Image-based methods for species identification offer cost-efficient solutions for biomonitoring. This is particularly relevant for invertebrate studies, where bulk samples often represent insurmountable workloads for sorting, identifying, and counting individual specimens. On the other hand, image-based classification using deep learning tools have strict requirements for the amount of training data, which is often a limiting factor. Here, we examine how classification accuracy increases with the amount of training data using the BIODISCOVER imaging system constructed for image-based classification and biomass estimation of invertebrate specimens. We use a balanced dataset of 60 specimens of each of 16 taxa of freshwater macroinvertebrates to systematically quantify how classification performance of a convolutional neural network (CNN) increases for individual taxa and the overall community as the number of specimens used for training is increased. We show a striking 99.2% classification accuracy when the CNN (EfficientNet-B6) is trained on 50 specimens of each taxon, and also how the lower classification accuracy of models trained on less data is particularly evident for morphologically similar species placed within the same taxonomic order. Even with as little as 15 specimens used for training, classification accuracy reached 97%. Our results add to a recent body of literature showing the huge potential of image-based methods and deep learning for specimen-based research, and furthermore offers a perspective to future automatized approaches for deriving ecological data from bulk arthropod samples.

Changing abiotic conditions can affect the phenology of animals and plants with implications for their reproductive output, especially in rapidly changing regions like the Arctic. For instance in arthropods, it was recently shown that females of the spider species Pardosa glacialis (Thorell, 1872) (Lycosidae) are able to produce two clutches within one growing season in years when snowmelt occur particularly early. This phenomenon could be widespread in northern latitudes, and here we investigated the voltinism of two other very abundant species of wolf spiders in the Low-Arctic, Pardosa hyperborea (Thorell, 1872) and Pardosa furcifera (Thorell, 1875), over the period 2015 -2017. While a bimodal pattern in the clutch size frequency distribution was only revealed for P. hyperborea, we were able to show that both species can produce a second clutch over the active season by using information on the embryonic stages. We also observed significantly larger first than second clutches. We argue that information about the embryonic stage can be critical for evaluating evidence of wolf spider populations producing more than one clutch in a season. Our study provides evidence that bivoltinism could be more widespread pattern than expected in Arctic wolf spiders. It remains to be investigated what the trophic consequences of such patterns are in a global warming context. We thus highlight the need for a coordinated framework for such further studies, integrating and relating various functional traits.

No abstract

The Arctic tundra is characterised by harsh conditions and environmental gradients are especially pronounced. Variation in functional traits along such gradients provide insights into the drivers of species abundance and distribution, and are particularly valuable in this region currently experiencing strong climate warming. Over three consecutive years, we analysed the interacting effect of two environmental factors, habitat and elevation, on the abundance, body size and clutch size in two common Low-Arctic invertebrate predators (Lycosidae, Araneae), Pardosa furcifera (Thorell 1875) and Pardosa hyperborea (Thorell 1872). Using generalised linear models, we firstly showed a habitat partitioning between P. furcifera, which dominated wet habitats like fens, and P. hyperborea, which was more associated with drier habitats like shrubs. Secondly, we found smaller body sizes at high elevation in P. hyperborea, a species that has a southern distribution in Greenland, and we identified season length as a major driver of the development in this species. In P. furcifera, a species likely more cold-adapted, we found no body size difference between elevations, suggesting that local conditions (e.g. prey availability or snowmelt timing) are more important in explaining body size variations. Finally, body size and clutch size were strongly correlated in both species, but clutch size was not affected by habitat or elevation. We argue that fecundity is likely influenced by trade-offs, and that considering additional complementary trait measurements would allow for a better understanding of the mechanisms underlying patterns in species life-history traits along environmental gradients.