Biodiversity and Conservation https://doi.org/10.1007/s10531-021-02168-w

ORIGINAL PAPER

Importance of foodplains for water diversity: a crucial habitat for the endangered beetle bilineatus in Southeastern Europe

Nataša Turić1 · Martina Temunović2 · Ildikó Szivák3 · Róbert Herczeg4 · Goran Vignjević1 · Zoltán Csabai5,6

Received: 16 April 2020 / Revised: 9 March 2021 / Accepted: 26 March 2021 © The Author(s), under exclusive licence to Springer Nature B.V. 2021

Abstract Natural foodplains are rapidly disappearing ecosystems worldwide, primarily due to changing hydrology. Continental harbours some of the largest remaining and best preserved natural riverine foodplains in Europe. To establish conservation priorities, we surveyed water beetle assemblages in three large foodplains situated in the , , and basins (Ramsar sites Kopački rit and Lonjsko polje, and Spačva riparian forest), with special focus on the red-listed beetle Graphoderus bilineatus. The main aim of our study was to investigate the key environmental factors driving variation of water beetle assemblage structure and occurrences of G. bilineatus. Environmental variables describ- ing habitat types and various habitat characteristics were measured and/or estimated for each sampling site. A total of 4.339 water beetle specimens belonging to 98 species were recorded. G. bilineatus was recorded at two out of the three investigated foodplains at 14 out of 30 sampling sites. We found signifcant diferences in water beetle assemblages among the three localities based on the total species dataset, accompanied by diferent environmental background and vegetation composition. The density of riparian vegetation, permanent water durability, shading, type of bank, and plant communities best explained the distribution of the water beetle species. Based on the regression model, human impact and shading have signifcant efects on the abundance of G. bilineatus. Sites with G. bilin- eatus presence have signifcantly higher species richness compared to the sites without the species. Co-existence analysis indicates various types of associations between G. bilineatus and other water beetle species.

Keywords Aquatic · Distribution · Co-existence · Threatened species · Conservation

Communicated by Jens Wolfgang Dauber.

* Nataša Turić [email protected] Extended author information available on the last page of the article

Vol.:(0123456789)1 3 Biodiversity and Conservation

Introduction

Floodplains of the lowland rivers are among the most productive and the most threat- ened ecosystems on Earth (Gopala and Junk 2000). The high biodiversity of foodplain and wetland ecosystems is considered to be a consequence of their position at the inter- face between land and water (Gopala and Junk 2000) where the hydrological regime is the major factor driving the patterns of aquatic community structure. Floodplains show large habitat diversity and dynamics, because of fuctuating water levels, related erosion, deposition processes, and varying levels of connectivity with the main river course (Ward et al. 2002). Major gaps still remain in the aquatic conservation knowledge of foodplains in Europe. More emphasis needs to be placed on the investigation of the ecological struc- ture and processes of such systems to ensure their efective management (Bayley 1991; Buijse et al. 2002). Information about the relative biodiversity value of diferent water body types is the main aim for many strategic conservation goals (Williams et al. 2003), includ- ing sustainable catchment management as required by the EC Water Framework Directive (2000/60/EC). Floodplains have long been neglected in the studies of the water beetle diversity, although they are one of the main groups of aquatic macroinvertebrates in freshwater and especially in wetland habitats. Temporary waters such as fooded meadows are highly important for water as these periodically created characteristic habitats provide abundant food resources and reproduction possibilities (Wiggins et al. 1980; Larson 1985; Vinnersten et al. 2009). The detailed study of the autecology of aquatic insects in the Dan- ube foodplain, especially water beetles, has revealed the existence of many species with narrow ecological requirements including adaptation to water level fuctuation (Turić et al. 2012). Water beetles are considered to be good bioindicators of habitat quality (Foster 1987) as they are sensitive to changes in various environmental factors such as vegeta- tion composition, water chemistry, hydroperiod, shading, size of water bodies, site age, and habitat heterogeneity (Nilsson and Söderberg 1996; De Szalay and Resh 2000; Fairchild et al. 2000, 2003; Lundkvist et al. 2001). Therefore, they are often used in the selection of prior habitats for conservation and management purposes in aquatic ecosystems (Foster 1987; Sánchez-Fernández et al. 2004, 2006). As a typical freshwater species, the dytiscid water beetle Graphoderus bilineatus repre- sents one of the few beetles in EU legislation, protected by both the EU and Bern Convention. In the IUCN Red List, it is classifed as Vulnerable, although the needs to be updated (Foster 1996). Graphoderus bilineatus is a Western Palaearctic species, distributed from Central Europe to western Siberia and from southern Scandinavia to northern parts of Southern Europe. However, in most European countries the species is very rare and is threatened locally in many regions of its global distribution range (Hendrich and Müller 2017). Recent investigations of the species occurrence and distribution were conducted to designate NATURA 2000 sites in most European countries, however, in some countries, populations have remained undetected (Cuppen et al. 2006; Vrazec et al. 2008; Hendrich and Balke 2002, 2005; Kálmán et al. 2007). In most countries at the European biogeographical level, its overall conservation status is “unfavourable-bad” with western and central European populations strongly declining (EIONET 2012). Until the recent past, only two historical records were known in Croatia from 1907 and 1943 (Guerguiev 1965; Šerić Jelaska et al. 2008) and two fndings from 1990 and 2005 (Kajzer 2001; Turić et al. 2012, 2015). According to Foster (1996), Croatia was not included in the range of Graphoderus bilineatus.

1 3 Biodiversity and Conservation

Graphoderus bilineatus is considered to be a stenotopic species with specifc habitat requirements. It prefers permanent standing and unshaded water bodies, usually shallow lakes, ponds, and channels with aquatic vegetation of medium density (Cuppen et al. 2006; Foster 1996; Klausnitzer 1984; Hendrich and Balke 2000), although it can be also found in a variety of smaller temporary stagnant waters. However, Iversen et al. (2013) have argued that G. bilineatus is a philopatric species that primarily depends on landscape connectiv- ity and structure, as well as on habitat stability and not on specifc habitat characteristics, suggesting that it is, in fact, a habitat generalist. Similar results are obtained from a study in the which pinpoints habitat connectivity as the key importance for G. bilineatus (Kolar and Boukal 2020). The life cycle of G. bilineatus is probably univoltine (Cuppen et al. 2006). It takes between two and three months and a new generation of adults appear from mid-summer. Breeding occurs in spring, eggs are laid in plant stems. Larvae pass through three instars and pupate out of water among mosses, plant debris, or stones. Both adults and larvae are predatory, and it is thought that larvae feed on cladocerans and other small invertebrates (Galewski 1975) and larvae frequently hunt in dense submerged vegetation. Anyhow, there are still many gaps in the current knowledge about the biology and the ecological requirements of the species which is a prerequisite for efective conservation management. One of them is a lack of knowledge of the habitat and feeding preferences of G. bilineatus larvae and their response to the changing environment (Kolar and Boukal 2020). To strengthen the information base for efective conservation of G. bilineatus and overall foodplain diversity, we carried out a detailed survey on the infuence of habitat fea- tures on G. bilineatus occurrence and abundance. We aimed to investigate the distribution patterns of G. bilineatus in the three largest foodplain areas of continental Croatia. Also, we tested whether the water beetle assemblages from the three localities difer from each other and we identifed the key environmental factors linked to the variation of water beetle assemblage’s structure. Finally, we tested the hypothesis that G. bilineatus occurs where the overall water beetle diversity is high, and we investigated which species G. bilineatus is associated with. To provide additional information useful for conservation and manage- ment actions in the future, we present an overview of the rare and endangered species in the studied area.

