1. INTRODUCTION

Agricultural orchards adopt various forms of protected cultivation, including plastic greenhouses and net-houses, in order to stabilize production and buffer crops from climatic or biotic stressors (Kittas et al. 2012;Villagrán et al. 2021). Although such structures can improve yield reliability, they also alter habitat openness and create physical barriers that limit arthropod movement between orchards and adjacent habitats (Lawson et al. 1994). These restrictions may reduce ecological interactions, constrain dispersal opportunities, and weaken trophic interactions. Protected cultivation can also generate distinct microclimatic conditions within enclosed spaces, and these conditions may vary depending on the specific management regime (Madzaric et al. 2018;Messelink et al. 2021).

Arthropods perform key ecological functions within orchard systems by contributing to herbivory regulation, predation, decomposition, and pollination (Tscharntke et al. 2012). Their assemblages respond sensitively to habitat structure, management intensity, and environmental disturbance, which makes them valuable ecological indicators in sustainable and ecofriendly agricultural systems (Tscharntke et al. 2005;Tuck et al. 2014). In addition, these ecological traits are often utilized within protected cultivation systems to support biological control and regulate pest populations (Chailleux et al. 2022). However, arthropod assemblages in such enclosed environments can become simplified or increasingly homogenized due to restricted movement, reduced connectivity, and altered microclimatic conditions (Tscharntke et al. 2012;Messelink et al. 2021). Despite the ecological implications of these changes, the community-level consequences of protected cultivation have not yet been examined in sufficient detail. Most previous research has focused on pest population dynamics or microclimatic responses, whereas relatively little attention has been given to shifts in biodiversity or the structural properties of arthropod assemblages under protected cultivation (Messelink et al. 2021;Villagrán et al. 2021;Chailleux et al. 2022).

In this study, we investigate how greenhouse structure influences arthropod diversity and assemblage composition in eco-friendly managed orchards. We compared diversity indices, species-estimation metrics, and assemblage differences among orchards that differ in enclosure. We expected that species richness decrease and overall assemblage structure would change under more enclosed greenhouse structures. We also anticipated that diversity among orchards would decline as structural enclosure increased, resulting in stronger assemblage homogenization. Understanding these structural effects is essential for interpreting arthropod-based ecological indicators and for informing management strategies that aim to maintain biodiversity and ecological function in eco-friendly orchard landscapes.

2. MATERIALS AND METHODS

2.1. Study site

Four eco-friendly managed orchards were selected for this study (Fig. 1). Two sites consisted of open field orchards, and the other two were greenhouse orchards. Three of the orchards cultivated grapes, located in Cheongyang (CY), Hampyeong (HP), and Damyang (DY), while the orchard in Yecheon (YC) cultivated apples (Table 1).

2.2. Survey method

Arthropods were collected using pitfall traps and yellow dish traps. At each orchard, five pitfall traps were installed at 5 m intervals, and three yellow dish traps were installed at 10 m intervals along a diagonal transect. Pitfall traps contained 70% ethanol and propylene glycol mixed preservative solution (7 : 3). Yellow dish traps were filled with a mixture of water, sugar, and neutral detergent, which acted as an attractant. After a three-day deployment period, all collected materials were transported to the laboratory, where arthropods were sorted from debris and transferred into 20 mL glass vials containing 70% ethanol. Sampling was conducted in June and September 2025, corresponding to the principal activity periods of arthropods in orchard environments.

2.3. Specimen processing and identification

Specimens were examined under microscope (SMZ- 645; Nikon corporation, Tokyo, Japan) and identified using taxonomic keys published by the National Institute of Biological Resources and associated literature (Yoon and Cheong 2012;Park et al. 2014;Kim and Lee 2018;Cho 2019;Seo 2019). Species-level identification was attempted for Araneae, Coleoptera, Hemiptera, Hymenoptera, and Diptera. Individuals that could not be identified to species due to morphological damage or insufficient diagnostic characters were assigned to morphospecies or identified to genus. Voucher specimens were deposited as Entomological Collection of the Organic Agriculture Division, National Institute of Agricultural Sciences, and stored in a 2°C refrigerator.

