Beta diversity has become essential for understanding ecosystem functioning and for determining biodiversity-conservation priority areas. However, the beta diversity patterns of invertebrates in tropical aquatic ecosystems are not well known, particularly in streams. Using data from low-order streams located in southern Brazil, we evaluated the beta diversity of Chironomidae. We tested the hypothesis that increased environmental heterogeneity increases beta diversity. The streams were grouped into two categories, rural and urban, according to the percentage of urbanization in their micro-basins. Our results showed that the heterogeneity of environmental variables can determine the beta diversity of Chironomidae, and the increased environmental heterogeneity caused by differences in the intensity of urbanization impacts can increase the beta diversity in urban streams. Therefore, the intensified impacts of anthropogenic activities in aquatic ecosystems can also increase beta diversity. Finally, we suggested that beta diversity can be an effective tool for not only designing measures and determining priority conservation areas, but also accurately identifying potentially degraded and priority sites that require water management plans.
The spatial variation in biodiversity has become a central theme in ecology (Kraft et al. 2011). Environmental variables play a fundamental role in determining the spatial patterns of species diversity (Hepp et al. 2012). In aquatic environments, the physical and chemical conditions of water can affect species distributions (Urban et al. 2006). Environmental changes, such as margin erosion, deforestation of riparian vegetation, and nutrient enrichment of soil near water bodies, can therefore determine the distribution of aquatic biota (Moreno et al. 2010; Cunico et al. 2012; Daga et al. 2012; Salvarrey et al. 2014) and consequently modify ecological processes (Buss et al. 2002; Milosevic et al. 2012).
One method of understanding these diversity patterns is to analyze beta diversity, which has become essential for understanding ecosystem functioning and for determining priority areas for the conservation of biodiversity (Margules et al. 2002; Nogueira et al. 2010). Beta diversity can represent the heterogeneous distribution of species in a given space due to environmental heterogeneity (Al-Shami et al. 2013). Because environmental impacts may be isolated events and may therefore increase ecosystem heterogeneity, the analysis of beta diversity in biomonitoring programs may help detect potentially degraded sites (Millán et al. 2011). However, few studies of diversity patterns in aquatic ecosystems, such as low-order streams, have focused on beta diversity (Clarke et al. 2008), particularly in tropical streams (Al-Shami et al. 2013).
Aquatic environments, such as streams, are particularly appropriate for analyzing patterns of beta diversity and associated causes because they are characterized by environmental gradients marked by natural or anthropogenic changes (Al-Shami et al. 2013). Anthropogenic activities associated with urbanization and agriculture can result in several environmental changes in streams, such as hydrographic changes, increases in nutrient and contaminant concentrations, altered channel morphology and stability, and decreased biodiversity (Paul & Meyer 2001; Meyer et al. 2005). In addition, environmental changes may vary depending on the degree of habitat modification and urbanization (Walsh et al. 2005).
The benthic macroinvertebrate community is very diverse in streams (Gutiérrez-Cánovas et al. 2013) and is frequently used as an indicator of environmental quality (Li et al. 2010) because it allows for the detection and evaluation of environmental impacts and helps ecologically characterize the status of conservation in aquatic ecosystems (Ruaro & Gubiani 2013). This bioindicator primarily relies on the response of organisms to physical and chemical changes in the habitat (Li et al. 2010). Environmental degradation gradients may therefore affect this community differently and form a gradient in the structure and distribution of these organisms in a hydrographic basin according to the impact type and intensity (Moreno et al. 2009).
The benthic macroinvertebrates belonging to the Chironomidae family (Insecta: Diptera) are very abundant and highly diverse in the majority of aquatic ecosystems (Epler 2001). This distribution of Chironomidae species may be associated with different degrees of tolerance to environmental changes (Szivak et al. 2013). Some species can be found in sites with extreme temperature, pH, concentration of dissolved oxygen and organic pollution levels, whereas others do not tolerate such conditions (Helson et al. 2006). Chironomidae is therefore an important tool for environmental monitoring and diagnosing and evaluating aquatic ecosystems (Armitage et al. 1995; Ferrington 2008) and for detecting environmental impacts caused by humans (De Bisthoven et al. 2005).
