This study set out to investigate several factors regarding the beetle families Cerambycidae and Elateridae. As well as investigate the potential spread of P. cylindrus. This study found new sites for threatened species and no signs of spread of the P. cylindrus. This study sought to find specific interactions between Cerambycids and Elaterids, as well as find how environmental factors affect species diversity and specific species. But also, how reliable species identification applications are when compared to an expert.
Species richness
Statistically, no connection between latitude and species richness was observed, in contradiction to established concepts. Despite some studies not reporting statistically significant declines, observable declines were still present. Supporting that a real decline was detected in this study. Studies with no significance had a lower latitudinal range. This is the case in this study with a small range from 55.5° to 60.8°. Stronger cases of latitudinal diversity could also be observed at larger scales than smaller ones. The results also showcase that the number of species is predicted to decline with a higher latitude, but these predictions are deemed mostly inaccurate. Dispersal ability plays a part in how steep the latitudinal diversity gradient line is. Beetles with higher dispersal ability will have a flatter gradient than beetles with a limited dispersal ability. This might explain why Cerambycids were observed further north than Elaterids. One reason why dispersal ability affects the northern limit is the most recent ice age. Which forced species to move south, and species are yet to move back and reach their northern limit.
Interaction matrix
The bipartite network does not give any indication of stronger connections between specific species. A very convoluted web shows that all Cerambycidae have interactions with all the saproxylic Elateridae. Indicating that no cerambycid is subject to larger predation pressure from any elaterid. The interaction matrix shows the wide and non-specialised predator-prey interactions between the families. Supported by the H2 value and also partially supported by the moderately high connectance value. Both nestedness values indicate a degree of interaction overlap, where specialist species interact with generalist species. The low modularity value and high number of modules suggest that this network is overall open but has many smaller loose groups. Higher connectance and nestedness values have been linked to more robust communities that better withstands disturbances, and since the values observed in this study are moderately high, the network is likely to be relatively resilient. But larger disturbances may still pose a threat and could pose a problem for local populations. Since the network has many smaller modules, it is not a threat on a larger scale though.
Land cover
Most species in both Cerambycidae and Elateridae are indifferent if forests are required, indicating that most of the tested species will be able to survive in outside forests. But most of the tested species have a positive relationship to forests. A two-year study found that a large number of species can persist in a fragmented landscape. Their results were comparable to findings from other regions, supporting the idea that cerambycids are capable of surviving in more fragmented forest habitats. The forest-dependent species Poecilium alni requires deciduous trees and branches that recently died for reproduction, which are more abundant in forests. Its dependence on forests means that only forests have the sufficient level of resources. The same logic applies for the other forest-dependent species as well. The statistically significant forest-avoiding species, Anaglyptus mysticus, is mainly connected to old hazel (Corylus avellana) trees. Old C. avellana trees in Sweden are mainly found in mixed deciduous forests and in zones between forests and open land. The latter environment would place C. avellana trees partially outside the forest land cover category thus creating an avoidance to forests. Larger forests have been shown to harbour more species, likely due to a higher amount of resources, which in turn explain the overall positive relationship with forest cover. Testing only 20 species poses a problem since it likely focuses on the more common species while overlooking others. More data and observations would be needed to be able to cover more species.
Tree species composition
Only the proportion of oaks showed a meaningful effect on species diversity. Oaks (Quercus sp.) are one of the most important species for biodiversity. The oaks are also very important for saproxylic species; they could work as safe havens for dependent species in an otherwise ill-suited habitat, such as spruce production forests. Oaks have a high level of biodiversity and could complement oaks in pastures, which themselves also have high levels of biodiversity, highlighting the oaks as keepers of biodiversity in several types of environments. The oaks in forests also showcase a higher amount of deadwood and are thus suitable for saproxylic beetles. This may although only be relevant in standing forests since saproxylic beetle species richness is negatively affected by forest growth, compared to oaks in open environments. Maintaining oaks in open landscapes remains important; however, oaks within standing forests also provide essential habitats. The larger buffer sizes allow more oaks within an area, thus increasing suitability. But it is most likely not relevant at buffer sizes larger than 2 500 metres. The connection between the number of saproxylic beetle species and large oaks is strongest at that distance. This in turn means that the oaks outside the buffer matter less.
Identification applications
Studies have shown that models can accurately identify the correct species, but their success rate depend on variance in appearance. This could explain why Artsorakel had a lower success rate for Elateridae, since they are overall more similar to each other Cerambycids. The success rate also depends on what the models could compare with and the training data available. Both websites use model services from the same source and the models is trained on user submissions. Given that both websites use models from the same source, it is interesting that their accuracy differs for Elateridae. While it is unclear whether they use the same training resources, this could explain the variation in performance. The amateur who used an identification key achieved the highest success rate due to how species keys are built. This success was further supported by the fact that many Cerambycids have distinct patterns and colouration. The average image quality likely had a negative impact on the success rate of the tested applications. These applications use pixel-based pattern recognition and compare input images with a reference database. It is therefore reasonable to assume that higher-resolution images would yield better results, but the images stored in the database also matters.