Carabid assemblage
The results show that the species assemblage of ground beetles in wooded pastures includes a large number of species common to both forests and non-wooded pastures. Hypothesis (i) is thus confirmed since ground beetle communities in wooded pastures largely encompass grassland and forest communities.
Our study confirms similar results obtained by Boutaud et al. (2022) in Germany and by Talvi et al. (1995) in Estonia. Even in the tropics, our results corroborate those of Ariza et al. (2021). Indeed, early successional habitats in Colombia, having undergone agricultural abandonment between 3 and 7 years, presented an assemblage of communities that overlapped with pasture and forest communities. The early successional habitat would be a mixed environment with ecological elements of pastures and forests, thus a heterogeneous environment that could be exploited by a subset of ground beetle species.
According to the results, it appears that the wooded pastures of the Jura mountains act as a transition zone between forests and non-wooded pastures. Indeed, some ground beetle communities find similar ecological niches in the WP as in the F and NWP. Ground beetle communities in WP tend towards grassland assemblages when the semi-open area is sparsely wooded. In contrast, WP tend towards a forest assemblage when they have a higher cover of trees (> 50% of TC). The ecological characteristics of wooded pastures are quite varied and can constitute a substitute habitat for certain eurytopic species. The heterogeneity of the WP, therefore, results in the assemblages of certain species that usually live in contrasting environments. Similar results for the family Geotrupidae have shown that wooded pastures support a dung beetle assemblage that is transitional between F and NWP assemblages (Somay et al. 2020). Thus, it seems that several families of beetles find suitable conditions in these semi-wooded areas. The change in species assemblage occurring at the boundary between different land use systems is sometimes accompanied by increased biodiversity, although there is no consensus on this (Kark and van Rensburg, 2006).
Wooded pastures probably allow improved connectivity between neighbouring land use systems, resulting in easier exchange of certain forest individuals between forests. The exchange of forest individuals would probably be reduced in a steep interface from F to NWP, without WP ensuring the transition between land use systems. Eggers et al. (2010) have effectively demonstrated that semi-open corridors, resembling wooded pastures, allow certain carabid forest species to migrate and guarantee a connection between open and forested areas. In addition, WP provide additional available surface habitat for a substantial proportion of ground beetles, which otherwise only survive in F or NWP. Moreover, the 48 ground beetle species collected in wooded pastures represent 14% of the 342 species of carabid beetles native observed in the Swiss Jura Arc (info fauna 2023). If this result is extrapolated to all wooded pastures in the Jura Arc, this percentage can increase considerably, making wooded pastures an important habitat for a diverse community of ground beetle species.
Nevertheless, some species were only found in NWP and F, indicating that some species cannot tolerate a semi-open environment, like wooded pastures. Structural attributes of vegetation (plant strata) that create environmental heterogeneity, valued by a wide range of species, may be perceived by other species as fragmentation of their habitat (Tews et al. 2004). Fragmentation can cause disruption of key biological processes, such as dispersal and resource acquisition (Saunders et al. 1991). This habitat discontinuity favors generalist species at the expense of specialist species of adjacent land use systems (Lövei et al. 2006), making wooded pastures not suitable for the preservation of strict open and forest specialist species.
Radio-tracking has shown that forest ground beetle species are reluctant to cross open areas (Riecken and Raths, 1996). This suggests that the forest-only species in our study are highly dependent on the size of their forest patch to persist and have a viable effective population size in the long term. De Vries (1994) has shown that the number of species with reduced dispersal ability increases with patch size. Furthermore, the survival of forest specialists also depends on connections to other alternative patches at short distances (Niemelä 2001). Species adapted to unstable conditions in open areas have generally good dispersal capacity (Niemelä 2001), suggesting that they are less affected by habitat fragmentation than forest species with reduced dispersal capacity.
Diversity parameters
We cannot confirm research hypothesis (ii), but we have nevertheless shown that WP and NWP have equivalent diversity indices, and that both have significantly higher diversity indices than F. However, the diversity indices are non significantly different between WP and NWP. These observations are consistent regardless of whether the relative abundances of species are considered, i.e. accounting mainly for common species (as with Hill-Simpson) or also for locally rare species (as with species richness). Furthermore, the asymptotes were already tending towards horizontality when considering relative abundances of species (with Hill-Shannon and Hill-Simpson), indicating that sampling efforts were sufficient for capturing the entire community of the common species in the three land use systems. Practically, we can say that open pastures and woody pastures in the study area contain twice as many effective numbers of species than forests (based on Hill-Shannon). This statement might be easily understandable for landscape managers and could facilitate decision-making in line with the promotion of ground beetle diversity. The choice of the diversity index has hence consequences for testing scientific hypotheses and for practice (Roswell et al. 2021).
