Chironomidae larvae in hypersaline waters of the Crimea: diversity, distribution, abundance and production

Abstract Chironomidae larvae may represent more than 70% of total Arthropoda numbers in hypersaline waters. Crimea, the largest peninsula of the Black Sea, has more than 50 hypersaline water bodies of marine and continental origin. Chironomidae larvae are common components of their ecosystems, but they still are poorly understood. This paper summarizes the results of a long-term study (2007–2016) of chironomids in Crimean hypersaline waters. More than 400 samples from 38 water bodies were used for analysis. The maximum salinity of water bodies containing Сhironomidae larvae was between 320 and 340 g/L. At first it was shown that Baeotendipes noctivagus (Kieffer, 1911) is the most halotolerant chironomid species in the world. Frequency of larvae occurrence varied and was negatively dependent on salinity. Four chironomid species were found: B. noctivagus, Cricotopus gr. cylindraceus (Kieffer, 1908), Tanytarsus gr. mendax Kieffer, 1925 and Paratanytarsus sp. Ceratopogonidae larvae were also found twice, at salinities of 150 and 270 g/L. B. noctivagus was the most common species, which occurred in 81% of samples with chironomids. Abundance of larvae fluctuated widely and reached high numbers: in plankton – to 8 thousand/m3, in floating green algae mats – up to 3 thousand/m2, and in benthos – up to 9 thousand/m2. Nonlinear dependence of chironomid abundance from salinity was observed; maximum abundance was at salinity levels of between 150 and 170 g/L. The average weight of larvae of 0.05–1.50 mm in length varied little in the samples; however, larvae of greater length had a significantly different average weight. Larvae of 8 mm in the samples had the average actual weight, which ranged from 0.750 to 2.203 mg.

As a component of diverse water ecosystems, the larvae of Chironomidae play an important functional role in the various inland waters and in the diet of water birds (Balushkina 1987;Armitage et al. 1995), can contribute to the spread of pathogenic organisms (Broza et al. 2005), and can cause allergic reactions in humans (Armitage et al. 1995). These factors indicate the importance of their comprehensive study. To date there are more than 15,000 described species of Chironomidae. Their larvae occupy a variety of aquatic habitats, including some very extreme, at latitudes ranging from 81°49ʹN to 68°00ʹS (Armitage et al. 1995). The maximum altitude at which there are representatives of chironomids is 5600 m above sea level (Himalayas), where Diamesa spp. live at subzero temperatures in water drainage channels in glaciers (Kohshima 1984;Saether & Willassen 1987). There are typical marine species, such as Clunio marinus Haliday, 1855 (Palmégn & Lindeberg 1959), or the genus of exclusively marine flightless midges Pontomyia Edwards, 1926 (Huang & Cheng 2011). Chironomidae larvae also inhabit hypersaline waters of different regions; it has been shown that larvae of chironomids can reach high abundance and dominate (in density and biomass) in water bodies at salinities of up to 70-120 g/L (Szadziewski & Hirvenoja 1981;Drake & Arias 1995;Balushkina et al. 2009;Zerguine 2014). Tanytarsus barbitarsis Freeman, 1961 dwells in Australian waters at a salinity of up to 177 g/L and can reach high abundanceup to 200,000 ind./m 2 (Kokkinn 1986).
