Aquatic Insects of Man-Made Habitats: Environmental Factors Determining the Distribution of Caddisflies (Trichoptera), Dragonflies (Odonata), and Beetles (Coleoptera) in Acidic Peat Pools

Abstract

As degradation of sensitive habitats like Sphagnum L. (Sphagnales: Sphagnaceae) peatbogs is endangering their invertebrate fauna, artificial peat pools may offer peatbog insect fauna a chance of survival. The entomofauna of seven peat pools in a peatbog and its surrounding natural marginal zone in SE Poland was investigated at the level of species, assemblages and faunistic metrics, indicating the key environmental drivers of the insect distribution and their implications for the biodiversity and potential conservation of these habitats. The species composition, specialists, and insect assemblages of the peat pools were linked with the fauna typical of both peatbogs and dystrophic pools with an open water surface. The most specialized fauna was found in the pools with the largest Sphagnum cover: only tyrphobionts, of all the ecological elements, significantly discriminated the fauna of peat pools and the marginal zone. Sphagnum cover was the key structural factor affecting the distribution of all the insects. Additionally, dragonflies were dependent on pH, beetles on temperature, and caddisflies on dissolved oxygen; however, structural factors—apart from Sphagnum cover—pool perimeter and emergent vegetation cover were predominant. Our results show that appropriate management of the structural factors of peat pools, especially Sphagnum cover, and the provision of different successional stages, can enhance biodiversity and help to maintain a valuable specialist fauna. Even along small environmental gradients and in a homogeneous area, the response of insects is highly differentiated. Dragonflies probably best represent the conservation value of the overall invertebrate fauna of Sphagnum bogs.

The large-scale excavation of peat for fuel on peatbogs in Europe began in the eighteenth and nineteenth centuries (Ilnicki 2002). Such operations are still continuing on different scales depending on the method of extraction. Peat cutting is environmentally damaging, because destroying substrate, plant cover and biota, and irreversibly changing water regimes. Where heavy machinery is used to extract peat, large areas of bogs are devastated, including surrounding areas, which are important buffer zones. Such activities are particularly destructive on Sphagnum bogs. These are among the most seriously endangered ecosystems worldwide (Spitzer and Danks 2006, Kato et al. 2009), because these habitats, which support highly specific flora and fauna, are globally rare. As a result, there are very good reasons for listing them in the International RAMSAR Agreement, Annex 1 of the EU habitats directive and in the Natura 2000 program (EU 2008, Beadle et al. 2015).

Despite the unique character of Sphagnum bog habitats, knowledge on their biota is far from satisfactory and very uneven; plants being much more known than animals (Kato et al. 2009). Moreover, researchers often completely ignore Sphagnum bogs, claiming that the corresponding fauna is species-poor relative to other habitats and that Sphagnum bog conditions are hostile to many taxonomic groups (Desrochers and van Duinen 2006). The vast majority of Sphagnum bog species are sensitive specialists. Their populations can collapse very quickly following exposure to stochastic processes. Every bog can be considered as a habitat island (Spitzer et al. 1999, Spitzer and Danks 2006) with a low species capacity and small populations of individual species. No much research has been done on the biodiversity, assemblages, or even microhabitat preferences of particular species in peatbogs (Kato et al. 2009, Drinan et al. 2013, Beadle et al. 2015, Brown et al. 2016). At present, hardly any Sphagnum bog is wholly free from human-induced disturbances (Spitzer and Danks 2006). Yet, too little is known about the interactions between invertebrate assemblages and environmental factors (natural gradients or specific anthropogenic transformations) in peatland pools, lakes, or pits (e.g., Smits et al. 2002, Baars et al. 2015). Thus, in the context of intensifying human pressure, such data are highly attractive, as they are key to the planning of restoration or conservation projects.

All the human interventions in particular environments have not solely negative consequences (Buczyński 2015, Buczyński et al. 2017), although in the case of sensitive Sphagnum bogs, caution is always called for. A specific example is the restoration of peatbogs in the British Isles: drainage ditches were dammed, giving rise to pools of water behind the dam which act as a habitat analogous to that of natural peat pools in the same area (Beadle et al. 2015). Similar results can be achieved by the controlled cutting of peat: this leads to the formation of tiny waterbodies – peat pools. Many studies of Sphagnum bogs have, thus, focused on newly created systems with the aim of evaluating the potential success of restoration using a variety of biological metrics based on the invertebrate fauna (e.g., Hannigan et al. 2011, Beadle et al. 2015, Elo et al. 2015, Brown et al. 2016). The detailed environmental relationships in old, pre-World War Two peat workings, in which succession has largely restored the original conditions, are poorly understood, especially in comparison with remnants of natural peatbogs since none of them are in the area surveyed. In south-eastern Poland, peat pools do not imitate natural nanotopes. They have rather introduced a new qualitative aspect to the landscape in so far as such pools do not exist in Sphagnum habitats in this part of the country. As a result, peat pools may take on considerable importance: they may become refugia for the fauna of peatbogs and their associated small waterbodies. Even in the case of very small or isolated peatbogs, they may become the sole reservoir of species for disturbed habitats. Their significance increases even more during periods of drought: relatively deep peat pools are then the most stable habitats for the peatbog fauna (Buczyński 2015). This is particularly important in the context of global climate warming (Batzer and Boix 2016). Obviously, if peat pools are playing such role, habitat conditions favorable to a particular animal group have to evolve and then require the time necessary for succession to take place.

In this paper, we analyze the occurrence of aquatic insects in diverse, former peatland habitats and in the natural habitat of a single Sphagnum peatbog. By choosing these particular study sites, we wished to avoid the effect of geographical differentiation of localities on the results. The selected localities lie close to one another and on one peatbog. They are subject to the same hydrographic, microclimatic, and catchment area regimes, and are colonized from the same local waters. The three orders of insects chosen for this study—Odonata, Coleoptera, Trichoptera—are regarded as good bioindicators of the state of the environment and biodiversity surrogates. Although their biology and ecology are relatively well known, their assemblages, their responses to specific environmental gradients/factors and even the microhabitat preferences of particular species (especially beetles and caddisflies) in peatbogs are poorly understood (e.g., Houghton 2004, Suriano et al. 2011, Ruhi and Batzer 2014).

The aims of this work were to 1) ascertain whether the peat pool fauna (analyzed at the levels of species, assemblages, and ecological metrics) is associated with the peatbog remnant (marginal zone) or whether a completely new faunistic system is appearing; 2) define the key physical and chemical properties of the water and the structural factors of peat pools governing the distribution of dragonflies, beetles, and caddisflies; and 3) specify the microhabitat preferences of particular species from the three orders. In the context of peatland conservation, the results will enable us to state whether artificial peat pools are a suitable surrogate (secondary) habitat for aquatic insects and how the biodiversity of the study area is linked to the occurrence of a specialized acidophilous fauna.

