Effects of anthropogenic impacts on benthic macroinvertebrates assemblages in subtropical mountain streams

The nature of the riparian and surrounding landscape has been modified by anthropogenic activities, which may subsequently alter the composition and functional structure of macroinvertebrate assemblages. The effect of these changes on function of benthic fauna is difficult to assess due to the scarce knowledge on functional structures in tropical streams. In this study we evaluate whether sites impacted and unimpacted by anthropogenic alterations differed in assemblage composition and density, richness and diversity of each functional feeding group. The selection of the sites was related to their distinct riparian characteristics, following the QBRy riparian quality index. Collector-gatherer was the dominant functional feeding group, comprising 91% of total density, whereas the proportion of shredders was very low, representing less of 0.5% of total density. Asemblage composition of macroinvertebrates differed between impacted and unimpacted sites. Predators were dominant in taxa number, representing about 60% of total taxa richness. In addition, the diversity and richness of collector-gatherers differed significantly between degraded and unimpacted sites, reflecting the sensitivity of this group to environmental changes and the utility to be used in the assessment of anthropogenic modifications. The results of this study reinforce the idea that riparian corridor management is critical for the distribution of macroinvertebrate assemblages as well as functional organization of lotic streams. KEywords. Anthropogenic changes, Argentina, lotic streams, mountain basin, riparian quality index. rEsumEn. Efecto de los impactos antrópicos sobre los ensambles de macroinvertebrados bentónicos en ríos subtropicales de montaña. Las actividades antrópicas han producido modificaciones en el paisaje ripario y el sistema terrestre adyacente, las cuales causan alteraciones en la composición y estructura funcional de los ensambles de macroinvertebrados. Asimismo, el efecto de estos cambios sobre la estructura funcional de macroinvertebrados es difícil de evaluar debido al escaso conocimiento de este parámetro en ríos tropicales. En este estudio se evaluó si sitios degradados y no degradados por impactos antrópicos diferían en la composición de los ensambles y en la densidad, riqueza y diversidad de cada grupo funcional. La selección de los sitios se relacionó con su calidad riparia, de acuerdo con el índice QBRy. Colectores-recolectores fue el grupo funcional dominante, comprendiendo el 91% de la densidad total, mientras que la proporción de trituradores fue muy baja, representando menos del 0.5% de la densidad total. La composición de los ensambles de macroinvertebrados difirió significativamente entre sitios degradados y de buena calidad riparia. Los predadores fueron dominantes en número de taxones, representando alrededor del 60% de la riqueza de taxones. La diversidad y riqueza de collectores-recolectores fue menor en sitios degradados, reflejando la sensibilidad de este grupo frente a cambios antrópicos. Los resultados de este estudio destacan la importancia del manejo sostenible del corredor ripario debido a su estrecha relación con la distribución de los ensambles de invertebrados y su estructura funcional. PAlAbrAs-clAvE. Argentina, cambios antrópicos, cuenca de montaña, índice de calidad riparia, ríos lóticos. Riparian vegetation influences the structure and functioning of stream macroinvertebrate communities through the provision of organic matter and by shading the stream (Hynes, 1975; Allan, 2004). Anthropogenic changes in the riparian corridor may subsequently alter the functional feeding group composition of macroinvertebrates by modifying the supply of food resources, and producing changes in habitat structure and quality (Dudgeon, 2006; Wantzen & Wagner, 2006). Theories of macroinvertebrate function have typically been derived from detailed studies of temperate systems (Merrit & Cummins, 1996; Cummins et al., 2005), and consideration of tropical systems has been either excluded, or included as an interesting anomaly, despite the tropics occuping the largest land area of the world’s climatic regions (Dobson et al., 2002). Information on factors affecting the functional structure of macroinvertebrates is not only vital for basic ecological understanding and biodiversity conservation, but also Mesa et al.

