Responses of the Molluscan Fauna to Environmental Variations in a Halodule Wrightii Ascherson Ecosystem from Northeastern Brazil

Responses of the molluscan fauna to environmental variations in a Halodule wrightii Ascherson ecosystem from Northeastern Brazil ABSTRACT This study observed the spatial and temporal distributions of molluscs in a Halodule wrightii meadow, verifying if they respond satisfactorily to seasonal changes in this seagrass ecosystem. Twenty-four species were identified. Chitons were rare, bivalves had greater number of species (11), followed by gastropods (9) which were also the most abundant class (73%). All classes were more abundant in the belowground. The most common species was Tricolia affinis, especially in aboveground. The occurrence of some species in both strata or out of the expected stratum may have been influenced by shallow layer of the sediment considered in this study, hydrodynamic, and low biomass of the studied meadow. According to univariate and multivariate analyses, despite of molluscan descriptors had been related to variables associated with rainfall, the seagrasses had an important role on the seasonal and vertical variations of the molluscan fauna. The biomass of the epiphyte Hypnea musciformis was correlated to temporal variations of the species from aboveground, indicating its secondary role for this community. The molluscs were sensible to environmental variations, and also reflected seasonal changes of the seagrass, showing that damages on these meadows reflect even at lower levels of the marine food web.


INTRODUCTION
Seagrass meadows are direct source of food for many marine organisms and they promote substrate for several epiphytic species, stability of the sediment and physicochemical variables, providing nursery, shelter from predators, and territory favorable to capture prey (Orth et al. 1984, Hall and Bell 1988, Phillips 1992, Marbà et al. 1996, Hemminga and Duarte 2000).
Among several communities associated with seagrasses, the molluscan fauna is one of the most abundant.The importance of molluscan grazer and scrapers of periphyton in these ecosystems has been reported for maintenance of the trophic web and control of epiphytes, favoring growth and productivity (Hootsmans and Vermaat 1985, Howard and Short 1986, Philippart 1995, Jernakoff and Nielsen 1997, Fong et al. 2000, Hemminga and Duarte 2000).
However, most of the studies which investigated relationships among molluscan fauna and seagrasses observed feeding preferences of grazer species and the effects of their activities on the leaves (Van Montfrans et al. 1982, Jensen 1983, http://dx.doi.org/10.1590/0001-37652013101212
Seagrasses are acutely responsive to environmental changes (Orth et al. 2006, Barros et al. 2013), inclusive climate global changes, which modify the distribution, productivity, and composition of seagrass communities (Short and Neckles 1999, Duarte et al. 2004, Orth et al. 2006).The large abundance in seagrass ecosystems and the sensitivity to environmental changes are essential requirements to use molluscs in environmental monitoring (Clarke and Ward 1994, García and Meneses 2000, Sánchez-Moyano et al. 2000).In short-term, studies on the effects caused by seasonal changes of seagrasses on specific communities may assist in diagnosis of local environmental changes.In long-term, these studies will even help in understanding of the global climate changes, since the associated communities reflect impacts along the trophic web.
The purpose of this study was to observe the spatial and temporal distributions of the molluscan fauna in a Halodule wrightii Ascherson meadow, verifying if they respond satisfactorily to seasonal changes in this seagrass ecosystem.

STUDy AREA
Goiabeiras Beach (03°41'31"S; 038°34'49"W) is bounded, on the west, by the mouth of Ceará River (Fig. 1), in which are found beach rocks covered by macroalgae and a small meadow of the seagrass Halodule wrightii Ascherson.The climate classification according to Köppen (1948) is Aw', in other words, rainy tropical climate with a long dry season.Morais (1980) recorded mild winds FIELD SAMPLING AND LABORATORy PROCEDURES Five random replicates were monthly sampled from April 2006 to July 2007 at low spring tides, using a 10-cm-diameter core, inserted at 10cm depth.Subsequently, the samples were sectioned in aboveground (shoots, leaves and associated fauna, exposed to hydrodynamic) and belowground (roots, rhizomes and associated fauna, inside the sediment).The aboveground samples were bagged, cut, and tagged.Thereafter, belowground was placed in another plastic bag and tagged.The samples were preserved in a 4% formalin solution and transported to the Zoobentos Laboratory of the Institute of Marine Sciences (Laboratório de Zoobentos do Instituto de Ciências do Mar, Universidade Federal do Ceará).
Sediment samples were also monthly collected in order to obtain granulometric analysis and OM content.Physical and chemical variables (i.e.water and air temperatures, pH, salinity, and dissolved oxygen) were also obtained with a multiparameter probe.Pluviometric precipitations, wind speed, waves, and period of the waves data were obtained from the Brazilian Institute of Space Research (Instituto Brasileiro de Pesquisas Espaciais -INPE) (2006/2007).Seagrasses (shoots and roots) and epiphytes were dried at 60°C in order to obtain biomasses (g dw.m -1 ).The molluscan fauna was preserved in 70% alcohol and identified under stereomicroscope with help of specific literature.The specimens were preserved at the Malacological Collection Prof. Henry Ramos Matthews of the Institute of Marine Sciences .