In the present study, we evaluated the beta diversity of Chironomidae in rural and urban streams and tested the hypothesis that increased environmental heterogeneity increases beta diversity. We expected the streams with greater heterogeneity in sediment composition and physical and chemical parameters to have a greater Chironomidae beta diversity. In addition, although rural streams may be impacted by agriculture, urbanization may result in high spatial environmental heterogeneity in urban streams, which would change the species composition at the impacted sites and consequently increase the beta diversity in urban streams. Finally, the novelty of this research is to test the efficacy of using beta diversity as an indicator of degraded areas, which would increase the accuracy of identifying priority sites that are in need of water resource management plans.
MATERIAL AND METHODS
The study was conducted in 10 low-order streams (sensuStrahler 1957) belonging to the Pirapó River hydrographic basin (22°30′ and 23°30′S; 51°15′ and 52°15′W – Figure 1). The predominant altitudes in the region vary between 380 and 540 m. Rainfall is concentrated during the summer months, and there is no defined dry season. The mean annual temperature is 21 °C (Queiroz 2003). The dominant landscape is a mosaic of agricultural activities and urban development, particularly in the Maringá metropolitan region, which is an important agro-industrial center of South-Central Brazil (Cunico et al. 2012).
The streams were grouped into two categories according to the percentage of urbanization of their micro-basins: rural streams (urbanization from 0 to 18.8%, n = 5) and urban streams (urbanization from 56.6 to 100%, n = 5). The urbanization percentage, adapted from Cunico et al. (2012), was calculated using high-resolution satellite images (Quickbird – panchromatic, 2005) and the vector editing tool of the Spring 4.3.2 software (Camara et al. 1996). The samples were collected in February 2008 (summer) along three longitudinal stretches in each stream (headwater, mid-catchment and mouth).
The composition of the benthic substrate was quantified using a quadrat (0.50 × 0.50 m) formed by a PVC tube subdivided with nylon thread into 25 subsections with areas of 0.10 m2. The presence/absence of benthic substrate types (silt/clay, sand, gravel, cobble, boulder and slab) and habitat structures (branches, twigs/leaves and artificial structures) were recorded for each subsection of the quadrat. The relative frequency of each category in the quadrat was calculated (Cunico et al. 2012). Five quadrats were sampled at each sampling site by the same researcher.
At the same sites, we measured the pH (DIGIMED DM-22), electric conductivity (μS cm−1 – DIGIMED DM-32), dissolved oxygen (mg l−1 – YSI 55D) and temperature (°C – YSI 55D) in the water. Water samples were collected to determine the total phosphorus (mg l−1) and nitrogen (mg l−1) contents; the samples were analyzed at the Laboratory of Agrochemistry and Hygiene of the State University of Maringá (Universidade Estadual de Maringá). All parameters were determined according to APHA (2000).
Chironomidae larvae were collected in triplicate using a Surber sampler (mesh = 250 μm; area = 0.09 m2). The sampler was placed against the current in stretches with boulders, cobbles and sand to standardize the substrate and minimize its effects on the faunal composition. The biological material was collected and fixed in the field in 4% formaldehyde buffered with calcium carbonate; the material was screened at the laboratory using a stereo microscope. For the morphology observations, the Chironomidae specimens were cleaned in 10% potassium hydroxide for 24 hours and were placed in semi-permanent slides with Hoyer solution, according to Trivinho-Strixino & Strixino (1995). The larvae were identified at the genus level using an optical microscope, identification keys from Trivinho-Strixino (2011) and Epler (2001), and consultations with specialists.
The possibility of spatial autocorrelation among the different sampling points in a single stream was tested using Moran's test (Sokal & Oden 1978). A principal component analysis (PCA) was performed to analyze the environmental heterogeneity among the streams in each category that resulted from different sediment compositions and the physical and chemical parameters of the water. A principal coordinate analysis (PCoA) was performed to analyze the species composition using a species density matrix. For both ordination analyses, the axes were retained according to the Broken-Stick criterion. Differences between the two types of streams (urban and rural) were tested with one-way analysis of variance (ANOVA) applied to the scores of the retained PCA and PCoA axes.