Material and Methods

Study Area and Sampling Procedures

The study was carried out in Western Polesie, a geographical region in the west of the East European Plain. The actual study plot was the Krugłe Bagno peatbog (area ca. 6.5 ha) (Fig. 1A) in the village of Zagłębocze (21°01′38″–21°01′59″E, 51°25′00″–51°25′12″N). This is a former continental raised bog, which by the mid-twentieth century had become almost completely exploited. The peat workings subsequently became densely overgrown with transition-bog-like Sphagnum mats, which now cover almost the entire area of the Krugłe Bagno. Nine peat pools, from 120 to 1,690 m² in area, have persisted to the present day. These are shallow waterbodies, no more than ca. 1-m deep, filled with Dy-type sediments. The Krugłe Bagno is surrounded by a 50- to 500-m wide belt of largely mixed woodland, with dominant Pinus sylvestris L. (Pinales: Pinaceae). Only its north-western part borders on buildings, fallow land, and single fields (Buczyński and Staniec 1998).

Fig. 1.

Study area and study sites: (A) the peatbog and its surroundings (MZ, marginal zone of the peatbog), 1–7: peat pools (Google Earth); (B) marginal zone of the peatbog, (C) peat pool No. 2 (photographs: Paweł Buczyński and Edyta Buczyńska).

Insects were sampled in seven peat pools and the marginal zone (natural remnant) (Fig. 1B and C) during six expeditions in spring (April, May), summer (June, July), and autumn (August, October) of 2013. In April, however, only two pools and the marginal zone could be investigated because the other five pools were still covered with ice. Three 4-m-long transects were marked out around the perimeter of each pool. The standardized sampling method for collecting larval stages involved sweeping along the transects using a 250-μm mesh hydrobiological sampler. Insects were sampled to a maximum depth of 0.5 m. The material was sorted in the laboratory and preserved in 70% ethanol. All the specimens were identified to species level, except for some larval stages of Coleoptera, which for lack of identification keys, could only be identified to genus level.

The following physical and chemical water parameters were measured at all sites in situ using a Hanna HI 9828 multiparameter portable meter: water temperature (°C), pH, oxygen reduction potential (mV), dissolved oxygen concentration (ppm), electrical conductivity (µS/cm), total dissolved solids (ppm), and salinity (PSU). The structural parameters distinguishing the pools were surface area – size (m²), pool perimeter (km) and Sphagnum cover (%). Sphagnum cover was calculated from photographs of the pools taken in different months. In addition, three parameters relating to vascular plants were calculated for each pool: i) helophyte cover (Typha angustifolia L., Typha latifolia L. (Poales: Typhaceae), and Eleocharis palustris (L.) Roem. & Schult. (Poales: Cyperaceae)) in the water – helo, ii) emergent vegetation cover (Carex spp. (Poales: Cyperaceae), Juncus spp. (Juncales: Juncaceae), and Eriophorum spp. (Poales: Cyperaceae)) on the pool shore – helo_w, and iii) the presence of trees on the shore – trees. Five classes describing these factors were distinguished, coded for further analysis as follows: 0, no plants/trees; 1, plants/trees poorly developed; 2, plants/trees moderately developed; 3, plants/trees well developed; and 4, plants/trees very well developed. The environmental features of each site are detailed in Table 1.

Table 1.

The environmental variables (mean value ± SD and range of values) at eight study sites situated in the peatbog

Variable	MZ	Peat pools
		1	2	3	4	5	6	7
Temp (°C)	15.2 ± 6.9 4.3 to 21.1	12.5±4.8 4.3 to 17.4	17.1 ± 3.8 11.0 to 23.3	14.8 ± 4.4 9.4 to 19.0	16.5 ± 4.2 10.2 to 21.2	17.7 ± 4.8 9.9 to 21.7	16.7 ± 5.6 9.7 to 23.1	16.1 ± 7.5 4.4 to 26.2
pH	4.68 ± 0.77 3.96 to 5.91	4.45 ± 0.23 4.11 to 4.70	4.45 ± 0.35 3.95 to 4.92	4.36 ± 0.19 4.10 to 4.54	4.43 ± 0.21 4.17 to 4.70	4.87 ± 0.32 4.56 to 5.35	4.62 ± 0.44 4.04 to 5.15	4.48 ± 0.18 4.28 to 4.81
ORP (mV)	−98.3 ± 57.3 −139.1 to 15.4	−147.1 ± 35.7 −209.7 to −106.5	−126.5 ± 33.3 −170.7 to −92.8	−148.3 ± 49.1 −205.9 to −88.4	−109.6 ± 29.1 −142.3 to −80.1	−133.0 ± 41.6 −195.9 to −92.6	−117.6 ± 40.5 −179.2 to −76.1	−106.6 ± 28.7 −137.2 to −74.0
DO (mg·d/m 3)*	2.58 ± 2.14 0.00 to 4.87	0.05 ± 0.12 0.00 to 0.31	0.58 ± 0.22 0.32 to 0.82	0.15 ± 0.24 0.00 to 0.55	0.54 ± 0.85 0.00 to 2.01	1.53 ± 2.27 0.00 to 5.56	1.89 ± 2.24 0.00 to 5.32	0.99 ± 0.48 0.47 to 1.81
EC (μS/cm)	270 ± 535 20 to 1362	51 ± 41 18 to 133	32 ± 27 2 to 64	31 ± 8 18 to 38	31 ± 14 16 to 45	29 ± 20 1 to 54	37 ± 23 17 to 74	40 ± 21 21 to 67
TDS (mg·d/m 3)	135 ± 268 10 to 681	25 ± 21 9 to 67	16 ± 14 1 to 32	16 ± 4 9 to 19	15 ± 7 8 to 23	15 ± 10 0 to 27	19 ± 12 8 to 37	20 ± 11 10 to 34
SAL (PSU)	0.13 ± 0.27 0.01 to 0.68	0.02 ± 0.01 0.01 to 0.06	0.01 ± 0.01 0.00 to 0.03	0.01 ± 0.01 0.01 to 0.02	0.01 ± 0.01 0.00 to 0.02	0.01 ± 0.01 0.00 to 0.02	0.01 ± 0.01 0.00 to 0.03	0.01 ± 0.01 0.00 to 0.02
SC (%)	–	10 ± 010	66 ± 17 40 to 80	3.4 ± 2.3 0 to 5	29.6 ± 3.1 5 to 30	30 ± 030	30 ± 030	100 ± 0 100
Size (m 2)	–	520	120	360	480	220	1690	120
PP (km)	–	0.122	0.049	0.189	0.120	0.070	0.430	0.450
trees*	–	2	3	2	3	2	2	4
Helo	–	0.67 ± 0.51 0 to 1	1.67 ± 1.03 0 to 3	0.50 ± 0.55 0 to 3	0.67 ± 0.52 0 to 1	0.83 ± 0.41 0 to 1	1.67 ± 0.82 0 to 2	3.00 ± 1.67 0 to 4
helo_w	–	0.83 ± 0.41 0 to 1	1.67 ± 0.82 0 to 2	1.67 ± 0.82 0 to 2	1.67 ± 0.82 0 to 2	2.50 ± 1.22 0 to 3	1.50 ± 0.84 0 to 2	0.83 ± 0.41 0 to 1

MZ, marginal zone of the peatbog.

