Reconciling biodiversity conservation with agricultural production requires a better understanding of how key ecosystem service providing species respond to agricultural intensification. Carabid beetles and spiders represent two widespread guilds providing biocontrol services. Here we surveyed carabid beetles and spiders in 66 winter wheat fields in four Northwestern European countries and analyzed how the activity density and diversity of carabid beetles and spiders were related to crop yield (proxy for land-use intensity), percentage cropland (proxy for landscape complexity) and soil organic carbon content, and whether these patterns differed between dominant and non-dominant species. Less than 17% of carabid or spider species were classified as dominant, which accounted for more than 90% of individuals respectively. We found that carabids and spiders were generally related to different aspects of agricultural intensification. Carabid species richness was positively related with crop yield and evenness was negatively related to crop cover. The activity density of non-dominant carabids was positively related with soil organic carbon content. Meanwhile, spider species richness and non-dominant spider species richness and activity density were all negatively related to percentage cropland. Our results show that practices targeted to enhance one functionally important guild may not promote another key guild, which helps explain why conservation measures to enhance natural enemies generally do not ultimately enhance pest regulation. Dominant and non-dominant species of both guilds showed mostly similar responses suggesting that management practices to enhance service provisioning by a certain guild can also enhance the overall diversity of that particular guild.

We used data on natural enemy communities in 66 paired winter wheat fields in four Northwestern European countries (Germany, the Netherlands, Sweden and United Kingdom) to investigate the response of natural enemy communities to landscape complexity, local land-use intensity and soil organic matter content, and specifically examined whether and how responses differ between dominant and non-dominant species. We focused on carabid beetles and spiders as they represent the two groups of natural enemies in arable fields in Northwestern European and widely used as bioindicators (Lang et al., 1999; Borchard et al., 2014). We used pitfall traps to collect carabids and spiders in field pairs that covered a gradient in land-use intensity and landscape complexity, with fields within pairs having contrasting soil organic carbon content. 

Pitfall traps (polypropylene beakers 155 mm high and 95 mm across) were used to survey ground-dwelling arthropods during the wheat flowering season (late May to early June). We placed one pitfall trap in the center of each treatment subplot at least 10 m from the field edge and filled it with 200 mL of a mixed solution of 2/3 water and 1/3 glycol and a drop of detergent to lower surface tension. A square aluminum plate was placed approximately 10 cm above each pitfall trap to prevent flooding by rain. Pitfall traps were opened for 10 days. All of the collected arthropods were stored in 70% ethanol solution for later identification. For the purpose of our study, the two most abundant species groups, carabid beetles (Carabidae) and adult spiders (Araneae), were selected as our bioindicators and they were counted and identified to species level using standard keys (Hackston, 2020; Nentwig et al., 2021). We determined the diet preference of each carabid beetle species based on Larochelle (1990) and the hunting strategy of all observed spider species based on Cardoso et al. (2011) following Gallé et al. (2019). Furthermore, because the arthropod communities will inevitably differ in composition between countries, we classified the carabids or spiders as nationally dominant and non-dominant species based on whether species made up respectively more or less than 5% of the total number of individuals caught of each species group in a country following Kleijn et al. (2015).