Controlling factors of benthic macroinvertebrates distribution in a small tropical pond , lateral to the Paranapanema River ( São Paulo , Brazil )

Aim: The aim of the present study was to examine the benthic fauna in a marginal pond lateral to the Paranapanema River and to identify the main controlling factors of its distribution. Considering the small size of the lacustrine ecosystem, we expected that seasonal variations of the benthic community attributes are more important than spatial variations; Methods: Two samplings, one in March and another in August, were carried out at nine sites in the pond. Sediment samples were obtained through a Van Veen grab for invertebrate sorting, granulometric analysis, and for quantification of organic matter in sediment. Other abiotic factors were measured, such as water transparency, dissolved oxygen, pH, electric conductivity, temperature, and depth of sediment sampling sites. Regarding the comparative analysis at spatial scale, no significant variations in density of the benthic invertebrate community were found. Results: In relation to the studied abiotic factors, only depth presented significant differences among sampling sites; All the measured environmental parameters presented significant differences among sampling months, except depth and the physical and chemical characteristics of the sediment. The abundance of Chaoboridae and Chironomidae was the unique attribute with a significant difference in comparing the two months. A higher abundance of taxa occurred in August, especially for Oligochaeta, Nematoda, Chaoboridae, and Chironomidae; Conclusions: Because of the low structural complexity of the studied pond, we concluded that the changes in benthic macroinvertebrate community attributes were mainly due to seasonal effects.

temporal variation, as compared with the organisms in environments with high structural complexity, which can persist in these sites for a longer time.According to Formigo (1997), composition and density of macroinvertebrate communities are relatively stable from one year to the next in nonperturbed systems.However, seasonal fluctuations linked to the dynamics of vital cycles of each species can result in extreme variations in community structure in some environments.
In lakes marginal to tropical rivers, as those located in flood plains, aquatic biota is mainly influenced by the regime of flood pulses, as it was observed in the High Paraná River (Higuti and Takeda, 2002) and in the High Paranapanema River (Davanso and Henry, 2006).
The aim of the present study was to examine composition and diversity of benthic macroinvertebrates in different sites of a pond which is marginal to a river.Relationships between the different taxa of benthic macroinvertebrates and environmental factors were determined.Because it is a pond with small size, we expected seasonal variations in benthic community to be more important than spatial ones, due to the structural homogeneity of the environment.

Material and Methods
The environment selected for the study (Mian Pond, Figure 1) is located near the mouth zone of the Paranapanema into the Jurumirim Reservoir (São Paulo, Brazil).It is a small pond (perimeter: 1,592 m; mean depth: 3.7 m; length: 749 m; maximum width: 115 m).The Mian pond presents permanent connectivity to the river and one of the margins is formed by a riparian forest while the other is a non-preserved area next to a soybean field, being partially protected from the wind.Nine sites distributed in three transects were selected for the samplings (Figure 1).
In each site, three sediment samples were collected in March and August 2009 with a 0.064 m 2 Van Veen grab for the analysis of benthic fauna and an additional three samples for the determination of granulometric composition and quantification of organic matter in sediment.Depth profiles of the lake bottom at sampling site transects are shown in Figure 2.
In the field, sampled sediment for benthic fauna analysis was washed in 250 µm mesh net and fixed with 4% formaldehyde.Afterwards, the organisms were sorted, identified, and counted under a stereoscopic microscope.They were identified up