Riparian vegetation influences the structure and functioning of stream macroinvertebrate communities through the provision of organic matter and by shading the stream (Hynes, 1975;Allan, 2004).Anthropogenic changes in the riparian corridor may subsequently alter the functional feeding group composition of macroinvertebrates by modifying the supply of food resources, and producing changes in habitat structure and quality (Dudgeon, 2006;Wantzen & Wagner, 2006).
Theories of macroinvertebrate function have typically been derived from detailed studies of temperate systems (Merrit & Cummins, 1996;Cummins et al., 2005), and consideration of tropical systems has been either excluded, or included as an interesting anomaly, despite the tropics occuping the largest land area of the world's climatic regions (Dobson et al., 2002).Information on factors affecting the functional structure of macroinvertebrates is not only vital for basic ecological understanding and biodiversity conservation, but also as a model for monitoring, restoring and maintaining the quality of stream ecosystems (Rosenberg & Resh, 1993;Palmer et al., 1997).
In the Northwestern Argentina, the ecological integrity of streams is being threatened by different anthropogenic changes.The expansion of agricultural and urban areas, overgrazing and the introduction of exotic species are some impacts that affect the riparian quality and integrity of biotic stream communities in this area (Fernández et al., 2008;Sirombra & Mesa, 2012).Although degradation of the riparian zone has significantly increased in tropical streams in the last few decades (Kasangaki et al., 2008), the consequences of these changes on the structure of macroinvertebrate assemblages are scarcely known (Nessimian et al., 2008).According to that, we asignate macroinvertebrate taxa of subtropical streams to functional feeding groups in order to assess whether sites of constrasting riparian quality differ in density, richness and diversity of functional feeding groups and the environmental variables related to these changes.We hypothesized that sites impacted and unimpacted by anthropogenic changes differ in assemblage composition and richness and diversity of functional feeding groups.
mAtEriAls And mEthods study area.We studied streams included in a basin of Tucumán province, located in northwestern of Argentina (26°36'S; 65°45'W) (Fig. 1).This region is characterized by the monsoons, with an annual precipitation of 1000 mm, falling mostly during summer (80%).High water period extends from November to April, with maximum precipitation values during January (250 mm approximately).The dry period extends from April to October with minimum rainfall values (11 mm).
The Yungas Phytogeographical province extends in the northwestern of Argentina from the frontier of Bolivia (23°) to the north of Catamarca (29°), with an area of 52.000 km 2 .Not only are these forests and their streams extremely important for the provision of water for irrigation, generation of electricity and human consumption throughout most of the Andean region, but they are also an extremely diverse and threatened ecosystems (Grau & Aragón, 2000;Brown et al., 2006).They are undergoing rapid change as a result of increasing human population with growing demands for water and food (Brown et al., 2006).The conversion of forests to pasture for cattle rearing is common and probably the most destructive change in the riparian zone of the Yungas streams, due to the radical modification of the vegetation, the reduced opportunities for regeneration of native species and because this activity creates the opportunity for the invasion of exotic species (Sirombra & Mesa, 2012).
Field methods.The selected sites were sampled during September 2005 and September 2006 (low water season).A Surber sampler (300 μm, 0.09 m 2 ) was used to collect the benthic samples.In the field, samples were preserved in 4% formaldehyde and packed for examination in the laboratory.Invertebrates were identified to the lowest possible taxonomic level (mostly genus) using the available keys (Domínguez & Fernández, 2009).Organic detritus were separated from benthic samples, cleaned of sediment and divided into fine (250 µm -1 mm) and coarse (>1 mm) particulate fractions (FPOM and CPOM, respectively).CPOM was separated into wood, leafs, and 'others' (seeds, fragments of roots, fruits and flowers).All fractions were dried (100 °C for 24 h) and weighed on an electronic scale to ± 0.01 mg.
Discharge was measured each sample date with a velocity meter (Global Water Flow meter).Water samples were analyzed in the laboratory for nutrient (nitrate) and dissolved oxygen following APHA (1992).Water temperature, pH (Methrom 704) and conductivity (Methrom E587) were measured in situ in each sampled site and date.Assignation of taxa to functional feeding groups.Aquatic insects have been assigned to feeding categories based on the kind of ingested food item obtained through gut content analysis, the behavioral strategy oriented to obtain resources and the nature of the morphological structure for alimentary acquisition.Gut contents of 1-5 individuals of each genus were subjected to examination of gut contents.The diet was analysed by removing the foregut and midgut content using ventral dissection of thorax (Peckarsky, 1996).The description and identification of ingested items were made under a microscope (200 X).The gut content of each specimen was mounted with glycerine in a customized slide with a central squared receptacle of uniform depth.The receptacle (20 mm x 20 mm) was divided into 625 equal-sized grid cells (0.64 mm 2 ), from which 15 were randomly selected.At each cell, the point interception method was employed for surveying diet composition.The scale bar of the microscope eye piece was used as a transect.Ten points were identified along the transect with a constant interval between them (Reynaga & Rueda Martin, 2010).Five categories of food resource were recognized: sediment particles, fine particulate organic matter (<1 mm), coarse particulate organic matter (>1 mm), microalgae and rests of aquatic invertebrates.The allocation of each taxon to the functional feeding group (FFG) depended mainly on the size, type and proportion of food items in gut contents (Cummins, 1973).Our own observations of feeding behavior during sampling and the observations of the mouthpart morphology in the laboratory were also helpful for FFG determination.Previous FFG allocations from literature were also consulted (Merritt & Cummins, 1996;Cummins et al., 2005;Tomanova et al., 2006;2007;Reynaga, 2009;Reynaga & Dos Santos, 2012) statistical analyses.Functional structure was described in terms of density (ind.m -2 ), species richness (using Rarefaction), diversity (Shannon diversity index) and evenness of the different FFGs.
Data of macroinvertebrates density (transformed at log 10 (x+1)) were explored using non-metric multidimensional scaling (NMDS) with Bray-Curtis distances (R program, package Vegan, Oksanen et al., 2006).This ordination was performed to assess if sites impacted and unimpacted by anthopogenic changes differed in assemblage composition.We correlated the density of each taxon with ordination axes in order to determine which best accounted for separation of sites in the ordination space using the Spearman correlation.One way similarities analysis was used to compare the variability in assemblage composition among impacted and unimpacted sites using the ANOSIM statistic test at α = 0.05 (R program, package Vegan, Oksanen et al., 2006).
Nonparametric Mann-Whitney test (Infostat) was used to evaluate the differences in physical and chemical variables, functional structure and density of dominant taxa of each FFG between impacted and unimpacted sites.