Material and methods

Study sites

We carried out our research at 30 sampling sites across three large foodplain complexes in Croatia situated in the Danube, Drava, and Sava basins: Kopački rit, Nature Park Lonjsko polje, and Spačva basin (Fig. 1). Kopački rit and Lonjsko polje are one of the largest remaining wetlands and best conserved natural riverine foodplains in the Danu- bian basin as well as in entire Europe. Kopački rit study site is located in the middle sec- tion of the River Danube (river 1410-1383) within the Nature Park covering approximately 17,390 ha (Mihaljević et al. 1999). Lonjsko polje Nature Park is situated in the alluvial plain of the central Sava river basin and covers an area of 50,650 ha (Klimo et al. 2008). The core phenomenon of both Nature Parks is fooding and associated lowland food- plain forests. Due to a specifc microrelief, all the habitats directly depend on frequency, duration, and level of fooding. The dominant wetland habitats are permanent freshwater 1 3 Biodiversity and Conservation

Fig. 1 Study area in continental Croatia with three investigated foodplains: NP Kopački rit (light grey) with sampling localities (dotted square), NP Lonjsko polje (dark grey) with sampling localities (black dots) and Spačva basin’ with sampling localities (dotted triangles) marshes, pools, ponds, oxbows, and swamps on organic soils with emergent vegetation, as well as large areas of open temporary fooded meadows. Both protected areas are of international importance, containing large areas designated as Ramsar sites, and are candidates for UNESCO′s World Natural Heritage List. Kopački rit is also the core zone of the recently designated transboundary Biosphere Reserve Mura-Drava-Danube. Spačva basin is one of the biggest still preserved complexes of lowland riparian for- ests in Europe, infuenced by the hydrological regime of the Sava River and its tributar- ies. The Spačva is the largest oak forest in Croatia, and one of the largest European Oak forests, covering an area of 40,000 hectares. The natural forest vegetation consists of mainly hardwood tree species such as the pedunculate oak (Quercus robur), narrow- leaved ash (Fraxinus angustifolia), common hornbeam (Carpinus betulus), and soft- wood species such as black alder (Alnus glutinosa), poplars (Populus sp.) and willows (Salix sp.). The development of these forests in history was determined by regular sea- sonal alternations of groundwater and periodic fooding. Most of the typical foodplain habitats depend on a specifc water regime, character- istic for foodplains in a natural or semi-natural state. This regime is mostly stated as a food pulse regime (Junk 1999). The rivers Sava and Danube have a pluvio-nival regime, with the discharge that peaks in winter and spring. The Drava River has a nivalo-pluvial regime, with peaks in summer. Kopački rit is a foodplain of the Danube and Drava rivers, with the Drava having a minor impact due to the high bank between the rit and the Drava River (Bonacci et al. 2002). Lonjsko polje and Spačva basin are foodplain complexes of the Sava river. Basic hydrological conditions varied during the study.

1 3 Biodiversity and Conservation

Generally, it is recognized that there is interannual variability in the seasonal or tempo- ral representation of hydrological periods. For example, fooding of the Kopački rit lasts from February to May and low water levels prevail from August to January (Bonacci et al. 2002). Hydrologically, 2010 was extremely wet on all three localities (Bakšić Pavlović et al. 2015; Turić et al. 2015). Extremely high foods have been reported and the consequent prolonged fooding that occurred during the spring–summer periods.

Water beetle sampling

The water beetles were sampled during 2010 from May to the end of October to cover the whole period of G. bilineatus activity. At each of the three localities, we selected 10 suitable sites for sampling, covering a wide range of freshwater habitats and vegetation types, including permanent and temporary ponds, fshponds, oxbows, canals, fooded lakes to large, fooded meadows and rivers. During this period, each sampling site was sampled fve times, with an interval of four weeks between two sampling dates. A brief descrip- tion of the sampling sites with Croatian National Habitat Classifcation (NKS) classes is given in Table S1. Water beetles were captured by sweeping a D-frame pond net (mesh size = 200 μm, mouth area = 23 ­cm2, depth = 12 cm) afecting both waterline, aquatic, and sub-aquatic vegetation of the littoral zone and surface of the substrate to cover all available microhabitats. Sampling time was 1 h on each occasion, including sweeping and removing adult beetles from the sample, so that diferent vegetation units, substrates, and open water areas were proportionally covered. This semiquantitative method and similar sampling time spent at each site allowed us to compare the structural parameters of the aquatic Coleop- tera assemblages between sites. In addition to sampling by hand net, two bottle traps with bait (conserved tuna fsh) were placed at each site on each sampling occasion (Turić et al. 2017). Bottle traps are a standard method for collecting water beetles (Koese and Cuppen 2006). The minimum distance between neighbouring traps was 3 m. Traps were set in the afternoon and picked up during the next day, after a period of 20 h in the water (one trap- ping night) considering the high-water temperatures and competition between the trapped species (Boukal et al. 2007). In total, 300 traps were set during the feld survey. Captured specimens were preserved in 70% ethanol, except large specimens that we could immedi- ately identify after sampling, especially specimens of G. bilineatus, which were released after identifcation in the feld. Only adult water beetles were collected and identifed in the laboratory using keys and descriptions by Csabai (2000), Csabai et al. (2002) and Nilsson (1996).

Environmental variables

We used two groups of environmental variables to describe the habitat features of the sampling sites. For the frst group we used habitat types based on the National Habitat Classifcation (NKS) (Nikolić 2011) that were recorded or estimated by an experienced botanist using plant species composition, percentage cover of plant species, and plant com- munities (Table S1). The dataset contained 61 plant species in total. For the second group, eleven environmental variables were used to further describe the habitat characteristics of each site (Table 1). Values of variables density of riparian vegetation (Veg) were described by four categories (without, small, medium, and dense), density of emergent and foating vegetation (EFVEG)—% of water surface covered by aquatic vegetation with four cate- gories (0–25%, 25–50%, 50–75%, > 75%), the density of submerged vegetation (SW) also 1 3 Biodiversity and Conservation 75%) 75%) > heavily disturbed) heavily Number of categoriesNumber 2 (Permanent or temporary) 2 (Stagnant or running)2 (Stagnant 4 (Without, small, medium, and dense) 4 (Without, 50%, and shaded—> 25%, medium—up to open—0%, small—up to 4 (Fully 4 (0–25%, 25–50%, 50–75%, 2 (Present or absent) 2 (Present 2 (Present or absent) 2 (Present 2 (Present or absent) 2 (Present 4 (Without, small, medium, and dense) 4 (Without, 2 (Steep or shallow) 2 (Steep 4 (Not visible, small disturbance, medium- partially disturbed and pronounced- and pronounced- medium- partially disturbed visible, small disturbance, 4 (Not covered by aquatic vegetation aquatic by covered Description Water permanence Water Stagnant or runningStagnant water Shadiness—% of water surface shaded by surrounding trees shaded by surface Shadiness—% of water The density of riparian vegetation Density of emergent and foating vegetation—% of water surface surface of water and foating vegetation—% Density of emergent Type of substrate—mud Type Type of substrate—soil Type Type of substrate—vegetation Type The density of submerged vegetation The density of submerged Bank slope Human impact Environmental variables used to describe the habitat characteristics at three investigated localities at the time of water beetle sampling beetle localities at the time of water describe used to variables the habitat at threeEnvironmental investigated characteristics 1 Table Variable Perm Stagfow Shad Veg EFVEG Submud Subsoil Subveg SW sh-st huimp

1 3 Biodiversity and Conservation four categories (without, small, medium, and dense). All densities were estimated using plant species data from plots around each sampling site (circle, 2 m diameter). Shadiness (Shad)—% of water surface shaded by surrounding trees was described with four catego- ries (fully open—0%, little—up to 25%, medium—up to 50%, and shaded—> 75%), human impact (huimp) with four categories (not visible, small disturbance, medium- partially dis- turbed and pronounced- heavily disturbed), water permanence and state are described with two categories—permanent or temporary and stagnant or running, respectively. Bank slope (sh-st) was described with two categories (steep or shallow). Type of substrate (Submud, Subsoil, and Subveg) was used as binary (present/absent) variable with a state of 1 or 0.

