Composition and diversity of the Chironomidae in subtropical streams : effects of environmental predictors and temporal analysis

Aims: The aims of this study were (i) to evaluate the spatial and temporal distribution of chironomids in subtropical streams and (ii) to investigate the importance of environmental predictors in the distribution of this group. Methods: Samples were collected in four annual seasons between 2010 and 2011, in 10 streams located in an agricultural matrix in southern Brazil, using a Surber sampler. Organisms were identified to the genus level. Some environmental variables were analysed in the sampling sites of the Chironomidae. The variations in abundance and diversity were compared by Analysis of Variance, whereas the community composition was analysed using Multivariate Analysis of Variance. A Redundance Analysis was used to evaluate the effect of environmental variables on the chironomid community. Results: 7,349 individuals were identified, belonging to three subfamilies (Chironominae, Tanypodinae and Orthocladiinae) and 57 genera. The abundance and richness of the Chironomidae were similar between sites and seasons. However, the composition varied spatially and temporally (p < 0.001). Water temperature, total organic carbon, total nitrogen and carbon:nitrogen ratio were variables that influenced the chironomid community. The genera Thienemanniela, Pentaneura, Paratanytarsus, Parapentaneura, Parametriocnemus, Hudsonimyia, Labrundinea and Larsia were present in summer, whereas Parakiefferiella, Paramerina, Metriocnemus were indicators for winter. Conclusions: The spatial distribution of chironomids was directly related to the environmental conditions of the sampling sites. The temporal variation in the community followed a pattern in relation to water temperature. The variables physical and chemical operate as environmental filters and thereby alter the chironomid community.


Introduction
Biological communities are structured based on interactions between species, with abiotic and biotic factors at different spatial and temporal scales (Peeters et al., 2004;Anjos et al., 2011).Variations in the composition of aquatic communities at small scales are usually explained by physicochemical variables of water quality, which combined with spatial data, contribute to explain the biological diversity and relationships of aquatic organisms with the environment (Hepp and Melo, 2013).Factors such as morphological characteristics of the environment (Silver et al., 2004;Chaib et al., 2013), substrate type (Hepp et al., 2012;König and Santos, 2013), water quality (Calle-Martínez and Casas, 2006;Tejerina and Malizia, 2012;Chaib et al., 2013), and availability of food resources (Peeters et al., 2004) are common variables that can affect the richness and diversity of aquatic invertebrates (Fesl, 2002).Moreover, anthropogenic disturbances in the environment as riparian vegetation removal and land use can affect the natural conditions of ecosystems and alter the structure of different communities, such as those of chironomids (Sensolo et al., 2012;König and Santos, 2013).
The Chironomidae is the most abundant and diverse group of macroinvertebrates in aquatic ecosystems, with 4147 described spcies in the world and 618 in Neotrpical region (Ferrington, 2008).In Brazil, 320 species have been described in limnic ecosystems, however, this number is underestimated, as studies on the biology, taxonomy and ecology of the Chironomidae in the country are scarce (Mendes and Pinho, 2006).Chironomidae family members confer ecological significance in the ecological and energy partition of resources in the environment, making them important in the biotic stability of aquatic ecosystems (Ferrington, 2008).The success of this group in different environments occur because physiological adaptations into inhabit differents places with unstable environmental variables (Entrekin et al., 2007).The Chironomidae possesses a greater fisiological ability to tolerate various environments than other groups of aquatic insects, since environmental variables such as altitude, water temperature, conductivity, and pH influence their temporal distribution (Tejerina and Malizia, 2012) and are excellent indicators of environmental conditions (Brodersen and Anderson, 2002).
The effect of seasonal patterns on the composition of chironomids can also be tested, as different climatic conditions might favour the emergence of species (García and Suárez, 2007).However, this might be a consequence of natural variability within the community and often be confused with decreases due to environmental disturbance (Milošević et al., 2012).Changes in rainfall promote common disorders in streams (Tejerina and Malizia, 2012).One of the main effects of spates is the removal of organic matter organic matter adhered to substrates and consequently to hinder the colonisation of invertebrates (Rosin and Takeda, 2007).
The effects of environmental conditions variation on aquatic communities can be clearly observed in chironomids, depending on the wide distribution that this group can achieve (Tokeshi, 1995;Takeda et al., 2004).Changes in the composition and structure of the chironomid community can elucidate the consequences of anthropogenic activities or the natural deterioration of stream ecosystems (Armitage et al., 1995).
Palavras-chave: insetos aquáticos, filtros ambientais, variáveis físicas e químicas, variação temporal.Composition and diversity of the Chironomidae … nutrient input that shape the parameters of aquatic biodiversity (Sensolo et al., 2012).These parameters can act as environmental filters, selecting species that occur at a given site due to limiting environmental conditions (Roque et al., 2010).Thus, only species that have attributes that give them the ability to get through such conditions will be able to survive in a particular location.
Knowledge concerning to the patterns of the chironomid community is essential to construct a database useful for biomonitoring programmes (Milošević et al., 2012).Few studies report the spatial and temporal distribution of this group in the subtropical region, mainly due to the difficulty in identifying larvae at the genus and species levels, and also due to their great taxonomic variability (Raunio et al., 2011).Thus, this study aims to (i) evaluate the spatial and temporal distribution of chironomids in subtropical streams, (ii) investigate the importance of environmental predictors in the distribution of this group, and (iii) contribute to the knowledge of the biodiversity of the group in subtropical streams.Our hypotheses are that the variability of chironomid community follows a gradient in relation to environmental predictors, including temperature and nutrients that act as environmental filters that favour the dominant groups, and that temporal variations exist in the composition of the community, especially among the summer and winter communitites (intraannual), due to variation in rainfall between these seasons.

