Marine Environmental Research 104 (2015) 10e19
Contents lists available at ScienceDirect
Marine Environmental Research journal homepage: www.elsevier.com/locate/marenvrev
Benthic assemblages on artificial reefs in the northwestern Adriatic Sea: Does structure type and age matter? Massimo Ponti a, *, Federica Fava a, b, Rossella Angela Perlini a, b, Otello Giovanardi c, Marco Abbiati a, d a
Dipartimento di Scienze Biologiche, Geologiche ed Ambientali (BiGeA) & Centro Interdipartimentale di Ricerca per le Scienze Ambientali (CIRSA), University of Bologna, UO CoNISMa, Via S. Alberto 163, I-48123 Ravenna, Italy in Romagna, Via Baccarini 27, I-48121 Ravenna, Italy Fondazione Flaminia e Per l'Universita c Istituto Superiore per la Protezione e la Ricerca Ambientale (ISPRA), Loc. Brondolo, I-30015 Chioggia, Italy d ISMAR, Consiglio Nazionale delle Ricerche e Istituto di Scienze Marine, Bologna, Italy b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 11 October 2014 Received in revised form 23 December 2014 Accepted 24 December 2014 Available online 25 December 2014
The use of artificial reefs is on the rise worldwide. While their fish aggregating effects are well known, the epibenthic assemblages have been poorly investigated. Two types of artificial reefs (pyramids of concrete slabs and bundles of concrete tubes) have been deployed out of the Po River Delta in 2006 and 2010. The epibenthic assemblages were investigated in 2009 and 2012. Benthic assemblages on both structure typologies were dominated by species tolerating high sedimentation rates. Dissimilarities were found among assemblages with different ages, and, in less extend, between reef typologies. Colonisation by Mytilus galloprovincialis and other major space occupiers did not follow a clear succession pattern and was not affected by reef typology. Species colonisation was likely driven by variability in environmental conditions and recruitment processes rather than by reef typology. This study suggests that environmental features of the deployment sites should be carefully considered in planning and designing artificial reefs, especially in eutrophic and turbid coastal waters, exposed to high river loads. © 2014 Elsevier Ltd. All rights reserved.
Keywords: Artificial habitats Benthic assemblages Coastal structures Fouling organisms Bio-construction Temporal variability Exotic species
1. Introduction Artificial reefs (ARs) are manmade structures deployed on sea bottoms with the primary purpose of protecting coastal habitats and increasing biotic resources by aggregating marine species and preventing trawling (Baine, 2001). ARs support sessile filter feeders, providing nourishment and refuges for motile species, and attracting bentho-nectonic fishes (Bohnsack and Sutherland, 1985; Baine, 2001). Since the mid-1800s, ARs were deployed in many regions of the world, including tropical and temperate areas, starting with the United States of America and Japan (Bohnsack and Sutherland, 1985). The materials used in their construction include natural rocks, concrete blocks and several discarded supplies, like tires, pipes, shells, barges, bundled solid waste, coal ash, vehicles, etc. (Feary et al., 2011). The first ARs along European coasts were
* Corresponding author. Laboratori di Scienze Ambientali, Via S. Alberto 163, I48123 Ravenna, Italy. Skype: massimo.ponti. E-mail address: [email protected] (M. Ponti). http://dx.doi.org/10.1016/j.marenvres.2014.12.004 0141-1136/© 2014 Elsevier Ltd. All rights reserved.
installed in the 1960s, most structures have been located in the Mediterranean Sea. Since this time more than 70 AR complexes, made by different materials, have been deployed along Italian coasts (Fabi et al., 2011). Traditionally, in the oligotrophic waters of the western Mediterranean Sea, the goals of ARs were to protect Posidonia oceanica meadows from illegal trawling, to increase habitat complexity and promote higher species diversity (Relini et al., 1994; Riggio et al., 2000; Gonzalez-Correa et al., 2005). Conversely, in the eutrophic waters of the central and northern Adriatic Sea, the main purpose was to increase fishery yields (Bombace et al., 1994; Ardizzone et al., 1996; Bombace et al., 1997). Regardless of the potential benefits of ARs, their increasing frequency worldwide has given rise to concerns regarding their possible negative impacts, especially the dumping of waste and the use of unsuitable materials. In response to these threats as well as international conventions addressing the issue, some regulations, guidelines and protocols have been drawn up (e.g. London Convention and Protocol/UNEP, 2009; for an overview see Fabi et al., 2011). As a result, most recently deployed subtidal artificial habitats have been designed for specific purposes. Concrete is the
M. Ponti et al. / Marine Environmental Research 104 (2015) 10e19
most common material, because it is cheap, versatile, allowing the realization of structures with different shapes and sizes, and may ensure long life, being resistant to the chemical and physical marine actions (Fabi et al., 2011). Physical properties of the concrete vary according to the reinforcements and additives included in the cement mixture. Several natural and synthetic admixtures are used as hardening accelerators and retarders, corrosion inhibitors, etc. Besides improving concrete properties, the choice of additives is dictated by economic and environmental considerations, including the reduction of greenhouse gas emissions and recycling of wastes. Unfortunately, manufacturers often do not make available the compositions and properties of the concrete products or the possible interactions with living organisms (e.g. patents: EP0134855 