Benthic communities in inland salinized waters with different salinities and nutrient concentrations and the ecology of Chironomus aprilinus ( Diptera : Chironomidae ) in the Czech Republic

The macrozoobenthos in saline pools at dumps in a former coal mining area was studied over a period of two years. Due to specifi c environmental conditions these pools are unique in the Czech Republic. Extremely high values of salinity (up to 11‰) along with a low concentration of dissolved phosphorus (0.01–0.1 mg.l–1) are typical of some of the water in this area. The pools were grouped into three categories based on their conductivity values and treated using cow dung, municipal wastewater treatment sludge and inorganic NPK (nitrogen-phosphorus-potassium) fertilizer at doses recommended for carp ponds. The application of fertilizer had a positive effect on the density and biomass of all the groups in the macrozoobenthos. The highest and the lowest increases in macrozoobenthos biomass were recorded after the application of NPK and cow dung, respectively. However, the application of fertilizer had no effect on the diversity of macrozoobenthos. Chironomus aprilinus, recorded in the Czech Republic for the fi rst time, inhabited all pools with conductivity ranges of between 5,000–16,000 μS.cm–1. The density of C. aprilinus larvae increased with increasing salinity reaching a maximum of about 17,083 ind.m–2 (biomass – 82 g.m–2). Analysis of C. aprilinus phenology revealed a bivoltine pattern with the summer generation of larvae reaching a maximum in June–July and the overwintering generation in October to November.


INTRODUCTION
The majority of aquatic diptera are restricted to freshwater and only a few species tolerate high salinities.Inland saline waters are unique habitats with typical communities of organisms tolerant of adverse and fl uctuating environmental conditions.These systems provide a unique opportunity to study the composition of their biological communities and how they respond to changes in their physico-chemical characteristics.The species diversity in the benthic communities in these waters is very low and often dominated by a few tolerant species, which can however occur in high numbers.
The pools were grouped into three categories according to their locality, conductivity and presence of littoral and submerged vegetation (Table 1).The conductivity recorded in the pools at the different localities did not overlap.The pools had a surface area of 15-45 m 2 at locality 1; 20 m 2 at locality 2 and 250-1000 m 2 at locality 3. The depth was about 30 cm at localities 1 and 2 and 50 cm at locality 3.
Application of fertilizers was used to determine the possibility of accelerating the biotic recovery of the pools studied.Cow dung, municipal wastewater treatment sludge or inorganic NPK fertilizer (commercial CERERIT, N -15%, P 2 O 5 -15%, K 2 O -15%) was added to particular pools in early August 2001.NPK fertilizer was also added to pools at locality 2 in early September as shown in Table 1.The dose of fertilizers was chosen so that the fi nal concentration of TP was 0.8 mg.l -1 in the pools to which fertilizers were added.The wastewater treatment sludge and cow dung were applied at the maximum dose of 5 t.ha -1 , which is the dose applied to semi-intensively managed carp ponds.

Environmental and biological sampling
Water samples, 100 ml each, were collected monthly from each pool.The pH, conductivity and oxygen concentration were measured in situ using a Multiline P4 device (WTE, Germany).Salinity (mg.l -1 ) was calculated from the conductivity measurements using conversion factor 8 as per Pitter (2009).In the laboratory, NH 4 + , NO 3 -, NO 2 -, PO 4 3-, SO 4 2-, Cl -, total N and total P concentrations were measured after fi ltration (GF/C fi lters), using a fl ow injection analyses method and a Tecator FIA 5042 Star instrument ties and the use of fertilizer for mitigating and accelerating the revitalization of abandoned mining areas.