2.4. Diversity and assemblage analyses

Based on the compiled species list, species richness, Shannon diversity, and inverse Simpson diversity indices were calculated for each site by pooling data from pitfall traps and yellow dish traps. Shannon and Simpson diversity were quantified using the Hill-number framework (Jost 2006). To evaluate differences in species diversity across cultivation structures, we employed interpolation and extrapolation procedures using the iNEXT package version 3.0.2 in R (Hsieh et al. 2016). iNterpolation and EXTrapolation (iNEXT) analysis integrates sample-size-based and coverage-based rarefaction and extrapolation to generate diversity estimates and standardized species accumulation curves. Extrapolated curves were generated with 95% confidence intervals.

Arthropod assemblage composition was analyzed using non-metric multidimensional scaling (NMDS) based on Bray-Curtis dissimilarity. Abundance data were square-root transformed to minimize the influence of highly dominant taxa. Differences in community composition among regions were tested using permutational multivariate analysis of variance (PERM ANOVA) based on Bray-Curtis dissimilarity. Analyses were conducted the “adonis2” function of the vegan package, with region included as the explanatory factor. Statistical significance was assessed using 999 permutations. Permutational analysis of multivariate dispersions (PERMDISP) was used to test for difference in multivariate dispersion among regions by comparing distances of samples to their group centroids based on Bray-Curtis dissimilarity. All statistical analyses were conducted in R version 4.5.1 (R Core Team 2025) using the vegan package version 2.7-2 (Oksanen et al. 2025).

3. RESULTS

A total of 1,961 individuals belonging to 141 species were collected across all sites. Among the major arthropod groups, Araneae (39 species, 241 individuals), Coleoptera (35 species, 75 individuals), Hemiptera (21 species, 252 individuals). Hymenoptera (17 species, 1,105 individuals), and Diptera (15 species, 89 individuals) accounted for the highest species richness. In addition, a small number of species from Dermaptera, Isopoda, Lepidoptera, and Orthoptera were also collected.

iNEXT analyses revealed substantial differences in species richness between the open and closed community groups. The observed species richness (q=0) was 121 species in the open group, approximately three times higher than the 41 species recorded in the closed group. Even when extrapolated to twice the current sample size, estimated richness remained notably higher in the open group than in the closed group. Sample coverage values were 97.6% for the open assemblage and 93.5% for the closed assemblage.

Asymptotic estimators further supported this pattern. The potential species richness of the closed assemblage was estimated at approximately 68 species, whereas the open assemblage was estimated at 168 species indicating that the closed assemblage represents only about 40% of the richness expected in the open assemblage. Shannon diversity (q=1) was slightly higher in the open group compared with the closed group. In contrast, inverse Simpson diversity (q=2) showed the opposite trend, with higher values in the closed assemblage than in the open assemblage. Overall, these results indicate that open-field assemblages exhibit markedly higher species diversity than closed assemblages (Table 2).

NMDS ordination based on Bray-Curtis dissimilarity showed clear compositional differences among the four orchard assemblages (stress≈ 0.17526; Fig. 3). The two greenhouse grape orchards (DY and HP) formed a tight cluster with substantial overlap, indicating high similarity between their assemblages. The partially sheltered vineyard (CY) was positioned between the greenhouse grape and open-field apple orchard. The open-field apple orchard (YC) was clearly separated from the vineyard sites and occupied a distinct region of the ordination space. These patterns indicate that both greenhouse structure and crop characteristics contributed to the observed variation in arthropod assemblage composition.

Following the NMDS, Bray-Curtis PERMANOVA confirmed that assemblage composition differed significantly among regions (F=2.99, p=0.001). The PERMDISP test indicated no significant difference in multivariate dispersion (F=0.038, p=0.990). The pairwise PERMANOVA analysis also revealed significant differences among most regional pairs after FDR correction, whereas the DY and HP comparison was not significant (Table 3), indicating that assemblage structure in these two regions has converged toward a similar assemblage, likely reflecting their shared condition of being managed under greenhouse cultivation with same crop.

4. DISCUSSION

Arthropod assemblages in this study exhibited pronounced differences between open-field and greenhouse orchard systems, with open-field sites consistently supporting higher species richness, abundance, and diversity indices. Rarefaction and extrapolation further indicated that open-field orchards contained substantially more estimated species than greenhouse orchards, underscoring the strong influence of structural openness and habitat heterogeneity on arthropod communities.