STATISTICAL ANALYSIS
The seasons were based on Euclidean distance of the abiotic variables, whose the groups of samples were submitted to variance analyses.Frequency of occurrence (F < 10% -Rare; 10% < F < 40% -Few Common; 40% < F < 70% -Common; F > 70% -More Common), density (ind.g-1), and descriptors of the molluscan fauna (number of species, diversity and evenness) were also obtained for both strata (below and above ground) and seasons (dry and rainy).Euclidean distance, number of species, diversity and evenness were obtained using Primer ® (Plymouth Routines in Multivariate Ecological Research), 6.1.6version.
Parametric variances (t-test) were performed in order to verify if biotic and abiotic variables were significantly different considering seasons.Relationships among biotic and abiotic variables were tested using univariate, non-parametric Spearman rank correlations, and multivariate Canonical Correlation Analysis (CCA).CCA observed the influence of the data set of environmental factors and macrophytes (seagrass and epiphyte) on the molluscan community variance.In order to observe responses of the molluscan community to macrophytes influences, CCA were performed twice, one without macrophytes biomasses and another one with macrophytes biomasses as environmental factors.Variables strongly correlated were excluded in order to both avoid multicollinearity and decrease inflation factors.Data were natural log-transformed and the down weighting rare species option was selected in order to avoid the influence of rare species leading to biased results.Monte Carlo test with 499 permutations, under reduced mode, was used to test the significance (p < 0.05) of the environmental variables influencing on the ordination axes.Spearman rank and t-tests were obtained using Statistica ® 7.0 and CCA was performed by CANOCO for Windows, 4.5 version.

ENVIRONMENTAL AND MACROPHyTES VARIATIONS
Euclidean distance (Fig. 2) determined two groups of samples associated mainly with presence and KCRISHNA V.In the first half of the year, both percentages of fine grains (silt/clay) and organic matter (OM) content varied according to pluviometric precipitations, while wind speed decreased.In the second half of the year, besides decrease in pluviometric precipitations, there was a progressive increase in wind speed, which peaked in September 2006 and October 2006.Waves and period of the waves varied (directly and inversely proportional, respectively) to wind speed.Pluviometric precipitations, percentages of fine grains, organic matter content, wind speed and waves were significantly different considering seasons (Table I).
These environmental changes also influenced H. wrightii.This seagrass was more developed during the rainy season, when the biomasses were higher.The average values of total biomass, in dry and rainy seasons, were 454 g dw.m -2 and 656 g.dw.m -2 , respectively.In contrast, the biomass of the one single epiphyte species, Hypnea musciformis Lamourox, increased in the dry season (Table II).However, no statistically significant differences were found regarding the seasons (Table II).
COMMUNITy DESCRIPTORS AND SPATIOTEMPORAL DISTRIBUTION OF THE MOLLUSCAN FAUNA Two hundred and thirty molluscs were captured, being 141 in the belowground and 89 in the aboveground.Gastropoda was the most abundant class, corresponding to 73% of the molluscan fauna, and Polyplacophora was the less abundant, corresponding to 4%.Twenty-four species were identified, and bivalves had greater number of species (11 species), followed by gastropods (9 species).All classes were more abundant in the belowground (Table III

). ENVIRONMENTAL INFLUENCES ON MOLLUSCS IN A H. wrightii bed
The species Tricolia affinis C. B. Adams (1850) was the most abundant, corresponding to 25.7% of the molluscan fauna from the belowground, and 60% of the molluscs from the aboveground.According to the frequency of occurrence, most of the molluscan fauna was considered few common or rare (Table III), especially in aboveground.In the belowground, T. affinis was common (F = 53.3%)as well as the bivalve Ctena orbiculata Montagu (1808) (F = 46.6%).In the aboveground, T. affinis was very common (F = 80%) and the other species were few common or rare.This gastropod was also the densest species in both strata (Table III), and still the dominant species in belowground (41.2%) (Fig 3a) and aboveground (61.2%) (Fig 3b).
Shannon's diversity and Pielou's evenness were greater in the rainy season.However, according to t-test, diversity was not significantly different between seasons, both in belowground (p = 0.180; df = 13) and aboveground (p = 0.148; df = 13).The indexes of evenness were significantly different in belowground (p = 0.032; df = 13), but was not different in aboveground (p = 0.173; df = 13).rainfall (pluviometric precipitations, organic matter content, and percentage of fine grain), whereas there were negative correlations among these descriptors and variables directly related to the dry season (winds speed, waves) (Table IV).Those descriptors were also positively correlated to H. wrightii biomasses, especially in aboveground, and the molluscan density was weakly correlated to H. musciformis biomass (Table V).The gastropod Tricolia affinis was the only species which was significantly correlated to some environmental variables (Table V).