B1, EP1193348 B1, WO2014125493 A1). The choice of the type of artificial structures and their placement should always take into account the expected, as well as possible undesired effects on coastal habitats. While fish aggregating effects of ARs are well known and the effectiveness of different structure typologies in this respect are well documented (Santos et al., 1997), much less attention has been dedicated to the benthic assemblages colonising the structures, despite the impact they may have on costal habitats, and the possible implications on ecosystem functioning, e.g. changes in species composition, species interactions and food webs (Ambrose and Anderson, 1990; Bertasi et al., 2007; Gallaway et al., 2009), alteration of population connectivity and genetic diversity (Cowen and Sponaugle, 2009; Fauvelot et al., 2009), facilitation of the spread of non-indigenous species (Airoldi et al., 2005; Bulleri and Airoldi, 2005; Glasby et al., 2007). Vagile and sessile species colonise ARs according to complex ecological processes affected by seasonal larval supply, water circulation, turbidity and nutrients, depths, orientation and physicalechemical features of the substrata (Anderson and Underwood, 1994; Relini et al., 1994; Riggio et al., 2000; Turner and Todd, 1993). Moreover, the interaction of abiotic and biotic factors may operate at different temporal and spatial scales (Glasby, 1998; Rodriguez et al., 1993). Water quality (e.g.: oligotrophic vs. eutrophic, clear vs. turbid) is considered a relevant factor in structuring benthic assemblages, as resulted by comparing ARs deployed in different locations at similar depths (Maughan, 2001; Relini et al., 1994; Riggio et al., 2000). Whereas within the same site and/or in very similar environmental conditions, materials (Anderson and Underwood, 1994; Glasby, 2000; Reyes and Yap, 2001), shapes (Bourget et al., 1994), orientation (Baynes, 1999; Genzano et al., 2011; Glasby and Connell, 2001; Ponti et al., 2002), shading and proximity to the seafloor (Glasby, 1999) are the main factors affecting the structure of benthic assemblages. Since 2001, thanks to regional and European funds, experimental concrete AR complexes have been deployed on sandy and
11
muddy bottoms along the western Adriatic coasts (Spagnolo et al., 2014). The aim of the present study was to analyse variability of macrobenthic assemblages in relation to the typology and age of ARs deployed off the Po River Delta (northern Adriatic Sea). Benthic assemblages on two types of artificial reefs, pyramids of slabs (Tecnoreef®, hereafter TR) and bundles of tubes (hereafter BT), differing in shape, concrete chemical composition, and age, were compared. 2. Materials and methods 2.1. Artificial reefs typologies and study site Two types of ARs have been investigated (Fig. 1): pyramids made of concrete slabs (TR e declared “sea-friendly” by the producer, Tecnoreef®, manufactured using only natural components, without synthetic additives; pH ~9; 1.8 and 2.4 m height); and bundles of common concrete tubes, BT, assembled in cubes laid on a concrete slab and retained by an iron cage (pH ~12; 1.8 m height). Despite similar size and surface rugosity, the two reef typologies differed in shape, surface inclination and material (e.g. type of concrete and pH). The ARs were deployed in 2 times: November 2006 (hereafter AR1) and March 2010 (hereafter AR2), 2 nautical miles offshore of the Po River Delta (northwestern Adriatic Sea, 44 540 N 12 33’ E), at 13e14 m depth and close to a longline mussel farm (Fig. 2). In AR1, TR pyramids and BT structures were arranged in two adjacent areas (100 200 m and 100 120 m) separated by 50 m. AR2 has a nucleus (100 200 m) of TR pyramids surrounded by BT structures deployed along the perimeter. TR pyramids deployed in AR2 differed from AR1 only by the smaller holes, strengthening the structures, which in AR1 was already damaged after two years. ARs have been deployed on a muddy bottom (silt > 75%) in an area affected by high freshwater and sediment inputs from the Po River, which has a mean flow of 1500 m3 s1 (period 1918e2006), with higher values in spring and autumn (Fig. 3). The combination of the thermohaline circulation and tide (up to 1 m) often results in strong currents. Frequently, water turbidity (mean Secchi disk 1.5 m) reduces penetration of solar radiation, while the superficial halocline and seasonal thermocline cause sharp stratification of the water column (Table 1). Moreover, effluents from the Po River are rich in nutrients, favouring growth of plankton (Aubry et al., 2012). 2.2. Sampling design and laboratory analyses With the aim to test for differences in benthic communities related to reef typology (TR vs. BT) and age (AR1 vs. AR2), macrobenthic assemblages were investigated in June 2009 on AR1 (~2.6 yrs after deployment; Fig. 3) and June 2012, both on AR1 and
Fig. 1. Example of ARs structures deployed: pyramids of slabs (2 floors, 1.8 m height), assembled with slabs of ‘sea-friendly’ concrete (Tecnoreef®), on the left and bundles of tubes, made by common concrete, laid on a concrete slab and retained by an iron cage (1.8 m height, on the right).
12
M. Ponti et al. / Marine Environmental Research 104 (2015) 10e19
Fig. 2. Geographic map of the study area and arrangement of the AR lots.
Fig. 3. Raw data (grey area) and cumulative sum series (bold line) of mean monthly flow of the Po River from 2006 to 2012 (courtesy from Hydro-meteorological Service of the EmiliaeRomagna Regional Environmental Agency). The reference value used to calculate the cumulative sum series is the long-term mean flow of the river (1500 m3 s1, period 1918e2006). Artificial reef deployments and sampling dates are reported by symbols.
