Conservation challenges in urban seascapes: promoting the growth of threatened species on coastal infrastructures


  • Shimrit Perkol-Finkel,

    Corresponding author
    1. SeArc – Ecological Marine Consulting, Tel Aviv, Israel
    • Dipartimento di Biologia Evoluzionistica Sperimentale, Centro Interdipartimentale di Ricerca per le Scienze Ambientali in Ravenna, University of Bologna
    Search for more papers by this author
  • Filippo Ferrario,

    1. Dipartimento di Biologia Evoluzionistica Sperimentale, Centro Interdipartimentale di Ricerca per le Scienze Ambientali in Ravenna, University of Bologna
    Search for more papers by this author
  • Valentina Nicotera,

    1. Dipartimento di Biologia Evoluzionistica Sperimentale, Centro Interdipartimentale di Ricerca per le Scienze Ambientali in Ravenna, University of Bologna
    Search for more papers by this author
  • Laura Airoldi

    1. Dipartimento di Biologia Evoluzionistica Sperimentale, Centro Interdipartimentale di Ricerca per le Scienze Ambientali in Ravenna, University of Bologna
    Search for more papers by this author

Correspondence author. E-mail:


  1. With nearly two-thirds of the human population concentrated along coastlines, coastal development and urbanized seascapes are inevitable. Proliferation of coastal and marine infrastructures, such as breakwaters, ports, seawalls and offshore installations, is associated with loss of natural habitats. This calls for new strategies aimed at elevating the ecological and biological value of coastal infrastructures, while minimizing their ecological footprint.

  2. We explored the feasibility of using coastal defence structures as a scaffold for the conservation of threatened marine species. We experimented with fucoids, canopy-forming algae on Mediterranean coasts, in the light of their declared conservation priority. We transplanted juveniles of Cystoseira barbata to a number of breakwaters and natural sites along the Adriatic Sea (Italy) and tested which factors could facilitate or inhibit its successful establishment.

  3. Survival of transplanted C. barbata was greater at most artificial and natural sites examined compared to the native sites where severe habitat loss was ongoing. Survival was greater at landward compared to seaward positions on the infrastructure, while no relevant effects of substratum characteristics (horizontal vs. vertical orientation, variable composition and increasing complexity) were observed. Lack of surrounding adult fronds did not impair the survival or growth of the transplants, suggesting a high transplantation potential also on novel infrastructures.

  4. Success of transplantation in areas remote from the source population was limited by biotic disturbance, which was more intense on coastal infrastructures in sedimentary environments compared to natural rocky sites.

  5. Synthesis and applications. Coastal and marine infrastructures can be harnessed to enhance desired species (such as threatened canopy-forming algae). A comprehensive understanding of the ecological functioning of these urban seascapes compared to natural habitats is required to minimize detrimental impacts, or potentially increase the ecological value, of coastal structures and efficiently incorporate such strategies into management and conservation actions. We investigated the influence of habitat type (including natural and artificial), surface complexity, herbivore exclusion, proximity to established populations and orientation on the transplantation success of threatened algae.


With nearly two-thirds of the world's population concentrated in coastal areas (Creel 2003), substantial coastal development is inevitable. The land–sea interface is exploited for various human uses including industry, transportation, energy and recreation (Airoldi & Beck 2007). These forms of coastal development are frequently associated with fragmentation and loss of natural habitats, damaged seascapes and reduced biodiversity (Airoldi & Beck 2007; Crain et al. 2009; Dugan et al. 2011).

It is known that coastal infrastructures do not function as surrogates to natural habitats (Bulleri & Chapman 2004; Jackson, Chapman & Underwood 2008). Their vertical profile compresses the intertidal zones, and their homogenous surfaces combined with high frequency of disturbances tend to favour impoverished assemblages dominated by opportunistic and invasive species (Bulleri & Airoldi 2005; Chapman et al. 2009; Airoldi & Bulleri 2011).