Study area
We conducted this study in 10 streams of 1 st to 3 rd order in the upper Uruguay River Basin (between 27°12'59" and 28°00'47"S; and 52°48'12" and 51°49'34"W (Faxinalzinho N=2, Marcelino Ramos N=2, Três arroios N=2 and Erechim N=4; Figure 1).The region has an altitude ranging from 400-800 m (Butzke, 1997), mean annual temperature is 17.6°C, and rainfall annual mean of 1,912.3mm (Bernardi and Budke, 2010).These elements characterise the climate of the area as belonging to the Köppen category Cfb (Moreno, 1961;Bernardi and Budke, 2010).The geology and soils consist of a basalt formation, and the soil is classified as "EC" (Oxisol Roxodistrófico).The vegetation is characterised by a mixed vegetation of Araucaria Forest and Subtropical Atlantic Forest (Oliveira-Filho et al., 2013).The region has an economy based on agriculture (soy, corn, wheat), which reflects about 70% of the landscape with anthropogenic uses.All studied streams are located in an agricultural matrix and have different percentages of vegetation in their riparian zone (ranging from 6% to 26%) (Decian et al., 2009;Sensolo et al., 2012).

Chironomidae collection
Chironomid larvae were collected in each stream at the different seasons of the year (spring, summer, autumn and winter) during 2010 and 2011.Three sub-samples were taken by working the Surber's metal frame (with an area 0.09 m 2 and net with mesh size of 250 µm) into the rocky stream bottom.The material was fixed in the field with 80% ethanol, placed in plastic vials and taken to the laboratory for separation of the individuals of the family Chironomidae from the other groups.The bleaching solution for larval identification consisted of 10% potassium hydroxide and was applied for 24 h.After creating semipermanent slides with Hoyer's solution, individuals were analysed with an optical microscope at a magnification of 1,000×.Identification was performed to the genus level using the identification key of Epler (2001) and Trivinho-Strixino (2011).The organisms identified were listed and deposited in the Collection of Benthic Invertebrates of the Regional Museum of the Upper Uruguay (MuRAU/URI -Erechim).

Water quality variables
Environmental variables (pH, dissolved oxygen, conductivity, turbidity, total dissolved solids, and water temperature) were analysed with a Horiba® U50 multiparameter analyser.Water samples were collected for quantification of the levels of total organic carbon (TOC) and total nitrogen (TN).These analyses were performed using a Determinator Organic Carbon TOC-VCSH (Shimadzu®).The methods for the analysis of these parameters are described in APHA (1998).