AR2 (i.e. ~5.6 and ~2.3 yrs after deployment, respectively; Fig. 3). Four structures per typology were randomly chosen at each sampling date for both AR1 and AR2. Assemblages were investigated by means of scraping and photographic sampling along sub-vertical external sides of the artificial structures. Each scraped sample had an area of 0.16 m2, a sampling size often adopted in studies on
benthic assemblages in the Adriatic Sea (e.g. Spagnolo et al., 2014). Due to the surface discontinuities and poor visibility, samples were obtained by pooling four adjacent quadrates (20 20 cm each) scraped from the concrete surface by hammer and spatula or trowel, using a custom aluminium frame equipped with a holder for satin cloth bags (bright yellow, highly visible and permeable; Video 1, online supplementary material). On each structure, six random photographs (21 28 cm) were taken using the Canon PowerShot G12 digital camera with an underwater housings (HSCN-G11, 10 Bar, Hong Kong), equipped with a S-TTL Strobe (D-2000, INON, Japan), and a removable custom steel frame (Video 2, online supplementary material). Supplementary video related to this article can be found at http://dx.doi.org/10.1016/j.marenvres.2014.12.004. In scraped samples, sessile and vagile species were identified to the lowest possible taxonomic level and their abundance was estimated as number of individuals per sample. In photographic samples organisms were identified by comparison with a species reference collection (voucher specimens and macro pictures) and assigned to morphological and ecological groups (Ponti et al., 2011). Percent cover of sessile organisms was estimated by superimposing a grid of 400 equal-sized cells (i.e. 0.25% each) and identifying all taxa visible within each cell (Benedetti-Cecchi et al., 1996). Areas of photos that were too dark, blurred or occupied by vagile organisms were excluded and percent cover was adjusted according to the total readable area of each image (Ponti et al., 2011). The endolithic bioeroder bivalve Rocellaria dubia (Pennant, 1777) was identified and quantified by counting its typical ‘8-shaped’ calcareous siphon holes. Percent
M. Ponti et al. / Marine Environmental Research 104 (2015) 10e19
13
Table 1 Annual water parameters in the study area (Q3: third quartile, data from the marine water-monitoring program, Veneto Regional Environmental Agency, reference year 2010). Parameter
Tsurface ( C) Tbottom ( C) Ssurface (psu) Sbottom (psu) pHsurface pHbottom O2 surface (%) O2 bottom (%)
Min 6.1 7.9 18.3 34.1 8.10 7.85 90.5 41.9
Max 26.8 22.3 34.3 37.6 8.40 8.27 111.4 99.1
Parameter 1
NeNH3 (mg l ) NeNO3 (mg l1) NeNO2 (mg l1) Ntot (mg l1) PePO4 (mg l1) Ptot (mg l1) Chl a (mg l1)
Median
Q3
Max
10.6 289 10.7 959 7.8 20.1 0.75
30.3 705 16.8 1439 13.0 39.1 1.15
44.7 2048 19.3 2998 51.0 86.1 1.87
cover of sediments was also estimated. 2.3. Data analyses Macrobenthic assemblages were analysed both in terms of densities (ind. sample1) in scraped samples, and percent cover, estimated in photos. Species richness (number of taxa, S), species diversity (Shannon's index with log base 2, H0 ) and the corresponding evenness component (Pielou index, J0 ) were calculated for each replicate sample (Magurran, 2004). Differences in assemblage structure, species abundance (density and percent cover) and diversity indices, between AR lots at different dates (Lot: AR1 in 2009 and 2012, AR2 in 2012) and AR typologies (Typ: TR and BT), were assessed by uni- and multivariate permutational analysis of variance (PERMANOVA; Anderson and Robinson, 2001; Anderson and ter Braak, 2003). Replication of the photographic sampling allowed an estimation of variability within the single structure (4 structures nested in Lot Typ; n ¼ 6). Univariate tests were run on Euclidean distances calculated on untransformed data (Anderson and Robinson, 2001). Multivariate analyses were based on BrayeCurtis similarity of square root transformed data (Clarke, 1993). Significant results were further analysed by ‘a posteriori’ pair-wise tests. When less than 999 unique values in the permutation distribution were available, asymptotical Monte Carlo p-values (pMC) were used instead of permutational pvalues. Similarity patterns were displayed by unconstrained ordination plots using the principal coordinate analysis (PCO, i.e. metric multidimensional scaling; Gower, 1966). Vectors superimposed on to the PCO plot represented the correlations of the abundances of the most relevant taxa with the PCO axes. Statistical analyses were performed using PRIMER 6 with PERMANOVA þ add-on package (Anderson et al., 2008). 3. Results 3.1. Density of sessile and vagile species A total of 116 taxa of macrobenthic invertebrates (3 sponges, 5 cnidarians, 35 polychaetes, 2 platyhelminthes, 1 nemertean, 1 nematode, 1 sipunculids, 27 molluscs, 27 crustacean, 3 bryozoans, 4 echinoderms, 7 ascidians) were found, 64 of which were identified to species level. Overall, the most abundant species were: the tube builder polychaetes Sabellaria spinulosa Leuckart 1849, Spirobranchus triqueter (Linnaeus 1758) and Serpula vermicularis Linnaeus 1767, the zoantharians Epizoanthus sp., the encrusting bivalve Anomia ephippium Linnaeus 1758, the mytilids Musculus subpictus (Cantraine 1835) and Mytilus galloprovincialis Lamarck 1819, the cosmopolitan bivalve Hiatella arctica (Linnaeus 1767), and the mud tube builder amphipod Monocorophium acherusicum (Costa, 1853). It is important to remark that macroalgae were nearly absent. Only filamentous algae forming a turf along with hydroids were
Fig. 4. PCO unconstrained ordination plot showing similarity relationships among sessile and vagile macrobenthic assemblages, in term of density data (sampling by scraping), according to AR typology (triangles TR: Tecnoreef® pyramids of slabs; quadrates BT: bundles of tubes) and AR lots, at different dates (white: AR1 in 2009: black: AR1 in 2012; grey: AR2 in 2012). Superimposed vectors represent correlations between the most important taxa abundance and PCO axes.