As coastal infrastructures are expected to proliferate alongside with human population (UN 2008), efforts should be made not only to minimize their detrimental impacts, but also to elevate their possible ecological value. This requires understanding of the types of assemblages or ecosystem functions that are desirable and feasible in these habitats. Initial steps in this direction have been made in highly urbanized areas in both temperate (Airoldi et al. 2005b; Chapman & Blockley 2009; Browne & Chapman 2011) and tropical environments (Perkol-Finkel et al. 2006, 2008). Nonetheless, the notion of combining ecological principles to urban infrastructure is rather new (Mitsch 1996; Bergen, Bolton & Fridley 2001) and to date has been scarcely applied in marine environments.

We examined the feasibility of facilitating the growth of threatened fucoid macroalgae of the genus Cystoseira on coastal defence structures. Fucoids and kelps form diverse, structurally complex and highly productive canopy habitats along many temperate rocky coasts (Steneck et al. 2002). Canopies are suffering widespread habitat loss at global scales (Airoldi & Beck 2007; Connell et al. 2008; Mangialajo, Chiantore & Cattaneo-Vietti 2008). Decline in the Mediterranean Sea is well documented, and today six Mediterranean species of Cystoseira are listed as threatened in the Bern Convention and in the Mediterranean Action Plan. In the Mediterranean, the proximate cause for loss of Cystoseira is anthropogenic disturbance, largely in the form of urbanization (Benedetti-Cecchi et al. 2001). Recent experiments have shown the potential for recovery of canopy-forming macroalgae through various approaches, including transplanting or seeding macroalgae back to their original habitat (Correa et al. 2006; Susini et al. 2007; Perkol-Finkel & Airoldi 2010), and the use of artificial reef for algal restoration is increasing (Terawaki et al. 2003; Falace, Zanelli & Bressan 2006; Park & Lee 2010). Here, we explored the alternative possibility of using coastal infrastructures as gardens for these important habitat-formers, and deploying them for other societal needs. This approach would enhance the ecological value of these infrastructures, without compromising their original function.

Relatively, few studies have attempted to transplant canopy-forming macroalgae onto artificial substrata (Terawaki et al. 2003; Jonsson et al. 2006), and little is known about the factors enhancing the success of these interventions. We transplanted juveniles of Cystoseira barbata Stackhouse C. Agardh onto a number of breakwaters and natural sites along the Italian North Adriatic Sea (Italy). Marine infrastructures offer atypical substrates for benthic assemblages in terms of orientation, exposure, structure and surface texture (Vaselli, Bulleri & Benedetti-Cecchi 2008; Burt et al. 2009; Bulleri & Chapman 2010), all of which are known to affect the recruitment, survival and growth of fucoids and other macroalgae (Harlin & Lindbergh 1977; Wells, Moll & Bolton 1989; Airoldi 2001; Jonsson et al. 2006). We tested whether the survival and growth of transplants differed between natural and artificial habitats, horizontal vs. vertical substrata, between landward vs. seaward sides of the breakwaters, or among substrata of different composition and increasing surface complexity. We also analysed whether lack of naturally occurring surrounding adult canopies on such infrastructures, which normally facilitate natural recruitment of canopies (Connell 2005; Irving & Connell 2006), limits successful transplantation. Finally, we used caging experiments to test the possible role of grazing pressure on success of transplantation, as this factor has been previously described as limiting for growth of macroalgae on coastal defence structures (Jonsson et al. 2006) and because pilot tests suggested the importance of this factor in our study system.

Materials and methods

Study area and species

The study was conducted at the Monte Conero promontory (43° 33′N, 13° 37′E) and the surrounding urbanized sandy coastline of the North Italian Adriatic Sea (Fig. 1). The rocky promontory hosts some of the last-remaining scattered populations of the threatened genus Cystoseira along the central-northern Italian Adriatic coast (Perkol-Finkel & Airoldi 2010). The fragmented state of these populations probably results from a synergistic effect of low substratum stability and competition with opportunistic species (Perkol-Finkel & Airoldi 2010). We sourced Cystoseira from two sites ‘Due Sorelle’ and ‘La Vela’. The algal assemblages at these sites were composed mainly of the species Cystoseira barbata C. Agardh (Fucales: Sargassaceae) that was found in varying densities from c. 2 to 5 m depth. A detailed description of the study area, the biology of the species and historical changes in the distributions of macroalgal canopies can be found in Perkol-Finkel & Airoldi (2010).