Crimea, which is the largest peninsula of the Black Sea (27,000 km 2 ), has more than 50 hypersaline water bodies of marine and continental origin (Kurnakov et al. 1936;Shadrin 2009) (Figure 1). Studies of biota, including Chironomidae, in hypersaline Crimean lakes, were previously carried out on eight lakes (Balushkina & Petrova 1989;Ivanova et al. 1994;Balushkina et al. 2009;Shadrin et al. 2010;Litvinenko & Shlyakhov 2011;Belmonte et al. 2012). It was shown that chironomid larvae are common components of the ecosystems of seven of them. The larvae were not found in Lake Koyashskoye, which has low productivity and where salinity did not fall below 160 g/L during the study period (Balushkina et al. 2009;Belmonte et al. 2012). The total number of chironomid larvae in seven lakes was very high in some cases, reaching 15,250 ind./m 2 (Balushkina et al. 2009) and even 69,000 ind./m 2 (Litvinenko & Shlyakhov 2011). Species identification of chironomid larvae was made in only two studies, and only Baeotendipes tauricus Tshernovskijjunior synonym of Baeotendipes noctivagus (Kieffer, 1911)was found (Balushkina & Petrova 1989;Balushkina et al. 2009). There is commercial harvesting of chironomid larvae in some Crimean hypersaline water bodies to use them as food for ornamental fish. Despite the important environmental and economic role of Chironomidae they are still poorly understood in the hypersaline waters of the Crimea, and have been studied in only a few water bodies.
The authors conducted chironomid larva research in the hypersaline waters of Crimea between 2007 and 2015. The goal of this paper is to describe, analyze and discuss the results of this long-term research, comparing them with published data from other regions of the planet.

Study area
Fifty relatively large and many smaller hypersaline water bodies are located in Crimea; they include Bay Sivash (the Sea of Azov) which is the largest lagoon in Europe (around 2560 km 2 ). By origin and ionic composition, the Crimean natural water bodies are divided into marine (talassohaline) and continental (atalassohaline-sulfate). In addition to natural water bodies, on the Kerch peninsula there are artificial water bodies that are increasing in salinity and becoming hypersaline due to increasing climatic aridity ( Figure 1; Kurnakov et al. 1936;Shadrin 2009;. All Crimean lakes are shallow and polymixic, and vary in size, range of abiotic factors (salinity, temperature, pH, etc.) and biotic composition. In 2000-2015, biota and ecology of saline water bodies in Crimea were studied and results of this were partially published (Zagorodnyaya et al. 2008;Balushkina et al. 2009;Belmonte et al. 2012;Shadrin & Anufriieva 2013a;etc.).

Sampling and processing
A total of 416 samples were collected from 38 water bodies in 2007-2015; 389 samples of zooplankton, 12 benthic samples and 15 samples of floating filamentous algae mats. Of these samples, 324 were collected from hypersaline water bodies and 92 from brackish waters. It has been shown in previous studies that in hypersaline conditions (Zagorodnyaya et al. 2008;Belmonte et al. 2012), most of the benthic animals, including chironomid larvae, are present in the water column rather than in the benthos; therefore, larvae were taken primarily from plankton samples. There are two reasons why benthic animals occupy the water column: (1) increased water density under high salinity; (2) the common presence of anoxic conditions near the bottom in hypersaline waters. Quantitative samples of zooplankton were collected by filtration of at least 50-100 L of water through a plankton net with a mesh size of 110 microns. Simultaneously with biota sampling salinity, temperature and pH were measured using a refractometer (Kelilong WZ212) and a pH meter (PHH-830). Samples were fixed with 4% formalin. A smaller part of the samples was qualitative. Benthic samples were collected in areas with a depth of 0.2-0.6 m by means of a benthic tube (area 0.018 m 2 ). Samples of floating filamentous algae mats were collected from an area of 0.25 m 2 . Samples were dried to air-dry mass, which was converted into absolute dry weight (adm) using a coefficient of 0.93 (Korelyakova 1977). The mass of sampled algae mat was determined on an electronic balance. The relative number of animals in the mats was determined by dividing the number of counted individuals by the weight of the mat piece. The samples were processed using an Olympus SZ-ST and LOMO MBS-9 stereo microscope. In 27 samples from hypersaline water bodies (Table I), larva species were identified using articles and keys for identification (Pankratova 1970(Pankratova , 1983Hirvenoja 1973;Wiederholm 1983;Makarchenko & Makarchenko 1999). Developmental stage was evaluated and size was measured. The length of the chironomid larvae was measured under a STEMI DV4 (Zeiss) stereo microscope with an ocular micrometer. The larval mass was determined by the weighing of larvae (predried on the filter paper) on a torsion balance WT-250. Large individuals were weighed individually; small individuals of similar size were weighed together and average mass was calculated.