The variable "Biodiversity Promotion Areas'' significantly influences the diversity indices and is positively correlated with ground beetle species diversity in our study. This confirms the result of Aviron et al. (2009) who showed that several taxa including ground beetles benefit from these environmental promotion measures in Switzerland, both in terms of species richness and species assemblage. These areas also improve ecosystem services such as pollination and natural pest control (Tschumi et al. 2016). Concretely, several extensive agricultural measures (Article 59 of the national direct payments ordinance) exist to promote biodiversity in agricultural lands, such as conserving indicator flora, maintaining favourable (micro-)structures such as undergrazed areas (grass left standing), bushes, stones, wetlands or bare ground, preserving isolated trees, keeping mosaic of areas with staggered use over time, etc. (Stäheli et al. 2010). All these measures in Biodiversity Promotion Areas are therefore particularly beneficial for the Carabidae, many of whose species are often dependent on specific small structures.
Our study confirms the results obtained by Molnár et al. (2001) and Boutaud et al. (2022), which demonstrate that species diversity is higher in open and semi-open areas or forest edges than in forests. Molnár et al. (2001) did not show any significant differences between the diversity of open and forest edge areas, which resemble a semi-open ecotone between grasslands and forests where microsites favouring higher diversity are found. Boutaud et al. (2022) established that the semi-open zone is slightly higher in diversity than the open zone.
The lower species richness and diversity in F than in open and semi-open areas is not easy to explain. The lower quality of the forests and the fact that the size of the forests is too small to support a higher species diversity may explain the lower forest species diversity in certain situations, but are not necessarily applicable in the case of the Jura mountain forests studied. This lower diversity in this land use system is widely observed by Luka et al. 2009, who showed that forests in Switzerland naturally contained fewer species of Carabidae than in other more open natural environments. Nevertheless, litter thickness seems to be a major factor linked to ground beetle composition and abundance. A forest floor with a high litter thickness decreases the abundance and species richness of ground beetles and increases the presence of strict forest species, as biotic interactions are probably more complex in this environment. In fact, litter would interfere with the movement and foraging of ubiquitous and species of more open habitats, which are not adapted to such an environment (Guillemain et al. 1997). Molnár et al. (2001) emphasise that the structure of the vegetation and its derived variations creating microclimates is one of the most important factors to determine the distribution of ground beetles.
The high species diversity of ground beetles in wooded pastures could be the result of the edge effect from the dispersal of individuals present in adjacent habitats (Magura et al. 2001). It is an arduous task to explain why diversity indices are not different between WP and a more homogeneous habitat, the NWP. According to the study by Lacasella et al. (2015), the distance estimation of 600 m in the heart of open pastures from the trap site to the first trees is necessary to avoid the appearance of forest species during surveys. In the case of our study, this buffer zone for NWP was not always respected, so there may have been an edge effect with other woodlands in the vicinity of NWP. Another hypothesis as to why NWP have such a high species diversity is that there may be a particularly wide variety of discrete micro-structures in this environment, such as stones, rocky outcrops, puddles, bare soil patches or unevenness of the soil created by livestock trampling, which might satisfy a large number of species (Chittaro et al. 2023, submitted). It should also be noted that open areas, such as NWP, have a naturally higher species potential than more wooded environments, without any apparent explanation linked to the quality and ecological features of the land use system (Luka et al. 2009).
Carabid biomass and average body size
We refute research hypotheses (ii) and (iii) about ground beetle biomass and average body size respectively and propose from the results that, in the context of the study, tree cover is significantly and positively correlated with ground beetle biomass and average body size. Consequently, the biomass and average body size is highest in F, followed by WP and the land use system with the lowest values is NWP.
Fernández-Tizón et al. (2020) shows that arthropod biomass in semi-natural grasslands is directly and positively related to primary production. Primary production could be positively correlated with the body size of some individuals, suggesting that the primary production of a habitat could lead to morphological evolution between taxa (Capellini and Gosling, 2007). In this logic, habitats with low primary productivity (e.g. dry pastures dotted with lapiez with a thin layer of organic matter above the bedrock) should favour species with smaller average body sizes (Sustek 1987). In forest environments, ground beetles have a larger body size on average than in other less forested environments based on values from Klaiber et al. (2017).