Study area
This study was carried out in the Sokolov basin, a brown-coal mining region in the west of the Czech Republic.Macrozoobenthos communities were sampled in aquatic habitats created for restoring habitats in the open-cast mine or in areas outside the mining area where overburden from the open-cast mine was dumped (Pecharová et al., 2001).
The Podkrušnohorská dump covers an area of 2,000 ha and is located 3 km north-west of the city of Sokolov (50°12´-50°14´N, 12°38´-12°42´E) (Fig. 1).It is the largest dump in the Czech Republic and consists of overburden from the Jiří, Družba and Lomnice brown coal mines.The overburden consists mostly of montmorillonite kaolinite-illite clays and claystones (Kříbek et al., 1998).Many pools developed spontaneously at this dump or are newly established as part of the hydric restoration of this former coal-mining region.
Waters in mining areas are often highly acidic due to the leaching by drainage water of compounds present in mine tailings.Thanks to the high neutralization ability of clay, acidifi cation by acid mine drainage (hereafter AMD) water does not play as important a role in the area studied as it does in other coal mining areas throughout the world (Nixdorf at al., 1998;Geller et al., 2000).The pH values of the water bodies we studied were close to neutral (Table 1).High concentrations of dissolved solids (conductivity up to 16,000 μS.cm -1 ), high alkalinity (8.5-15.5 mmol.l - ) and low concentrations of dissolved phosphorus (0.06-0.7 mg.l -1 ) are typical of water at the Podkrušnohorská dump.Salinity of 3-11‰ corresponds to slightly brackish (mesohaline) water (Healy, 1997;Pitter, 2009).Unlike in seawater, the pre-  (Parsons et al., 1984).The concentrations of Ca, Mg, K, Na and Fe cations were determined using the AAS method and a Varian SpectrAA-640 instrument (Varian Inc., http://www.varianinc.com).Alkalinity was measured by potentiometric titration with 0.1 M HCl.
Samples of macrozoobenthos were collected monthly from July until November 2001 (except August) and from February till October 2002 (except August).Each sample consisted of 10 subsamples pooled together.Individual quantitative subsamples were collected using a plastic cylindrical corer (working area 96 cm 2 ).Samples were washed through a sieve (1.0 mm mesh size) and fi xed using 6% formaldehyde.The remaining benthic invertebrates on the sieve were sorted by hand in the laboratory.Biomass was assessed as formalin wet weight.In each sample, the biomass and species composition of the macrozoobenthos were determined.The larvae of C. aprilinus were identifi ed morphologically (Webb & Scholl, 1985) and cytotaxonomically using the polythene chromosomes in the salivary glands (Keyl & Keyl, 1959;Keyl, 1962) of a sample of unpreserved material.Size data for all the C. aprilinus larvae were assessed in terms of body length (measured from the anterior margin of the labrum to the end of the abdomen) and larval instar based on head width (at the widest point of head capsula) (Saether, 1980).

Statistical analyses
Direct gradient redundancy analysis (RDA) in Canoco for Windows 4.5 software (Ter Braak & Šmilauer, 1998) was used to evaluate the association of all the environmental variables recorded (vegetation cover, alkalinity, type of fertilization, chlorophyll a, conductivity, Cl -, NH 4 + , PO 4 3-and seasonality as individual month of observations) with the macrozoobenthos community [total abundance of zoobenthos, abundance and biomass of C. aprilinus, the percentage of C. aprilinus, Shannon-Wiener index H′ = -Σi pi lnpi (H -calculated using natural logarithms of relative densities of particular taxa)] and equitability index E = H′/lnS, where S is the number of taxa in a sample.Statistical signifi cance of each analysis was tested using the Monte Carlo permutation test with 499 permutations as a time series and pools set as covariates.The analysis was visualized using CanoDraw software (Ter Braak & Šmilauer, 2002).Level of signifi cance was set to α ≤ 0.05 in all statistical comparisons.
For analysis of differences in species abundance/numbers between localities two-way ANOVA and for analyses of statistical comparison of average C. aprilinus abundance in different treatments one-way ANOVA in STATISTICA software version 12.0 (Statsoft Inc., 2008) was used.
The association between presence/absence of C. aprilinus and environmental factors (vegetation cover, age of reservoirs, water electric conductivity) and their interactions was tested using a generalized linear model with a binomial error distribution and a logit link.
The abundance of particular groups of macrozoobenthos (Chironomidae, Ceratopogonidae and others) differed signifi cantly between localities (Two-way ANOVA, F (4, 849) = 4.824, p = 0.00075) .The highest species richness was recorded at locality 1 where the other categories (Ephemeroptera, Odonata, Coleoptera, Diptera: Stratiomyidae) made up about 10% of total numbers.At the other two localities they made up less than 1% of the total numbers (Fig. 2).Similarly the number of chironomid taxa differed signifi cantly between localities (Two-way ANOVA, F (8,1415) = 18.890, p < 0.00001).At localities 2 and 3 C. aprilinus made up about 70% of all chironomids and at locality 1 only 5%.On the other hand, larvae of Procladius made up 50% of all chironomids at locality 1 and only 23% at localities 2 and 3.
The highest species diversity was recorded in pools with the lowest conductivity at locality 1 (4,000-6,400 μS.cm -1 ).Larvae of Procladius, Tanypus and Tanytarsus made up more than 90% whereas Chironomus larvae made up only 4% of all chironomid larvae and were found in only 5 of the 12 pools studied (Fig. 3).An exceptionally high density of 10,417 ind.m -2 (biomass 21 g.m -2 ) of C. aprilinus was recorded in June 2002 at one pool with a conductivity of 5,700 μS.cm -1 .The numbers were lower by an order of magnitude in the other pools at locality 1.The diversity index H´ was signifi cantly positively correlated with the presence of aquatic vegetation (Fig. 6).

