The associations between ground beetle (Coleoptera: Carabidae) communities and environmental condition in floodplain forests in the Pannonian Basin

We studied assemblages of carabids in eight similar habitats, fi ve in Slovakia and three in Serbia. The ground beetles were caught by pitfall traps from February 2015 until November 2016. We compared the incidence of Carabidae in fl oodplain forests and ecotones alongside the River Danube in Slovakia and the Rivers Tisza and Begej in Serbia. We determined their association with anthropogenic effects,diversity of plants in the different vegetation layers, cover of vegetation layers (herbaceous plants, shrubs and trees), area of forest stands, circumference of forest stands, distance to forest edge, age of forest stands, depth of leaf litter and physico-chemical properties of soil and leaf litter (conductivity, pH, relative content of H, C, P and N). In total, 2,495 adult individuals of 110 species of carabids were collected. The total epigeic activity of the carabids was signifi cantly and positively associated with the number of species of plants in E3 vegetation layer and the relative content of N, and negatively with the cover of the E1 layer. Species richness was signifi cantly positively associated with the number of species of plants in the E3 layer and the pH of leaf litter, but an opposite trend in evenness.


INTRODUCTION
Currently, there are very few areas of fl oodplain forests left in Europe. The vast majority have been destroyed and many of the remaining fragments are in a poor condition. According to Ábrahámová et al. (2014), they are among the most endangered natural ecosystems in Europe. Therefore, it is important to record the current conditions in these habitats over the widest possible area using important bioindicators such as Carabidae. This was the main reason for studying the alluvial soils of three rivers with analogous habitat characteristics, situated at the same latitude and approximately 400 km apart.
For a better comparison, three locations in Serbia geomorphologically and botanically similar to those monitored in Slovakia ( Fig. 2) were selected. The Begej River (a tributary of the Tisza) is the smallest of the rivers monitored. The mean anuual temperature in the proximity of Zrenjanin (from 1991 to 2010), which is close to these areas, was 11.9°C, the average annual relative humidity 73.0% and mean annual total precipitation 607.0 mm (Republic Hydrometeorological Service of Serbia, 1999. Two sites (S6 and S7) on the alluvium of the Tisza River within the Ritovi Donjeg Potisja Special Nature Reserve, a protected area 5 km from the village of Aradac, close to the town of Zrenjanin, were studied.
The third site in Serbia (S8) is located in the protected area Carska Bara Special Nature Reserve, near the village of Belo Blato, close to the town of Zrenjanin, approximately 20 km from the sites studied along the Tisza River.
The aim of this study is to determine the specifi cities of carabid communities in alluvial forests and the main ecological and environmental factors associated with their diversity and dynamics. For these reasons, we compared similar types of habitats on the alluvial soils of three rivers (the Danube in Slovakia and the Tisza and the Begej in Serbia) in different geographical regions, by analysing environmental parameters, such as anthropogenic effect, plant species diversity and cover of herbaceous plant layer (E 1 ), shrub layer (E 2 ) and tree layer (E 3 ), size of area, edge of area, circumference of area, age of trees, depth of leaf litter, pH of the soil and leaf litter, conductivity, content of P, N, C, H in the soil and leaf litter and their association with particular ground beetle communities.

Sites studied
This research was conducted at eight sites located on the alluvial soils of three rivers: the Danube in Slovakia and the Begej and the Tisza in Serbia. Of these, fi ve sites (S1-S5) are situated on the outskirts of Bratislava in Slovakia (Fig. 1). The mean annual temperature in the area of Bratislava (from 1991 to 2015) was 10.2°C, average annual relative humidity 72.2% and mean annual total precipitation 676.2 mm (Lapin et al., 2019). S1 -located near the overpass at Bajkalská Street (48°8´22.31˝N, 17°8´57.03˝E, 138 m a.s.l.) is a fragment of fl oodplain forest of the Salici-Populetum association, with a semi-open habitat and a slightly moist microclimate.
S2 -located in the settlement Malé Pálenisko (48°8´11.54˝N, 17°9´14.45˝E, 132 m a.s.l.) is a semi-open habitat with a moist microclimate, situated near an old arm of the Danube River, with a fl uctuating water level.
S4 -is near the waste incinerator in Bratislava (48°6´24.12˝N, 17°10´9.68˝E, 132 m a.s.l.) and is a semi-open xerothermic habitat on a gravel bar, with a shallow layer of soil (up to 30 cm).  More detailed information on the vegetation at these sites is presented in Litavský et al. (2018Litavský et al. ( , 2021.

