Department of Physics, Chemistry and Biology (IFM)

The outcome of this study revealed that grazing pressure and percentage canopy cover both significantly influenced carabid diversity in the semi-natural grasslands of Tinnerö. Generally, when sward height increased, both numbers of species and individuals showed a positive response, while canopy cover showed a negative correlation with the number of species and individuals, (table 1 and 2). These findings are similar to past research such as that by Kruess & Tscharntke (2002) who tested contrasting responses of plant and insect diversity to variation in grazing intensity. Species diversity indices (H’ and J’) (fig. 3) were highest in sites that had low grazing pressure. This makes sense as high grasses are essential for high carabid activity providing better areas for hibernating, refuge, foraging and reproducing (Lövei & Sunderland 1996), while high carabid diversity in low canopy cover was an indication of high activity due to increased metabolic rates resulting from high exposure to sunlight, a phenomenon described by Lövei (2008).

Carabid responses to sites

Contrary to the general pattern of high-grass/low-canopy resulting in higher diversity and vice-versa, certain sites as Edhaga A, Edhaga C, Fjärilskullen A and Fjärilskullen B showed otherwise. The causality of such an observation in the said sites is not fully understood, however, the correspondence analysis indicates that canopy cover and grazing pressure had quite similar effects on variability of species occurrence. So, this can possibly be an indication that other factors such as plant diversity, moisture, temperature and plant litter biomass, and the presence of prey and competition would be important (Lovei and Sunderland 1996). Also, the Shannon’s diversity index by sites (fig. 3), and the weights of some species to certain sites in the ordination (CA) (fig. 4) can be said to be an indication of site preference by carabids. We did not delve on the individual species habitat preferences; however, we saw several species clearly indicating preference to certain sites, possibly due to effects from unchecked parameters as mentioned above.

At sites with high grazing pressure, few species were present, and the populations were generally sparse as indicated by the low Shannon’s diversity index estimation. Sites such as Granhagen, Tinnerö-nörra and Sveden-A showed the lowest diversity and low evenness (fig. 3), a characteristic of high disturbance (Connell 1978), which also leads to dominance by certain species, such as opportunistic and possibly generalist species. Besides, all singletons were found in short sward height, this along with low species richness estimations in some sites, further supports findings that high grazing would lead to smaller populations. The CA analysis (fig. 4) also shows that these sites had quite similar species composition as well as similar low sward and open canopies. True to this, these sites were highly dominated by Poecilus species, which are subtly categorised as medium size, and therefore supporting findings that intense disturbance such as high grazing pressure can eliminate large species from an assemblage (Magura et al. 2005; Sustek 1987).

Differences in canopy cover could result in variation in plant species diversity and composition. The correlation is that lower canopy cover leads to an increase in plant species diversity such as grasses and herbs. Plant diversity been subsequently linked to higher invertebrate diversity that in can support, for example, (Zou et al. 2013) found that herb diversity positively correlated with beetle abundances. Thus, a denser canopy cover results in lower plant diversity (Dáttilo and Dyer 2014), thereby limiting support for invertebrates like carabids. Interestingly, we found high canopy cover to favour Nebria brevicollis, as it is during the summer these species complete their larval stages leading to a preference for habitats with high canopy cover to minimise the risk of drying out (Lovei and Sunderland 1996).

Followings the H’ and J’ analyses discussed earlier, no clear patterns were observed with regards to effects of canopy cover. A reason to this can be because canopy cover per trap within a site varied largely, for example, a site could have 0% canopy over a trap and more than 50% over another trap. Also, such seemingly random patterns can result from canopy effects on ground beetles’ diversity depending on factors such as species of trees involved, the stage of development and even the stand structure (Toigo et al. 2013), whose effect was not tested.

Species responses

Carabus species displayed a possibility of habitat preference, as abundance showed more change along grass height and not much change along canopy cover. These large flightless species were more abundant in intensively grazed patches. This is contrary to Cole et al. (2006) who found that the occurrence of larger carabids particularly Carabus were high in extensively grazed areas. Moreover, Cole et al. (2002) also showed that Carabus, like other larger species, preferred larger prey like the leafhoppers (Family: Cicadellidae), species often found in high swards with vegetation that is structurally diverse in unrestored grasslands. However, large Carabus species could prefer exposed ground e.g. as the parasite burden might be lower (Lovei and Sunderland 1996), because pursuit of prey is easier or that they are more active perhaps because they need to pursue prey.