Data processing
To assess the frequency of occurrence of chironomid larvae in different ranges of salinity, all 416 samples from brackish and hypersaline water bodies were used: where Kc is the number of samples in a certain range of salinity, which contained chironomids; and K is the total number of samples in the interval. To analyze the relationship between chironomid abundance and environmental factors, only quantitative samples (Table I) were used. Taking into account the temperature correction coefficient Q10 = 2.25, production was calculated using the formula for growth of chironomid larvae (Balushkina 1987): where Pproduction, J/m 2 ·day or J/m 3 ·day, Nabundance, ind./m 2 or ind./m 3 , Windividual mass, mg. Data were subjected to standard statistical processing. The regression equations between chironomid abundance and environmental factors were calculated by the least squares method in the standard program MS Excel 2007. The significance of differences in mean values was evaluated by Student's t-test, and the confidence level of the correlation coefficients was determined (Müller et al. 1979).

Results
Chironomids are common inhabitants of the Crimean hypersaline lakes, among which there are water bodies both maritime and continental in origin. Chironomids are also abundant in Bay Sivash and in saline/hypersaline artificial ponds. The maximum salinity at which Chironomidae larvae were found was between 320 and 340 g/L. Frequency of larva occurrence at different salinities varied and was negatively dependent on salinity if higher than 30-50 g/L ( Figure 2). Dependency may be reliably approximated by a linear equation (R = 0.935; p = 0.0005): where Y is the frequency of occurrence (%); and S is the salinity (g/L; average of interval). Thus, 87% of probability to find chironomid larvae in plankton at a salinity of above 50 g/L was determined by salinity. It should be noted that if salinity was below 50 g/L, Chironomidae larvae were usually 5-20 times less abundant in plankton samples than in the benthos or floating mats. There were also benthic or mat samples without chironomid larvae, despite their presence in the plankton. With a rise of salinity above 100-110 g/L, the portion of cases in which the most part of chironomid larvae lived in the plankton increased. The largest number of samples (127) was collected in Lake Chersonessus (near Sevastopol) from 2007 to 2015. The seasonal occurrence of larvae in plankton was studied; chironomid larvae were absent in samples collected from November to March. In April they were present, but not every year. Chironomid larvae were not observed during these months in other lakes.
In total, four chironomid species were found in 27 samples from hypersaline waters: B. noctivagus, Cricotopus gr. cylindraceus (Kieffer, 1908), Tanytarsus gr. mendax Kieffer, 1925 and Paratanytarsus sp. Ceratopogonidae larvae were found twice, at salinities of 150 and 270 g/L, but were not identified to the species group level. B. noctivagus was the most common chironomid species, which was represented in 81% of chironomid samples at salinities between 25 and 280 g/L. Cricotopus was represented in 22% of samples at salinities between 30 and 65 g/L. Tanytarsus was in 8% of chironomid samples at salinities between 30 and 180 g/L, and Paratanytarsus occurred once at a salinity of 58 g/L ( Figure 3). As a rule, only one chironomid species was represented in a sample, and only in two samples was B. noctivagus found together with other species: once with Cricotopus at a salinity of 65 g/L, when Cricotopus contributed 66% of the total number of chironomids; the other with Tanytarsus at a salinity   of 180 g/L, when Tanytarsus represented about 1% of the total abundance. Cricotopus and Tanytarsus were found together in one sample at a salinity of 30 g/L. Abundance of Chironomidae larvae fluctuated widely (Figure 4(a, b; Table II) and reached high values: in plankton, up to 8 thousand/m 3 or 3 thousand/m 2 (calculated taking into account depth), in floating mats, up to 3 thousand/m 2 , and on the benthos, up to 9 thousand/m 2 . As an example, in August 2015 in Bay Sivash at salinity between 61 and 65 g/L and temperature of 30°C, the abundance of Chironomidae larvae in plankton was 14 ind./m 3 (B. noctivagus), in floating mats it was 2820 ind./m 2 (Cricotopus) and in the benthos it was 1667 ind./m 2 (B. noctivagus). In this case, low abundance of Chironomidae larvae in plankton was probably due to a high concentration of juvenile fish -Knipowitschia caucasica (Berg, 1916) and Atherina No.sample number from Table I; *abundance, ind./m 2 ; **biomass, mg/m 2 ; ***production, J/m 2 •day; ****qualitative sample. boyeri Risso, 1810 (Shadrin et al. 2016). It should be noted that the average weight of animals in biotopes at this time also varied: 0.40 mg in plankton, 0.25 mg in mats and 0.75 mg on the bottom. Typically, the abundance of chironomid larvae was higher in the floating mats and in the benthos than in the plankton, though often there were some cases when chironomids massively presented in the plankton but had very low numbers in the mats or on the bottom.