The beta diversity of Chironomidae was analyzed using a test of homogeneity of dispersion (PERMDISP; Anderson et al. 2006), which tests the variability in the Chironomidae species composition between rural and urban streams. The test calculates a centroid for each type of stream and determines the Bray–Curtis distance of each sampling site to the centroid. A greater mean Bray–Curtis distance to the centroid corresponds to a greater dissimilarity in the species composition, suggesting an increase in the beta diversity. The significance (p < 0.05) of the differences in the beta diversity (mean of the Bray–Curtis distance to the centroid) between the two types of streams was tested using a residual least-squares permutation test with 999 permutations.
There was no spatial autocorrelation among the different sampling points in each stream according to Moran's test (Moran's I =0.227; p = 0.971), indicating that the sampling points can be considered independent.
In terms of the environmental factors, the first two PCA axes were retained according to the Broken-Stick criterion. The PCA analysis indicated that there were differences in the environmental variables between the two types of streams, as observed on axis 1. This difference was confirmed by one-way ANOVA (F1, 28 = 30.30; p < 0.001). The urban stream sampling sites displayed a greater dispersion on the graph, suggesting a greater environmental heterogeneity among the sampling sites (Figure 2). Higher values of total water phosphorus and nitrogen content were observed in these streams, and the sediment was characterized by large particles, such as boulders, cobbles and twigs/leaves.
Chironomidae was represented by a total of 21,761 larvae belonging to three subfamilies and 32 genera (see Appendix, available online at http://www.iwaponline.com/wst/071/112.pdf). The subfamily Chironominae was the most abundant (14,053 individuals or 64% of the total), followed by Orthocladiinae (6,253 individuals or 29%) and Tanypodinae (1,455 individuals or 7%). The first two PCoA axes were retained according to the Broken-Stick criterion. The PCoA clearly showed differences in the species composition between the two types of streams, and these differences were statistically significant (F1, 28 = 125.53; p < 0.001; Figure 3). Chironomus, Goeldchironomus, Parachironomus, Polypedilum and Rheotanytarsus species were representative of the urban streams and were negatively associated with the PCoA axis 1. Parametriocnemus and Corynoneura species dominated in the rural streams, and they were positively associated with the PCoA axis 1 (Figure 3).
The pattern observed for the environmental variables and the species composition was also observed for beta diversity. The mean distance to the centroid was greater for urban streams (distance to the centroid = 0.52) than for rural streams (distance to the centroid = 0.41). The difference in species composition variability (beta diversity) between the two types of streams was significant (PERMDISP, F1, 28 = 4.81; p = 0.03), indicating a higher dissimilarity in the Chironomidae species composition in urban streams (Figure 4).
The patterns of invertebrate beta diversity are not well known in tropical aquatic ecosystems, particularly in streams (Al-Shami et al. 2013). In addition, the relationship of environmental variables with beta diversity in streams has been relatively less studied than the relationship with alpha diversity (Costa & Melo 2008). Understanding the processes that contribute to increasing the spatial variability in diversity therefore continues to be a fundamental question of contemporary ecology (Tuomisto et al. 2003). The present results show that environmental variables are different between urban and rural streams and are also more heterogeneous among urban streams. Consequently, the invertebrate species composition is different for different types of streams, and it responds to the environmental heterogeneity. Species with restricted occurrences were observed for each type of stream. Beta diversity was therefore higher in the urban streams in response to the high environmental heterogeneity found in these environments, and the hypothesis that an increased environmental heterogeneity increases Chironomidae beta diversity was not rejected.