Methods
This research was carried out from February 2015 until November 2016. Ground beetles were captured using pitfall traps (Stašiov, 2015) consisting of plastic cups (diameter of opening 9 cm and volume 0.5 l) containing 1% formaldehyde as a fi xative. At each site, fi ve traps were set in a line at a distance of 5 m from each other. The traps were operational from February 2015 to November 2016 and emptied and refi lled approximately at halfmonthly intervals. The contents of the fi ve traps at each site were pooled, which formed the sample for that date. The samples were sorted in the laboratory and the ground beetles subsequently identifi ed to species according to Trautner & Geigenmüller (1987), Hůrka (1996) and Müller-Motzfeld (2004). Specimens were fi xed in 75% ethyl alcohol and deposited in the collection of the Faculty of Natural Sciences of the Comenius University in Bratislava. A list of species was drawn up following Lorenz (2019). In Table  1, for each species, we added its ecological characteristics. Farkač et al. (2006) classify carabids into three groups: relict (R), adaptable (A) and eurytopic (E), based on their ecological valence and association with a particular habitat. Following Šustek (2000, 2004a, 2010, 2012), the humidity preferences of ground beetles were classifi ed as either; 1 -strongly xerophilous, 2-3 -intermediate between strongly xerophilous and mesohygrophilous, 4-5mesohygrophilous, 6-7 -intermediate between mesohygrophilous and strongly hygrophilous, or 8 -strongly hygrophilous. In terms of dispersal ability, species were divided into three groups: (i) non-fl ying, (ii) fl ying and (iii) brachypterous, which occasionally fl y (Hůrka, 1992(Hůrka, , 1996Matalin, 2003;Šustek, 2012).
Samples of soil and leaf litter for chemical analyses were collected on June 8th 2016. More detailed information on the methods of analysis of the soil and leaf litter samples is given in Litavský et al. (2018Litavský et al. ( , 2021.
We defi ned four levels of anthropogenic effect on the study area: level 1 -with minimum disturbance; level 2 -with grazing; level 3 -with infrequent movement of vehicles and minor solid waste pollution; level 4 -with solid waste pollution, frequent mowing and cutting and building activities.
Of the environmental factors that could affect the composition of carabid communities we recorded species richness in individual layers, stand canopy of individual layers, anthropogenic effect, area of fragment, distance from another fragment (edge), length of the periphery of the fragment, age of trees at the sites and average depth of leaf litter, which are presented in Litavský et al. (2018Litavský et al. ( , 2021. During the vegetation period, we recorded plant species diversity in three vegetation layers (E 1 , E 2 , E 3 ), as well as the covers of the tree, shrub and herbaceous plant layers, in both countries and within sites of constant size (400 m 2 ). More information on the method of investigation of the vegetation is given in Litavský et al. (2021).

Data analysis
We calculated the diversity of the carabid communities in terms of species richness, Shannon diversity (Spellerberg & Fedor, 2003) and evenness (Jost, 2006;Tuomisto, 2012;Fedor & Zvaríková, 2018). Species epigeic activity was the mean number of individuals caught per trap per day. Carabid community composition is summarized in the species-by-site matrix of epigeic activities.
Due to the relatively low number of sites sampled and the high number of environmental variables, we did not build models describing diversity and species composition, but adopted an ex-ploratory approach and investigated all species-environment associations.
In order to investigate the diversity characteristics of communities (total epigeic activity, species richness, Shannon diversity and evenness), we produced a matrix of pairwise Spearman's correlation coeffi cients between environmental variables and the diversity characteristics.
Non-metric multidimensional scaling (NMDS) based on Bray-Curtis distance was used to visualize similarities in the composition of carabid communities. The vectors of environmental variables were projected onto ordinations in the directions of their maximum correlations with the confi guration of sites. The statistical signifi cance of fi tted environmental vectors was evaluated using permutation tests (10,000 permutations).
This exploratory approach allowed us to investigate all speciesenvironment associations without the risk of missing any important information caused, e.g., by the removal of highly correlated variables. Indeed, the results of such a heuristic analysis should be considered carefully and regarded as a process for generating hypotheses rather than a generalization to wider populations.
The analyses were performed in R (R Core Team, 2016) using the Hmisc (Harrell, 2016) and the Vegan libraries (Oksanen et al., 2016).