Factors affecting abundance and biomass
The application of fertilizers had a positive effect on zoobenthos at all three localities.The highest average abundance of C. aprilinus was recorded in pools fertilized with NPK at all three localities in 2002 (Fig. 4).The difference between the densities of C. aprilinus recorded in pools fertilized with NPK and those treated with sludge and the control pools that were not treated with fertilizers was signifi cant (One-way ANOVA, F (3, 272) = 11.22;p < 0.001) (Fig. 5).
Factors affecting the abundance and biomass of C. aprilinus are presented in Fig. 6.The closer the explanatory arrows and explained variables are the closer they are positively correlated.Opposite direction of arrows indicates a negative correlation.The variability in the abundance and biomass of C. aprilinus and total diversity (H´) is best explained by the time of year -June (4% of total variance explained by all factors, F = 11.99,p = 0.002).After including this variable in the model, most of the residual variance  is explained by the presence of vegetation (3%, F = 9.53, p = 0.002), November (2%, F = 7.21, p = 0.002), chlorophyll a (2%, F = 8.88, p = 0.008), application of NPK (2%, F = 7.02, p = 0.002) and one percent of the variability can be explained by the following factors: July, alkalinity, February, March, control.Further variables proved to be insignifi cant.
The results of the GLM (Table 2) revealed a negative correlation between the occurrence of C. aprilinus and the presence of vegetation in pools but, this relationship is probably indirect as the presence of vegetation is primarily correlated with the salinity of habitats.

Seasonal dynamics
Average densities and biomass of C. aprilinus larvae changed seasonally.Maxima were recorded during autumn (September to November) and summer (June to July) (Figs 4 and 6).The summer maximum biomass was higher than that recorded in autumn.Length analysis (Fig. 7) and changes in abundance and biomass (Fig. 4) indicated a bivoltine cycle in the population.Larvae of the 3 rd instar (length 3-9 mm) dominated in September to October.They grew during the winter and the highest occurrence of 4 th instar larvae was recorded in February and March.The emer- gence of the overwintering generation started in March/ April at locality 1, depending on weather.Mass emergence took place in May at localities 2 and 3.After that low numbers of C. aprilinus larvae were recorded at all localities in May (Fig. 4).Larvae of the 1 st and 2 nd instars were regularly found in zooplankton samples taken at all three localities in March to July.The emergence of the summer generation was more protracted and occurred during August and Sepember.In September the larvae of the 4th instar disappear from the benthos and the population was dominated by 3 rd instar larvae of the overwintering generation (Fig. 7).The two generations were well separated at localities 1 and 2, whereas they overlapped at locality 3.