Streams are highly heterogeneous habitats, as reflected in the variation in their faunal assembly composition (Hepp & Melo 2013). However, the micro-basin urbanization percentage of each stream could increase the species spatial heterogeneity and, consequently, the differences in species composition. Generalist and opportunistic species become more dominant in urbanized areas (Magura et al. 2008; Jones & Leather 2012). Environmental heterogeneity is an important factor that affects local diversity, particularly for benthic macroinvertebrates, because it increases the variability in resources and refuges for organisms (Costa & Melo 2008; Mykra et al. 2011). In general, the observed pattern of environmental variables suggests a higher incidence of degradation points in urban streams, which directly affect the Chironomidae species composition.
Chironomidae larvae occupy different biotopes and display a high sensitivity to environmental conditions (Bhattacharya et al. 2006). Therefore, different urbanization percentages, which potentially reflect different levels of disturbance and/or impact, may determine the predominant tolerant species (generalists) present in urban streams and the predominant sensitive species (specialists) present in rural streams. The observed predominance of genera such as Chironomus, Thienemanniella, Rheotanytarsus and Polypedilum in areas subjected to urban impacts seems to be consistent (Day et al. 2006; Helson et al. 2006). In contrast, genera such as Corynoneura, Parametriocnemus and Lopescladius, which are present in rural streams, may be associated with better preserved areas with a higher presence of riparian vegetation, which contains species that are less tolerant to environmental impacts (Henriques-Oliveira et al. 1999; Trivinho-Strixino 2011).
Chironomus and Polypedilum display a high range of tolerance to pollution, and these species are widely recognized as indicators of organic pollution in lotic ecosystems (Armitage et al. 1995). Although they are generalists, Chironomus larvae were abundant in the streams located near highly urbanized areas. Polypedilum larvae occurred in practically all the streams, but some were abundant in areas with a higher urbanization influence, whereas others were more abundant in areas with increased agricultural practices (e.g., use of agrochemicals). In contrast, some genera occurred exclusively in rural streams, such as Cladotanytarsus and Paramerina. These genera are rare and are present in areas with microhabitats in well-preserved streams, where plants, plant debris and leaf litter are the predominant substrates (Trivinho-Strixino 2011).
The availability of organic matter and the type and size of the particles that compose the substrate may determine the benthic macroinvertebrate community in this type of ecosystem (Costa & Melo 2008). A wide range of environmental factors may therefore affect the distribution of organisms within streams (McCulloch 1986), and the high heterogeneity of these environments allows species with different ecological requirements to occur in different locations within the same stream (Al-Shami et al. 2013). The effect of environmental heterogeneity on species composition (tolerant and sensitive) resulted in higher beta diversity in streams with a high incidence of environmental impacts (urban streams). Therefore, the intensification of the impacts caused by anthropogenic activities in aquatic ecosystems could also contribute to an increase in the beta diversity (Chase & Leibold 2002).
Our results coincide with those obtained by Astorga et al. (2014), who suggest that local environmental heterogeneity may be the strongest determinant of beta diversity of stream invertebrates. Therefore, the heterogeneity produced broader habitat differences between streams and thus higher beta diversity (Jost 2007). In addition, our results confirm the previous findings of Johnson & Angeler (2014), who concluded that a disturbance in the environment results in a loss of (sensitive) taxa, while the diversity of tolerant taxa remains high (β-diversity). The results of the present study indicate that measuring beta diversity can lead to a better understanding of the functioning of ecosystems (Legendre et al. 2005) and that beta diversity helps determine degradation processes and areas for biodiversity conservation (Overton et al. 2009; Clarke et al. 2010; Anderson et al. 2011).
In summary, our results show that different intensities of urbanization impacts in streams may increase the heterogeneity of environmental variables, which determines Chironomidae species composition and possibly elevates the beta diversity of this invertebrate group, particularly in urban streams. Beta diversity is therefore an effective tool for designing measures and determining priority conservation areas and for determining potentially degraded sites. We suggest that beta diversity, a neglected metric in applied ecological studies, should be quantified in biomonitoring programs and included in environmental impact evaluation indexes to identify priority streams in need of water resource restoration and management plans and that Chironomidae beta diversity in particular is an important attribute when evaluating aquatic environments subjected to human influence.