Class
In the CCA without macrophytes biomasses (Fig. 4a), the Monte Carlo test does not indicated significant correlation of environmental variables to variance of the molluscs abundance, both considering the first canonical axis (eigenvalue = 0.357; F = 1.514; p = 0. 694) and all canonical axes (Trace = 0.988; F = 0.975; p = 0.516).In this case, abiotic variables explained 20.4% of all canonical axes; species-environment correlation explained 64.2% of the two first axes variances, and 36.1% of the axis 1 variance.The environmental variables more strongly correlated to the axis 1 was fines percentage (-0.536),followed by sand percentage (-0.506)and gravel percentage (0.483).Organic matter content (-0.689) was the variable more strongly correlated to the axis 2. The samples dispersion showed similarities among samples from belowground and their respective samples from aboveground.
Considering the CCA with macrophytes biomasses as environmental factors (Fig. 4b), the Monte Carlo test indicated significant correlation between environmental variables and the variance of the samples, both for the first canonical axis (eigenvalue = 0.999; F = 4.422; p = 0.002) and all canonical axes (Trace = 2.011; F = 1.726; p = 0.002), although some variables have increased the inflation factor, probably because of correlations among them.Nevertheless, these factors were within acceptable limit (Table VI).Abiotic variables explained 41.5% of all canonical axes; species-environment correlation explained 65% of the two first axes variances, and 49.7% of the variance of the axis 1.The environmental variable more strongly correlated to the axis 1 was sand percentage (-0.999),followed by Halodule wrightii biomass (-0.868) and Hypnea musciformis biomass (0.497).Organic matter content (-0.689) was, once more, the variable more strongly correlated to the axis 2. The inclusion of the macrophytes biomasses positioned belowground samples on the negative fraction of the Axis 1 and aboveground samples on the positive fraction of this Axis.

TABLE IV (continuation) TABLE V Spearman rank correlations among descriptors of the molluscan community with Halodule wrightii and Hypnea musciformis biomasses in Goiabeiras
Beach, Fortaleza-Ceará, Northeastern Brazil.