Figure 1.

Map of the study sites along the north Adriatic Sea, and insert map of locations at the Monte Conero promontory.

Other rocky habitats in the area exist only in the form of detached breakwaters, two of which, at the localities named Urbani and Numana (Fig. 1), were used for the experiments. According to preliminary surveys, the natural bedrocks surrounding these breakwaters had a very sporadic appearance of naturally recruited C. barbata. We also transplanted juveniles onto breakwaters at the localities of Marotta, Lido Adriano and Punta Marina (c. 50, 140 and 150 km north of Monte Conero), where no Cystoseira naturally occurs.

Transplantation experiments

We transplanted juveniles (2–3 months old, 5 cm high) of C. barbata collected from loose boulders at Due Sorelle and La Vela in June 2008. Previous studies showed that recruits in these habitats have low survival probability because of severe substratum instability (Perkol-Finkel & Airoldi 2010). The boulders were broken into small fragments holding 1–2 individuals that were transplanted onto the substrate into the new habitats using epoxy putty (Subcoat S. Veneziani) to form experimental plots comprising five transplanted individuals.

Such plots were transplanted in four habitat types (hereafter ‘Habitats’). These included (i) ‘Native habitat’, that is, the loose boulder fields where the few juveniles naturally occurring in the area are found and from which they were initially collected, and three additional habitats (hereafter ‘Other Habitats’) where natural recruitment of juveniles was not observed, including (ii) Natural bedrock habitat (i.e. stable bedrocks represented by boulders >10m3 in size or eroded rocky platforms), (iii) Artificial habitat – seaward side, and (iv) Artificial habitat – landward side. For each habitat, two replicated areas (hereafter ‘Areas’) were established. For the Native and Natural habitats, one area was set at La Vela and another at Due Sorelle. For the artificial habitats, one area was set at Urbani and another at Numana. Within each area, four plots with transplants were created in each of the following positions (hereafter ‘Positions’): (i) horizontally surrounded with naturally occurring adults (HA), (ii) horizontally without surrounding adults (HW) and (iii) vertically with no surrounding adults (V). Vertical positions with surrounding adults were not included because of the natural scarcity of adults on vertical surfaces. There were no comparable positions in the native habitat, which was represented by small, irregular, loose boulder fields with no consistent relief.

As the main goal of this experiment was to test the feasibility of enhancing fragmented communities of C. barbata by transplantation onto artificial structures and identify optimal conditions (i.e. position) for such transplantations, there was no need to include transplantation methodological controls (normally used when transplantation is used to explore aspects of the ecology of the species), as any effect of transplantation would be part of the hypothesis of interest. Transplantation into the original native habitat and into stable rocky bedrocks served as a comparison to understand how successful would be the transplantation in artificial conditions compared to more natural conditions at natural bedrocks.

The height of the juveniles was recorded at the time of transplantation, and growth was monitored along with survival rates in September and October 2008 and in February 2009. At each date, we also measured the size of unmanipulated C. barbata juveniles naturally occurring at La Vela and Due Sorelle to explore whether transplanted juveniles had different growth rates from unmanipulated ones. For this, all juveniles were carefully removed from one random 6 × 6 cm plot on each of three randomly selected boulders, for subsequent measurements in the laboratory. Survival of unmanipulated juveniles from native habitats was known to be virtually nil (Perkol-Finkel & Airoldi 2010), and no formal comparison was included.