Variables are given as codes (see Study area and sampling procedures) with units. Variables with statistically significant differences are denoted by an asterisk (*).

The character of the peatbog’s immediate surroundings was analyzed in the field and on the basis of aerial photographs available from Geoportal (http://geoportal.gov.pl/); the area and perimeter of each peat pool were calculated using the tools available on the same website.

Data Analysis

Faunistic Metrics

The following faunistic metrics were used: species richness (S), abundance (N), the Shannon–Wiener diversity index (H), Buzas-Gibson’s evenness (E), and the Simpson dominance index (D). Calculation of the dominance structure for every peat pool and the marginal zone was based on the division into classes proposed by Biesiadka (1980).

The categories of specialists/nonspecialists are commonly used to describe peatbog fauna (Spitzer and Danks 2006). To assess whether the fauna was specialized and typical of the habitat or was composed of nonspecialists (eurytopes), coefficients of naturalness (Wns, quantitative formula; Wni, qualitative formula) were calculated for each study site by substituting the stenotopy index (Wze) (taking values from 0 to 16 for each species) into Fischer’s (1996) formula, modified by Czachorowski and Buczyński (1999). Values from 16 to 4 defined specialists: tyrphobionts (Wze = 16), species occurring exclusively in peat bogs, and tyrphophiles (Wze = 8 or 4), species typical of peat bogs, but not strictly confined to such habitats. The lowest values (Wze = 2 or 1) defined non-specialists—eurytopes and random species. The Wze values were adopted after Czachorowski and Buczyński (1999) and Klausnitzer (1996) or, if data for a particular species were unavailable, defined individually by the authors. The Mann–Whitney and Kruskal–Wallis tests (in combination with Dunn’s post hoc test) were used to detect significant differences in faunistic indices between the sites (habitats) on the sample matrix. The calculations were performed in STATISTICA 11.

Similarity Analysis: Species and Habitats

Two-way indicator species analysis (TWINSPAN) was used to create the cladogram distinguishing the insect assemblages and to identify the indicator species associated with the acidic waters studied. TWINSPAN was performed in WinTWINS 2.3 (Hill and Šmilauer 2005); species with less than three occurrences (given as frequency) were excluded from the matrix for this analysis.

Hierarchical agglomerative clustering with the Unweighted Pair Group Method with Arithmetic Mean was applied to define the faunistic similarities between the seven peat pools and the marginal zone. A dendrogram based on the Bray–Curtis index was constructed in the PAST 3.18 program (Hammer et al. 2001).

Insect Assemblages and Species Versus Environmental Factors

Variation partitioning analysis based on canonical ordination (CCA) was used to display the effects of physical and chemical as well as structural environmental variables in the species composition of each insect order. To construct the matrix only data relating to the species of which there were >1 individual in a sample were used. Stepwise selection with 499 test permutations was employed to filter the significant (P < 0.05) parameters primarily responsible for species variation. This showed whether and to what extent species variability within the particular insect assemblages was governed by physical and chemical factors, structural factors, and their potential shared effect (Šmilauer and Lepš 2014). The statistical significance of each fraction (primary factors, secondary factors, and their sum including shared effects) was also evaluated. On this basis, CCA biplots were generated for each group, with significant factors only being taken into account.

Generalized linear models (GLMs) were applied to identify the optimal conditions for the most abundant insect species. Based on the environmental variables that proved to be the most significant for a given insect order, these predictive models indicated the relevant tolerances and optima that might be useful for distinguishing the ecological requirements of a species. In accordance with the recommendations for this analysis (Šmilauer and Lepš 2014), we used a Poisson distribution to construct the models.

Variation partitioning analyses (CCAs) and GLMs were calculated and plotted using CANOCO 5 statistical software (ter Braak and Šmilauer 2012).

Results

The Insect Fauna of Peat Pools: Species Richness and Diversity Indices

In total, 1,837 insects from 76 species were caught. Beetles were the species-richest assemblage (51), followed by dragonflies (16), and caddisflies (9) (Table 2). The marginal zone supported 56% of all the species recorded, the individual pools 25–44% and the pools combined 92.5%. About 9% of the species were exclusive to the marginal zone, from 1 to 6% were found in the individual pools, but as many as 44% in the peat pools combined. The dominance structures were the most evenly balanced among the beetles, but distinctly less so among the dragonflies and caddisflies, though at roughly the same level. The largest numbers of taxa—predominantly dragonflies and beetles—were recorded in the marginal zone. The diversity index was also the highest there. Taxa richness was the highest in pools 1 and 4, and the lowest in pool 7. Peat pools 1 and 7 exhibited the greatest faunal diversity; the diversity indices for the other pools were much the same. The dominance in peat pools 1–6 was very similar; this value was very low only in pool 7 and in the marginal zone. The evenness index was the highest in the most densely vegetated peat pool 7, identical in the marginal zone and pool 5, and slightly lower in the other pools (Table 2).

Table 2.

List of the insect species collected. Wze, stenotopy index; 1–7, peat pools; MZ, marginal zone of the peatbog. %N, percentage of a species in the total number of individuals. Eudominants (>10%) are given in bold.