Introduction
Comparative studies of small water bodies have been used to examine the relationships between environmental factors and the structure of aquatic communities (Heino, 2000).Benthic macroinvertebrates present great importance in some ecological processes, such as energy fluxes and nutrient cycling (Leal et al., 2003;Henry and Santos, 2008).Bioturbation of sediment surface and fragmentation of leaves from riparian vegetation are some of the processes of nutrient release to water carried out by benthic organisms (Caliman et al., 2007;Callisto et al., 2009).
Considering the benthic macroinvertebrates, insects are the predominant taxonomic group in abundance and biomass in the majority of tropical lakes (França and Callisto, 2007).In all these environments, a predominance of Diptera, Chironomidae (Stenert et al., 2004;Roque et al., 2004), and Chaoboridae (Fukuhara et al., 1987) has been recorded.
According to Jonasson (1996), the distribution, composition, and diversity of macroinvertebrates, as well as of other aquatic communities, are affected by abiotic and biotic factors and also by mutual interactions among the organisms.Thus, benthic macroinvertebrate communities clearly indicate the ecological conditions of inhabited aquatic ecosystems.According to Kownacki et al. (2000), benthic fauna composition in aquatic environments depends mainly on factors such as substratum type, water trophic status, and hydro-period.
Oxygen and depth also constitute essential factors to macroinvertebrate distribution (Santos and Henry, 2001).Density of these organisms is remarkably lower at great depths, but the existence of some species tolerant to low oxygen concentrations is evidenced at these sites (Hirabayashi and Hayashi, 1994).Other important factors for benthic species distribution are the availability of food resources (Sanseverino et al., 1998) and the interspecific trophic interactions, such as competition and predation (Walker, 1998).
Habitat complexity can determine the composition of a local community.According to Barreto (1999), diversity of biological communities in more complex sites tends to increase due to the presence of environments with minor stress, ample availability of shelter against predators, and protection against physical disturbances, which serve to assist in survival, recovery, and persistence of the organisms.Therefore communities in habitats with low complexity usually present great 1969), electrical conductivity (through a Hach Mod.2511 conductivimeter) and pH (through a Micronal Mod.322 pHmeter).All of them, except the transparency, were measured at the surface and 10 cm of the bottom.Sediment granulometric composition was determined according to the Wentworth scale (Suguio, 1973) and organic matter content, through calcination loss (in a furnace at 550 °C for 1 hour).
The following water physical and chemical variables were determined: temperature (by mercury thermometer), transparency (by Secchi disk), dissolved oxygen (through the Winkler method, modified by azyde addition, Golterman and Clymo,  and sediment organic matter and granulometric composition).
All the statistical analyses were carried out using the Statistica 6.0 software (Statsoft, 2002).

Results
A small variation in rainfall (<3 mm) was observed, between the two sampling months (Figure 3).Considering that in July there was a high level of rainfall, atypical for the dry season, no significant difference in depth was found in comparing the two studied periods (Figure 4).
In almost all the sites, silt/clay was dominant in the sediment, followed by very fine and fine sands the measured environmental variables and for the benthic groups (Chaoboridae, Chironomidae, Nematoda and Oligochaeta) that presented great total abundance of organisms (>4%).Next, data were log (x + 1) transformed and submitted to a two way variance analysis (ANOVA) involving sampling sites (n = 9) and periods (n = 2).Normality and variance homogeneity were obtained.A 0.1 level was used for testing the significance of ANOVA.
Pearson correlations were computed to assess significant relationships (p > 0.03) between abundance of benthic taxa and environmental data (water temperature, depth, transparency, pH, dissolved oxygen and electrical conductivity,  by site 5 (19.76%) in August, and the lowest value (1.99%) was found at site 9 in August.