Due to the small number of floating mat and sediment samples, subsequent analysis of abundance dependence on salinity was made only for plankton samples (Figure 4a). In general, for all lakes nonlinear dependence of chironomid larvae abundance from salinity was observed; maximum abundance was in the salinity range of 150-170 g/ L. In the samples taken only from Lake Chersonessus, a significant trend of an increase of larva abundance in plankton up to 120 g/L was noted. Dependence is close to linear (R = 0.689; p = 0.005), and up to 48% of the total variability in the chironomid larva numbers in the lake plankton can be explained by the variability of salinity. The dependence of the population density on temperature was also nonlinear (Figure 4b), and the largest number of individuals was observed in the range of 16-29°C. Abundance did not correlate with pH. In the period when the larvae were represented in plankton there was a general trend of increase in the average abundance for all lakes and every year from April to July, with a subsequent decrease in October ( Figure 5). In Lake Chersonessus the peak of the chironomid larva abundance occurred also in July.
B. noctivagus larva length varied among the samples from 1.5 to 12.0 mm, and the body weight from 0.036 to 4.880 mg (Table III). Average weight of larvae of 0.05 to 1.50 mm varied little among the samples; larvae of greater length had significantly (p = 0.001) different average weight with pairwise comparison of samples. The analysis of the entire set of samples showed that there is a non-significant trend of a weight decrease of same-length larvae with a salinity increase. As an example, in Lake Kiyatskoye in August 2014 and 2015, salinity was similar at 180 g/L and 185 g/L, respectively. However, individual average mass of same-length larvae was significantly different (p = 0.001) in 2014 and 2015. In 2014, larvae of 4 mm had a mass of 0.445 mg, 6 mm -1.044 mg and 8 mm -2.333 mg; and in 2015, length of 4 mm -0.244 mg, 6 mm -0.567 mm and 8 mg -0.750 mg. Dependence of weight on length is described by equations such as:  where Wbody weight (mg); Llength (mm); and a and b are coefficients. Paired comparison showed that equation coefficients "b" in most cases were significantly (p = 0.001-0.0001) different and the coefficients "a" were not significantly different; the average value of the coefficient "a" was calculated. The "b" coefficients were recalculated for all of the equations using this average value of "a" (Table III). The biomass and production of chironomid larvae was calculated taking into account all data (Table II). The maximum biomass of larvae in the plankton reached 2560 mg/m 3 , and dependence on salinity had a dome-shaped form (Figure 6a). Maximum production reached 37.5 J/m 3 day, and also had a dome-shaped form of dependence on salinity (Figure 6b).

Discussion
As seen from new and previously published (Ivanova et al. 1994;Zagorodnyaya et al. 2008;Balushkina et al. 2009;Belmonte et al. 2012) data, Chironomidae larvae are a common component of ecosystems in Crimean hypersaline waters and play an important role in production processes. Larvae with a wide range of sizes from 2.5 to 9.5 mm were abundantly represented in water bodies of salinity up to 280 g/L, indicating the active state of larval chironomid populations. Even though larvae have been found at salinities between 320 and 340 g/L, this does not mean that they can function properly in such salinity. It has been shown previously that larvae of some Chironomidae species may spend a period of time in a dormant state, anhydrobiosis (Suemoto et al. 2004;Cornette & Kikawada 2011). It can be assumed that some B. noctivagus stages can move into anhydrobiosis due to drying or very high salinity.