Taxon	Code	Wze	Dominance at study sites [%]								%N
			Peat pools							MZ
			1	2	3	4	5	6	7
ODONATA
Lestidae
Lestes dryas Kirby	Les_dry	2	–	–	–	–	–	–	–	0.6	0.1
Lestes sponsa (Hansem.)	Les_spo	2	1.5	–	0.7	0.9	–	0.7	–	10.3	2.2
Lestes virens (Charp.)	Les_vri	4	57.6	17.4	41.6	20.2	11.8	27.7	–	17.4	28.0
Coenagrionidae
Coenagrion hastulatum (Charp.)	Coe_has	4	20.5		8.8	5.2	8.8	2.0	–	0.6	5.8
Coenagrion puella (L.)	Coe_pue	1	11.4	0.7	13.9	15.9	–	17.6	–	0.6	10.0
Coenagrion pulchellum (Vander L.)	Coe_pul	2	0.8	–	–	–	–	–	–	–	0.1
Nehalennia speciosa (Charp.)	Neh_spe	8	–	4.7	0.7	–	44.1	1.4	–	7.1	3.6
Aeshnidae
Aeshna cyanea (O.F. Müll.)	Aes_cya	1	–	–	–	0.4	–	–	–	–	0.1
Aeshna subarctica Walk.	Aes_sub	16	–	0.7	–	–	–	–	–	–	0.1
Corduliidae
Cordulia aenea (L.)	Cor_aen	2	–	–	–	3.0	–	–	–	–	0.7
Libellulidae
Leucorrhinia dubia (Vander L.)	Leu_dub	16	1.5	5.4	2.2	1.7	–	2.7	–	2.6	2.5
Leucorrhinia pectoralis (Charp.)	Leu_pec	4	5.3	34.9	28.5	44.2	26.5	42.6	–	38.1	33.5
Leucorrhinia rubicunda (L.)	Leu_rub	8	1.5	33.6	2.2	7.7	5.9	3.4	25.0	13.5	10.3
Libellula quadrimaculata L.	Lib_qua	2	–	0.7	1.5	0.4	–	2.0	75.0	7.1	2.1
Libellulidae: Sympetrinae juv.	–	–	–	1.3	–	–	–	–	–	–	0.2
Sympetrum danae (Sulz.)	Sym_dan	8	–	0.7	–	0.4	2.9	–	–	0.6	0.4
Sympetrum vulgatum (L.)	Sym_vul	1	–	–	–	–	–	–	–	1.3	0.2
N=	–	–	132	149	137	233	34	148	4	155	992
COLEOPTERA
Gyrinidae
Gyrinus sp.	–	–	–	–	1.4	–	–	–	–	–	0.1
Haliplidae
Haliplus ruficollis (De Geer)	Hal_ruf	2	–	–	–	–	1.1	–	–	–	0.1
Haliplus wehnckei Gerh.	Hal_weh	4	–	–	–	1.1	–	–	–	–	0.1
Noteridae
Noterus clavicornis (De Geer)	Not_cla	2	1.0	–	–	–	–	–	–	>0.1	0.1
Noterus crassicornis (O.F. Müll.)	Not_cra	4	19.2	19.0	32.4	22.1	18.3	31.1	15.8	12.4	19.6
Dytiscidae
Acilius canaliculatus (Nic.)	Aci_can	2	6.1	–	–	1.1	–	–	–	–	1.0
Acilius sulcatus (L.)	Aci_sul	4	2.0	–	5.4	4.2	–	–	–	–	1.4
Agabus sp.	–	–	4.0	–	4.1	–	–	–	–	2.8	1.7
Bidessus unistriatus (Goeze)	Bid_uni	8	–	2.4	1.4	–	2.2	–	1.8	–	0.8
Cybister lateralimarginalis (De Geer)	Cyb_lat	2	6.1	–	–	2.1	–	2.2	–	–	1.2
Dytiscus circumflexus Fabr.	Dyt_cir	2	1.0	–	–	–	–	–	–	–	0.1
Dytiscus dimidiatus Bergstr.	Dyt_dim	4	2.0	–	–	–	–	6.7	–	–	0.7
Dytiscus marginalis L.	Dyt_mar	2	6.1	1.2	5.4	9.5	–	–	–	1.7	3.2
Graphoderus austriacus (Sturm)	Gra_aus	2	2.0	–	1.4	2.1	4.3	2.2	–	12.4	4.4
Graphoderus cinereus (L.)	Gra_cin	2	18.2	–	5.4	11.6	4.3	26.7	–	7.3	8.6
Hydaticus seminiger (De Geer)	Hyd_sem	2	–	2.4	–	–	–	–	–	–	0.3
Hydaticus stagnalis (Fabr.)	Hyd_sta	2	–	–	–	–	–	–	–	0.6	0.1
Hydaticus transversalis (Pontopp.)	Hyd_tra	2	–	–	–	–	–	–	–	1.7	0.4
Hydroglyphus geminus (Fabr.)	Hyd_gem	1	1.0	–	–	1.1	–	–	3.5	–	0.6
Hydroporus angustatus Sturm	Hyd_ang	8	–	–	1.4	–	1.1	–	3.5	2.8	1.2
Hydroporus erythrocephalus (L.)	Hyd_ery	8	2.0	1.2	1.4	–	–	–	–	2.3	1.1
Hydroporus incognitus Sharp	Hyd_inc	8	–	–		–	–	–	3.5	2.3	0.8
Hydroporus obscurus Sturm	Hyd_obs	16	–	1.2	1.4	–	–	–	7.0	1.1	1.1
Hydroporus scalesianus Steph.	Hyd_sca	16	–	–	–	–	–	–	1.8	–	0.1
Hydroporus striola (Gyll.)	Hyd_str	8	–	–	–	–	–	–	–	10.7	2.6
Hydroporus tristis (Payk.)	Hyd_tri	8	–	–	–	–	1.1	–	–	–	0.1
Hydroporus sp.	–	–	–	–	–	–	–	–	–	4.0	1.0
Hygrotus decoratus (Gyll.)	Hyg_dec	8	3.0	7.1	5.4	1.1	–	–	15.8	1.7	3.6
Hygrotus impressopunctatus (Schall.)	Hyg_imp	2	–	4.8	–	1.1	2.2	–	–	0.6	1.1
Hygrotus inaequalis (Fabr.)	Hyg_in	2	4.0	10.7	–	9.5	6.5	6.7	3.5	1.7	5.0
Hygrotus versicolor (Schall.)	Hyg_ver	2	1.0	–	1.4	–	–	–	–	0.6	0.4
Taxon	Code	Wze	Dominance at study sites [%]								%N
			Peat pools							MZ
			1	2	3	4	5	6	7
Hyphydrus ovatus (L.)	Hyh_ova	2	–	–	–	–	–	–	–	0.6	0.1
Ilybius sp.	–	–	–	–	–	1.1	–	–	–	>0.1	0.1
Laccophilus hyalinus (De Geer)	Lac_hya	1	–	–	–	1.1	>0.1	–	–	>0.1	0.1
Laccophilus poecilus Klug	Lac_poe	8	12.1	23.8	16.2	24.2	37.6	8.9	–	12.4	17.7
Porhydrus lineatus (Fabr.)	Por_lin	4	1.0	–	2.7	1.1	–	2.2	–	–	0.7
Rhantus suturalis (MacL.)	Rha_sut	2	1.0	–	–	–	1.1	–	–	–	0.3
Helophoridae
Helophorus granularis (L.)	Hel_gra	4	–	1.2	–	–	–	–	1.8	–	0.3
Hydrochidae
Hydrochus crenatus (Fabr.)	Hyd_cre	2	–	2.4	–	1.1	–	2.2	–	1.1	0.8
Hydrophilidae
Anacaena limbata (Fabr.)	Ana_lim	2	–	1.2	–	–	1.1	–	–	–	0.3
Anacaena lutescens (Steph.)	Ana_lut	8	–	–	–	–	–	–	3.5	2.3	0.8
Anacaena sp.	–	–	–	–	–	2.1	–	–	–	0.6	0.4
Cercyon analis (Payk.)	Cre_ana	2	1.0	–	–	–	–	–	–	–	0.1
Coelostoma orbiculare (Fabr.)	Coe_orb	2	–	1.2	–	–	–	–	–	1.1	0.4
Enochrus affinis (Thunb.)	Eno_aff	8	1.0	3.6	4.1	–	–	2.2	8.8	2.8	2.5
Enochrus coarctatus (Gredl.)	Eno_coa	8	–	–	–	–	–	–	–	0.6	0.1
Enochrus melanocephalus (Ol.)	Eno_mel	8	–	1.2	–	1.1	1.1	2.2	7.0	1.1	1.4
Enochrus ochropterus (Marsh.)	Eno_och	8	–	–	–	–	3.2	–	–	–	0.4
Enochrus quadripunctatus (Herbst)	Eno_qua	2	1.0	–	–	1.1	–	–	3.5	0.6	0.7
Enochrus testaceus (Fabr.)	Eno_tes	2	–	1.2	–	–	–	>0.1	–	–	0.1
Enochrus sp.	–	–	1.0	–	–	–	–	2.2	–	0.6	0.4
Helochares obscurus (O.F. Müll.)	Hel_obs	2	3.0	13.1	8.1	–	5.4	2.2	12.3	3.4	5.4
Hydrobius fuscipes (L.)	Hyb_fus	2	–	–	–	–	–	–	1.8	>0.1	0.1
Hydrochara caraboides (L.)	Hyd_car	2	–	–	1.4	–	–	–	1.8	1.1	0.6
Laccobius minutus (L.)	Lac_min	2	–	–	–	–	1.1	–	–	>0.1	0.1
Laccobius sp.	–	–	–	1.2	–	–	>0.1	–	–	5.1	1.4
Scirtidae
Elodes sp.	–	–	–	–	–	1.1	–	–	–	>0.1	0.1
Microcara testacea (L.)	Mic_tes	8	–	–	–	–	1.1	–	–	>0.1	0.1
Scirtes sp.	–	–	–	–	–	–	5.4	–	3.5	–	1.0
Curculionidae
Tanysphyrus lemnae (Payk.)	Tan_lem	1	–	–	–	–	2.2	2.2	–	–	0.4
N=	–	–	99	84	74	95	93	45	57	177	724
TRICHOPTERA
Polycentropodidae
Cyrnus insolutus McL.	Cyr_ins	8	2.7	–	–	–	–	–	–	–	0.8
Holocentropus dubius (Ramb.)	Hol_dub	8	35.1	–	90.2	56.3	–	50.0	–	–	60.3
Leptoceridae
Oecetis furva (Ramb.)	Oec_och	2	13.5	–	–	–	–	–	–	–	4.1
Limnephilidae
Limnephilus stigma Curt.	Lim_sti	8	–	–	–	–	–	10.0	–	–	0.8
Phryganeidae
Agrypnia obsoleta (Hag.)	Agr_obs	8	–	–	–	–	–	10.0	–	–	0.8
Agrypnia varia (Fabr.)	Agr_var	4	45.9	–	7.3	37.5	100.0	–	–	–	27.3
Oligotricha striata (L.)	Oli_str	4	–	–	–	3.1	–	–	–	–	0.8
Phryganea bipunctata Retz.	Phr_bip	1	–	–	–	–	–	20.0	–	–	1.7
Phryganea grandis L.	Phr_gra	4	2.7	–	2.4	3.1	–	10.0	–	–	3.3
N=	–	–	37	0	41	32	1	10	0	0	121
Taxa richness (S)			37	29	30	36	26	28	20	43
Number of individuals (N)			268	233	252	360	128	203	61	332
Shannon-Wiener index (H)			2.75	2.48	2.54	2.58	2.59	2.31	2.71	3.1
Buzas-Gibson’s evenness (E)			0.42	0.41	0.42	0.36	0.51	0.36	0.75	0.51
Dominance index (D)			0.11	0.12	0.11	0.12	0.12	0.16	0.08	0.06