observed in photographic samples. Benthic assemblages sharply differed between AR1 and AR2, as shown by the distance between points on the PCO plot (Fig. 4). Statistical analysis revealed that on both AR typologies assemblages differed according to deployment and sampling dates (pair-wise test: pMC < 0.01; Table 2). No differences were found between assemblages on TR pyramids and BT structures in AR1 in 2009. Significant differences between reef typologies were found in 2012 in both AR1 and AR2 (pair-wise test: pMC < 0.05). The results of the PERMANOVA on the twelve most abundant species and on diversity indices are reported in Table 3. Seven species showed significant differences between AR1 and AR2 at different dates, regardless of reef typology, further four species were affected by the interaction between reef typologies and lots at different dates. In particular, Epizoanthus sp. was already present in 2009 and increased significantly in 2012, reaching an average of 676 ± 112 (s.e.) polyps per sample in AR1, while they were almost absent in AR2 in 2012 (Fig. 5). The burrowing bristleworm Polydora ciliata (Johnston, 1838) was particularly abundant in AR1 in 2009 (on average 66 ± 13 s.e. ind. sample1; pair-wise tests: p < 0.01). The calcareous tube builder S. vermicularis was significantly more abundant in the younger assemblages (i.e. 55 ± 19 s.e. ind. sample1
Table 2 PERMANOVA test on assemblage similarities (sampling by scraping) according to lots, at different dates (Lot), and ARs typology (Typ). Source
df
SS
MS
Pseudo-F
p (perm)
Unique perms
Lot Typ Lot Typ Res Total
2 1 2 18 23
12,668 2587 1451 7325 24,031
6334 2587 725 407
15.564 6.357 1.782
0.0001 0.0001 0.0212
9934 9934 9908
14
M. Ponti et al. / Marine Environmental Research 104 (2015) 10e19
Table 3 Summary of PERMANOVA test on selected species densities and diversity indices according to lots, at different dates (Lot), and AR typologies (Typ). Lots and dates (Lot)
Typology (Typ)
Lot Typ
MS
MS
MS
Pseudo-F2,18 p
Epizoanthus sp. 1,046,300 Polydora ciliata 10,088 Sabellaria spinulosa 21,381 Spirobranchus triqueter 733,250 Serpula vermicularis 4634 Serpula concharum 7219 Musculus subpictus 13,999 Mytilus galloprovincialis juv. 68,701 Anomia ephippium 42,421 Rocellaria dubia 813 Hiatella arctica 42,764 Monocorophium acherusicum 310,600 S J0 H0
29.94 21.48 0.89 86.32 3.68 4.84 4.17 9.70 10.71 11.49 61.56 5.54
487.170 20.16 0.035 12.94 1.166 10.26
0.0001 0.0003 0.4561 0.0001 0.0364 0.0028 0.0274 0.0029 0.0014 0.0013 0.0001 0.0129
Pseudo-F1,18 p
*** 40,837 *** 551 ns 119,290 *** 384,810 * 3174 ** 3675 * 864 ** 25,091 ** 7957 ** 1320 *** 2542 * 952,420
0.0001 *** 0.0001 *** 0.0003 ***
140.170 0.004 0.288
1.17 1.17 4.94 45.30 2.52 2.46 0.26 3.54 2.01 18.66 3.66 16.98 5.80 1.38 2.54
0.3485 0.3068 0.0281 0.0002 0.1359 0.0842 0.6205 0.0686 0.1770 0.0008 0.0699 0.0008
ns 25,073 ns 273 * 6140 *** 250,710 ns 1546 ns 4384 ns 1011 ns 6465 ns 10,701 *** 460 ns 2193 *** 351,300
0.0294 * 0.2526 ns 0.1338 ns
Res Pseudo-F2,18 p 0.72 0.58 0.25 29.51 1.23 2.94 0.30 0.91 2.70 6.51 3.16 6.26
55.167 0.017 0.452
2.28 6.34 3.98
0.5931 0.5821 0.8144 0.0001 0.3290 0.0166 0.7606 0.4342 0.0956 0.0083 0.0654 0.0083
MS ns 34,947 ns 470 ns 24,141 *** 8495 ns 1259 * 1491 ns 3359 ns 7081 ns 3962 ** 71 ns 695 ** 56,091
0.1346 ns 0.0098 ** 0.0424 *
24.167 0.003 0.114
Significant levels were indicated by the following symbols: ns ¼ not significant, * ¼ p < 0.05; ** ¼ p < 0.01; *** ¼ p < 0.001.