The massive presence of Chironomidae larvae in the hypersaline waters of Crimea is not an exception; this phenomenon has been observed in different regions (see Table IV, where a list is given of Chironomidae species the larvae of which are able to live in hypersaline waters). Approximately 38 species belonging to different subfamilies may exist at a salinity of more than 35 g/L, and 16 species among them occur at a salinity of more than 100 g/L. Only three species were found at a salinity of 150 g/L and higher. E. K. Suworow (1908) found active stages of chironomid larvae (Chironomus sp.?) in Lake Bulak (near Caspian Sea) at a salinity of 285 g/L. Now B. noctivagus can be considered the most halotolerant Chironomidae species in the world, inhabiting in an active state hypersaline waters with a salinity of up to 280 g/L, and possibly up to 340 g/L. It is likely that Suworow found B. noctivagus larvae in Lake Bulak, which is within the area of distribution of this species.
A question arises: what adaptations allow some chironomid species to exist at high salinity? It is known that arthropods can use two strategies of osmo-adaptation to exist in hypersaline habitats (Khlebovich & Aladin 2010). Osmo-regulating animals use mechanisms of active hypo-osmotic salt regulation in the body fluids, keeping a lower concentration of salts in these fluids than in the environment. Osmo-conforming animals do not have mechanisms for salt regulation in body fluids; osmo-adaptation is carried out at the cellular level by the synthesis of compatible osmolytes and/or by obtaining them from the outside with accumulation in cells. Compatible osmolytes are low-molecularweight organic compounds (polyols, some amino acids and methylamines, etc.); they protect proteins under conditions of osmotic stress and do not impact on the normal course of metabolic processes (Yancey 2001). Among the Diptera, including Chironomidae, there are species that use one or the other of these strategies (Sutcliffe 1960;Neumann 1961;Bradley 1987;Herbst & Bradley 1988;Patrick Figure 6. Dependence of (a) biomass and (b) production of Chironomidae larvae on salinity in plankton of Crimean saline waters.  & Bradley 2000;Renault et al. 2016). Some species can use both osmo-adaptation mechanisms together; Culex tarsalis Coquillett, 1896 can accumulate various compatible osmolytes not only in the cells but also in the body fluids (Patrick & Bradley 2000). Osmo-adaptations of the Chironomidae species found during this study have not been studied yet; however, some assumptions can be made. Using data available in literature, primarily Kokkinn (1986), it can be assumed that the species found during the studies, especially B. noctivagus, have quite effective osmo-regulation mechanisms that can ensure the existence of larvae at salinities up to 90-120 g/L. At higher salinities these mechanisms become insufficient; the animals also begin to accumulate osmolytesosmo-protectants in cells and body fluids. Osmolytes can be accumulated in the body due to the breakdown of proteins (amino acids-alanine, proline, etc.), or come from food (glycerol, betaine, etc.); this was indicated for several Diptera species (Patrick & Bradley 2000;Yoder et al. 2006;Renault et al. 2016). Mechanisms of osmo-regulation, as well as the synthesis of osmolytes, require a lot of energy, so obtaining osmolytes from the external environment (Yoder et al. 2006) can significantly reduce the energy costs for osmoadaptation processes. It is likely that the smaller average body size and mass of individuals in the water bodies with higher salinity (Table III) are due to higher energy costs for osmo-adaptation processes which reduce the efficiency of use of assimilated energy for the growth of larvae. It was noted previously (Kokkinn 1986) that T. barbitarsis in Australia, at salinity between 100 and 177 g/L, was found only in highly productive waters where blooms of algae were observed. In Crimea, the authors found chironomids at the highest salinity also only in waters where microalgal blooming was observed. Previously, this phenomenon was noted for the copepods in the hypersaline lakes of Crimea (Shadrin & Anufriieva 2013b;Anufriieva 2015). The presence of microalgal blooms may ensure the presence of chironomids not only by providing necessary energy, but also by providing osmolytes. In Crimea, blooming of the green unicellular alga Dunaliella