Statistical analysis

Water beetle assemblages and infuence of habitat features

We examined the diferences in the water beetle assemblages and the habitat features between three investigated localities (Nature Park Kopački rit, Nature Park Lonjsko polje, and Spačva basin) using Linear Discriminant Analysis (LDA) (Fisher 1936). Before the analysis species dataset was log­ 10(x + 1) transformed. We also tested all environmental variables for autocorrelation using Spearman rank-correlation. Consequently, one envi- ronmental variable (density of riparian vegetation) and one habitat type (coded as A3221) were omitted from further analyses due to strong collinearity (r > 0.7). We assessed three separate LDA models: (1) species abundance data, (2) the environmental variables describ- ing habitat characteristics, and (3) the appearance of habitat types (dummy variables) were used as variables, and the sampling sites were used as objects. We applied three a priori groups according to investigated localities. Centroids of these groups, as a mean of x and y coordinates of objects (sampling sites) belonging to the same group (sampling localities), were calculated and marked on the ordination plots. We also conducted a cross-validation procedure within LDA which provides membership probabilities of each object for the dif- ferent a priori groups based on the retained discriminant functions. To clarify the mean- ings of discriminant functions within LDA, Spearman rank-correlations were calculated between the discriminant axes and the variables. To evaluate the diferences in the composition of water beetle assemblages and the habi- tat features between three investigated localities we conducted permutational multivariate analyses of variance using distance matrices (ADONIS) (Oksanen 2011). We used sam- pling sites (10 sampling sites × 3 localities) as objects, the abundance data of the water beetle species or the environmental parameters as response variables, and sampling loca- tions (3 levels) as grouping variables within one model. We handled the species dataset and two environmental variable groups (habitat types and habitat characteristics) separately. A total of 999 runs were performed on the Bray–Curtis distance matrix for the species data- set, on the Kendall distance matrix for the habitat characteristics, and on the Jaccard dis- tance matrix for the habitat types. These statistical analyses were performed with software R using the packages ‘MASS’ for LDA and ‘vegan’ for ADONIS (R Core Team 2013). Canonical Correspondence Analysis (CCoA) was applied to assess the relationship between the water beetle assemblages and measured environmental variables (ter Braak and Smilauer 2002). Standard options of CANOCO were used except for the species abundance data which were ­log10(n + 1) transformed before the analyses. Rare taxa were not removed from analyses but down-weighted automatically (Cao et al. 1998). Forward selection procedures were carried out to ranking environmental variables by 1 3 Biodiversity and Conservation the conditional efects of assemblages. Variables were individually tested within their groups (habitat types and habitat characteristics) and selected according to the follow- ing criteria: explaining more than 4% of the total variance in biotic datasets and having a signifcant (α = 0.05) efect on the distribution patterns of assemblages. Unrestricted random Monte Carlo permutation tests (999 permutations) were performed to deter- mine the signifcance of individual environmental variables and canonical axes. To explore the relative infuence of the measured environmental variables (habitat types and habitat characteristics), the total variance was partitioned to the proportion attributed to habitat environmental variables and vegetation characteristics (habitat types) and their interactions (partial Canonical Correspondence Analysis—pCCoA) (Bocard et al. 1992). Several CCoA and pCCoA were performed: (1) CCoA with the habitat characteristics as variables, (2) CCoA with the habitat types as variables (3) pCCoA with the habitat characteristics as variables and the habitat types as covari- ables, and (4) pCCoA with the habitat types as variables and the habitat characteristics as covariables. The (3) and (4) steps (pCCoA) allowed us to separate the pure efect of habitat types and habitat characteristics and the efect of their interaction. The sum of eigenvalues of each analysis gives the percentage of variation explained by each group of variables. The CCoA and pCCoA procedures were carried out with software CANOCO ver. 4.5.

Infuence of habitat features on Graphoderus bilineatus

We used linear regression to explore the efects of the environmental variables on the abundance of G. bilineatus. These variables are interpreted on the ordinal scale there- fore we used the Helmert contrast coding system. Helmert coding compares each level of a categorical variable to the mean of the subsequent levels. For instance, the frst contrast compares the mean of the dependent variable for level 1 of human impact with the mean of all the subsequent levels of human impact (levels 2–3–4). The “full model” contained all variables. We reduced the full model with a stepwise algorithm based on the AIC value (Akaike 1973). The variables which remained in the reduced model were used for further analysis. For this analysis, we used R and the lm function from the stats package (R Core Team 2013). We compared species diversity of sampling sites with and without G. bilineatus occurrence using Mann–Whitney U-test to test our hypothesis that G. bilineatus occurs at sites supporting higher overall water beetle diversity. Finally, we performed a co- occurrence analysis to assess which species are positively or negatively associated with G. bilineatus. Briefy, the analysis is based on a C score that measures pairwise species co-occurrence by calculating the number of checkerboard units between all possible species pairs in a matrix (Stone and Roberts 1990). The observed C score is then com- pared to a null model (distribution of random matrices) to test the signifcance of asso- ciations. If the observed C score is higher than expected by chance the association is negative suggesting competition, while C scores lower than expected by the null model suggest positive association (Stone and Roberts 1990; Schmera et al. 2007). We com- pared observed patterns of G. bilineatus co-occurrence at all sampling sites with spe- cies recorded with at least 51 specimens (number of recorded G. bilineatus specimens) to a null model based on 1000 randomized matrices of the same size.

1 3 Biodiversity and Conservation

Results

Water beetle assemblages and Graphoderus bilineatus distribution patterns

A total of 4.339 specimens belonging to 39 genera and 98 species (Haliplidae, Noteridae, , Gyrinidae, Spercheidae, Hydrochidae, Helophoridae, and Hydrophilidae) were recorded at 30 sites during the survey (Table S2). The highest number of species was deter- mined in Lonjsko polje (79), followed by 61 species in Kopački rit, while the lowest num- ber of species was recorded in Spačva Basin (46). Graphoderus bilineatus was recorded at 14 out of the 30 surveyed sites with a total of 51 specimens. All specimens of G. bilinea- tus were caught at two out of three localities: in the Nature Park Kopački rit (21) and the Nature Park Lonjsko polje (30). On the other hand, we have not found any specimens in the Spačva basin forested foodplain. By analyzing the total species dataset of NP Kopački rit, NP Lonjsko polje, and Spačva using ADONIS signifcant diferences in water beetle assemblages were found among the three localities (F = 2.73, p = 0.001) (Table 2). This result was supported by the LDA plot based on the species dataset (Fig. 2a), where the centroids of the three localities were signifcantly distanced and only small marginal over- laps were observed between the groups (93.3% of objects were grouped correctly in mul- tidimensional space based on cross-validation procedure). Accordingly, we successfully elucidated diferences in assemblage patterns of three investigated localities, hence further analyses were carried out to identify which environmental variables explain the observed diferences in the water beetle assemblage structure.

Habitat features of the investigated localities and their relationship with the water beetle assemblages

The results of LDA models based on habitat types and habitat characteristics showed sig- nifcant structure between the three localities confrming that they also difer in their habi- tat features (Fig. 2b, c). Assessing the dataset of habitat characteristics globally, 93.3% of objects were grouped correctly in multidimensional space based on cross-validation pro- cedure, while this value reached 96.6% in the case of habitat types. The outcome of the ADONIS was similar. Marginally signifcant diferences were found among three localities based on habitat characteristics (F = 2.39, p = 0.07) and signifcant diferences were found among localities based on habitat types (F = 4.02, p = 0.001) (Table 2). Once we revealed

Table 2 2 Results of ADONIS Source of variation df F R p assessing the efect of geographical location on the Species composition composition of water beetles’ assemblages and the diferent Location 2 2.732 0.17 0.001** types of habitat features Residuals 27 0.83 Environmental variables Location 2 2.399 0.15 0.070** Residuals 27 0.85 Habitat types Location 2 4.019 0.23 0.001** Residuals 27 0.77

p values indicate statistical signifcance: **p < 0.01 1 3 Biodiversity and Conservation

Fig. 2 Using LDA, diferences are shown in a composition of water beetle assemblages b environmental variables describ- ing habitat characteristics, and c habitat types between sampling sites. Sampling sites were plot- ted on the ordination diagram. Diferent symbols (colours) represent the sampling sites at the three diferent localities: open circle—KP: Nature Park Kopački rit, open square—LP: Nature Park Lonjsko polje, + – SB: Spačva basin. The centroids of sites were marked with ‘ + ’ and the confdence intervals are drawn around the centroids (A) Water beetle assemblages