RESULTS AND DISCUSSION
During this research, a total of 2,495 adult carabids belonging to 22 tribes, 48 genera and 110 species were recorded. Seventy-nine species of ground beetles are recorded at the sites in Slovakia and 58 species at those in Serbia (Majzlan & Litavský, 2017), with 27 of the species occurring in both countries. The values of the total epigeic activity of ground beetles captured at individual sites during this research are shown in Table 1 (967), which is a semi-open habitat with a moist microclimate and where the highest species richness of ground beetles was recorded (38 species). The lowest species richness of carabids was recorded at S4, which is a semi-open xerothermic habitat (23 species). In addition, we divided the ground beetles into different groups according to several criteria (Table 1). Based on their ecological valence and association with a particular habitat, we recorded 10 relict, 38 eurytopic and 62 adaptable species of carabids. In terms of their humidity preferences 24 were strongly xerophilous, 39 mesohygrophilous and 47 strongly hygrophilous species. With regard to their ability to fl y, we recorded 20 brachypterous species that occasionally fl y, 20 non-fl ying species and 70 species of ground beetles capable of fl ying.  (2017); EV -ecological valence of carabids and their association with habitats: R (relict), A (adaptable), E (eurytopic); HP -humidity preferences of carabids recorded: 1 -strongly xerophilous, 2-3 -intermediate between strongly xerophilous and mesohygrophilous, 4-5 -mesohygrophilous, 6-7 -intermediate between mesohygrophilous and strongly hygrophilous, 8 -strongly hygrophilous; FL -ability to fl y: F (fl ying), N (non-fl ying), B (brachypterous, occasionally able to fl y).