Wze, stenotopy index; 1–7, peat pools; MZ, marginal zone of the peatbog.%N, percentage of a species in the total number of individuals.

Eudominants (>10%) are given in bold.

Cluster Analysis: Habitats, Insect Assemblages, and Indicators Within Divisions

The faunistic similarities between several peat pools and the marginal zone were relatively high (Bray–Curtis index > 0.5). The highest similarity (>60%) was between pools 4 and 6, then between pools 3 and 1, and also between pool 2 and the marginal zone (Fig. 2). The first four of these pools formed a separate cluster, connected with the densely vegetated pool 2 and the marginal zone. The very densely vegetated pool 7 and pool 5 were faunistically the most distinct. The analysis of qualitative faunistic similarities (Jaccard index) between the fauna of all the peat pools and the peatbog remnant showed that 45% of the taxa were common to the whole area, whereby this percentage consisted only of dragonflies and beetles, as caddisflies were absent from the marginal zone. The Jaccard index was higher for dragonflies (up to 62.5%) and lower for beetles (41.6%).

Fig. 2.

Dendrogram from Unweighted Pair Group Method with Arithmetic Mean cluster analysis for the study sites based on the numbers of individual insect taxa (Bray–Curtis index). MZ, marginal zone of the peatbog; 1–7, peat pools.

TWINSPAN yielded three assemblages of species (Fig. 3): assemblage A1 was dominated by beetles, A2 contained similar numbers of odonates and coleopterans, whereas assemblage B comprised all three insect orders. At higher levels (divisions), these assemblages consisted of species with both steno- and eurytopic habitat requirements or life strategies. In lower divisions, species pairs or groups with more strictly defined requirements were formed. The indicator species of the first division included members of the three orders, whereas divisions 2–6 consisted exclusively of dragonflies and beetles, the former being prevalent (Fig. 3). For example, Hygrotus decoratus defined the habitat type opposite to that of Graphoderus cinereus (division 2), whereas there were two indicator dragonflies for division 4: the tyrphophilous Leucorrhinia pectoralis versus the eurytopic Libellula quadrimaculata. In division 5, the opposing species were Leucorrhinia pectoralis and Lestes virens. The most important indicator species in the lowest divisions of the study area was Leucorrhinia pectoralis, which was replaced by Leucorrhinia rubicunda and key representatives of the other two orders in the higher divisions.

Fig. 3.

TWINSPAN cladogram showing insect assemblages (A1, A2, B) typical of acidic habitats. [C], Coleoptera; [O], Odonata; [T], Trichoptera. Indicator species (positive and negative) for divisions 1–6 are provided.

Specialization of the Insect Fauna: Naturalness Indices and Distribution of Specialists

The fauna most typical of Sphagnum peatbogs, as demonstrated by the overall value of the naturalness indices, inhabited peat pools 2 and 7 with the largest Sphagnum mats (Table 3). The deciding factors in this respect were the dragonflies in pool 2 and the beetles in pool 7; pool 5 also had a highly specific odonate fauna (Wns and Wni > 6). The least habitat-specific fauna colonised pools 1, 4, and 6. The most highly specialized caddisfly species were present in pool 6, and with regard to the species composition, in pools 3 and 4. In general, our results show that the best development conditions for each insect order prevailed in different pools. Sometimes, these conditions were at opposite extremes: for example, pool 5 was a key area for odonates but was unattractive to trichopterans, whereas pool 7 was important to beetles but unfavourable or nonoptimal for the other orders. The values of both indices for the marginal zone were only moderately high, and lower than for some of the pools, although in the case of the peatbog remnant, one could have expected higher values (Table 3). The differences between Wns and Wni at the sample level between the sites studied were statistically significant (Kruskal–Wallis test for Wns: P = 0.024, for Wni: P = 0.018). In turn, the differences between the marginal zone and all the peat pools combined were not statistically significant (Mann–Whitney test for Wns: P = 0.929, for Wni: P = 0.578).

Table 3.