(B) Environmental variablesdescribing habitat characteristics

(C) Habitat types

1 3 Biodiversity and Conservation the key environmental factors linked to the variation of water beetle assemblage structure of the three localities, we aimed to specify the position of G. bilineatus in the community represented by the total species pool. Aquatic habitats at all three localities varied in permanence (permanent, temporary), the density of riparian, emergent, submerged, and foating vegetation, type of substrate, shad- ing, bank slope, and human impact (Fig. 2; Table 3). The canonical correspondence analysis (CCoA) was initially performed on the envi- ronmental variables, habitat types, and full species datasets to explain the distribution of species and position of G. bilineatus in the community (Fig. 3). The frst four axes of the CCoA together explained 30% of the total variance in the dataset (Table S3). Envi- ronmental variables density of riparian vegetation (Veg), shadiness (Shad), water perma- nence (Perm), bank slope (sh-st) and NKS classes (A. 4.1.2.6.) communities with Carex versicaria (As. Caricetum vesicaria), (A.3.2.1.4.) communities with Spirodela polyrrhizae and Salvinia natans (As. Spirodelo-Salvinietum natantis) best explained the distribution of the water beetle species (Table S4). The position of G. bilineatus suggests that the species prefers habitats with substrate soil, overgrown with Carex versicaria, and increased sub- merged vegetation. Variance partitioning within CCoA analysis shows that environmental variables account for 27.9%, habitat types for 13%, and their interaction for 10.3% of the total variability. Unexplained variability was 48.8% in this data. The linear regression analysis showed that the reduced model had lower AIC values than the full one (70.758 and 78.856, respectively). In addition, based on the F-test the models were not diferent (F = 0.347, p = 0.931) which also confrms that the reduced model is more appropriate to describe the efects of environmental variables. The restricted model contained six variables: water permanence (Perm), shadiness (Shad), type of sub- strate—soil (Subsoil), the density of submerged vegetation (SW), bank slope (sh-st), and human impact (Huimp). Based on the variance analysis, human impact and shading had

Table 3 Spearman correlation between environmental variables (habitat characteristics and habitat types) and the frst two linear discriminant axes (LD1, LD2) Habitat char. LD1 LD2 Habitat type LD1 LD2

Perm 0.14 0.04 A4141 0.29 0.11 Stagfow 0.05 0.23 A3212 − 0.35* 0.27 EFVEG − 0.13 − 0.43* A3213 − 0.02 0.37* Shad − 0.61*** 0.61*** A4119 − 0.33 − 0.02 Submud 0.04 − 0.45* A3331 − 0.49** -0.21 Subsoil − 0.64*** 0.05 A3335 0.15 0.19 Subveg − 0.02 0.31 A3232 0.01 0.25 SW − 0.05 − 0.22 A3336 0.15 0.11 sh-st 0.01 − 0.23 A3214 0.77*** − 0.56** huimp − 0.09 0.29 A4114 − 0.36* − 0.22 Habitat type A4116 − 0.42* − 0.32 A4126 − 0.01 0.44* A42 − 0.27 − 0.23 A3312 − 0.24 0.18 A3231 − 0.25 − 0.2 A4111 0.3 0.09 p values indicate statistical signifcance based on Spearman correlation ***p < 0.001; **p < 0.01; *p < 0.05. The full names of variables are listed in Table 1 and Table S1

1 3 Biodiversity and Conservation

Fig. 3 0 Partial canonical cor- . respondence analysis (pCCoA) 1 KP1 KP4 shows variables that best explain KP2 the distribution of water beetle A4126 species. Convex hulls represent Veg sampling sites with (purple) and KP7 without (yellow) G. bilineatus Shad SB1 independently from the localities. SB56 LP7 SB10 Other water beetle species are KP8 represented by x. The position Gra bil A4116 LP10 Subsoil A3214 of G. bilineatus (Gra bil) is Subveg emphasized. The frst axis of KP3 A4111 huimp SB4 LP3 SB2 LP2 the pCCoA had an eigenvalue SB9 LP6 KP5 LP8 of 0.237, but the frst four axes shst KP9 KP10KP6 SB3 SB8 together explain 30% of the LP1 LP9 LP5 Perm total variance in the dataset. For LP4 abbreviations see Table 1. and Table S1

SB7 - 0.8 -1.0 1.0 signifcant efects on the abundance of G. bilineatus (Table S5). Linear regression showed a signifcant diference between the frst level of the human impact and the subsequent lev- els (Table S6). The diference between the remaining two cases is decreasing as the human disturbance increases which indicate the negative efects of the anthropogenic disturbance on the abundance of G. bilineatus. The shading difered only between the last two levels. Besides, Perm and Subsoil also difered in their levels and thus the intercept of permanent durability and subsoil were higher than the second level (Table S6).

Co‑existence analysis of Graphoderus bilineatus

According to the Mann–Whitney U test for two independent samples, sites with G. bilin- eatus had signifcantly higher species richness (p = 0.00074) compared to the sites without (Fig. 4). Co-existence analysis was performed to test for associations between G. bilin- eatus and other water beetle occurrences, thus investigating the role of this species in the organization of the assemblages. The obtained results are based on species captured with a higher number of individuals (N > 53). Five species showed a signifcantly negative asso- ciation, nine species showed a signifcantly positive correlation, while fve species showed a neutral relationship with G. bilineatus (i.e., the value of their co-existence C score falls between the lower and upper 2.5 percentile range of the randomized index values; Fig. 5).

Discussion

Lowland riverine foodplains are the most dynamic and heterogeneous water bodies (Tock- ner and Stanford 2002). Despite their high spatio-temporal heterogeneity, foodplains cre- ate suitable conditions for numerous water beetles and for threatened species that depend on these river-foodplain habitats, such as Graphoderus bilineatus and Berosus geminus (Turić et al. 2015). Water beetles are excellent surrogates of overall macroinvertebrate 1 3 Biodiversity and Conservation

Fig. 4 Man-Whitney U test between sites with G. bilinea- tus and without G. bilineatus showing signifcant diferences indicating that G. bilineatus occurs where the overall species richness is higher

Fig. 5 The null model on association (expressed as co-existence index) with 19 pairs of species. Dots show the observed index value, while whiskers show the upper and lower 2.5 percentile ranges. The signifcant negative deviation is shown by the asterisk diversity in freshwater habitats and have been used in conservation and protection man- agement procedures and as well as a suitable umbrella species (Sánchez‐Fernández et al. 2006; Kolar and Boukal 2020). Therefore, the high species richness of foodplain habitats in the region reveals their importance for the maintenance of aquatic biodiversity, at both local and regional scales. Previous studies on Graphoderus bilineatus and its specifc habi- tat requirements have provided contradictory results that difer across the European spe- cies range (Galewski 1971; Holmen 1993; Nilsson and Holmen 1995; Hendrich and Balke,