in AR1 in 2009 and 35 ± 13 s.e. ind. sample1 in AR2 in 2012; pairwise tests: p < 0.05). The tiny mussel M. subpictus was significantly less abundant in AR1 in 2009, compared to AR2 in 2012 (31 ± 15 s.e. vs. 114 ± 27 s.e. ind. sample1; pair-wise test: p < 0.05). Juveniles (
Contents lists available at ScienceDirect
Marine Environmental Research journal homepage: www.elsevier.com/locate/marenvrev
Benthic assemblages on artificial reefs in the northwestern Adriatic Sea: Does structure type and age matter? Massimo Ponti a, *, Federica Fava a, b, Rossella Angela Perlini a, b, Otello Giovanardi c, Marco Abbiati a, d a
Dipartimento di Scienze Biologiche, Geologiche ed Ambientali (BiGeA) & Centro Interdipartimentale di Ricerca per le Scienze Ambientali (CIRSA), University of Bologna, UO CoNISMa, Via S. Alberto 163, I-48123 Ravenna, Italy in Romagna, Via Baccarini 27, I-48121 Ravenna, Italy Fondazione Flaminia e Per l'Universita c Istituto Superiore per la Protezione e la Ricerca Ambientale (ISPRA), Loc. Brondolo, I-30015 Chioggia, Italy d ISMAR, Consiglio Nazionale delle Ricerche e Istituto di Scienze Marine, Bologna, Italy b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 11 October 2014 Received in revised form 23 December 2014 Accepted 24 December 2014 Available online 25 December 2014
The use of artificial reefs is on the rise worldwide. While their fish aggregating effects are well known, the epibenthic assemblages have been poorly investigated. Two types of artificial reefs (pyramids of concrete slabs and bundles of concrete tubes) have been deployed out of the Po River Delta in 2006 and 2010. The epibenthic assemblages were investigated in 2009 and 2012. Benthic assemblages on both structure typologies were dominated by species tolerating high sedimentation rates. Dissimilarities were found among assemblages with different ages, and, in less extend, between reef typologies. Colonisation by Mytilus galloprovincialis and other major space occupiers did not follow a clear succession pattern and was not affected by reef typology. Species colonisation was likely driven by variability in environmental conditions and recruitment processes rather than by reef typology. This study suggests that environmental features of the deployment sites should be carefully considered in planning and designing artificial reefs, especially in eutrophic and turbid coastal waters, exposed to high river loads. © 2014 Elsevier Ltd. All rights reserved.
Keywords: Artificial habitats Benthic assemblages Coastal structures Fouling organisms Bio-construction Temporal variability Exotic species
1. Introduction Artificial reefs (ARs) are manmade structures deployed on sea bottoms with the primary purpose of protecting coastal habitats and increasing biotic resources by aggregating marine species and preventing trawling (Baine, 2001). ARs support sessile filter feeders, providing nourishment and refuges for motile species, and attracting bentho-nectonic fishes (Bohnsack and Sutherland, 1985; Baine, 2001). Since the mid-1800s, ARs were deployed in many regions of the world, including tropical and temperate areas, starting with the United States of America and Japan (Bohnsack and Sutherland, 1985). The materials used in their construction include natural rocks, concrete blocks and several discarded supplies, like tires, pipes, shells, barges, bundled solid waste, coal ash, vehicles, etc. (Feary et al., 2011). The first ARs along European coasts were
* Corresponding author. Laboratori di Scienze Ambientali, Via S. Alberto 163, I48123 Ravenna, Italy. Skype: massimo.ponti. E-mail address: [email protected] (M. Ponti). http://dx.doi.org/10.1016/j.marenvres.2014.12.004 0141-1136/© 2014 Elsevier Ltd. All rights reserved.
installed in the 1960s, most structures have been located in the Mediterranean Sea. Since this time more than 70 AR complexes, made by different materials, have been deployed along Italian coasts (Fabi et al., 2011). Traditionally, in the oligotrophic waters of the western Mediterranean Sea, the goals of ARs were to protect Posidonia oceanica meadows from illegal trawling, to increase habitat complexity and promote higher species diversity (Relini et al., 1994; Riggio et al., 2000; Gonzalez-Correa et al., 2005). Conversely, in the eutrophic waters of the central and northern Adriatic Sea, the main purpose was to increase fishery yields (Bombace et al., 1994; Ardizzone et al., 1996; Bombace et al., 1997). Regardless of the potential benefits of ARs, their increasing frequency worldwide has given rise to concerns regarding their possible negative impacts, especially the dumping of waste and the use of unsuitable materials. In response to these threats as well as international conventions addressing the issue, some regulations, guidelines and protocols have been drawn up (e.g. London Convention and Protocol/UNEP, 2009; for an overview see Fabi et al., 2011). As a result, most recently deployed subtidal artificial habitats have been designed for specific purposes. Concrete is the
M. Ponti et al. / Marine Environmental Research 104 (2015) 10e19