Values of qualitative naturalness (Wns) and quantitative naturalness (Wni) indices for each peat pool (1–7) and the marginal zone of the peatbog (MZ)

Index	Site
	Peat pools							MZ
	1	2	3	4	5	6	7
All insect orders
Wns	4.06	5.81	5.46	4.00	4.92	4.85	5.95	4.84
Wni	3.98	5.80	4.82	4.23	5.64	3.99	6.00	4.90
Odonata
Wns	5.13	7.44	5.44	4.73	6.00	5.44	5.00	5.00
Wni	3.86	6.28	3.92	3.96	6.12	3.93	3.50	4.74
Coleoptera
Wns	3.41	5.00	5.50	3.33	4.59	4.00	6.06	4.77
Wni	3.65	4.96	4.94	4.22	5.47	3.67	6.18	5.05
Trichoptera
Wns	5.20	–	5.33	5.00	4.00	5.80	–	–
Wni	5.24	–	7.61	6.23	4.00	6.20	–	–

The highest scores are given in bold.

Comparison of the proportions of ecological groups—in terms of species richness—in individual peatbogs showed a very similar distribution pattern over the study area (Fig. 4A). The proportions of ecological groups in the three orders of insects (Fig. 4B) showed a total absence of tyrphobionts and a large percentage of tyrphophiles among trichopterans, a high proportion of tyrphophilous odonates (82% of the total dragonfly fauna) and similar proportions of tyrphophiles and nonspecialists in terms of species richness of Odonata and Coleoptera assemblages. At the sample level, the most significant differences (Kruskal–Wallis test, P ≤ 0.03) were in mean abundance (P = 0.03), species richness (P = 0.01), and proportions of ecological groups in terms of species richness or abundance (P = 0.01), applied almost exclusively to the tyrphobionts.

Fig. 4.

Proportions of ecological groups (tyrphobionts, tyrphophiles, and nonspecialists) at particular study sites (A) and within the three insect orders (B). S, species richness; I, abundance.

The Effects of Environmental Factors on the Distribution of the Entomofauna

Three environmental factors—one physical and chemical as well as two structural—were significant for each of the insect groups (Table 4). Both sets of variables were equally important for dragonflies and beetles, whereas structural factors were more important for caddisflies. The shared effects were different: extremely low for beetles, higher for odonates, and the highest for trichopterans.

Table 4.

Results of variance partitioning analysis (CCAs) for Odonata, Coleoptera, and Trichoptera

Taxon/variance		Environmental variable	%C	P	pseudo-F
Odonata
TVE (%)	66
VEFC (%)	35.3	pH	40	0.002	4.5
VEST (%)	36.3	Sphagnum cover	21.8	0.026	2.3
		Pool perimeter	20.7	0.024	2.4
SE (%)	5.6
Coleoptera
TVE (%)	52.3
VEFC (%)	25.7	Temperature	22	0.04	1.6
VEST (%)	26.7	Sphagnum cover	29.1	0.004	2.3
		Coverage of emergent vegetation	24.6	0.004	2
SE (%)	0.1
Trichoptera
TVE (%)	88.6
VEFC (%)	41.5	Dissolved oxygen	32.2	0.049	2.5
VEST (%)	54.4	Sphagnum cover	35.5	0.008	3.6
		Pool perimeter	24.3	0.046	3.1
SE (%)	10.9

TVE, total variance explained; VE_FC, variance explained by physical and chemical factors; VE_ST, variance explained by structural factors; SE, shared effect; %C, percent contribution of a particular variable; P, significance level of the effect tested by Monte Carlo permutation test; pseudo-F, value of the pseudo-F ratio statistic

For dragonflies, the key factors were pH, Sphagnum cover, and pool perimeter (Fig. 5A, Table 4). The occurrence of the three Leucorrhinia species was most closely related with pH and Sphagnum cover. They were all closely associated with intermediate pH. In addition, Leucorrhinia rubicunda was the most strongly associated with moderately overgrown peat pools. Coenagrion hastulatum inhabited the pools with the longest perimeter; and Cordulia aenea with intermediate perimeter.

Fig. 5.

Ordination CCA biplots showing the distribution of Odonata (A), Coleoptera (B), and Trichoptera (C) species in peat pools versus significant environmental factors (pH; SC, Sphagnum cover; PP, pool perimeter; helo_w, emergent vegetation cover on peat pool shores; Temp, water temperature; DO, dissolved oxygen concentration). The taxon codes are given in Table 2.

Beetles were significantly dependent on temperature, Sphagnum cover and emergent vegetation cover (Fig. 5B, Table 4). Among species strongly associated with the first factor, Laccophilus poecilus and Tanysphyrus lemnae belong to more thermophilous species in contrast to Hygrotus inaequalis and Acilius sulcatus. Anacaena lutescens, Helophorus granularis, Hydaticus seminiger, and Helochares obscurus were strictly associated with larger Sphagnum cover, whereas Graphoderus cinereus and Noterus crassicornis were related to poorly overgrown sites. Tanysphyrus lemnae, Noterus crassicornis, Hygrotus versicolor, and Acilius sulcatus were associated the most with the emergent vegetation cover: the first species with the densest cover, the second and third with moderate, and the last species with the thinnest cover.

Dissolved oxygen, Sphagnum cover, and pool perimeter were the most significant factors for caddisflies (Fig. 5C, Table 4). Since acidic habitats favor a rather species-poor caddisfly fauna, few close relationships between these variables and species were found: Holocentropus dubius was strongly related with intermediate oxygen levels and waterbodies with intermediate perimeter, whereas Agrypnia varia and Phryganea grandis were more typical of lower oxygen levels and pools with the shortest perimeter.

Occurrence Patterns of Particular Insect Species Versus Their Key Habitat Factors

Among the dragonflies, Leucorrhinia pectoralis displayed the broadest and Lestes virens the narrowest pH tolerance (Fig. 6). Leucorrhinia rubicunda and Lestes virens clearly preferred low pH, their numbers falling with decreasing acidity. The abundance of Nehalennia speciosa increased with pH (optimum pH, 5.78), whereas the abundance of Leucorrhinia pectoralis was almost the same over the observed pH range. The two Leucorrhinia species preferred pools with large Sphagnum cover: the optimum value for Leucorrhinia rubicunda was 61% (tolerance 17%) and 46% for Leucorrhinia pectoralis (tolerance 23%). Nehalennia speciosa exhibited the same optimum as Leucorrhinia pectoralis, but with a lower tolerance (14%). The two Coenagrion species associated with thinly vegetated habitats had optima of 16% (Coenagrion hastulatum) and 23% (Coenagrion puella). Apart from Leucorrhinia rubicunda and Nehalennia speciosa, the optimum peat pool perimeter for dragonflies was similar and ranged between 0.22 and 0.3 km. Tolerance ranges were also similar, the widest being for Leucorrhinia pectoralis (0.4) and the narrowest for Coenagrion hastulatum (0.09).

Fig. 6.

Response curves of the most frequent Odonata, Coleoptera, and space is lacking Trichoptera species to significant environmental variables using GLM (pH; SC, Sphagnum cover; PP, pool perimeter; helo_w, emergent vegetation cover on peat pool shores; Temp, water temperature; DO, dissolved oxygen concentration). The taxon codes are given in Table 2.

