Variations of the macrobenthic community in a seagrass transplanted area of the Lagoon of Venice *

The importance of seagrass within coastal marine and lagoon ecosystems has been widely described (Den Hartog, 1977; Larkum et al., 1989). When compared with neighbouring areas, the meadows reveal higher animal and algal biomasses and abundances, with greater specific richness (Homziak et al., 1982; Edgar et al., 1994; Boström and Bonsdorff, 1997). The main factors contributing to this improvement in biodiversity may be summarised as increased: i) availability of microhabitat, ii) protection from predators, iii) trophic resources, iv) sediment settling, v) hydrodynamic force reduction (Lewis, 1984). Meadows do strongly affect local water flows by significantly reducing current speed (Harlin et al., 1982; Fonseca et al., 1983) and by producing a typical relative increase in the sediment fine fraction (Orth, 1977; Peterson et al., 1984). Moreover, these habitats often act as nursery areas for many fish species, since they offer protection and shelter from predators (Blunden and Kennedy, 1982; Connolly, 1994). In the last decade, the whole ecosystem of the Lagoon of Venice has been subjected to increasing anthropogenic stress, mainly due to ‘free-access’ exploitation of the short-necked clam Tapes philippinarum (Pranovi and Giovanardi, 1994; Giovanardi and Pranovi, 1999), which has produced a reducSCI. MAR., 64 (3): 303-310 SCIENTIA MARINA 2000


INTRODUCTION
The importance of seagrass within coastal marine and lagoon ecosystems has been widely described (Den Hartog, 1977;Larkum et al., 1989).When compared with neighbouring areas, the meadows reveal higher animal and algal biomasses and abundances, with greater specific richness (Homziak et al., 1982;Edgar et al., 1994;Boström and Bonsdorff, 1997).The main factors contributing to this improvement in biodiversity may be summarised as increased: i) availability of microhabitat, ii) protection from predators, iii) trophic resources, iv) sediment settling, v) hydro-dynamic force reduction (Lewis, 1984).Meadows do strongly affect local water flows by significantly reducing current speed (Harlin et al., 1982;Fonseca et al., 1983) and by producing a typical relative increase in the sediment fine fraction (Orth, 1977;Peterson et al., 1984).Moreover, these habitats often act as nursery areas for many fish species, since they offer protection and shelter from predators (Blunden and Kennedy, 1982;Connolly, 1994).
In the last decade, the whole ecosystem of the Lagoon of Venice has been subjected to increasing anthropogenic stress, mainly due to 'free-access' exploitation of the short-necked clam Tapes philippinarum (Pranovi and Giovanardi, 1994;Giovanardi and Pranovi, 1999), which has produced a reduc-tion in seagrass.In an attempt to test procedures which may prove to be useful tools in future 'restoration' projects aiming at counteracting these stress effects, an experimental transplant of Cymodocea nodosa (Ucria) was carried out in the southern basin of the Lagoon.The purpose of the study was to assess the feasibility of transplanting C. nodosa by means of the 'turf technique' and to describe the changes and developments occurring in phyto-and zoobenthos in the artificial meadow.

MATERIALS AND METHODS
The study site is located in the southern basin of the Lagoon of Venice (Fig. 1), on a 2-metre deep silty-sandy bottom (Barillari, 1981), in an area where the hydrodynamic pattern has recently been modified by restoration of neighbouring channels.The experimental area, located close to a natural Cymodocea nodosa meadow, typically contained pleustophytic algae, which have a clear seasonal cycle, with low biomass in autumn and winter and very abundant biomass in spring and summer (Curiel et al., 1995).
The C. nodosa transplant, made during spring 1995, involved an area of roughly 30 m 2 and was performed using the 'turf technique' (Phillips, 1974), 56 pieces of turf (∅=23 cm, height=30 cm) being inserted into the bottom.