1 3 Biodiversity and Conservation

2000; Cuppen et al. 2006; Bameul 2013; Iversen et al. 2013). Croatia lies on the south- eastern margin of G. bilineatus distribution in Europe and there are still many gaps in the current knowledge about the biology and ecological/habitat requirements of the species. Therefore, it was an urgent task to conduct a study of this kind to collect the data necessary for its efective conservation measures within the framework of the EU Habitat Directive and the NATURA 2000 network. During this study, we identifed 98 species of water beetles (Coleoptera), indicating high species diversity in the investigated foodplains. This represents an increase in the total number of species reported in a previous 6-year study at Kopački rit (Turić et al. 2015). Overall, our study contributed records of 14 new species for the water beetle fauna in the investigated area. Among them, there are several valuable species, rarely found in aquatic habitats throughout Croatia and typically captured as single specimens. Therefore, it seems that their lasting presence in Croatia wildlife is threatened. These species include Hydropo- rus umbrosus, Hydrochus megaphallus, Helophorus minutus, Crenitis punctatostriata, Rhantus consputus, and Cercyon bifenestratus. Several species that have been assigned as regionally extinct or under threat of regional extinction in Ireland or the UK (Foster et al. 2009; Foster 2010) were collected in the investigated foodplains in high or even very high numbers. These are, for example, Laccophilus hyalinus, Hygrotus versicolor, Clem- nius decoratus, Hydrochara caraboides, and Berosus signaticollis. Our study indicated the presence of species Hydaticus aruspex in the Nature Park Lonjsko polje. On the territory of neighbouring countries, it has been recorded only in at a small number of sites. The Croatian presence of H. aruspex is very surprising, and it probably represents a relict population from warmer periods of postglacial time (Bameul 1997). It is a species with a very fragmented distribution, although it is unknown whether the present rarity is the result of natural processes (i.e., it is a true relict) or human activity (habitat loss and fragmenta- tion). It is a potential candidate to be considered for inclusion in the EU legislation, in both EU Habitats Directive and Bern Convention as it fulflls all requirements: it is a large and conspicuous beetle, and in comparison, with the two species of aquatic Coleoptera already included (Dytiscus latissimus and Graphoderus bilineatus) it has a more restricted distribu- tion, fewer populations are known, and none is at present known to sustain a large number of specimens. More detailed studies are necessary both to assess its status and to design action plans to provide its persistence. The investigated foodplains difered in species richness and abundance as well as in the structure of the water beetle assemblages. The number of species (above 50) indicates that all three areas have valuable and well-preserved aquatic habitats for water beetles following Bameul (1994). The structure of water beetle assemblages of all three foodplains seems to refect the appropriate conditions for the conservation of biological diversity in terms of species richness of this group of insects (Eyre and Foster 1989; Hufnagel et al. 1999). Apart from marginal overlaps, there were remarkable diferences in species pools and abundances between localities. Nature Parks Kopački rit and Lonjsko polje have higher species abundance and diversity than Spačva Basin. This could be due to the increased number of ecological niches available for rare species provided by dynamically fooded areas in the two Nature Parks (Turić et al. 2015). Both Nature Parks are best-preserved naturally fooded areas in the Danube basin with high complexity and diversity of aquatic habitats where biotic communities are directly linked to water levels and refected by hydrological connectivity (Tockner et al. 2000). A 6-year study of fow variability and hydrological connectivity in a Danubian foodplain Kopački rit has revealed diferences in aquatic diversity and community composition between hydrologically contrast- ing periods indicating that diferent phases of hydrological connectivity produce distinct 1 3 Biodiversity and Conservation aquatic insect assemblages (Turić et al. 2015). On the other hand, the forests in the Spačva Basin have an almost stable water system, usually with spring and autumn foods. Also, the present state of the Spačva forest is strongly infuenced by intensive exploitation and numerous hydro-ameliorative activities in the past which have disrupted natural hydrology (Cestarić et al. 2017). Signifcant diferences between three foodplains were also found based on habitat char- acteristics (Fig. 2b) and even more based on habitat types with only marginal overlaps between the groups (Fig. 2c). Diferent environmental background of the three localities was accompanied by signifcant diferences in water beetle assemblages. The ecological traits and distribution of water beetles were primarily defned by environmental features (fow of water, the density of emergent and foating vegetation, shading, type of substrate, type of bank, human impact) and habitat types. Relation of water beetles’ assemblages and abundance to environmental features and habitat types (such as vegetation, water chemistry, hydroperiod, shading, size of water bod- ies, site age, and habitat homogeneity) has been supported by many studies (Eyre et al. 1990; Nilsson and Söderberg 1996; De Szalay and Resh 2000; Fairchild et al. 2000, 2003; Lundkvist et al. 2001; Bosi 2001; Oertli et al. 2002; Pakulnicka et al. 2015). However, environmental fow characteristics such as seasonal food pulses and fooding duration are perhaps among the main factors infuencing water beetle assemblages in lowland food- plains (Death 2008; Pakulnicka and Nowakowski 2012; Greenwood and Booker 2015; Turić et al. 2015). Unexplained variability in CCoA was higher probably due to the infu- ence of food regime components on water beetle assemblages. Recent food dynamics research on water beetle diversity in the Danubian foodplain has emphasized the impor- tance of diferent degrees of connectivity for high diversity because the changes in river fow alter the extent, duration, and frequency of foodplain inundation (Turić et al. 2015). Likewise, landscape factors such as landscape structure and type, landscape utilization, and geomorphology are important for aquatic environments and macroinvertebrates inhabiting them (Lundkvist et al. 2002; Iversen et al. 2013). In Spačva Basin, where the landscape is dominated by forests, non-overgrown permanent and anthropogenic water bodies have with no doubt a strong impact on the water beetle assemblages and their diversity. Shading from trees is an additional factor limiting the migration of beetles between water bodies in Spačva. Water beetle assemblages in diferent water bodies of three foodplains were also infuenced by the diferent types of substratum as well as the type of bank which both had a signifcant efect on the presence and number of water beetles. These fndings correspond to former studies where the substratum and condition and structure of valley slopes could indicate the natural environmental factors which are important for the aquatic invertebrate fauna (Ward 1992; Bereczki et al. 2012; Szivák and Csabai, 2012; Pakulnicka et al. 2015). The high degree of dependency of water beetle fauna and water body overgrowth by emer- gent and foating vegetation is highlighted by LDA analysis. For example, macrophytes density is higher along the shoreline and many authors emphasize the positive infuence of dense aquatic vegetation on water beetle’s diversity and abundance as vegetation provides shelter, food, and oviposition sites (Ambrožić et al. 2018; De Szalay and Resh 2000; Valla- dares et al. 2002; Molnár et al. 2009; Gioria et al. 2010; Pakulnicka et al. 2015). Structural complexity defned by plant density or confguration is a component of habitats hypoth- esized to afect species interactions, population densities and community composition in a variety of aquatic insects dominated systems (Yee 2010). Populations of G. bilineatus were recorded at 14 sampling sites along the wetlands of Lonjsko Polje and Kopački rit Nature Parks with a surprisingly high number of individuals (51 in total). Within the genus Graphoderus, besides the studied species, two additional 1 3 Biodiversity and Conservation species were recorded—G. cinereus and G. austriacus. G. austriacus had the lowest abun- dance in all three investigated areas, indicating that this species may be rarer and prob- ably more endangered than the G. bilineatus. The same results were confrmed in Switzer- land (Carron 2005). The G. bilineatus sex ratio difered in the areas where the species was sampled. In Lonjsko Polje the number of males and females was equal, while in Kopački rit males dominated. Considering the results of gender analysis in other groups of aquatic insects, it can be concluded that when the ratio is not 1:1, it is hard to determine what environmental factors infuence it. If females are dominant in certain habitats, all speci- mens were likely collected in micro-habitats with vegetation, where, according to exist- ing knowledge of their behaviour, the largest number of females can be expected (Nilsson 1996). One explanation for the decrease in the number of females is that the dominance of males may indicate the emergence of females or the territorial behaviour of males. Our study shows that the majority of G. bilineatus occurred in fooded-meadows and canals or ox-bows with stagnant waters, overgrown with Carex versicaria, and increased submerged vegetation. Only a few specimens were found in fshponds and temporary ponds. In other countries, G. bilineatus is reported from isolated waters such as large and natural ponds and lakes of various types as well as artifcial man-made ponds ditches, and canals (Cuppen et al. 2006; Iversen et al. 2013). The linear regression analysis revealed weak relationships between species presence, environmental features, and habitat types. Our fnding corroborates an earlier study that showed that G. bilineatus inhabits a wide range of habitats which is in contrast with the widespread perception of the species as being stenotypic (Iversen et al. 2013). In other words, none of the environmental variables, usually highlighted as most important (water permanence, water quality, water depth, veg- etation type) in the literature (Hendrich and Balke 2000; Cuppen et al. 2006), had a distinct efect on the presence of G. bilineatus. According to our analysis, only human impact and shading have signifcant efects on the abundance of G. bilineatus, which indicates the neg- ative efects of the anthropogenic disturbances. These results refect fndings of other stud- ies (Iversen et al. 2013, 2017; Turić et al. 2015), highlighting the importance of integrating landscape characteristics, connectivity, and stability to determine population viability of G. bilineatus. Today, the species distribution range is infuenced by strong anthropogenic impacts, leaving a less stable and unfavourable habitat structure for G. bilineatus. Iversen et al. (2013) documented that for G. bilineatus as species with low dispersal ability, estab- lishment and maintenance of a large range is likely due to generality in habitat demands (broader ecological niche than previously thought) and landscape connectivity. In lowland riverine foodplains survival of G. bilineatus mainly depends on hydrological connectiv- ity and habitat heterogeneity providing a broad selection of ecological niches (Turić et al. 2015). The foodplain environment contains well-mixed populations and many interact- ing species where some species occur together in communities, while others tend to avoid each other. Factors thought to explain co-existence patterns of water beetles vary (Vamosi and Vamosi 2007). Water beetles are a highly diverse group, both in their habitat requirements, feeding strategies, and reproduction. They represent diferent functional groups (guilds), comprised of predatory beetles, omnivores, collectors, sap- rophages feeding on dead organic matter, etc. The co-occurrence of many species is sometimes a consequence of their specialization or is a result of random events, i.e. an ecological drift (Pakulnicka et al. 2013). Although the diet of G. bilineatus is not yet sufciently known (Cuppen et al. 2006), it is known to be a predatory species and is thought to feed mostly on small crustaceans, larvae, and adults of other aquatic insects (Chironomidae, Diptera). We here identifed that G. bilineatus is commonly found in 1 3 Biodiversity and Conservation sites where the overall species richness of water beetles is signifcantly higher and some patterns of co-occurrence with other water beetle species were observed. Five species show a signifcant negative association with G. bilineatus. For example, a negative asso- ciation was frst identifed with Acilius sulcatus. Iversen et al. (2017) examine how fight ability might vary across seasons and between two genera, Acilius and Graphoderus with similar geographical ranges, life histories, and dispersal-related morphology. They found diferences in fight ability between species and that fight ability is important for profound diferences in local occurrences between these two species. Within study sites, Graphoderus is often found in large aquatic habitats, contra Acilius which occupies both large and small water habitats. Further, negative associations were found with Hydro- chara caraboides. Adult Hydrochara is frequently found in company with other large water beetles. G. bilineatus is species most likely to predate on, or compete with Hydro- chara like other large water beetles (of the genera Hydrophilus, Dytiscus, Acilius and Colymbetes) (Boyce 2004). Larvae of Graphoderus, like some other large water beetles, are extremely aggressive generalist predators. Ecology studies of Hydrochara in captiv- ity are shown that larvae of this species are a generalist ambush predator, whose choice of prey is largely dictated by the range of suitable prey species within its preferred size range that are most abundant in the microhabitats it occupies through its development (Boyce 2004). Consecutive species with negative association were species from genera Haliplus (Haliplus fuviatilis), and family Dytiscidae (Cybister lateralimarginalis and Lacco- philus minutus). In the negative association with Haliplus fuviatilis and Laccophilus minutus there are diferences in body sizes, prey composition, swimming behaviour, and habitat requirements. Yee (2010) documented that structural complexity and predation among genera Dytiscus, Graphoderus, and Rhantus can have important efects on exclu- sion and co-existence in aquatic habitats, especially if these predators act as intraguild predators. Species from genera Dytiscus are ferce predators feeding on practically any that they can kill and these groups likely have a generally negative efect on other dytiscids, as evidenced by the high predation rates on other genera (Yee 2010). The result suggests that G. bilineatus does not occur with just a few species while co- occurrence with other species is more common. In summary, our study shows distinct diferences in species richness and abundance of water beetles among three foodplain areas accompanied by diferent environmental backgrounds and habitat types. At the same time, naturally fooded areas in the Dan- ube basin harbour a rich water beetle fauna and may provide refuges for many rare and endangered species, like G. bilineatus which depend on these river-foodplain habitats. Our results indicate that globally threatened species G. bilineatus may have a favourable conservation status in Croatia which lies towards the southeastern distribution limit. Likewise, results support previous fndings emphasizing the importance of integrating landscape connectivity and hydrological connectivity with identifcation of the environ- mental requirements of highly valued species into conservation strategies of species to achieve long term persistence (Iversen et al. 2013, 2017; Turić et al. 2015).