The respective thermal optima of the beetles Hygrotus inaequalis and Helochares obscurus were 13.2 and 16.2°C, both displaying similar tolerances (3.8 and 4.5°C) (Fig. 6). Laccophilus poecilus and Graphoderus cinereus were the most thermophilous species, their abundance increasing with water temperature. Noterus crassicornis had a thermal optimum of 16°C but also the broadest thermal tolerance (7.7°C). Hygrotus inaequalis and Laccophilus poecilus preferred the pools with large Sphagnum cover (61 and 49%) and exhibited similar tolerances (≈25%). In turn, Graphoderus cinereus had the narrowest tolerance and optimum at 19%. The abundance of Noterus crassicornis dropped with increasing Sphagnum cover. The reverse applied to Helochares obscurus, which was more abundant in more densely vegetated habitats. The emergent vegetation cover affected Hygrotus inaequalis and Helochares obscurus, which exhibited similar occurrence patterns with respective optima of 2.2 and 3. Laccophilus poecilus displayed a distinct preference for the most developed emergent vegetation, Noterus crassicornis was most abundant where the cover of this vegetation was moderate and Graphoderus cinereus where it was thin.

Among caddisflies, Holocentropus dubius exhibited a far greater demand for oxygen than Agrypnia varia: the optimum and tolerance of the former were 2.53 and 2.31 ppm, respectively, those of the latter were 0.85 and 1.12 ppm (Fig. 6). Holocentropus dubius distinctly preferred peat pools without Sphagnum mats; its abundance falling with increasing moss cover. Agrypnia varia clearly avoided both open pools and pools with dense plant cover, exhibiting an optimum at 18% with a tolerance of 7.4%. By colonizing pools of varying sizes, and hence with different perimeters, Holocentropus dubius achieved its optimum at a value of 0.27 km. In contrast, Agrypnia varia preferred smaller peat pools (optimum of 0.12 km).

Discussion

General Features of the Insect Fauna of Peatland Habitats

The species richness of beetles and dragonflies in the study area can be regarded as typical of peatlands (Smits et al. 2002, Drinan et al. 2013, Elo et al. 2015, Brown et al. 2016), but the caddisfly fauna was species-poorer in comparison to similar studies (Smits et al. 2002, Hannigan and Kelly-Quinn 2012), despite having the highest proportion of specialists (tyrphophiles). Dragonflies and beetles were represented by both tyrphobionts and tyrphophiles, but more than half of their species were nonspecialists. Habitat-dependent insect species are key to this habitat type, constitute also a species pool for the colonisation of new sites (van Duinen et al. 2007, Drinan et al. 2013), and are the best bioindicators of successional changes (Spitzer and Danks 2006). The presence of habitat specialists is important, but the proportion between them and eurytopic species (opportunists) is crucial. Their proportions in the peat pools studied here were evenly balanced, possibly indicating that these relatively old waterbodies have achieved a state of faunistic equilibrium. The predominance of opportunists on the Sphagnum bogs, with high richness and diversity, might even be a consequence of negative processes (pollution, deforestation of drainage basin, inflow of nutrients, increase in trophy, lowering groundwater table, and drying out), taking place in these habitats (Bernard et al. 2002). Diversity metrics alone are not a sufficient tool for reliably assessing Sphagnum bog fauna, especially in the context of peatland conservation. Such metrics must be complemented by i) a greater knowledge of the basic life history of insect assemblages as well as of rare and specialist peatbog species or (ii) the use of additional metrics (van Duinen et al. 2007, Drinan et al. 2013, Brown et al. 2016). Our study indicates that naturalness indices may be useful tools for analyzing the specialization potential of peat pools. In comparison with the values observed in other Sphagnum bogs in Poland (Czachorowski and Buczyński 1999), those obtained in this study were high for caddisflies and intermediate for dragonflies, underscoring the crucial part played by such pools in the survival of the specialist peatbog fauna.

Even among small and rather homogeneous sites, like those explored here, distinct differences in the fauna have emerged from species richness and composition, with major implications for biodiversity. Even if each kind of site and microhabitat had its own contribution to biodiversity, the peat pools combined—with 44% of exclusive species and 15 specialists—made the most valuable contribution to the biodiversity of aquatic insects on the peatbog studied. On the basis of these data, one can risk stating that if the marginal zone becomes degraded or dries out—even as a result of climate warming—the permanent pools should provide refugia for almost the entire peatbog fauna, an aspect also emphasized by Batzer and Boix (2016).

This study has demonstrated that peat pools may be as valuable as primary peatbog habitats for the occurrence of specialist insects (with regard to species composition) or even more so (naturalness indices), since particular assemblages or species are dependent on different stages of succession (Buczyński 2015). This also applies to the early successional stages: peat pools with open water and floating-leaved and/or submerged vegetation provide an opportunity for the restoration of assemblages typical of small lakes surrounded by peatbogs, with both limnophilous and tyrphophilous species. In turn, the pools with the most extensive Sphagnum cover were faunistically the most specialized (maximum values of Wns and Wni) and similar to the marginal zone (especially with regard to odonates), and hence to a natural habitat. As a result, the peat pool insect fauna to a large extent relates to the last remnants of this habitat, whereas habitat diversity promotes acidophilous and acid-tolerant species characteristic of dystrophic pools.

Environmental Factors Affecting the Distribution of Insects in Peat Pools

The Sphagnum cover of the pools was the most significant factor in this man-made habitat governing the distribution of all the studied insect groups. The same key factor was found for aquatic macroinvertebrates in Scragh bog (Hannigan and Kelly-Quinn 2012), with a rather similar percentage of explained variation in fauna (23 vs. 24.5%). In accordance with Heino (2010), moss cover increases the number of predatory species. Heino’s explanation of this phenomenon can also be extrapolated to our peat pools: the lentic macroinvertebrates of such habitats may be more dependent on the quantity of terrestrial-based rather than aquatic-based detritus. The trees on the pool shores (although significantly different at the various sites) were not crucial for insects as providers of either shade or allochthonous detritus.

Pool perimeter was a significant factor for odonates and trichopterans, although, interestingly, the actual size of a peat pool had not a significant effect for any of the three orders—this corresponds with the results of Hannigan and Kelly-Quinn (2012). The perimeter, however, may be significant for territorial insects like dragonflies (Corbet 1999): a longer perimeter will provide a potentially greater number of territories for males and will enhance their reproductive success. For caddisflies, it will provide a larger food base, with more detritus for case-bearing species and more favorable microhabitats for net attachment of predatory species. The significant association of beetles with vegetation of waterbodies, including peat pools, has frequently been evoked: for example, the herb and shrub cover as well as the number of plant species were crucial for coleopterans in a raised bog in Belarus (Sushko 2007). The present results also confirmed such relationships: beetles were dependent on the emergent vegetation cover. Many species preferred moderately or poorly developed helophytes which may be related to easier dispersion and foraging of coleopterans within this vegetation zone (Klausnitzer 1996).