The variation in density of taxonomic groups is shown in Figure 7. Groups with the highest density were Chaoboridae, Oligochaeta, Nematoda, and Chironomidae.Chaoboridae (exclusively composed of Chaoborus sp.) was the predominant group (relative abundance corresponding to 44% of total abundance) in relation to the other organisms.Chaoboridae constituted more than 50% of the organisms in four sites (sites 1 and 6, in both months and sites 3 and 7, only in March) and was absent only at site 9 in August.The maximum density (344 individuals.m - ) was observed at site 4 in March.Oligochaeta predominated in three sites (site 2 in March and sites 8 and 9 in August).The highest densities were recorded in August (maximum of 240 individuals.m - ) and rose to around 21% of all the organisms in the entire study.Nematoda occurred in almost all the sampled sites and presented the highest relative abundance (50% of the total density of organisms) at site 9 in March.The highest density (286 individuals.m - ) occurred at site 2 in August.Considering both months, Nematoda density corresponded to approximately 17% of the organisms.
Chironomidae was a group with low abundance (≈2%) in the first month of the study and appeared only in two sampling sites.However, in August, Chironomidae specimens were observed in almost all the sites (except in site 9) and the highest density (260 individuals.m - ) was recorded in site 3.
Only depth presented a significant difference among sampling sites (F = 3.30; p = 0.04).Comparing the two periods of the year, ANOVA showed a significant difference for water temperature (highest value in March), water electrical conductivity at surface and bottom (highest values in August), water pH at surface and bottom (highest values in August), dissolved oxygen in water at bottom (highest concentration in August), and water transparency (highest value in March) (Figure 8).Regarding the abundance of benthic macroinvertebrates, significant differences were recorded for Chaoboridae (highest value in March) and Chironomidae (highest value in August).
Significant Pearson correlations between biological and environmental variables are presented in Table 1.Chaoboridae density showed positive correlation with water temperature and silt/clay (the finest fraction of sediment) and negative correlation with larger particles in sediment (very coarse, (Figure 5).Higher values of silt/clay were observed in March.In site 4, sediment was composed of different fractions at the same levels as in March and August (Figure 5).In site 9, it was observed the greatest change in sediment composition from March to August.No variation pattern, regarding organic matter content in the sediment, was evidenced between sites when comparing the sampling months (Figure 6).The highest value (21.46%) was recorded in March at site 3, followed were not detected on a spatial scale, since depth was the only factor, considering all chemical and physical variables, with significant difference among sampling sites.Nevertheless, it was not likely to cause significant differences in ecological attributes of the community in these local.These observations evidenced great influence of seasonality on organization of the macroinvertebrate community.
In relation to benthic macroinvertebrates recorded in lakes of the same study region (Davanso andHenry, 2006, 2007), Mian Pond was the only that didn't show significance difference in spatial distribution of the fauna.This finding can be explained by the reduced size, simple shape (no dendritic), and its relatively homogeneous sediment composition, morphological characteristics which offer low habitat complexity for the colonization and permanence of new species.Therefore, the simple structure of the pond must be considered an important factor in the homogeneity of spatial composition registered in the water body, resulting in few local alterations in abiotic factors and, consequently, in benthic community.coarse, and mean sand).Chironomidae density was associated negatively with water temperature and transparency and positively with water electrical conductivity, pH (of both the depths), and dissolved oxygen at bottom.