most common material, because it is cheap, versatile, allowing the realization of structures with different shapes and sizes, and may ensure long life, being resistant to the chemical and physical marine actions (Fabi et al., 2011). Physical properties of the concrete vary according to the reinforcements and additives included in the cement mixture. Several natural and synthetic admixtures are used as hardening accelerators and retarders, corrosion inhibitors, etc. Besides improving concrete properties, the choice of additives is dictated by economic and environmental considerations, including the reduction of greenhouse gas emissions and recycling of wastes. Unfortunately, manufacturers often do not make available the compositions and properties of the concrete products or the possible interactions with living organisms (e.g. patents: EP0134855 B1, EP1193348 B1, WO2014125493 A1). The choice of the type of artificial structures and their placement should always take into account the expected, as well as possible undesired effects on coastal habitats. While fish aggregating effects of ARs are well known and the effectiveness of different structure typologies in this respect are well documented (Santos et al., 1997), much less attention has been dedicated to the benthic assemblages colonising the structures, despite the impact they may have on costal habitats, and the possible implications on ecosystem functioning, e.g. changes in species composition, species interactions and food webs (Ambrose and Anderson, 1990; Bertasi et al., 2007; Gallaway et al., 2009), alteration of population connectivity and genetic diversity (Cowen and Sponaugle, 2009; Fauvelot et al., 2009), facilitation of the spread of non-indigenous species (Airoldi et al., 2005; Bulleri and Airoldi, 2005; Glasby et al., 2007). Vagile and sessile species colonise ARs according to complex ecological processes affected by seasonal larval supply, water circulation, turbidity and nutrients, depths, orientation and physicalechemical features of the substrata (Anderson and Underwood, 1994; Relini et al., 1994; Riggio et al., 2000; Turner and Todd, 1993). Moreover, the interaction of abiotic and biotic factors may operate at different temporal and spatial scales (Glasby, 1998; Rodriguez et al., 1993). Water quality (e.g.: oligotrophic vs. eutrophic, clear vs. turbid) is considered a relevant factor in structuring benthic assemblages, as resulted by comparing ARs deployed in different locations at similar depths (Maughan, 2001; Relini et al., 1994; Riggio et al., 2000). Whereas within the same site and/or in very similar environmental conditions, materials (Anderson and Underwood, 1994; Glasby, 2000; Reyes and Yap, 2001), shapes (Bourget et al., 1994), orientation (Baynes, 1999; Genzano et al., 2011; Glasby and Connell, 2001; Ponti et al., 2002), shading and proximity to the seafloor (Glasby, 1999) are the main factors affecting the structure of benthic assemblages. Since 2001, thanks to regional and European funds, experimental concrete AR complexes have been deployed on sandy and
11
muddy bottoms along the western Adriatic coasts (Spagnolo et al., 2014). The aim of the present study was to analyse variability of macrobenthic assemblages in relation to the typology and age of ARs deployed off the Po River Delta (northern Adriatic Sea). Benthic assemblages on two types of artificial reefs, pyramids of slabs (Tecnoreef®, hereafter TR) and bundles of tubes (hereafter BT), differing in shape, concrete chemical composition, and age, were compared. 2. Materials and methods 2.1. Artificial reefs typologies and study site Two types of ARs have been investigated (Fig. 1): pyramids made of concrete slabs (TR e declared “sea-friendly” by the producer, Tecnoreef®, manufactured using only natural components, without synthetic additives; pH ~9; 1.8 and 2.4 m height); and bundles of common concrete tubes, BT, assembled in cubes laid on a concrete slab and retained by an iron cage (pH ~12; 1.8 m height). Despite similar size and surface rugosity, the two reef typologies differed in shape, surface inclination and material (e.g. type of concrete and pH). The ARs were deployed in 2 times: November 2006 (hereafter AR1) and March 2010 (hereafter AR2), 2 nautical miles offshore of the Po River Delta (northwestern Adriatic Sea, 44 540 N 12 33’ E), at 13e14 m depth and close to a longline mussel farm (Fig. 2). In AR1, TR pyramids and BT structures were arranged in two adjacent areas (100 200 m and 100 120 m) separated by 50 m. AR2 has a nucleus (100 200 m) of TR pyramids surrounded by BT structures deployed along the perimeter. TR pyramids deployed in AR2 differed from AR1 only by the smaller holes, strengthening the structures, which in AR1 was already damaged after two years. ARs have been deployed on a muddy bottom (silt > 75%) in an area affected by high freshwater and sediment inputs from the Po River, which has a mean flow of 1500 m3 s1 (period 1918e2006), with higher values in spring and autumn (Fig. 3). The combination of the thermohaline circulation and tide (up to 1 m) often results in strong currents. Frequently, water turbidity (mean Secchi disk 1.5 m) reduces penetration of solar radiation, while the superficial halocline and seasonal thermocline cause sharp stratification of the water column (Table 1). Moreover, effluents from the Po River are rich in nutrients, favouring growth of plankton (Aubry et al., 2012). 2.2. Sampling design and laboratory analyses With the aim to test for differences in benthic communities related to reef typology (TR vs. BT) and age (AR1 vs. AR2), macrobenthic assemblages were investigated in June 2009 on AR1 (~2.6 yrs after deployment; Fig. 3) and June 2012, both on AR1 and
Fig. 1. Example of ARs structures deployed: pyramids of slabs (2 floors, 1.8 m height), assembled with slabs of ‘sea-friendly’ concrete (Tecnoreef®), on the left and bundles of tubes, made by common concrete, laid on a concrete slab and retained by an iron cage (1.8 m height, on the right).
12
M. Ponti et al. / Marine Environmental Research 104 (2015) 10e19
Fig. 2. Geographic map of the study area and arrangement of the AR lots.
Fig. 3. Raw data (grey area) and cumulative sum series (bold line) of mean monthly flow of the Po River from 2006 to 2012 (courtesy from Hydro-meteorological Service of the EmiliaeRomagna Regional Environmental Agency). The reference value used to calculate the cumulative sum series is the long-term mean flow of the river (1500 m3 s1, period 1918e2006). Artificial reef deployments and sampling dates are reported by symbols.