salina (Dunal) Teodoresco, 1905 in water bodies with salinity of more than 200 g/L is a common phenomenon; the concentration of glycerol (osmolyte) in their biomass can reach 80% of organic matter at that salinity (Shadrin & Anufriieva 2013b). All this leads to the conclusion that the physiological capabilities do not define the upper level of salinity at which Chironomidae naturally inhabit water bodies, and the biotic environment (algal concentration and composition) is just as important. Both elements -salinity and microalgae developmentare changing in the water bodies and determine the site-to-site and temporal variability of chironomid larva composition and density in hypersaline waters, as was shown for brackish waters (Cañedo-Argüelles & Rieradevall 2009). Hypersaline water bodies in Crimea are characterized by high variability; many of them are partially temporal or dry completely. The successful existence of animal species in these water bodies is ensured by the ability to endure, in a resting state, conditions which are incompatible with active life; the existence of dormant stages has been demonstrated for almost all crustaceans living in hypersaline lakes in Crimea (Moscatello & Belmonte 2009;Shadrin et al. 2015). Chironomid larvae Polypedilum vanderplanki Hinton, 1951 are able to remain in anhydrobiosis for up to 17 years, and to restore normal activity when released into the water (Cornette & Kikawada 2011). The ability of the larvae to be in an inactive state (anhydrobiosis) for quite a long time has been revealed in some other сhironomid species (Suemoto et al. 2004;Jones 2009). With respect to B. noctivagus this issue has not been studied, but it can be assumed that the larvae of this species are also capable of long-term anhydrobiosis because they are present in significant numbers in ephemeral ponds also.
The study results clearly show that dependence of the length of the body on body weight varies depending on environmental factors, so averaging may lead to significant distortion of the actual situation. The following generalized equation was calculated for all of our samples: W ¼ 0:0116 Á L 2:328 (5) Using Equation (5), the larva mass was calculated for individuals of 6 mm in length -0.752 mgbut the average weight of larvae of such length ranged from 0.567 mg to 1.044 mg in the samples. When calculating (Equation 5) the mass of 6-mm larvae, it was overestimated or underestimated by nearly 30% compared with the real values. In the case of 8-mm larvae, the calculation gave 1.468 mg, but the average actual weight of this size larva ranged from 0.750 to 2.203 mg in the samples. Overestimation or underestimation of the average actual weight can exceed 40%. The issue requires more detailed study and discussion in future. In this study, biomass and production were calculated using actual average individual mass. According to the data above, biomass and production are dependent on salinity but other factors such as food are likely to play an equally important role. Salinity determines the maximum possible performance. The question of interaction of factors in determining the structural and functional characteristics of larvae is interesting, but there is not enough data on the state of the environment and Chironomidae biology for an in-depth discussion. However, some things may be assumed.
Calculations of chironomid production are most likely overestimating the actual values because a generalized Equation (2) of chironomid growth was used. It was shown for some Chironomidae species that the growth rate decreases and development duration increases with a salinity increase (Kokkinn 1990;Cartier et al. 2011). On this basis, it can be assumed that calculated production values are being used which may be 30-60% higher than the real values in the hypersaline lakes of Crimea. Further research is needed in this direction.

Conclusions
Chironomid larvae are a common and abundant component of the Crimean hypersaline waters, playing an important role in the functioning of the ecosystems. B. noctivagus is the most common and abundant species among them. It is likely that it is the most halotolerant Chironomidae species in the world. Despite this, the adaptation mechanisms providing for its existence in the harsh conditions of hypersaline waters are not known. Further physiological, biochemical, genetic and ecological research is needed in this direction.