Although open-field orchards exhibited markedly higher species richness and Shannon diversity, their inverse Simpson diversity (q=2) was lower than that of greenhouse orchards. Because inverse Simpson diversity in the Hill-number framework corresponds to the effective number of dominant species and is more sensitive to the relative abundances of common taxa than to species richness per se (Jost 2006), this pattern may indicate a relatively stronger dominance structure in openfield assemblages. The higher connectivity of openfield orchards with surrounding habitats could facilitate the influx of highly mobile or opportunistic taxa (Hendrickx et al. 2007;Tscharntke et al. 2012), potentially leading to numerical dominance by a subset of species despite high overall richness. In contrast, greenhouse orchards support fewer species overall, but environmental filtering associated with enclosure may result in more even relative abundances among the taxa that persist (Gossner et al. 2016).

Greenhouse orchards showed a marked degree of assemblage homogenization. The physical boundaries created by greenhouse structures restrict arthropod movement across the orchard-landscape interface, limiting both external immigration and internal dispersal (Messelink et al. 2021;Chailleux et al. 2022). Such constraints reduce opportunities for trophic interactions, habitat connectivity, and competitive encounters, filtering community composition toward taxa capable of persisting under reduced connectivity or toward assemblages already present within the enclosed system (Madzaric et al. 2018;Bertellotti et al. 2023). This effect was clearly reflected in our results, particularly in the nearcomplete overlap of DY and HP in the NMDS ordination, along with the markedly lower species diversity and population sizes observed in greenhouse orchards compared to open-field sites. Additionally, the structural enclosure of greenhouses may generate environmental simplification that imposes physiological or behavioral pressures on arthropods, which in turn could further reinforce the community convergence observed in these systems (Kittas et al. 2012;Gossner et al. 2016).

The grape open-field orchard (CY) occupied an intermediate position between the apple open-field and greenhouse grape orchards. This intermediate positioning likely reflects a combination of crop identity shared with the greenhouse vineyards and the partially exposed conditions of CY, which permit greater arthropod influx than fully enclosed systems. In the same context, a distinct pattern emerged in the apple open-field orchard (YC), which separated from all vineyard sites. This differentiation likely reflects substantial structural differences between apple and grape cultivation systems, including canopy architecture, tree form, bark characteristics, and understory species composition (Sullivan et al. 2023). In addition to these factors, previous research in orchard systems has shown that variation in fruit type can influence soil physical and chemical properties as well as arthropod community characteristics, due to plant-soil feedbacks and associated vegetationsoil interactions (Wee et al. 2023). The combined influence of crop identity and fully open-field conditions appears to have produced the unique assemblage observed in YC.

Collectively, our findings suggest that structural openness exerts a stronger influence on arthropod assemblages. Protected cultivation represents a form of landuse intensification, and such intensification can reduce biodiversity and impair ecosystem services by simplifying habitats and limiting ecological interactions (Hendrickx et al. 2007;Tuck et al. 2014). Despite the clear patterns observed, this study has limitations. The number of sampling sites and traps was relatively small, and features such as rain shelters and crop identity were not quantified in detail. Microclimatic variables were also not measured, and any interpretation of their influence remains speculative. However, greenhouse systems generally differ from open orchards in terms of microclimatic and soil conditions, including higher temperatures, reduced wind exposure, and lower temporal variability, as wells as a tendency for nutrient accumulation under intensive management (Kittas et al. 2012;Villagrán et al. 2021). These structural differences can environmental filters shaping arthropod community composition (Gossner et al. 2016;Madzaric et al. 2018).

In this context, the differences detected across diversity indices and multivariate analyses indicate that the structural openness plays a pivotal role in shaping arthropod communities within orchard environments (Messelink et al. 2021). Protected cultivation environments promote simplified and more uniform assemblages due to restricted connectivity with surrounding habitats, whereas open-field orchards support richer and more heterogeneous communities. This contrast is consistent with broader evidence showing that habitat configuration, land-use intensity, and connectivity strongly mediate arthropod diversity patterns in agricultural landscapes (Hendrickx et al. 2007;Gossner et al. 2016;Bertellotti et al. 2023). These patterns also suggest the potential for trade-offs between production stability and biodiversity conservation in orchard systems, highlighting the need to identify management strategies that balance both goals (Bommarco et al. 2013;Seufert and Ramankutty 2017). While protected systems can enhance environmental control and management efficiency, reduced habitat connectivity and structurally buffered conditions may constrain the diversity of arthropod communities and associated ecological functions (Madzaric et al. 2018;Messelink et al. 2021). Broader surveys encompassing additional orchard types, structural configurations, and landscape contexts with long-term monitoring will be essential for evaluating these trade-offs and refining biodiversityfriendly management approaches.