Data analysis
The abundance and richness of organisms were estimated in each sample.To evaluate differences in abundance and richness of the Chironomidae genera between the sampling sites and seasons, an analysis of variance (one-way ANOVA) was performed.To evaluate differences in Chironomidae genera composition between the sampling localities and annual seasons a multivariate analysis of variance (MANOVA) was carried out.To identify indicator genera for different periods of the year and sampling sites, we used the indicator species analysis method proposed by Dufrêne and Legendre (1997).To evaluate the influence of environmental variables on the distribution of organisms, we performed a redundancy analysis (RDA) using an array of biological abundance data transformed by log(x+1).The second array used in RDA comprised environmental data.Analyses were conducted using the "vegan" package (Oksanen et al., 2012) in R software (R Foundation for Statistical Computing, 2010).

Results
We identified 7,349 organisms in all study sites, distributed into three subfamilies (Chironominae, Tanypodinae, and Orthocladiinae) and 57 genera (Table 1).The Chironominae showed the greatest abundance, representing 49% of the total and 23 genera.The Orthocladiinae represented 39% of the total sample and was represented by 11 genera.The Tanypodinae represented 12% of the total and 11 genera.The most abundant genera during the study were Polypedilum (20% of the total) and Rheotanytarsus (19% of the total).The abundance and richness of the Chironomidae were similar spatially and temporally (p>0.05; Figure 2).
We found spatial and temporal variations in the composition of chironomid communities.Among the sites, the most striking differences were between Erechim streams in relation to the others (MANOVA: F 3,64 = 3.303, p = 0.001).Differences between the seasons in terms of temperature became clear (summer and winter; F 3,64 = 3.009, p = 0.001).Intermediate seasons showed a similar composition to each other, and the similarity between the following seasons (e.g., spring with summer and autumn with winter).The indicator species analysis revealed indicator taxa for summer: Thienemanniela, Pentaneura, Paratanytarsus, Parapentaneura, Parametriocnemus, Hudsonimyia, Labrundinea and Larsia, whereas Parakiefferiella, Paramerina, Metriocnemus showed potential as winter indicator genera.
The distribution of organisms was related to water temperature, and to eutrophication indicator variables (TOC, TN, and C:N ratio; Table 2; Figure 3).However, in the RDA analysis, the most abundant genera were negatively influenced by TOC variables, TN, and the C:N ratio, indicating some sensitivity of the most abundant genera (Figure 3).The higher water temperature-related collections were made in the summer and spring, whereas the streams of Marcelino Ramos and Três Arroios had a higher concentration of TN and a higher C:N ratio.