AR2 (i.e. ~5.6 and ~2.3 yrs after deployment, respectively; Fig. 3). Four structures per typology were randomly chosen at each sampling date for both AR1 and AR2. Assemblages were investigated by means of scraping and photographic sampling along sub-vertical external sides of the artificial structures. Each scraped sample had an area of 0.16 m2, a sampling size often adopted in studies on
benthic assemblages in the Adriatic Sea (e.g. Spagnolo et al., 2014). Due to the surface discontinuities and poor visibility, samples were obtained by pooling four adjacent quadrates (20 20 cm each) scraped from the concrete surface by hammer and spatula or trowel, using a custom aluminium frame equipped with a holder for satin cloth bags (bright yellow, highly visible and permeable; Video 1, online supplementary material). On each structure, six random photographs (21 28 cm) were taken using the Canon PowerShot G12 digital camera with an underwater housings (HSCN-G11, 10 Bar, Hong Kong), equipped with a S-TTL Strobe (D-2000, INON, Japan), and a removable custom steel frame (Video 2, online supplementary material). Supplementary video related to this article can be found at http://dx.doi.org/10.1016/j.marenvres.2014.12.004. In scraped samples, sessile and vagile species were identified to the lowest possible taxonomic level and their abundance was estimated as number of individuals per sample. In photographic samples organisms were identified by comparison with a species reference collection (voucher specimens and macro pictures) and assigned to morphological and ecological groups (Ponti et al., 2011). Percent cover of sessile organisms was estimated by superimposing a grid of 400 equal-sized cells (i.e. 0.25% each) and identifying all taxa visible within each cell (Benedetti-Cecchi et al., 1996). Areas of photos that were too dark, blurred or occupied by vagile organisms were excluded and percent cover was adjusted according to the total readable area of each image (Ponti et al., 2011). The endolithic bioeroder bivalve Rocellaria dubia (Pennant, 1777) was identified and quantified by counting its typical ‘8-shaped’ calcareous siphon holes. Percent
M. Ponti et al. / Marine Environmental Research 104 (2015) 10e19
13
Table 1 Annual water parameters in the study area (Q3: third quartile, data from the marine water-monitoring program, Veneto Regional Environmental Agency, reference year 2010). Parameter
Tsurface ( C) Tbottom ( C) Ssurface (psu) Sbottom (psu) pHsurface pHbottom O2 surface (%) O2 bottom (%)
Min 6.1 7.9 18.3 34.1 8.10 7.85 90.5 41.9
Max 26.8 22.3 34.3 37.6 8.40 8.27 111.4 99.1
Parameter 1
NeNH3 (mg l ) NeNO3 (mg l1) NeNO2 (mg l1) Ntot (mg l1) PePO4 (mg l1) Ptot (mg l1) Chl a (mg l1)
Median
Q3
Max
10.6 289 10.7 959 7.8 20.1 0.75
30.3 705 16.8 1439 13.0 39.1 1.15
44.7 2048 19.3 2998 51.0 86.1 1.87
cover of sediments was also estimated. 2.3. Data analyses Macrobenthic assemblages were analysed both in terms of densities (ind. sample1) in scraped samples, and percent cover, estimated in photos. Species richness (number of taxa, S), species diversity (Shannon's index with log base 2, H0 ) and the corresponding evenness component (Pielou index, J0 ) were calculated for each replicate sample (Magurran, 2004). Differences in assemblage structure, species abundance (density and percent cover) and diversity indices, between AR lots at different dates (Lot: AR1 in 2009 and 2012, AR2 in 2012) and AR typologies (Typ: TR and BT), were assessed by uni- and multivariate permutational analysis of variance (PERMANOVA; Anderson and Robinson, 2001; Anderson and ter Braak, 2003). Replication of the photographic sampling allowed an estimation of variability within the single structure (4 structures nested in Lot Typ; n ¼ 6). Univariate tests were run on Euclidean distances calculated on untransformed data (Anderson and Robinson, 2001). Multivariate analyses were based on BrayeCurtis similarity of square root transformed data (Clarke, 1993). Significant results were further analysed by ‘a posteriori’ pair-wise tests. When less than 999 unique values in the permutation distribution were available, asymptotical Monte Carlo p-values (pMC) were used instead of permutational pvalues. Similarity patterns were displayed by unconstrained ordination plots using the principal coordinate analysis (PCO, i.e. metric multidimensional scaling; Gower, 1966). Vectors superimposed on to the PCO plot represented the correlations of the abundances of the most relevant taxa with the PCO axes. Statistical analyses were performed using PRIMER 6 with PERMANOVA þ add-on package (Anderson et al., 2008). 3. Results 3.1. Density of sessile and vagile species A total of 116 taxa of macrobenthic invertebrates (3 sponges, 5 cnidarians, 35 polychaetes, 2 platyhelminthes, 1 nemertean, 1 nematode, 1 sipunculids, 27 molluscs, 27 crustacean, 3 bryozoans, 4 echinoderms, 7 ascidians) were found, 64 of which were identified to species level. Overall, the most abundant species were: the tube builder polychaetes Sabellaria spinulosa Leuckart 1849, Spirobranchus triqueter (Linnaeus 1758) and Serpula vermicularis Linnaeus 1767, the zoantharians Epizoanthus sp., the encrusting bivalve Anomia ephippium Linnaeus 1758, the mytilids Musculus subpictus (Cantraine 1835) and Mytilus galloprovincialis Lamarck 1819, the cosmopolitan bivalve Hiatella arctica (Linnaeus 1767), and the mud tube builder amphipod Monocorophium acherusicum (Costa, 1853). It is important to remark that macroalgae were nearly absent. Only filamentous algae forming a turf along with hydroids were
Fig. 4. PCO unconstrained ordination plot showing similarity relationships among sessile and vagile macrobenthic assemblages, in term of density data (sampling by scraping), according to AR typology (triangles TR: Tecnoreef® pyramids of slabs; quadrates BT: bundles of tubes) and AR lots, at different dates (white: AR1 in 2009: black: AR1 in 2012; grey: AR2 in 2012). Superimposed vectors represent correlations between the most important taxa abundance and PCO axes.