The pH value was crucial for dragonflies. This parameter has very often been shown to be one of the main drivers of aquatic invertebrate distribution in peatlands (Smits et al. 2002, Więcek et al. 2013, Baars et al. 2015), including Odonata (Elo et al. 2015). Many odonate species are perfectly adapted to low pH. Leucorrhinia dubia and Leucorrhinia rubicunda are regarded as species largely restricted to acidic habitats (Rychła et al. 2011), which corresponds with our CCA results. One of the reasons why certain dragonfly species are relatively abundant at low pH is the zero or limited predation pressure from fish under these conditions (Corbet 1993). Fish were present in these peat pools, but their influence was attenuated by the presence of thick Sphagnum mats, providing to larvae effective shelter from predators (Henrikson 1993).

Temperature was a key factor of distribution for beetles, as was water dissolved oxygen content for caddisflies. This result is fairly typical, given that trichopterans are regarded as the most sensitive of the three orders analyzed here to oxygen deficits (Lock and Goethals 2012). The links between beetles and specific thermal conditions may be due to their diverse requirements during larval development (Nilsson and Svensson 1994). The most temperature-dependent species in our pools preferred low to moderate values (4.3–11°C). Natural waterbodies on Sphagnum bogs are typically subject to marked temperature fluctuations, from low to very high (Sternberg 1993). Deeper and larger peat pools may be refuges for species preferring low temperatures.

Structural factors, which are easy to control simply by managing the succession, offer interesting prospects for the active conservation of peat pools. The results of CCAs indicated that certain species preferred different degrees of vegetation cover. For instance, when creating a peat pool, it can be modified according to the needs of a particular species. Leucorrhinia rubicunda, for example, distinctly preferred the most densely vegetated peat pools, in contrast to other Leucorrhinia spp. Moreover, some physical and chemical properties of peat pools can be manipulated. For example, water temperature, crucial for beetles, depends to a large extent on shading, i.e., on the presence or absence of trees lining the pool shore. Smits et al. (2002) drew attention to a similar, practical aspect of using their results (underlining, like us, the importance of pH, vegetation structure and open water) for restoration management in Estonia, whereby their research focused on an intact raised bog system. The present findings and recommendations can be regarded as complementary to those describing systems of the following type: restored (rewetted) pools versus natural undisturbed peatland pools or artificial remnant sites (van Duinen et al. 2003, Hannigan et al. 2011, Elo et al. 2015, Brown et al. 2016). Our case extends these systems to include the remnants of a natural habitat (transition bog) versus ca. 80-yr-old peat pools, as a new type of anthropogenic lentic habitat in the study area.

The Species Level: From Habitat Preferences to Conservation Aspects

The distribution, species richness, rarity, and description of habitat preferences of particular species or assemblages may be useful in the conservation context of Sphagnum bogs. The presence of legally protected and endangered species raises the natural value of an area and facilitates formal and practical conservation actions. In our study area, we found just three of the 15 dragonfly species legally protected in Poland (Regulation of the Minister of the Environment 2016): Aeshna subarctica, Leucorrhinia pectoralis, and Nehalennia speciosa. The last-mentioned species is also on the IUCN Red List of Threatened Species (category EN; Bernard et al. 2009). Aeshna subarctica was recorded only in the peat pools, and Nehalennia speciosa and Leucorrhinia pectoralis in both the marginal zone and the peat pools.

Even though the invertebrate species of Sphagnum bogs in Europe are quite well known (Batzer and Boix 2016), the search for the best indicators of this kind of habitat continues, as the requirements and responses of particular species to specific factors remain still unknown. Promising results for Sphagnum bogs have been achieved for carabid beetles and ants as indicators for conservation planning (Brigić et al. 2017), and also spiders, whose assemblages have been found to depend on peatbog successional stages (Buchholz 2016). The first indicator level (TWINSPAN) in the present results pointed out two opposite groups: one dominated by eurytopes preferring the open water surface and the other by tyrphophiles from the well-vegetated peat pools. Thus, as in the case of the spiders, the stage of succession was a key factor to distinguish different indicators representative of all the insect orders in the peat pools. Odonates and beetles were lower division indicators, mostly replicating the tyrphophile—eurytope opposition. As in Brigić et al. (2017), the discriminant species were tyrphophiles; these authors failing to record any tyrphobiont species. In contrast, we found this element to be the rarest, but no species from this category turned out to be suitable as an indicator. Also, worth emphasizing is the fact that TWINSPAN designated species from all the orders as first-level indicators, which has confirmed that assessment of Sphagnum bogs should be based on the widest possible taxonomic range of insects. In the present case, each of the three insect orders had an important contribution. Nevertheless, dragonflies were the best indicators, as they are generally regarded as good indicators of changes in Sphagnum peatbogs and are well used in restoration monitoring studies (Elo et al. 2015). Specifically, the tyrphophilous Leucorrhinia pectoralis and Leucorrhinia rubicunda were key species at all levels. GLM analysis revealed that these two species had different pH and perimeter requirements, as well as different optima with regard to Sphagnum cover. Among the beetle species, only Hygrotus decoratus (tyrphophile) and G. cinereus (eurytope) were of any use for discriminating waterbodies, whereby the absence of endangered/protected species in this group makes them, in our opinion, less suitable indicators for the conservation of this type of habitat.

Umbrella species—another faunistic tool used in conservation practice—indicate important habitats potentially worth protecting and define the various processes/changes taking place in these areas, especially the negative ones. Our GLM results have corroborated the good umbrella species status of Leucorrhinia pectoralis for Sphagnum bogs and waterbodies bounded by Sphagnum, including anthropogenic peat pools (Bernard et al. 2002). In our study area, it was the most numerous species and exhibited the broadest tolerance among the odonates toward significant habitat factors governing its populations. Our detailed data on the habitat requirements of this species in peat pools closely resemble those given by Wildermuth (2001) regarding the optimal habitat conditions created for Leucorrhinia pectoralis in fens. They indicate that active conservation of this species on Sphagnum bogs can be just as effective. Nehalennia speciosa is one of the prime target species protected along with Leucorrhinia pectoralis, since the optima and tolerances of both species are convergent, especially with respect to Sphagnum cover. Worth underlining is the fact that Nehalennia speciosa, a rare IUCN red-listed species, was present in five of the eight peat pools (including the marginal zone), which shows that it can function well in both primary and secondary habitats (Bernard and Wildermuth 2005). The factor that might limit its occurrence in our peat pools is pH. Our results show that its response curve is quite different from that of the other odonates, its optimal pH being 5.8 (tolerance = 0.5). The optimum pH range for its Polish populations is 4.0–5.5 (Bernard and Wildermuth 2005).

No analysis of the fauna of Sphagnum bogs at the species level can ignore questions of genetic diversity and conservation. Climate change is adversely affecting the network of surface waters and waterlogged habitats in many parts of Europe (Dokulil 2013), which may increase the isolation of habitat islands like peatbogs and cause many of them to disappear. With a too small number of localities and when a certain critical isolation threshold is exceeded, metapopulations, together with genetic material (e.g., Sternberg 1995, Korkeamäki and Suhonen 2002), may be seriously altered. Detrimental demographic processes may then ensue, causing numbers to fall below the Minimum Viable Population level (Traill et al. 2007).