Discussion
In this study, low rainfall variation was observed in the two sampling months, selected to be representative of the rainy and dry seasons.This observation can characterize an "atypical" year, considering that the precipitation in August was different from the previous occurrence in the region (Martins and Henry, 2004;Henry, 2003Henry, , 2005;;Davanso andHenry, 2006, 2007).This certainly affected the environmental variables of the aquatic ecosystem, except water temperature.
Despite effects of precipitation regime on benthic fauna were slightly detectable, significant seasonal alterations in community structure were observed, mainly due to variation in temperature, dissolved oxygen and sediment composition between the seasons.Alterations in benthic fauna   In addition to the morphometric characteristics of the pond, another aspect that contributed to the high homogeneity of the environment, especially in sediment composition, is the absence of macrophytes.According to Corbi andTrivinho-Strixino, (2002), andBeckett et al. (1992), the highest densities of benthic macroinvertebrates are recorded in environments rich in macrophytes.maintenance of Chaoboridae and Chironomidae during the entire year.However, in our study, Chironomidae showed a significant negative correlation with water temperature, similarly to that observed by Davanso and Henry (2006) in a pond near the Mian Pond.On the other hand, a significant association between Chaoboridae organisms and high temperature was recorded in the present study, a condition also verified by Cleto-Filho and Arcifa ( 2006) who found highest densities of Chaoborus in hot periods, in the Monte Alegre Lake.
According to Roque et al. (2004) andBrito Junior et al. (2005), Chironomidae larvae are the most representative and abundant group of the benthic macroinvertebrates due to their high capacity to adapt to different environmental conditions which many other groups cannot.Therefore, anoxic conditions in March may have caused the absence of these organisms, since Chironomidae densities present significant positive correlation with bottom oxygen concentrations.With the increase in oxygen concentration in August, these aquatic insects were recorded.
Low pH values in March, significantly different from those observed in August, were probably related to the intense degradation of organic matter, since low water acidity is derived from ions released during decomposition.This fact seems to cause some disadvantaged to some taxa in the month of March, such as Chironomidae, that showed positive correlation with water pH.
Transparency was a secondary factor responsible for seasonal variation in density of these two benthic groups as it depends on the rain intensity.The rainfall peak in July, one month before sampling, was the factor responsible for re-suspension of fine material from the bottom to the water column through the continuous mixing of water.Thus, significantly higher water transparency in March, when compared to August, seems to have negatively affected the fauna.Again, Chironomidae was most affected, which was negatively correlated to the variable.Leech and Johnsen (2009) concluded that changes in transparency of freshwaters may cause alteration in species depth distribution and affect predator-prey behavior, then, increases in transparency may benefits visual predators.Possibly in this study, high transparency facilitates the visualization of Chironomidae by predators, such as fishes.
Regarding the granulometric composition of pond sediment, silt/clay was dominant in almost Macrophytes can increase the structural complexity of the environment, as the patches produces high amount of organic matter in sediment that lies just below them and in their surrounding area.Through this, some spatial differences may be produced in benthic composition since the patches are unequally distributed along the ponds.
Low values of dissolved oxygen in the pond bottom were found in both sampling months, especially in March (in all the sampling sites <3 mg.L -1 ).According to Hepp (2002), values <4 mg.L -1 are enough to cause the mortality of many invertebrates not adapted to hypoxic conditions near sediment.The variation in pond depth appears not to directly influence the observed alterations in oxygen concentration, since the Mian Pond is a shallow environment.Differences in oxygen among sites were insignificant and appear to be unimportant for spatial distribution of the organisms.Oxygen concentrations at the bottom were lower in the first month sampled.This fact can be explained by the high amount of allochthonous matter introduced in the pond due to the high level of water flux and discharge of the Paranapanema River in January and February 2009.Organic matter deposition at the pond bottom increased biological processes, especially decomposition, producing high oxygen consumption.Then oxygen was probably the determining factor for the lower values of benthic fauna density in March (except for Choboridae).
Chaoboridae, Oligochaeta, Nematoda and Chironomidae were predominant in benthic fauna.According to Pamplin et al. (2006) and Higuti and Takeda (2002), the Diptera order, including the Chaoboridae and Chironomidae families, and the Oligochaeta class, represent the most noticeable and relevant organisms of macroinvertebrate benthic assemblies.
Considering the two sampling periods, there were significant alterations in the densities of both groups, with the Chaoboridae being the most significant in total abundance in March and Chironomidae in August.These modifications in densities can be due to significant environmental variations among the months of the study, especially in the water temperature.This observation is evidence of the powerful influence of seasonality on the distribution of these organisms.Strixino and Trivinho-Strixino (1980) observed that aquatic ecosystems presenting low mean depths (approximately 3 m) are able to maintain a high water temperature, thus enabling a proliferation and all the sampled sites, followed by very fine and fine sand fractions.This characteristic is another factor that enhanced the great Chaoboridae development in Mian Pond, especially in March when fine fractions was higher, as densities presented positive correlations with silt and negative correlations with coarse sand fractions.
Considering the associations between benthic macroinvertebrate distribution in the aquatic environment and abiotic factors described previously, we concluded that reduced spatial variation of benthos occurred in the Mian Pond.Despite the fact that in 2009, no evident hydrologic variations were observed between rainy and dry seasons due to "atypical" rainfall in July, the seasonality effects on benthic fauna were more expressive than spatial ones, mainly because of this community response to some environmental factors, like water temperature, dissolved oxygen at deep zones, and characteristics of bottom sediment.Even though there was a wide variation in depth among the sampled sites, it was not able to change the structure of benthic community, reinforcing the idea that in environments with low structural complexity, physical and chemical factors of water and sediment are more homogeneous, having smaller effect on fauna.

Figure 2 .
Figure 2. Bathymetric profile of Mian Pond bottom, with the respective depths and sampling sites at the transects (the numbers 54, 64, and 68 correspond to pond widths at the site of the three transects).

Figure 4 .
Figure 4. Depth values at the nine sampling sites of Mian Pond in March and August 2009.

Figure 6 .Figure 5 .
Figure 6.Mean (n = 3) organic matter content in the sediment in sampled sites and months.

Figure 8 .
Figure 8. Mean (± standard error) values of environmental variables and Chironomidae and Chaoboridae abundance in March and August in the Mian Pond(**significance level: 0.1).

Table 1 .
Significant Pearson correlations (p < 0.03) between biological and water abiotic and sediment variables (Temp: temperature; Con/B: electrical conductivity at bottom; pH/B: pH at bottom; O 2 /B: oxygen at bottom; Tr: transparency; VCS: very coarse sand; CS: coarse sand; MS: mean sand; S+C: silt and clay).