Data collection
In March and September 1996, samples of phytoand zoobenthos were collected inside the transplanted area (treated plot) and all around it outside (control area).
The shoot density of the transplanted meadow was measured in situ using a 30 x 30 cm 2 square in 10 replicates.Ten shoots of C. nodosa were collected, to evaluate the number of leaves per shoot and the length of leaves.The photosynthetically-active surface (LAI -leaf area index) was determined by multiplying the mean surface area of one-shoot leaves (only one face) by meadow shoot density.
Seagrass biomass (g d.w.m -2 ) was measured from samples collected with a manual corer (∅ 23 cm) in 5 replicates, by separating leaves, roots and rhizomes, after removing sediment.For every core, 10 shoots were also collected in order to study epiphytic macroalgae and to estimate their biomass and coverage percentage.
Macroalgal biomass (g d.w.m -2 ) and coverage percentage were measured by collecting material inside a 1m x 1m frame in 5 replicates per plot.The diversity index based on Shannon's formula modified by Boudouresque (1970) was calculated.
Macrobenthos samples were collected with a manual corer (∅ 23 cm, 5 replicates per plot) and separated from sediment through a 1-mm-mesh sieve.Organisms were classified to their lowest possible taxonomic level, and wet biomass was determined for each taxon.
The micro-mesobenthos of the surface sediment layer was sampled with an air lift sampler with 3 replicates per plot.

Data analysis
Species diversity in treated and control areas was estimated by the Shannon Weaver (H') and Margalef (d) indices.
Changes in the number of species, total number of individuals, diversity indices and total wet biomass were analysed by one-way ANOVA.
The Abundance Biomass Comparison -ABCmethod (Warwick, 1986) was applied in order to evaluate the level of disturbance of the macrobenthic community in the treated and control areas.Differences among macrobenthic assemblages were assessed using the multidimensional scaling technique, based on the Bray-Curtis similarity matrix (Clarke and Warwick, 1994).Differences in assemblage composition between treatments and data were assessed by ANOSIM, a randomization/permutation procedure which tests for differences in the ranked similarity matrix (Clarke and Green, 1988).

Macroalgae
The algal component of the two plots showed many differences in taxon composition and specific richness (Table 1).The total number of recorded species in the treated plot was 18, but only 6 in the control area.
The temporal pattern of specific richness also showed an opposite trend in the two plots: in the treated plot it increased from 10 taxa in March to 16 in September, whereas in the control area it decreased from 6 to 4 (Table 1).This also influenced the Shannon Weaver index, which was relatively stable in the control area (0.95 in March and 1.09 in September) but increased from 1.83 (March) to 2.33 (September) in the treated plot.
In the control area, the highest values of biomass and covering percentage were recorded in March (Table 2), when macroalgae, mainly Ulva rigida, Gracilaria verrucosa and Chaetomorpha linum, formed a homogeneous and contiguous mat at least 10 cm thick on the bottom.In September, lower values were recorded as the consequence of a remark-MACROBENTHIC COMMUNITY IN A SEAGRASS TRANSPLANTED AREA 305 able reduction in U. rigida biomass, due to summer degradation processes.
In the treated plot, the vegetal compartment was dominated by C. nodosa, and the total macroalgal biomass and covering percentage reached negligible values (Tables 2-3).The highest values of density, biomass and LAI of C. nodosa were recorded in September, at the end of intense summer growth; instead, the leaf component proved to be very scarce in March (Table 3) because of very low winter levels of shoot density, leaf size and leaf number per shoot, typical of environments at intermediate latitudes (Rismondo et al., 1997).
Comparisons between the transplant area and a natural C. nodosa bed (Rismondo unpubl. data) showed that, although the growth parameters recorded in the transplanted meadow fell within normal range, they did not reach maximum values, mainly in shoot density and leaf biomass (Table 3).

Micro-meiobenthos
Qualitative analysis of micro-meiobenthic samples (Table 4) revealed the presence of vegetal detritus from the seagrasses in the treated plot and mainly Ulva in the control area.In both areas the highest amounts were detected in autumn.Moreover, benthic filamentous microalgae and typically Ciliate interstitial forms (Remanella, Geleia) tended to prevail in the treated plot, with abundant bacterial flocculent material in the control area, in which a complete lack of interstitial forms was noted.In September, all meiobenthic groups (Ciliates, Nematoda, micro-Polychaetes, Ostracoda) were two or three times more abundant in the control area than in the treated plot.