DISCUSSION
Saline pools at the dumps in a former coal mining area are unique habitats with a fauna found nowhere else in the Czech Republic.The inland saline pools differed from coastal waters in their water chemistry.The salinity was due mainly to sulphates and corresponds to mesohaline conditions.Chironomus aprilinus was recorded at these localities in the Czech Republic for the fi rst time.Besides C. aprilinus, other halophilous organisms were also recorded, e.g., the rotifer Hexarthra fennica (Levander) and larvae of Ephydra cf.riparia (Diptera: Ephydridae).
The macrozoobenthos was dominated by chironomids in both recent and old pools, and the occurrence of C. aprilinus was positively correlated with salinity.Strenzke (1960) reports females only depositing eggs in brackish water with a chloride content of 4000-13700 mg l -1 under experimental conditions.The same author found this species in natu-   ral conditions in Germany at salinities of 1,200-16,100 mg l -1 .Tourenq (1975) reports incidental occurrence in fresh water but gave 1,500-6,000 mg l -1 as the normal range.He found, however, larvae of this species in places where the chloride content was as high as 15,000-22,000 mg l -1 .Steenbergen (1993) cites a mean of 5,800 mg l -1 and rarely recorded larvae below 1,000 mg l -1 .Vallenduuk in Eggermont et al. (2008) reports conductivity limits of 3,000-5,000 S.cm -2 for the occurrence of C. aprilinus.
These results are in good agreement with the salinity range recorded at the pools studied (3,200-13,040 mg l -1 ).However, at the locality with the lowest salinity (3,200-6,400 mg l -1 ) not all pools were inhabited by C. aprilinus.
The biomass of benthic and planktonic communities depends on the primary production of autotrophic organisms, which is limited mainly by the availability of phosphorus and nitrogen in the water column.The availability of nutrients, along with high conductivity, is the limiting factor of productivity in water biotopes.As a result, poor development of phytoplankton and littoral macrophytes was recorded at the different localities studied.The application of fertilizers is the fastest method, but not the most longlasting, to support primary producers in these ecosystems and subsequent development of the macrozoobenthos.The genus Chironomus is tolerant of eutrophic conditions (Marquez et al., 1999).This was confi rmed in our study.Larvae of the sub-family Chironominae showed a significant increase in abundance in response to organic enrichment.The larvae of Chironomus dominated in fertlized pools at all salinities.
Chironomid larvae play a pivotal role as potential drivers of nutrient dynamics.The role of chironomids is not only as an important source of food for benthic fi sh, but also as a major contributor to the phosphorus fl ux from sediments in ponds with a residual organic load (Biswas et al., 2009), especially in pools without fi sh.
The analysis of the phenology of C. aprilinus revealed a bivoltine cycle with possibly a partial third generation.These fi ndings are in agreement with the results of other studies.Palmén & Aho (1966) report two generations of C. aprilinus in southern Finland, but only one in cold environments and possibly a partial third generation in warm years.In the delta region of the Netherlands, Krebs (1982) recorded two generations a year (sometimes three), emerging from the end of March until early November.Larvae overwinter in the third or fourth instar.Spring and autumn emergence of Chironomidae is typical in temperate wetlands (Armitage at al., 1995;Moeller-Pillot, 2009).In warmer climatic conditions there are likely to be more generations per year and the emergence of the overwintering generation can take place earlier.Fuentes et al. (2005) found high densities of 3 rd and 4 th instar larvae in May, which may correspond to the fi rst summer generation.Similarly, Rajabipour et al. (2011) show that May, June and July are the best months for collecting egg clutches of C. aprilinus.At our localities we recorded larval minima in May.Due to the sampling method used, the small larvae of the fi rst two instars passed through the sieve.Moreover, it is diffi cult to sample the fi rst two instars as they only take a few days to complete their development (Matěna, 1989).
The total absence of another halophilous to halobiont species C. salinarius Kieffer is interesting.This species is regularly reported as syntopic with C. aprilinus along seashores worldwide (Parma & Krebs, 1977;Drake & Arias, 1995;Kondo, 1998).According to Cartier et al. (2006) C. salinarius can tolerate variation in salinity with very few individuals surviving in environments with a salinity level over 35 g l −1 , high survival at low salinities (0 and 5 g l −1 ) and moderately high salinities (20, 25 and 30 g l −1 ) and poor survival at intermediate levels of salinity (10 and 15 g l −1 ).So the salinities of the pools investigated were within the range tolerated by this species.
We can suggest two reasons for the absence of C. salinarius in the pools examined in this study.The species is either bound to habitats close to the seashore where the salinity is determined by a high content of chlorides and is unable to cope with high concentrations of sulphates or only C. aprilinus is capable of dispersing from coastal areas because it can tolerate freshwater habitats and so use such places as stepping stones to fi nd more suitable habitats.These hypotheses require further study.

Fig. 1 .
Fig. 1.Map of the Velká Podkrušnohorská (VP) dump with an inset showing its location within the Czech Republic.Numbers 1 to 3 indicate the position of localities studied.

Fig. 3 .
Fig. 3. Percentage composition of chironomid larvae at localities 1-3.Mean percentage abundance for all pools at each locality is presented.

Fig. 7 .
Fig. 7. Length frequency distributions of C. aprilinus larvae over time.Upper panel -locality 1, middle -locality 2, bottom panel -locality 3. Bold horizontal bars indicate the average length for a given month, narrow bars indicate numbers of larvae in a given length class.Each bean plot shows the length frequency recorded in each month and is smoothed using a Gaussian kernel.

Fig. 5 .
Fig. 5. Statistical comparison of average abundance of C. aprilinus in the different treatments.One way ANOVA, different letters above boxes indicate signifi cant differences (p < 0.001).

Table 1 .
Characteristics and chemical parameters recorded at the localities studied and carp ponds in the Czech Republic (highest values are in bold).
Fig. 2. Percentage composition of macrozoobenthos at localities 1-3.Mean percentage abundance for all the pools at each locality is presented.