Supplementary Information The online version contains supplementary material available at https://​doi.​ org/​10.​1007/​s10531-​021-​02168-w.

Acknowledgements The authors express their gratitude to Stjepan Kemfelja, Krešimir Mikulić, Edin Lugić, Petra Kutleša, Dajana Hmura, and Željko Zahirović for extensive help during feldworks. Thanks also for the valuable and constructive comments of two anonymous reviewers.

1 3 Biodiversity and Conservation

References

Akaike H (1973) Information theory as an extension of the maximum likelihood principle. In: Petrov BN, Csaki F (eds) Second International Symposium on Information Theory. Akadémiai Kiadó, Budapest, Hungary, pp 267–281 Ambrožič Š, Gaberščik A, Vrezec A, Germ M (2018) Hydrophyte community structure afects the presence and abundance of the water beetle family Dytiscidae in water bodies along the Drava River. Ecol Eng 120:397–404 Bakšić N, Pavlović N, Duić Ž (2015) Analysis of the Groundwater Series in the Area of Lowland Forests in the Spačva Basin. Doctoral symposium in civil engineering. Zagreb Bameul F (1994) Les Coléoptères aquatiques des Marais de la Perge (Gironde), témoins de la f n des temps glaciaires en Aquitaine. Bulletin de la Societe Entomologique de 99:301–321 Bameul F (1997) Hydaticus aruspex Clark rediscovered in France, and notes on the identifcation of the French species of Hydaticus Leach (Coleoptera, Dytiscidae). Bulletin de la Societe Entomologique de France 102(5):419–437 Bameul F (2013) Disparition de Graphoderus bilineatus (De Geer, 1774) (Coleoptera, Dytiscidae) des marais de la Perge cause par l’Ecrevisse américaine à patte rouges. Bulletin de la Societe Ento- mologique de France 118:133–136 Bayley PB (1991) The food pulse advantage and the restoration of river-foodplain systems. Regul Rivers 6:75–86 Bereczki C, Szivak I, Mora A, Csabai Z (2012) Variation of aquatic insect assemblages among seasons and microhabitats in Hungarian second-order streams. Aquat Insects 34(1):103–112 Bonacci O, Tadić Z, Moržan A, Radeljak I (2002) Park prirode Kopački rit - Plan upravljanja - Sektorska studija „Hidrologija i meteorologija“. Hidroing d.o.o. za projektiranje i inženjering, Osijek, Croatia, p 177 Borcard D, Legendre P, Drapeau P (1992) Partialling out the spatial component of ecological variation. Ecology 73:1045–1055 Bosi G (2001) Abundance, diversity and seasonal succession of dytiscid and noterid beetles (Coleoptera: Adephaga) in two marshes of the Eastern Po Plain (). Hydrobiologia 459:1–7 Boukal DS, Boukal M, Fikáček M, Hájek J, Klečka J, Skalickỳ S, Štastinỳ J, Trávniček D (2007) Catalogue of water beetles of the Czech Republic (Coleoptera: Sphaeriusidae, Gyrinidae, Haliplidae, Noteri- dae, Hygrobiidae, Dytiscidae, Helophoridae, Georissidae, Hydrochidae, Spercheidae, Hydrophilidae, Hydraenidae, Scirtidae, Elmidae, Dryopidae, Limnichidae, Heteroceridae, Psephenidae). Klapalekiana (Suppl) 43:289 Boyce DC (2004) A study of the distribution and ecology of the lesser silver water beetle Hydrochara car- aboides on the Somerset Levels. English Nature Research Reports. Peterborough. pp 69 Buijse AD, Coops H, Staras M, Jans LH, Van Geest GJ, Grifts RE, Ibelings BW, Oosterberg W, Roozen FCJM (2002) Restoration strategies for river foodplains along large lowland rivers in Europe. Freshw Biol 47:889–907 Cao Y, Williams DD, Williams NE (1998) How important are rare species in aquatic community ecology and bioassessment? Limnol Oceanogr 43:1403–1409 Carron G (2005) Kommentierte Checkliste der Dytiscidae und Noteridae (Coleoptera) der Schweiz. Mit- teilungen der Entomologischen Gessellschaft Basel 55(3):121–131 Cestarić D, Škvorc Ž, Franjić J, Sever K, Krstonošić D (2017) Forest plant community changes in the Spačva lowland area (E Croatia). Plant Biosyst 151(4):584–597 Csabai Z (2000) Vízibogarak kishatározója, I. kötet. Vízi természet- és kornyezetvedelem. 15. kötet, Környezetgazdálkodási intézet, Budapest, pp 267 Csabai Z, Gidó Z, Szél G (2002) Vízibogarak kishatározója II. [A Guide for the Identif cation of Water Beetles of Hungary II.] In: Vízi természet- és környezetvédelem. 16. kötet. [Water, Nature and Envi- ronment Protection. Vol. 16.] Environmental Management Institute, Budapest, pp 206 Cuppen J, Koese B, Sierdsema H (2006) Distribution and habitat of Graphoderus bilineatus in the Nether- lands (Coleoptera: Dytiscidae). Nederlandse Faunistische Mededelingen 24:29–40 De Szalay FA, Resh VH (2000) Factors infuencing macroinvertebrate colonization of seasonal wetlands: responses to emergent plant cover. Freshw Biol 45:295–308 Death RG (2008) Efects of foods on aquatic invertebrates. In: Lancaster J, Briers RA (eds) Aquatic insects: challenges to population. Proceedings of the Royal Entomological Society’s 24th Symposium, CAB International, Preston, UK, pp 103–122 EIONET (2012) Graphoderus bilineatus. European Environment Agency, 1–9. https://​bd.​eionet.​europa.​ eu/​artic​le17/​repor​ts2012/​speci​es/​summa​ry/?​group=​Arthr​opods​&​period=​3&​subje​ct=​Graph​ oderus+​bilin​eatus