Species
Abbr. EV HP FL Sites Σ specimens S1 S2 S3 S4 S5 S6* S7* S8* Abax carinatus (Duftschmid, 1812) Abca The associations between diversity characteristics and environmental variables were evaluated using correlation analysis. The total epigeic activity of carabids was significantly and positively associated with the number of species of plants in the E 3 vegetation layer and the relative content of N in leaf litter, and negatively with the cover of the E 1 layer (Table 2). Species richness was signifi cantly positively associated with the number of species of plants in the E 3 layer and the pH of the leaf litter, while the trend in evenness was exactly the opposite.
NMDS ordination revealed that the composition of carabid communities was signifi cantly associated with the number of species of plants in the E 3 layer and pH of the leaf litter (Fig. 3) The results of our two-year research indicate that the fl oodplain forests and their ecotones alongside the Danube, the Tisza and the Begej Rivers provided appropriate conditions for ground beetle communities, based on their high species richness. Šustek (1994b) states that about 25-35 species of carabids are usually present at any one time at one locality under natural conditions, but in the best conserved fl oodplain forests the number of carabid species exceeds 40. We recorded the highest number of carabid species at S2 (38 species), S5 (36 species) and S6 (34 species) (typical fl oodplain forests near oxbow lakes), indicating that they are high quality habitats for carabids. In total, we recorded 110 species of ground beetles during this study. For comparison, Šustek (2004b) records 60 species of carabids using pitfall trapping in the Jurský Šúr Nature Reserve close to Bratislava. A survey of beetles in the Litovelské Pomoraví Protected Landscape Area, within the fl oodplain forests on the Morava River, Nakládal (2008) revealed 93 species of ground beetles at six sites in . Nakládal's (2008 objective was to record as many species as possible and therefore he used a variety collecting methods, such as beating trees and shrubs, individual sampling, sweeping, pitfall trapping, examina-  Although we used only pitfall traps to collect carabids at only eight sites, it is clear from the above that we recorded a higher total species richness of carabid beetles in our study. There was a statistically signifi cant negative association between total epigeic activity of ground beetles and the cover of the herbaceous plant layer (E 1 ). This can be explained by the fact that density of vegetation, especially that of the herbaceous plant layer, can affect predator activity (the majority of carabids), by making it diffi cult for them to move on the surface of the soil. The decrease in epigeic activity of soil-dwelling beetles with increase in the density of vegetation is also reported by Heydemann (1957), Honek (1988), Humphrey et al. (1999) and Thomas et al. (2006). For example, Honek (1988) reports that some staphylinids and most species of carabids prefer sparse rather than dense stands. Zou et al. (2013), however, report that the density of herbaceous plants had little effect on the beetle activity in the Changbai Mountains.
The signifi cant positive association between both total epigeic activity and species richness of ground beetles, and plant species richness in the E 3 vegetation layer is probably due to the longer period of shading (earlier budding of different species of trees and late leaf fall), which ensure more stable and humid microclimatic conditions for a longer period, as well as a greater food supply as a high tree diversity results in a richer leaf litter. Pearce et al. (2003) also point out that some ground beetles may also benefi t from increased tree species richness. Vehviläinen et al. (2008) confi rm that carabids differ signifi cantly in their preferences for stands composed of particular, yet different, species of trees. During a study of the effects of river and fl oodplain restoration on riparian ground beetles, Januschke & Verdonschot (2016) revealed that the  Table 1.  (2005), carabid communities in forest ecosystems consist of heliophobic species that prefer areas shaded by trees or at least by dense shrub vegetation. It is cooler in the shade, which slows down the drying out of the surface of the soils and leaf litter, reduces evaporation and improves the water balance in the stand. For this reason, the association of individual species with the presence of tree cover are to some extent positively associated directly with the species' moisture requirements.
In addition to the effect of plants, ground beetle communities are also affected by properties of the soil and leaf litter. Nitrogen is an essential element for organisms. This is supported by the statistically signifi cant positive associations between the total epigeic activity of carabids and the nitrogen content of leaf litter. The nitrogen content of the leaf litter may indirectly affect (through saprophagous prey of ground beetles) the food supply of carabid beetles. Leaf litter rich in N is an attractive food for saprophages that use the nitrogen in their own physiological processes (Vician et al., 2018). Therefore, this material decomposes more rapidly due to the activity of saprophages (Wittich, 1942(Wittich, , 1943. Dunger (1958) stresses the importance of nitrogen as the main element determining animal production and sources of food for invertebrates, including ground beetles. Vician et al. (2018) also report statistically signifi cant associations between the content of N in the leaf litter and species richness, Shannon diversity and species composition of carabid beetle communities at nine stands in the Borová Hora Arboretum (Central Slovakia).
The composition of carabid beetle communities was also associated with the pH of the leaf litter, with a signifi cant positive association between species richness of ground beetles and leaf litter pH, but the opposie trend in species evenness (Table 2). Vician et al. (2018) note that a fl oodplain forest in which leaf litter had a high pH also had a high species richness of ground beetles. Magura et al. (2003) also point out that leaf litter has a positive effect on carabid species richness.
In terms of the carabids' ecological valence and their association with a particular habitat, most (62) (56.4%) were adaptable, 38 (34.5%) were eurytopic and 10 (9.1%) were relict species. Relict species of carabid were mostly recorded near the oxbow lakes on the Danube (S5), the Tisza (S6) and the Begej Rivers (S8). This indicates these forest stands are ecologically stable. In comparison, Igondová & Majzlan (2015) did not record any relict species of ground beetles in the carabid communities during a one-year study of the Šuja peat bog (northern Slovakia). In relation to their ability to fl y, of the species of carabid recorded (70) (63.6%) were able to fl y, 20 (18.2%) were brachypterous and occasionally fl y, and 20 (18.2%) were non-fl ying. Arndt & Hielscher (2007) conclude that most species of forest ground beetles are unable to fl y or do not regularly fl y. Šustek (2012) state that species of carabids that inhabit unstable riparian habitats are able to fl y and successfully colonize anthropogenic ecosystems, such as arable land or vegetation in human settlements. Nevertheless, up to 70% of the species we recorded were able to fl y and were present in closed forest stands that were mostly little or unaffected by human activity.
We determined how ground beetle communities vary in the different habitats in fl oodplain forests in Serbia and Slovakia and found that the total epigeic activity of carabids was signifi cantly positively associated with the number of species of plants in the tree layer and the relative content of N in the leaf litter, and negatively with the cover of the herbaceous plant layer. Species richness was signifi cantly positively associated with the number of species of plants in the tree layer and pH of the leaf litter, while evenness showed the opposite trend. Based on these results (Fig. 3), we selected several species of carabids, which can serve as bioindicators. We conclude that A. micans, A. fuliginosum, P. assimilis, A. fl avicollis, D. globosus, P. atrorufus, P. rufus and N. brevicollis, which prefer forest stands with a high number of species of plants in the tree layer, can be used as bioindicators of the presence of high tree species richness in fl oodplain forests. We also found that T. quadristriatus, O. azureus, C. melanocephalus and C. coriaceus preferred stands in which the pH of the leaf litter is low and could be used as bioindicators for assessing changes in landscape structure caused by human activity resulting in soil acidifi cation. Therefore, more information on these associations might be helpful in further elucidating how carabids respond to vegetation, soil and microclimatic conditions, and how these conditions vary in the various types of fl oodplain forests.