rEsults
Mean values of environmental characteristics relative to impacted and unimpacted sites were shown in Table I.Dry channel width was significantly higher in unimpacted sites (Mann-Whitney test, p<0.05), whereas water temperature and conductivity were higher in degraded sites (Mann-Whitnet test, p<0.05) (Tab.I).
Total abundance, richness, diversity and evenness of each FFG were similar between sites, except for collector-gatherers.Richness, diversity and evenness of collector-gatherers were significantly higher in unimpacted than impacted sites (Mann Whitney test, p<0.01) (Tab.II).In addition, density of Tipulidae was significantly higher in unimpacted sites, whereas Oligochaeta was dominant in degraded sites (Tab.II).

discussion
The dominance of collector-gatherers could represent an adaptation to highly disturbed enviroments, as suggested by other tropical and subtropical studies (Callisto et al., 2001;Buss et al., 2002;Boyero et al., 2006;Tomanova et al., 2006;Uieda & Motta, 2007).In disturbed streams in terms of discharge, the supply and persistence of a particular food item is very variable.Hence the ability to exploit changing resources may potentially maintain population stability against natural fluctuations (Dobson et al., 2002).Organic matter is a non-limited resource in tropical streams.Lotic streams receive large quantities of particulate organic matter from riparian zone by erosion of banks and by drift from upper zones.In addition, leaf decomposition to fine detritus in warm tropical streams is fast (Mathuriau & Chauvet, 2002;Dobson et al., 2002) and continuous throughout the year.Hence, the occurrence of this food item in the guts contents would be mainly due to its high availability in the habitat (Henriques-Oliveira et al., 2003) rather than to feeding specialization.Indeed, a community dominated by taxa with high affinity Tab.I. Mean values (SD) of environmental variables relative to impacted and unimpacted sites, and results of Mann-Whitney tests used to compare these characteristics among the groups of sites, Tucumán province, northwestern of Argentina.Significant differences are showed in bold (n.s, no significant).to the same food resource implies less inter-specific competition for the resource, and hence a relatively high resource availability.