1 3 Biodiversity and Conservation

Eyre MD, Foster GN (1989) A comparison of aquatic Heteroptera and Coleoptera communities as a basis for environmental and conservation assessments in static water sites. J Appl Entomol 108(1–5):355–362 Eyre MD, Foster GN, Foster AF (1990) Factors afecting the distribution of water beetle species assem- blages in drains of eastern England. J Appl Entomol 109:217–225 Fairchild GW, Faldus AM, Matta JF (2000) Beetle assemblages in ponds; efects of habitat and site age. Freshw Ecol 44:523–534 Fairchild GW, Cruz J, Faulds M, Short AEZ, Matta JF (2003) Microhabitat and landscape infuences on aquatic beetle assemblages in a cluster of temporary and permanent ponds. J N Am Benthol Soc 22:224–240 Fisher RA (1936) The use of multiple measurements in taxonomic problems. Ann Eugen 7:179–188 Foster GN (1987) The use of Coleoptera Records in Assessing the Conservation Status of Wetlands. In: Luf ML (ed) The use of invertebrates in site assessment for conservation. The University of New- castle upon Tyne, Newcastle upon Tyne, pp 8–17 Foster GN (1996) Graphoderus bilineatus (De Geer, 1774). In: Van Helsdingen PJ, Willemse LPM, Speight MCD (eds) Background information on invertebrates of the Habitats Directive and the Bern Conven- tion. Part 1 - Crustacea, Coleoptera, and Lepidoptera. European Invertebrate Survey, Leiden, pp 40–48 Foster GN (2010) A review of the scarce and threatened Coleoptera of Great Britain Part (3): Water bee- tles of Great Britain. Species Status 1. Joint Nature Conservation Committee, Peterborough Foster GN, Nelson BH, Connor Á (2009) Ireland Red List No. 1-Water beetles. National Parks and Wild- life Service, Department of Environment, Heritage and Local Government, Dublin, Ireland Galewski K (1971) A study of morphobiotic adaptations of European species of the Dytiscidae. Polskie Pismo Entomol 41:488–702 Galewski K (1975) Descriptions of the unknown larvae of the genera Hydaticus Leach and Graphoderus Dejean (Col., Dytiscidae) with some data on their biology. Annales Zoologici, Warszawa 32:249–268 Gioria M, Schafers A, Bacaro G, Feehan J (2010) The conservation value of farmland ponds: predicting water beetle assemblages using vascular plants as a surrogate group. Biol Conserv 143:1125–1133 Gopal B, Junk WJ (2000) Biodiversity in wetlands: an introduction. In: Gopal B, Junk WJ, Davis JA (eds) Biodiversity in wetlands: assessment, function and conservation, vol 1. Backhuys Publishers b.V., Leiden, pp 1–10 Greenwood MJ, Booker DJ (2015) The infuence of antecedent foods on aquatic invertebrate diversity, abundance, and community composition. Ecohydrology 8(2):188–203 Gueorguiev VB (1965) Deuxieme contribution a la connaissance des Coleopteres Hydrocanthares de Yougoslavie. Acta ento Musei Nationalis Pragae, Prag 36:489–500 Hendrich L, Balke M (2000) Verbreitung, Habitatbindung, Gefährdung und mögliche Schutzmaβnahmen der FFH- Arten Dytiscus latissimus Linnaeus, 1758 (Der Breitrand) und Graphoderus bilineatus (De Geer, 1774) in Deutschland (Coleoptera: Dytiscidae). Insecta Berlin 6:98–114 Hendrich L, Balke M (2002) Breitrand (Dytiscus latissimus) und Schmalbindiger Tauch-fügelkäfer (Graphoderus bilineatus). Berichtspfichten in Natura-2000-Gebieten – Empfehlungen zur Erfas- sungder Arten des Anhangs II und Charakterisierung der Lebensraumtypen des Anhangs I derFFH- Richtlinie. Angewandte Landschaftsökologie 42:301–305 Hendrich L, Balke M (2005) Graphoderus bilineatus (De Geer, 1774) (Coleoptera: Dytiscidae). In: Petersen B, Ellwanger G, Biewald G, Hauke U, Ludwig G, Pretscer P, Schröder E, Sysmank A (Eds.): Das europäische Schutzgebietssystem Natura 2000. Ökologie und Verbreitung von Arten der FFH-Richtlinie in Deutschland. Band 1: Pfanzen und Wirbellose. Schriftenreihe für Land- schaftspfege und Naturschutz 69(1):388–396 Hendrich L, Müller R (2017) Rote Liste und Gesamtartenliste der Wasserkäfer von Berlin (Coleoptera: Hydradephaga, Hydrophiloidea part., Hydraenidae, Elmidae und Dryopidae). 38 pp. In: Der Landes- beauftragte für Naturschutz und Landschaftspfege / Senatsverwaltung für Umwelt, Verkehr und Kli- maschutz (Eds.), Rote Listen der gefährdeten Pfanzen, Pilze und Tiere von Berlin. https://doi.​ org/​ 10.​ ​ 14279/depos​ itonce-​ 5851​ Holmen M (1993) Fredere insecter I . Del 3: Biller knytett till vand. Entomologiske Meddelelser 61:117–134 Hufnagel L, Bakony G, Vásárhelyi T (1999) New approach for habitat characterization based on species lists of aquatic and semiaquatic bugs. Environ Monit Assess 58:305–316 Iversen LL, Rannap R, Thomsen PF, Kielgast J, Sand Jensen K (2013) How do low dispersal species establish large range sizes? The case of the water beetle Graphoderus bilineatus. Ecography 36:770–777 Iversen LL, Rannap R, Briggs L, Sand-Jensen K (2017) Time-restricted fight ability infuences dispersal and colonization rates in a group of freshwater beetles. Ecol Evol 7(3):824–830 Junk WJ (1999) The food pulse concept of large rivers: learning from the tropics. Large Rivers 3:261–280