observed in photographic samples. Benthic assemblages sharply differed between AR1 and AR2, as shown by the distance between points on the PCO plot (Fig. 4). Statistical analysis revealed that on both AR typologies assemblages differed according to deployment and sampling dates (pair-wise test: pMC < 0.01; Table 2). No differences were found between assemblages on TR pyramids and BT structures in AR1 in 2009. Significant differences between reef typologies were found in 2012 in both AR1 and AR2 (pair-wise test: pMC < 0.05). The results of the PERMANOVA on the twelve most abundant species and on diversity indices are reported in Table 3. Seven species showed significant differences between AR1 and AR2 at different dates, regardless of reef typology, further four species were affected by the interaction between reef typologies and lots at different dates. In particular, Epizoanthus sp. was already present in 2009 and increased significantly in 2012, reaching an average of 676 ± 112 (s.e.) polyps per sample in AR1, while they were almost absent in AR2 in 2012 (Fig. 5). The burrowing bristleworm Polydora ciliata (Johnston, 1838) was particularly abundant in AR1 in 2009 (on average 66 ± 13 s.e. ind. sample1; pair-wise tests: p < 0.01). The calcareous tube builder S. vermicularis was significantly more abundant in the younger assemblages (i.e. 55 ± 19 s.e. ind. sample1
Table 2 PERMANOVA test on assemblage similarities (sampling by scraping) according to lots, at different dates (Lot), and ARs typology (Typ). Source
df
SS
MS
Pseudo-F
p (perm)
Unique perms
Lot Typ Lot Typ Res Total
2 1 2 18 23
12,668 2587 1451 7325 24,031
6334 2587 725 407
15.564 6.357 1.782
0.0001 0.0001 0.0212
9934 9934 9908
14
M. Ponti et al. / Marine Environmental Research 104 (2015) 10e19
Table 3 Summary of PERMANOVA test on selected species densities and diversity indices according to lots, at different dates (Lot), and AR typologies (Typ). Lots and dates (Lot)
Typology (Typ)
Lot Typ
MS
MS
MS
Pseudo-F2,18 p
Epizoanthus sp. 1,046,300 Polydora ciliata 10,088 Sabellaria spinulosa 21,381 Spirobranchus triqueter 733,250 Serpula vermicularis 4634 Serpula concharum 7219 Musculus subpictus 13,999 Mytilus galloprovincialis juv. 68,701 Anomia ephippium 42,421 Rocellaria dubia 813 Hiatella arctica 42,764 Monocorophium acherusicum 310,600 S J0 H0
29.94 21.48 0.89 86.32 3.68 4.84 4.17 9.70 10.71 11.49 61.56 5.54
487.170 20.16 0.035 12.94 1.166 10.26
0.0001 0.0003 0.4561 0.0001 0.0364 0.0028 0.0274 0.0029 0.0014 0.0013 0.0001 0.0129
Pseudo-F1,18 p
*** 40,837 *** 551 ns 119,290 *** 384,810 * 3174 ** 3675 * 864 ** 25,091 ** 7957 ** 1320 *** 2542 * 952,420
0.0001 *** 0.0001 *** 0.0003 ***
140.170 0.004 0.288
1.17 1.17 4.94 45.30 2.52 2.46 0.26 3.54 2.01 18.66 3.66 16.98 5.80 1.38 2.54
0.3485 0.3068 0.0281 0.0002 0.1359 0.0842 0.6205 0.0686 0.1770 0.0008 0.0699 0.0008
ns 25,073 ns 273 * 6140 *** 250,710 ns 1546 ns 4384 ns 1011 ns 6465 ns 10,701 *** 460 ns 2193 *** 351,300
0.0294 * 0.2526 ns 0.1338 ns
Res Pseudo-F2,18 p 0.72 0.58 0.25 29.51 1.23 2.94 0.30 0.91 2.70 6.51 3.16 6.26
55.167 0.017 0.452
2.28 6.34 3.98
0.5931 0.5821 0.8144 0.0001 0.3290 0.0166 0.7606 0.4342 0.0956 0.0083 0.0654 0.0083
MS ns 34,947 ns 470 ns 24,141 *** 8495 ns 1259 * 1491 ns 3359 ns 7081 ns 3962 ** 71 ns 695 ** 56,091
0.1346 ns 0.0098 ** 0.0424 *
24.167 0.003 0.114
Significant levels were indicated by the following symbols: ns ¼ not significant, * ¼ p < 0.05; ** ¼ p < 0.01; *** ¼ p < 0.001.
in AR1 in 2009 and 35 ± 13 s.e. ind. sample1 in AR2 in 2012; pairwise tests: p < 0.05). The tiny mussel M. subpictus was significantly less abundant in AR1 in 2009, compared to AR2 in 2012 (31 ± 15 s.e. vs. 114 ± 27 s.e. ind. sample1; pair-wise test: p < 0.05). Juveniles (