Disturbed agricultural ecosystems, such as pastures, experience large fluctuations in their total energy value (Blake et al. 1994), i.e. the nutrient resources exploitable by organisms. Disturbances also occur in forests in the context of their management, but these are in principle sporadic activities in a short period of time compared to the management of agricultural areas that are affected by grazing and the addition of fertilisers and livestock dung. Moreover, these ephemeral energy sources favour species with smaller bodies (Schoener and Janzen, 1968). A habitat with a fluctuating energy source is more easily exploited by widespread, dispersive individuals capable of withstanding large variations in population size, and such species are more likely to be characterised by small body size (Gaston and Lawton, 1988).
Wing morphology
Our results confirm that tree cover has a significant influence on the wing morphology of ground beetles. Brandmayr (1991) shows that brachypterous species are few in number in the early stages of a habitat succession, but their numbers increase as open areas begin to be colonised by trees and shrubs. Individuals with low dispersal ability, such as brachyptera or wingless being mainly medium to large in size, are more likely to be found in forest-dominated environments (Shibuya et al. 2014). Moreover, most of them hardly leave the forest canopy or only move a few tens of meters away from it (Della Rocca et al. 2021). It is therefore important for wingless and brachypterous ground beetles that their habitat is disturbed as little as possible, in order to guarantee that resources are available in the long term. Nevertheless, timber extraction causes a lot of disturbance, but logging activities within a stand are typically not very frequent, although this depends on the type of forest management. The same applies to natural disturbances such as storms or fires, which cause major disturbances (Bouget and Duelli, 2004), but these events are sporadic and in principle rare.
This is also confirmed by Gobbi et al. (2007), who studied the assemblage of terrestrial beetles during glacier retreat. Indeed, areas recently freed from the ice, with unstable, stony and loose soil, have low numbers of brachyptera, but a higher number of macropterous species, whereas areas that have been ice-free for a longer period of time, with a higher vegetation cover and more mature soils, have more brachypterous species. In such temporary environments (pioneer, ruderal, meadows, etc.) which can be quickly reforested, ground beetles, mostly macropterous, need to disperse continuously to find favourable biotopes, and logically have a higher (re-)colonisation capacity (Den Boer 1970). These continually disturbed open environments, particularly by animal grazing and agricultural and forestry work, in the case of our study, slows down the advance of woody plants. The macropterous species therefore have a high dispersal capacity, guaranteeing their survival thanks to the flight induced by their developed wings (Roff 1990; Shibuya et al. 2014). This flight capacity gives them a clear advantage over brachyptera for living in open areas.
Certain studies relate habitat structural stability or litter thickness to wing morphology (Guillemain et al. 1997; Shibuya et al. 2014). In our case, the environmental variable "tree cover" implies a certain stability of the soil, due to the rooting of the land, and guarantees a thickness of litter constantly renewed by the leaf litter fall or by the biological activity induced by the trees, that can vary according to species type (hardwood or softwood) and forest management. Although ground beetles are mainly epigeous, they are dependent on tree cover, which partially and indirectly structure the habitat and influences ecological and edaphic conditions.
Moisture preference
Humidity and light are known to be limiting factors for ground beetle assemblage (Lövei and Sunderland, 1996). Tree cover directly influences temperature and light at the site scale and their interaction varies moisture level (North et al. 2005). According to the study by Kagawa and Maeto (2014), soil moisture had a marked effect on the abundance and distribution of ground beetle species. Forest species prefer generally moist soils, while intermediate and open areas species tend towards drier conditions.
In our study, tree cover influences also the composition of ground beetles according to their preference for moisture. Hygrophilc and steno-hygrophilic species are closely linked to the habitat with a high tree cover that is most likely associated with high moisture content in our study. Mesophilic and mesophilic-hygrophilic individuals tend towards slightly more open environments but are still linked to some presence of woody plants. In contrast, xerophilic individuals are found in open areas and are more correlated with the variables Biodiversity Promoting Areas and dry grasslands and pastures. These results make sense since DGP and BPA have higher values in areas that tend to be less densely covered with trees. According to our results, the tree cover, and thus the shading induced by it, apparently conserves moisture and produces a set of ground beetles with hygrophilic habits. Although it should be mentioned that many other biotic and abiotic factors contribute to the variation in humidity of an environment.