for eDNA research and calls for increased implementation of field control samples.

Here, we investigate the evolutionary history of five northern dragonfly species to evaluate what role the last glaciation period may have played in their current distributions. We look at the population structure and estimate divergence times for populations of the following species: Aeshna juncea (Linnaeus), Aeshna subarctica Walker, Sympetrum danae (Sulzer), Libellula quadrimaculata Linnaeus and Somatochlora sahlbergi Trybom across their Holarctic range. Our results suggest a common phylogeographic pattern across all species except for S. sahlbergi. First, we find that North American and European populations are genetically distinct and have perhaps been separated for more than 400,000 years. Second, our data suggests that, based on genetics, populations from the Greater Beringian region (Beringia, Japan and China) have haplotypes that cluster with North America or Europe depending on the species rather than having a shared geographic affinity. This is perhaps a result of fluctuating sea levels and ice sheet coverage during the Quaternary period that influenced dispersal routes and refugia. Indeed, glacial Beringia may have been as much a transit zone as a refugia for dragonflies. Somatochlora sahlbergi shows no genetic variation across its range and therefore does not share the geographic patterns found in the other circumboreal dragonflies studied here. Lastly, we discuss the taxonomic status of Sympetrum danae, which our results indicate is a species complex comprising two species, one found in Eurasia through Beringia, and the other in North America east and south of Beringia. Through this study we present a shared history among different species from different families of dragonflies, which are influenced by the climatic fluctuations of the past.

A breeding colony of ivory gulls was discovered on an ice floe in northeast Greenland in August 2008. The ice floe resembled nearby islands in that it was covered with a thick layer of gravely moraine, and furthermore its position was fixed throughout most of the breeding season as the surrounding first year ice only broke up in mid August when most of the gull chicks had fledged.