Unimpacted
Predators were highly dominant in taxa number, representing about 60% of total taxa richness.This finding coincides with other studies in tropical streams (Callisto et al., 2001;Cheshire et al., 2005), showing the importance of this FFG for the biodiversity of these lotic systems.On the other hand, shredders were scarce in the studied streams, comprising about less than 0.5% of total density of taxa.This pattern was in acordance with other reports relating with the scarcity of shredders in tropical streams (Dobson et al., 2002;Mathuriau & Chauvet, 2002;Wright & Covich, 2005;Wantzen & Wagner, 2006).It has been suggested that shredding may be less important in tropical systems because there are alternative decomposition pathways for leafs, such as faster microbial processing due to higher temperatures (Irons et al., 1994).In addition, the low quality of tropical leafs in terms of roughness and high tannin and low nutrient contents may also explain the low numbers of this group in comparison to temperate systems (Stout, 1989;Wantzen et al., 2002).Shredders tend to have long life-cycles and are slow colonizers (Jacobsen & Encalada, 1998), life traits that make them unsuitable for living in frequently disturbed streams.
In relation to the effects of land use, the higher value of nitrate in degraded streams could be associated with livestock and agricultural activities, factors of increasing expansion in the riparian zone of the studied streams (Brown et al., 2002;Sirombra & Mesa, 2012).Modifications in cover structure and quality of the riparian vegetation, factors associated with the remotion and trampling of native species and the introduction of exotic vegetation, could increase the amount and intensity of light reaching the stream surface, increasing water temperature.The increase of light reaching the stream surface would increase the biomass of algae in impacted sites.In addition, a higher leaf biomass characterized degraded sites.According with our observations, whole leafs of Citrus and Acacia appeared in most samples of impacted reaches, constituting a significant proportion of the total organic matter biomass in these sites.Leafs of these exotic species have high cellulose, low nitrogen content and chemical inhibitors, factors that negatively affect microbial invasion and subsequent processing and degradation (Alonso et al., 2010).In addition, these leafs could be used mainly as substrate and indirect source of food for several macroinvertebrates, such as collectors.
This study found that sites of different riparian quality differed in the assemblage composition of macroinvertebrates, and this result was in accordance with other studies (Miserendino & Masi, 2010;Arnaiz et al., 2011).In addition, richness, diversity and evenness of collector-gatherers differed among undegraded and degraded sites.This result was in accordance with other works (Rawer-Jost et al., 2000;Stepenuck et al., 2002;Compin & Céréghino, 2007), showing the utility of the structure of functional feeding groups of macroinvertebrates as indicators of anthropogenic impacts.Native forested sites having good ecological conditions supported more diversity and richness of collector-gatherers, whereas sites degraded by anthropogenic impacts showed a significant decrease in these parameters.This could be related with that 33 of a total 40 taxa relative to this FFG corresponded to ephemeropteran, trichopteran and coleopteran species.A great proportion of these groups are composed by taxa intolerant to anthropogenic changes, such as increase of nitrate and water temperature, determining their absence and consequently, the lower richness of collector-gatherers in degraded sites.In contrast, Oligochaeta had a higher density in impacted streams, and this was related with the tolerance of this group to degraded environmental conditions (Rosenberg & Resh, 1993).
The results of this study showed the utility of collectorgatherers as a tool in the assessment of anthropogenic changes in subtropical streams.In addition, this work reinforces the idea that riparian corridor management is critical for the distribution of macroinvertebrate assemblages as well as functional organization of lotic streams.The maintenance of good conditions of this corridor makes possible the availability of habitats necessaries for maintain a high biodiversity in stream communities.

Fig. 1 .
Fig. 1.Localization of the study area and distribution of reaches of good (white dots), and poor/bad quality (grey dots) of their riparian zone, Tucumán province, northwestern of Argentina.

Fig. 2 .
Fig. 2. Non-metric multidimensional scaling ordination based on density of benthic macroinvertebrates collected in impacted (black circles) and unimpacted sites (white circles), Tucumán province, northwestern of Argentina.