Macrobenthos
Details of the 35 macrobenthic taxa collected during the surveys, diversity measures (total number of individuals, Shannon and Margalef indices) and the total wet biomass are listed in Table 5.In the treated plot, all considered parameters (except the Margalef index) increased from March to September, whereas they showed a decrease in the control area.
The one-way ANOVA carried out using the diversity measures recorded in the treated and control areas (T March, T September, C March, C September) showed significant differences in total number of species, total number of individuals and total wet bio-mass (Table 6).The LSD post hoc test highlighted the fact that these differences were always due to the samples collected in the treated plot in September.The trophic structure of the macrobenthic community, based on abundance data (Fig. 2a), showed small changes in the control area, whereas the structure changed in autumn in the treated plot, with a more homogeneous distribution of the trophic groups: deposit feeders, almost reaching 50% in spring, decreased to 30%, while an increase was observed in the groups of herbivores (from 3% to 17%) and 'others' (all species whose trophic guild could not be identified).
Based on biomass data (Fig. 2b), both plots showed temporal variations in trophic structure: in September filter feeders were dominant (35% in control area and 87% in treated plot).The overdominance in the latter was mainly due to the phylum Porifera, which found a suitable substratum for growth on Cymodocea leaves.
The ABC plots of the control area (Fig. 3) showed a slightly stressed community, the two curves intersecting in both sampling seasons (spring W = 0.221, autumn W = 0.136).In the treated plot the macrobenthic community appeared to be quite stressed in spring (W=0.07)but was undisturbed (W=0.440) in September, a situation typical of 'mature' and 'stable' environments, with a large dominance of k-strategy species prevailing.
Multidimensional scaling of samples is shown in Figure 4.The T Sep sample group is well separated from the others and is located not far from a reference natural bed of C. nodosa (Rismondo, unpubl. data).The other three sample groups (T Mar, C Mar, C Sep) are partially mixed and more widespread.The distribution observed in the F igure 4 is confirmed by the ANOSIM test (p<0.001).

DISCUSSION
Our data show that seagrass transplant produces significant changes in the whole benthic community: -the high values of macroalgal biomass and coverage recorded in the control area were not observed in the treated plot, where the specific richness of algal taxa trebled, mainly due to epiphytic species growing on Cymodocea leaves; -as a preliminary qualitative observation, microand meiobenthos in the treated plot reflected a condition with a lower organic sediment load than in the control area; this is especially clear when considering the evolution of Cilate microfauna and the subsurface interstitial colonization pattern; -comparisons between treated and control areas showed differences in specific richness, diversity indices and trophic structure of the macrobenthic community, differences which became more pronounced in September.This pattern in the treated plot may be due to the seasonality of C. nodosa leaf growth and the effects of meadow density on macrobenthic organism distribution.
The structure of the macrobenthic community in the treated plot changed from early spring to autumn, with a statistically significant increase in diversity and better distribution of organisms among the trophic guilds.This was probably due to recovery of the canopy which, after winter reduction, brought the community to a 'well-structured' status, as also highlighted by the ABC results.
Instead, stress signals recorded in the macrobenthic community in the control area (see ABC results) were probably due to the detrimental effects produced on soft-substrate fauna by contiguous macroalgal mats (mainly Ulva rigida and Chaetomorpha linum), as also demonstrated by Raffaelli et al. (1989) and Norkko and Bonsdorff (1996).
The complete development of the canopy in September also caused increased homogeneity of the spatial distribution of organisms (Irlandi et al., 1995), which were more widespread and patchily distributed in March (treated plot) and in samples collected in both seasons in the control area, as highlighted by the MDS ordination.
Comparisons between the transplanted meadow and a nearby natural C. nodosa bed showed that, only two vegetative seasons after the experimental transplant: -seagrass performance was similar in the two meadows.The growth rate, leaf number per shoot and shoot density of the transplanted meadow fell within the range of values recorded in the natural bed; -in autumn, the structure of the macrobenthic community approached that of a natural bed.As demonstrated by Webster et al. (1998), the main factors best explaining macrobenthic community structure are leaf number per shoot and shoot density.Thus, just two vegetative seasons after transplant seem to be sufficient to build up structural complexity similar to that of a natural meadow.Another important structuring factor is the heterogeneity produced by the root-rhizome matrix (Webster et al., 1998), which may be very important in C. nodosa meadows subject to seasonal canopy reduction; our data confirmed that the belowground biomass (roots, rhizomes) of the treated plot approached that of the natural bed.

CONCLUSIONS
The experiment confirmed the feasibility and interest in Cymodocea nodosa transplants using the 'turf technique' in the Lagoon of Venice: the performance of the transplanted meadow quickly reached that of natural beds.Moreover, the possibility of creating or re-establishing a functioning seagrass meadow (sensu Kikuchi and Pérèe, 1977) could be evaluated.
Our data also revealed the key-species role played by seagrass in structuring the marine coastal ecosystem (Mazzella et al., 1992).This role was clearly magnified in a 'sensitive area' such as the Lagoon of Venice, subjected to growing anthropic pressure and negative impact.

TABLE 6 .
-One-way ANOVA results on macrobenthos data (number of species, total number of individuals, Margalef and Shannon indices, total wet biomass).