Poecilus species, particularly versicolor were found to be the most abundant species in most sites (thirteen out of the eighteen sites). In the other three sites where the Poecilus was not abundant, larger zoophagous/omniphagous species such as the Pterostichus dominated, a possible indication of predator competition, which possibly also leads to to temporal niche separation since the abundance peak for Poecilus was earlier than Pterostichus. We found that 87% of Poecilus species occurred in open to very-low canopy and short sward height or highly grazed areas, interestingly opposite to the general result indicated by the GLMM, however, this was expected as species in this group mainly occur in open habitats Lindroth (1986).

The second most diverse genus in this study was Pterostichus with eight identified species. The species showed a clear preference to both open canopy and very low sward in line with past research. In the Melanarius species, for example, Wallin (1986) found that Pterostichus melanarius preferred open habitats where reproduction can take place in summer. This further supports Pterostichus melanarius as a species of open habitat that inhabits even in areas that are intensively managed, as was found by (Marggi, 1992). Wallin also found that during high disturbance, the species moved into nearby forests, a possibility in this case, though not tested.

The findings of this study support Kotze et al. (2011) and de la Pena et al. (2003) proposal of managing semi-natural grasslands with low-intensity grazing as they may be essential in providing refuge for carabids. Besides, intermittently managing grasslands overall enhances diversity (Kruess and Tscharntke 2002).

Carabus were almost exclusively most active in highly grazed, short-sward grounds. We believe these species are driven into such suboptimal habitats due to their ability to shift periods in which they are active, thereby, reducing interspecific competition (Kaltsas et al. 2013). Overall, high grazing led to lower resources availability for carabids causing a reduction in species number.

Most of the species identified are widely distributed in Sweden (Lindroth 1986), and overall, with viable communities, and because carabids can suppress pest outbreak in agriculture application, their effectiveness can be improved by taking into consideration the condition of the carabid habitats close to sites of interest. Beetle banks, for example, have been shown to be useful in this regard as they provide habitat for polyphagous predators (Macleod et al. 2004). Tinnerö grasslands could act as possible refuge and recolonization sites. The ordination analysis (fig. 4) showed that species configurations were different in different sites based on the level of grazing and canopy present, therefore, further supporting that carabids are reliable in conservation for indicating changes in a habitat. Since all species identified in this study are viable according to the red-list (Westling 2015) and considering that grazing has been going on for quite some time in Tinnerö nature reserve, we conclude that the current practice of grazing different sites at different intensities is ideal for the continued support of carabid communities in these grasslands, as small populations can be replenished from the more carabid favourable habitats.

If the aim of management is to increase biodiversity of carabids, it may be important to consider strategies to reduce the impact of grazing on carabid populations in semi-natural grasslands. Because we found a positive correlation between grazing pressure and carabid populations, investigating a possible threshold of grazed grass height so that grazing can occur but still providing a safe habitat for carabids, can be important. For future research of this kind, we ask that you keep the following in mind. The different sites studied here are not isolated from the adjacent areas (e.g. grazed, ungrazed, forest, agricultural). Thus, the possibility of movement between the adjacent areas should not be ignored. Many carabid species have been found to disperse in the adjacent areas, e.g. forest interior and grass. Because of dispersion between forest interiors and grass, Niemelä (1988) highlights that dispersal between adjacent patches is important in species assemblage studies. In addition, competition has been found to exist between carabids and ants, therefore, neglecting ants in carabid community assemblage studies can in some cases also lead to imprecise conclusions (Lovei and Sunderland 1996).

Like many other field studies, this research had some ethical repercussions on the environment including biodiversity. Using pitfall trapping, other insect species were inevitably caught. However, none of the species were rare or threatened, therefore, no threats were caused to the local populations. It is also understood that few traps were lost in the ground, further degrading the land as this apparatus is non-biodegradable. For future research of this kind, it is important to devise methods that minimize the risks of catching species not of interest, and of course less damaging to the environment, in line with the 15^th UNs sustainable development goal to protect, restore and promote sustainable use of terrestrial ecosystems and to halt and reverse land degradation and halt biodiversity loss (United Nations 2017). It is important to understand how to maintain semi-natural grasslands as they provide numerous ecosystem services such as pollination and air purification, etc. Also, we humans have a moral obligation to reduce, or stop the extinction of species. Therefore, studies as this equip us with the knowledge to effectively implement these objectives.