1 3 Biodiversity and Conservation

Kajzer A (2001) Prispevek k poznavanju vodnih hroščev (Coleoptera: Hydrocanthares) Slovenije in dela Bal- kana. Acta Entomologica Slovenica 9(1):83–99 Kálmán Z, Kálmán A, Soós N, Csabai Z (2007) Possibilities and restrictions of using bottle trap method for investigations on diving beetle populations I. Acta Biologica Debrecina Supplementum Oecologica Hun- garica 16:77–87 Klausnitzer B (1984) Käfer im und am Wasser mit 1 04 Abbildungen. Die Neue Brehmbücherei 567:148 Klimo E, Hager H, Matić S, Anić I, Kulhavý J (2008) Floodplain forests of the temperate zone of Europe. Zlín, Lesnická Práce, p 623 Koese B, Cuppen J (2006) Sampling methods for Graphoderus bilineatus (Coleoptera: Dytiscidae). Neder- landse Faunistische Mededelingen 24:41–48 Kolar V, Boukal DS (2020) Habitat preferences of the endangered diving beetle Graphoderus bilineatus: impli- cations for conservation management. Insect Conserv Divers. https://doi.​ org/​ 10.​ 11111/​ icad.​ 12433​ Larson DJ (1985) Structure in temperate predaceous diving beetle communities (Coleoptera: Dytiscidae). Ecog- raphy 8:18–32 Lundkvist E, Landin J, Milberg P (2001) Diving beetle (Dytiscidae) assemblages along environmental gradients in an agricultural landscape in southeastern Sweden. Wetlands 21:48–58 Lundkvist E, Landin J, Karlsson F (2002) Dispersing diving beetles (Dytiscidae) in agricultural and urban land- scapes in south-eastern Sweden. Ann Zool Fenn 39:109–123 Mihaljević M, Getz D, Tadić Z, Živanović B, Gucunski D, Topić J, Kalinović I, Mikuska J (1999) Kopački rit Research Survey and Bibliography. Hazu, Zavod za Znanstveni rad u, Osijeku, p 188 Molnár A, Csabai Z, Tóthmérész B (2009) Infuence of fooding and vegetation patterns on aquatic beetle diver- sity in a constructed wetland complex. Wetlands 29:1214–1223 Nikolić T (2011) Flora Croatica Database. http://​hirc.​botan​ic.​hr/​fcd. Prirodoslovno-matematički fakultet, Sveučilište u Zagrebu Nilsson AN (1996) Coleoptera, introduction. In Nilsson A (ed) Aquatic insects of North Europe. A Taxonomic Handbook Strenstrup. Apollo Books, pp 115–122 Nilsson AN, Holmen M (1995) The Aquatic Adephaga (Coleoptera) of Fennoscandia and Denmark: II. Dytisci- dae Fauna, Entomologica Scandinavica (No. 32). EJ Brill, Leiden, The Nilsson AN, Söderberg H (1996) Abundance and species richness patterns of diving beetles (Coleoptera, Dytiscidae) from exposed and protected sites in 98 northern Swedish lakes. Hydrobiologia 321:83–88 Oertli B, Joye DA, Castella E, Juge R, Cambin D, Lachavanne JB (2002) Does size metter? The relationship between pond area and biodiversity. Biol Conserv 104:59–70 Oksanen J (2011) Multivariate analysis of ecological communities in R: vegan tutorial. R Documentation, R Foundation for Statistical Computing. http://www.R-​ proje​ ct.​ org​ Pakulnicka J, Nowakowski JJ (2012) The efect of hydrological connectivity on water beetles fauna in water bodies within the foodplains of a lowland river (Neman river, Belarus). Oceanol Hydrobiol Stud 41(2):7–17 Pakulnicka J, Górski A, Bielecki A, Buczyński P, Tończyk G, Cichocka JM (2013) Relationships within aquatic beetle (Coleoptera) communities in the light of ecological theories, Fundamental and Applied Limnology. Arch Hydrobiol 18(3):249–258 Pakulnicka J, Górski A, Bielecki A (2015) Environmental factors associated with biodiversity and the occur- rence of rare, threatened, thermophilous species of aquatic beetles in the anthropogenic ponds of the Mas- urian Lake District. Biodivers Conserv 24:429–445 R Core Team (2013) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, . http://www.R-​ proje​ ct.​ org/​ Sánchez-Fernández D, Abellán P, Velasco J, Millán A (2004) Selecting areas to protect the biodiversity of aquatic ecosystems in a semiarid Mediterranean region. Aquat Conserv Mar Freshw Ecosyst 14:465–479 Sánchez-Fernández D, Abellán P, Mellado A, Velasco J, Millán A (2006) Are water beetles good indicators of biodiversity in Mediterranean aquatic ecosystems? The case of the Segura river basin (SE Spain). Biodiv- ers Conserv 15:4507–4520 Schmera D, Heino J, Podani J, Erös T, Dolédec S (2017) Functional diversity: a rewiev of methodology and current knowledge in freshwater macroinvertebrate research. Hydrobiologia 787:27–44 Šeric Jelaska L, Temunović M, Durbešic P (2008) Popis vodenih kornjaša podreda Adephaga iz zbirke Entomološkog odjela Gradskog muzeja Varaždin. In: Vargović E, Bregović A (eds.) Zbornik radova sa znanstvenog skupa „Franjo Košcec i njegovo djelo 1882.-1968. Zagreb- Varaždin: HAZU-Zavod za znan- stveni rad u Varaždinu, pp 163–172 Stone L, Roberts A (1990) The checkerboard score and species distributions. Oecologia 85(1):74–79 Szivák I, Csabai Z (2012) Are there any diferences between taxa groups having distinct ecological traits based on their responses to environmental factors? Aquat Insects 34(1):173–187 Ter Braak CJF, Smilauer P (2002) CANOCO Reference manual and CanoDraw for Windows User’s guide: Software for Canonical Community Ordination (ver. 4.5). Biometris, Wageningen Ceské Budejovice, pp 499 1 3 Biodiversity and Conservation

Tockner K, Stanford JA (2002) Riverine food plains: present state and future trends. Environ Conserv 29:308–330 Tockner K, Baumgartner C, Schiemer F, Ward JV (2000) Biodiversity of a Danubian foodplain: structural, functional, and compositional aspects. In: Gopal B, Junk WJ, Davis JA (eds) Biodiversity in wetlands: assessment, function, and conservation, vol 1. Backhuys Publishers, Leiden, pp 141–159 Turić N, Merdić E, Hackenberger Kutuzović B, Jeličić Ž, Vignjević G, Csabai Z (2012) Structure of aquatic assemblages of Coleoptera and Heteroptera in relation to habitat type and food dynamic structure. Aquat Insects 34:189–205 Turić N, Temunović M, Radović A, Vignjević G, Sudarić Bogojević M, Merdić E (2015) High food pulses drive the temporal dynamics of assemblages of aquatic insects (Heteroptera and Coleoptera) in a temper- ate foodplain. Freshw Biol 60(10):2051–2065 Turić N, Temunović M, Vignjević G, Antunović Dunić J, Merdić E (2017) A comparasion of methods for sampling aquatic insects (Heteroptera and Coleoptera) of diferent body sizes, in diferent habitats using diferent baits. Euro J Entomol 114:123-132 Valladares LF, Garrido J, Garcia-Criado F (2002) The assemblages of aquatic coleoptera from shallow lakes in the northern Iberian Meseta: infuence of environmental variables. Eur J Entomol 99:289–298 Vamosi JC, Vamosi SM (2007) Body size, rarity, and phylogenetic community structure: insights from diving beetle assemblages of Alberta. Divers Distrib 13:1–10 Vinnersten TZP, Lundström J, Petersson E, Landin J (2009) Diving beetle assemblages of fooded wetlands in relation to time, wetland type, and Bti-based mosquito control. Hydrobiologia 635:189–203 Vrazec A, Pirnat S, Kapla A, Denac D (2008) Zasnova spremljanja stanja populacij izbranih ciljnih vrst hroščev vključno z dopolnitvijo predloga območij za vključitev v omrežje NATURA 2000, Projektna naloga, Končno poročilo, Bioloski Institut Jovana Hadžija ZRC SAZU, Ljubljana, pp 101 Ward JV (1992) Aquatic insect ecology 1 Biology and habitat. Wiley, New York Ward JV, Tockner K, Arscott DB, Claret C (2002) Riverine landscape diversity. Freshw Biol 47:517–539 Wiggins GB, Mackay RJ, Smith IM (1980) Evolutionary and ecological strategies of in annual tempo- rary pools. Arch Hydrobiol 58:97–206 Williams P, Whitfeld M, Biggs J, Bray S, Fox G, Nicolet P, Sear D (2003) Comparative biodiversity of rivers, streams, ditches, and ponds in an agricultural landscape in Southern England. Biol Conserv 115:329–341 Yee DA (2010) Behavior and aquatic plants as factors afecting predation by three species of larval predaceous diving beetles (Coleoptera: Dytiscidae). Hydrobiologia 637:33–43

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations.

Authors and Afliations

Nataša Turić1 · Martina Temunović2 · Ildikó Szivák3 · Róbert Herczeg4 · Goran Vignjević1 · Zoltán Csabai5,6

1 Department of Biology, University Josip Juraj Strossmayer of Osijek, Cara Hadrijana 8/A, 31000 Osijek, Croatia 2 Department of Forest Genetics, Dendrology and Botany, Faculty of Forestry, University of Zagreb, Svetošimunska 25, 10000 Zagreb, Croatia 3 Department of Hydrozoology, Balaton Limnological Institute, MTA Centre for Ecological Research, Klebelsberg Kuno 3, Tihany 8237, Hungary 4 Bioinformatics Research Group, Szentágothai Research Centre, University of Pécs, Ifúság útja 6, Pécs 7624, Hungary 5 Department of Hydrobiology, University of Pécs, Ifúság útja 6, Pécs 7624, Hungary 6 Department of Botany and Zoology, Faculty of Sciences, Masaryk University, Kotlářská 2, 602 00 Brno, Czechia

1 3