The introduction of non-native tree species has become a global concern and may disrupt native communities and related ecosystem functions. Soil food webs regulate organic matter decomposition and nutrient cycling in forests with their feeding activities, but evaluating the consequences of tree species introduction on soil invertebrates is challenging due to the complex trophic structure and wide range in body size of soil invertebrates. Here, we employed an energetic food web approach and estimated the energy flux in soil food webs using a four-node model including soil meso- and macrofauna decomposers and predators. We examined pure and mixed stands of native European beech (Fagus sylvatica), introduced Douglas fir (Pseudotsuga menziesii), and native range-expanding Norway spruce (Picea abies) across site conditions. Compared to native forests, introduced tree species reduced the total mass of macrofauna predators by 92% at sandy sites but not that of decomposers, suggesting trophic downgrading in soil food webs by Douglas fir. The energy flux in mixed forests was intermediate between respective monocultures, suggesting that tree mixtures mitigate potential negative impacts of introduced tree species on food web functioning. Across size classes, soil macrofauna responded more sensitively to changes in environmental conditions than soil mesofauna. Despite the lower total mass, the energy flux through mesofauna outweighed that through macrofauna when considering energy loss to predators, highlighting the importance of mesofauna for decomposition processes in forest soil food webs. Additionally, total energy flux positively correlated with species richness, pointing to the significance of soil biodiversity for trophic functionality. Overall, the study emphasizes the critical role of tree species composition, site conditions, and soil biodiversity in driving energy flux through soil food webs and maintaining forest ecosystem functions.

Soil samples were taken to evaluate the population density of soil meso- and macrofauna between November 2017 and January 2018. At an equidistance from trees of the same (pure stands) or different species (mixed stands), two small (⌀ 5 cm) and two large (⌀ 20 cm) soil cores were taken at a distance of approximately 10 m close to the center of each plot. Soil cores were separated into litter (OL/F), 0–5 cm, and 5–10 cm depth and transported to the laboratory for extraction of soil meso- and macrofauna using heat. Mesofauna (Acari, Collembola, Diplura, Pauropoda, Protura, Symphyla) were extracted from small cores, and Pseudoscorpionida and macrofauna were extracted from large cores. The extracted animals were collected in 50% diethylene glycol and then transferred into 70% ethanol for determination.

Animal taxa were first sorted and assigned into functional groups. Adults of soil microarthropods (Collembola, Mesostigmata, and Oribatida) were identified to species level, and other soil invertebrates were identified to taxonomic levels sufficient for functional guild assignment using a stereomicroscope and a microscope. In particular, Coleoptera were identified to families, including juveniles, allowing trophic assignment to predators or decomposers. Species diversity was estimated based on one soil core for soil Collembola, Mesostigmata, and Oribatida, and two soil cores for Pseudoscorpionida and macrofauna (Araneae, Coleoptera, Isopoda, Myriapoda, Opiliones, Wenglein et al. in preparation). Decomposer groups included Collembola, Diplopoda, Diplura, Elateridae, Isopoda, Lumbricidae, Oribatida, Pauropoda, Protura, Scarabaeidae, and Symphyla; predator groups included Araneae, Cantharidae, Carabidae, Chilopoda, Coccinellidae, Dermestidae, Mesostigmata, Opiliones, Pseudoscorpionida, and Staphylinidae.

For microarthropods, species-specific body sizes were used to estimate their fresh mass. Exceptions were juveniles of Oribatida and Mesostigmata, which were grouped into size classes. Different allometric equations were used for adults and juveniles of mites, because no species key exists for mite juveniles, and only body length was estimated. Individual body size of other animal groups was measured from one soil core in each plot using a camera system (Zeiss Stemi 2000-CS with Zeiss AxioCam ICc1). Where there were more than ten individuals present in each morphospecies, ten individuals were randomly selected for body size measurements. Individual biomass was estimated from body size using group-specific power equations based on either body length or both length and width (Table S2). Individual metabolic rate was estimated based on group-specific allometric scaling equations, using mean annual air temperature in each plot. Total fresh mass and community metabolism were summed up from individual-estimated data and used for energy flux models.