Discussion
The Chironomidae is a heterogeneous, diverse and important family for structuring aquatic communities (Pinder, 1986;Armitage et al., 1995;Merritt et al., 2008;Ferrington, 2008).In our study, we found two-fold more genera (57 genera) than other studies conducted in streams in the subtropical region (29 genera recorded by Sensolo et al., 2012;20 genera by Battistoni et al., 2013; 29 genera by König and Santos, 2013).For tropical streams, Aburaya and Callil (2007) identified 27 genera.Fesl (2002) studied streams in temperate regions and found 43 genera distributed among five subfamilies.Calle-Martínez and Casas (2006) also found 23 genera (three subfamilies) in temperate streams.We sampled three subfamilies in our study and observed a great amount of genera compared with other studies, both in tropic and temperate regions.
Polypedilum and Rheotanytarsus together accounted for approximately 40% of the total organisms, revealing their importance in structuring the community.Polypedilum is often indicated as tolerant to pollutants, such as diffuse contamination arising from agricultural activities (Macdonald and Taylor, 2006).This genus is cosmopolitan, comprising species with multiple dietary habits (Merritt et al., 2008), including tolerant to a broad range of environmental conditions species (Silva et al., 2013).Polypedilum immatures are easily found in many types of substrates and sediments and have great adaptability (Serpa-Filho et al., 2007).Rheotanytarsus is a filltering-collector and is abundant among other genera of the Chironomidae (Gonçalves Junior et al., 2003;Biasi et al., 2013;König and Santos, 2013).Rheotanytarsus larvae can also tolerate a wide range of environmental disturbances such as nutrient enrichment, a factor that is common in aquatic environments with an agricultural influence, as Polypedilum (Roque et al., 2000;Gresens et al., 2007), as in this study.
Our results indicate that variations in the Chironomidae community depend on seasonal and spatial factors.Approximately 50% of the variation in the composition of the the Chironomidae community can be explained by physico-chemical variables.The streams studied had an environmental gradient, particularly in the concentrations of TOC, TN, and the C:N ratio.This effect on the community in local scale might be driven by large-scale spatial factors, such as landscape attributes.This is probably because agricultural activities (predominantly in the vicinity of streams) contribute to the diffuse pollution of water bodies   (Sensolo et al., 2012).Thus, in our study, TOC, TN and the C:N ratio variables might operate as environmental filters and thereby promote changes in the chironomid community, favoring dominant genera as Polypedilum e Rheotanytarsus.The streams located in Marcelino Ramos and Três Arroios showed a higher concentration of TOC.
The spatial distribution of chironomids is directly related to the environmental conditions of the sites, physical and chemical characteristics of streams define the water quality, and influence the distribution of this group (Fesl, 2002).Other studies have also identified variability in the chironomid community that is generated by spatial factors.Roque et al. (2010) found that physicchemical parameters of water and other factors such as substrate type and attributes of the landscape are important modulators of the chironomid community in subtropical streams.Tejerina and Malizia (2012) studied streams in Argentina and found variations in chironomids following an altitudinal gradient and in relation to environmental variables such as temperature, conductivity and pH gradient.A similar pattern to that in our study, was observed by Calle-Martínez and Casas (2006) and Milošević et al. (2012), who found a relationship between the Chironomidae community with variables indicative of eutrophication (e.g., nitrogen and phosphorus), emphasising the importance of water quality for community composition.
Temporal variability in the communities monitored was mainly related to water temperature.We observed a seasonal variation in the composition of the Chironomidae, where some genera were representative in summer (Thienemanniela, Pentaneura, Paratanytarsus, Parapentaneura, Parametriocnemus, Hudsonimyia, Labrundinea, and Larsia) and others in winter (Parakiefferiella, Paramerina, and Metriocnemus).These seasonal variations might be driven by climatic and hydrological factors that are altered in certain seasons (Tejerina and Malizia, 2012).Milošević et al. (2012) found many variations in the chironomid community in contrasting seasons such as summer and winter, because different weather conditions can promote the emergence of some species in particular.However, some studies have reported disturbances arising from hydrological variations in the chironomid community, especially in richness and abundance, where some species might disappear from the stream after heavy rainfall; others might decrease in abundance, whereas others persist.
These arguments highlight the importance of understanding the temporal and spatial distribution of chironomid communities, so that the natural variability of the community is considered in biomonitoring studies and leads to accurate results.Knowledge concerning the distribution of chironomids is an important ecological tool for biological studies; as well as being a prominent group in aquatic ecosystems, chironomids can represent indicators of the level of pollution in freshwater ecosystems in monitoring programmes.

Table 1 .
Composition and diversity of the Chironomidae … Mean values (± standard deviation) of the abundance of Chironomidae genera collected in streams during Spring (P), Summer (V) Autumn (O) and Winter (I) in cities Erechim, Marcelino Ramos, Faxinalzinho and Três Riachos, Rio Grande do Sul, Brazil.

Table 1 .
Continued... Composition and diversity of the Chironomidae …

Table 2 .
Mean values (± standard deviation) of environmental variables measured during spring (P), Summer (V) Autumn (O) and winter (I) in the cities of Erechim, Marcelino Ramos, Faxinalzinho and Três Arroios, Rio Grande do Sul, Brazil.