DISCUSSION
The molluscan fauna of the studied Halodule wrightii ecosystem seemed use these seagrass as its main substrate, although many groups also widely found in these ecosystems as nematodes, polychaetes, crustaceans and fishes, are not directly related to seagrasses (cf.Corbisier 1994, Garcia et al. 1996, Garcia and Vieira 1997, da Rocha et al. 2006).However, several studies have indicated that even little seagrasses may provide greater density and diversity of macrofauna than non-vegetated Despite the environmental influences recorded, this studied seagrass meadow had a significant role on the spatial and temporal distributions of the molluscan fauna.
The most representative molluscan species of the studied meadow was Tricolia affinis.The literature reports this microgastropod in beach rocks, coral reefs and macrophytes, occurring from intertidal zone to 50m depth, with higher dry mass between 1-2m in subtidal zone, and on the base of algae, where greater accumulation of sediment occurs.This species is herbivorous, grazer-scrapers of periphyton, and the epiphytic diatoms are its main supplementary feeding (Mountouchet 1979, Rios 1994, Alves and Araújo 1999, Széchy and Paula 2000, Fernandes et al. 2006).Because of its significant presence in many phytal ecosystems, T. affinis has attracted attention of marine ecologists.Its geographic distribution, anatomy, physiology, behavior (cf.Marcus and Marcus 1960), ecology, population dynamics, and spatial distribution on macroalgae (cf.Braga 1983, Széchy and Paula 2000, Pereira et al. 2010) have been studied.Relationships of this species with other seagrasses and macrophyte species also have already been recorded.Bandel and Wedler (1987) observed the common presence of T. affinis associated with Syrigodium filiforme Kützing and Thalassia testudinum Banks and Sol.ex K.D. Koenig, in Caribbean Sea (Colombia).Pereira et al. ( 2010) did not observe a clear pattern of seasonal variation of T. affinis in the macroalgae Sargassum spp.At the studied H. wrightii meadow, correlations and variances of descriptors indicated explosion of T. affinis during rainfall, when doubled its density.Furthermore, correlations between T. affinis abundance and H. wrightii biomass also showed that this macrophyte were the main substrate for this gastropod, even considering secondary contribution of H. musciformis as substrate for all molluscan fauna.This epiphyte allowed the maintenance of the molluscs when the seagrass biomass decreased, during the dry season.The complexity of stems provides habitat availability, protection against predators and hydrodynamic, and retains food particles, increasing abundance and density of fauna (Hall and Bell 1988, Leite and Turra 2003, Chemello and Milazzo 2002, da Rocha et al. 2006, Pereira et al. 2010).The algae also provide an excellent microhabitat for molluscs, rather than as a source of food (ávilla 2003).The architecture of H. musciformis operates as an environmental extension for establishment of species and retains particles necessary for feeding (Hall and Bell 1988, Leite and Turra 2003, da Rocha et al. 2006).The algae architecture may be more important for abundance of fauna than its biomass (Pereira et al. 2010).Studying effects of the architecture of five algae species on abundance, richness and diversity of molluscan fauna, Chemello and Milazzo (2002) observed that greater structural complexity, greater abundance and diversity species.Similarly to present study, Leite and Turra (2003) noted positive correlations between abundance of fauna and biomass of Sargassum sp., and also between abundance and Sargassum + Hypnea complex.This may explain the correlations of epiphyte biomass both to the density of molluscs and variances of the samples.
Some studies observed synchronism between biomass of seagrasses and density of molluscan communities (Mukai 1976, Alves andAraújo 1999, Barros et al. in press), and between the life cycle of molluscs and seasonal fluctuations of macrophytes (Toyohara and Nakaoka 1999).Spatiotemporally, relationships among molluscs and seagrasses were observed through variations according to the seasonality (Alves and Araújo 1999, Barros et al. in press), above and below strata (Barros et al. in press) and depth (Creed and Kinupp 2011) of H. wrightii.Mainly in the aboveground, positive correlations between community descriptors and wind speed may be related to windy influences on hydrodynamic, which may causes horizontal transport of organisms (i.e. algae and fauna) from adjacent areas for inside the meadow.
Belowground, however, seemed more stable than aboveground as seem in other studies on seagrasses (Gambi et al. 1995) and associated fauna (Williams and Heck 2001), in these ecosystems.Meanwhile, the shallow part of the sediment considered in this study (10cm), due to the presence of reef rocks under the meadow, may be a part of the sediment more vulnerable to hydrodynamic.Barros andRocha-Barreira (2009/2010) observed vertical and temporal segregation of benthic macrofauna, but these authors suggested, in addition to seagrass effects, influences of environmental variables on the ecosystem.In fact, the biomass of H. wrightii was not the exclusive factor to cause a vertical segregation of the species.Barros et al. (in press) observed that chiton population, commonly found in rocky environments, were found in H. wrightii established on reef rocks.As well as all molluscan fauna studied here, these authors verified that chiton species were more abundant during the rainy season, and recorded Ischnochiton niveus Ferreira, 1987 for the first time in Brazil, in the root system of these meadow.
Although the molluscan fauna had been qualitatively similar in both strata, the analyses showed temporal variation due to environmental influences, through of weak, direct Spearman correlations with variables associated with the rainfall, probably because of their positive influence on H. wrightii.Also, multivariate analyses showed vertical segregation of strata (when considered macrophytes influences).
Significant influences of environmental variables on the samples variance and vertical segregation were found only when the macrophytes biomasses were considered as environmental variables.This indicates that, although some species have occurred on both strata and the environment has significant effects on the molluscan community, the seagrasses caused different effects in each stratum.As observed by Nakaoka et al. (2001), studying epifaunal communities in meadows of Zostera marina L. and Zostera caulescens M., variations in these communities is not determined by a single or some strong external factors, but by complex interactions of multiple factors operating differently for each component species.
Thus, although is considered influences both of the environment and macrophytes on molluscs, seagrasses were the main substrate for this community and were also determinant for the spatial and temporal distributions of the species.This show the importance of seagrasses for this community structure, even is evidenced the secondary role for the temporal distribution of aboveground species played by epiphytes.The occurrence of some species in both strata, and species out of their original stratum, may have been influenced by the shallow part of the sediment considered in this study, low values of H. wrightii biomass, and strong influences of external factors on the meadow.Thus, besides environmental influences on molluscan fauna, these species reflected also seasonal changes of the seagrass, showing that damages on these meadows reflect even at lower levels of the marine food web.
S. BARROS and CRISTINA A. ROCHA-BARREIRA absence of pluviometric precipitations.Samples from April 2006 to June 2006, and from February 2007 to May 2007 were associated with rainfall.The samples associated with the dry season (from July 2006 to January 2007, and July 2007) were most homogeneous.

Figure 2 -
Figure 2 -Euclidean Distance dendrogram of the samples, based on environmental variables obtained throughout this study in Goiabeiras Beach, Fortaleza-Ceará-Northeastern Brazil.

TABLE VI Factors inflation in Canonical Correlation Analysis without and with the macrophytes biomasses as environmental variables, acting on the molluscan community from Goiabeiras Beach, Fortaleza-Ceará, Northeastern Brazil.
Corbisier 1994, Lee et al. 2001, Casares and Creed 2008, Rosa and Bemvenuti 2007)nuti 2007).