Abstract:

Aims: Environmental DNA (eDNA) metabarcoding technology offers a novel approach for assessing soil arthropod diversity. This study aimed to compare the performance of three eDNA sampling methods for detecting soil arthropods. 

Methods: Soil samples were collected from eight crop fields. Three eDNA sampling methods were employed: direct soil extraction (SOL), air filtration (AIR), and soil water filtration (WAT). Arthropod community composition was analyzed using high-throughput sequencing. 

Results: Species richness was significantly higher in WAT than in SOL and AIR (P < 0.05), with no significant difference between SOL and AIR. The three methods also differed considerably in beta diversity. Permutational multivariate analysis of variance (PERMANOVA) revealed highly significant differences in soil arthropod community structures among the three sampling methods (R² = 0.2816, P = 0.001), indicating that the sampling method explained 28.16% of the total variation in community composition. Principal coordinates analysis (PCoA) further showed that samples from the three methods formed three distinct clusters along the ordination plot. SOL detected the highest number of operational taxonomic unit (OTUs; 3,314), followed by AIR (1,023) and WAT (871). At the family level, the relative abundance of soil-dwelling groups such as Isotomidae detected by the SOL was higher, but the differences among treatments were not significant (P > 0.05). WAT exhibited the highest relative abundance of Chironomidae and AIR demonstrated superior detection efficiency for Scarabaeidae. Among natural enemy groups, Staphylinidae and Empididae were detected by all three methods. Network analysis revealed dense OTU clusters around SOL nodes in SOL–AIR and SOL–WAT comparisons, while in the AIR–WAT comparison, high-abundance OTUs were largely method-specific. Rank occurrence curves indicated that SOL detected the most OTUs (169) and showed the most even distribution, followed by WAT (113) and AIR (83), with AIR dominated by a few abundant OTUs. 

Conclusion: The distribution of eDNA in the solid, liquid, and gas phases of soil is closely related to the ecological niche and behavioral types of arthropods, exhibiting heterogeneous distribution. SOL offers stable and broad-spectrum detection of soil-dwelling groups and various natural enemy taxa. AIR shows advantages in detecting groups such as Diptera and Scarabaeidae. WAT is particularly effective in enriching taxa like Chironomidae that prefer moist environments. For biodiversity assessment of farmland soil, the SOL is recommended as the foundation. If the focus is on specific taxa such as Diptera or Coleoptera, the AIR can be added as a supplement to SOL. If the study involves taxa that prefer moist environments, such as Chironomidae, the WAT should be additionally incorporated. This study provides a scientific basis for understanding the spatial distribution of eDNA in soil ecosystems and for the optimizing methods for monitoring soil arthropod diversity in agricultural fields.

Key words: environmental DNA, soil arthropods, agroecosystem, biodiversity monitoring, DNA metabarcoding