Differences in the average survival and sizes of transplanted juveniles between habitats and positions were tested using asymmetrical permutational anovas, including three factors: Habitat type (where the Native habitat was confronted with the three Other habitats: Natural bedrock, Artificial seaward and Artificial landward; fixed factor), Area (two areas, nested in each habitat; random factor) and Position [horizontal surrounded with adults (HA), horizontal without adults (HW) and vertical (V); fixed factor]. These asymmetrical analyses involved partitioning components of variation through two sub-analyses (see: Winer 1971). First, we ran two-way analyses testing for differences among the four habitats and areas, and contrasting the native habitat with the three other habitats irrespective of the possible different positions at the other habitats. Second, we ran three-way analyses, testing for effects of habitat, positions and areas at the other habitats only. We used the statistical package permanova+ for primer (Anderson, Gorley & Clarke 2008) to partition the variability and obtain F-statistics on a matrix of Euclidean distances calculated from the original raw data and calculated all P-values using 9999 random permutations of the appropriate exchangeable units and Type III sums of squares to cope with the unbalanced design (Anderson, Gorley & Clarke 2008). We used permanova (as opposed to a classic anova test) due to ease of use with complex unbalanced design and to avoid the usual normality assumptions. The analyses were performed on data retrieved in October 2008, as this was the last date for which all plots remained intact; after this date, some areas (one Natural bedrock and one Artificial seaward) were damaged by a storm in December 2008. Both survival and size data had homogeneous variances [Levenes’ (1960) univariate test run using permisd (Anderson, Gorley & Clarke 2008)], and there was no need for transformation. The average size of all surviving transplants at the end of the experiment, in February 2009, was also compared to the average size of naturally occurring juveniles using a t-test.

To test whether the conditions identified as optimal for the growth of Cystoseira also applied to more remote coastal infrastructures in sedimentary environments, we ran two additional transplantation experiments. The first was set at the seaward and landward sides of two breakwaters located at Punta Marina and Lido Adriano, simultaneously with the experiment set in the Monte Conero promontory (Fig. 1). Juveniles were transported by car to these locations as quickly as possible in 100-L tanks. At each side of the breakwaters, four plots (with five individuals in each) were transplanted at each of the following positions: (i) Horizontally without surrounding adults (HW) and (ii) Vertically with no surrounding adults (V). Some individuals were kept in tanks on land for approximately the same time of transportation and transplanted back at the original source location at Due Sorelle as procedural controls. All juveniles transplanted to breakwaters showed 100% mortality within a week of transplantation, and no further sampling was performed.

Such a rapid loss of transplants was inconsistent with the results from the experiments carried out on breakwaters in the Monte Conero region, and not related to the transplantation procedure, therefore, the following year (June 2009) we ran a second experiment at the locality of Marotta (Fig. 1), which presented water conditions more similar to those at Monte Conero than the other two stations. Four small boulders (c. 0·1 × 0·1 m) holding numerous recruits of Cbarbata were transplanted from Due Sorelle and established horizontally without surrounding adults (HW) at the landward sides of two replicated breakwaters. Here too, zero survival was recorded, as all transplants disappeared within 3 days.

Caging experiments

We used caging experiments to explore whether the loss of transplants observed at artificial habitats set on sedimentary shorelines was related to environmental factors (e.g. lower water quality or excess sedimentation along a sedimentary shoreline), biotic factors (i.e. pressure from grazers or other sources of biological disturbance) or a combination of both. In June 2009, we collected 32 small boulders (c. 10 cm diameter) densely covered with naturally occurring juveniles of C. barbata from La Vela. The boulders were attached to the breakwaters, using epoxy putty, at each of two sites randomly selected at Due Sorelle (natural sites on a stable bedrock) and on two breakwaters at Marotta (artificial sites on a sandy bottom setting). We did not include a comparison with artificial habitats in a rocky setting as we had already demonstrated in the prior transplantation experiment that survival and growth of transplants in this habitat was similar to that of transplants on natural bedrocks. We predicted that if loss of transplants at artificial sandy sites was related to biotic factors, their survival would increase below cages, which limit access to potential grazers. Conversely, differences in survival between the study locations would persist below cages under the prevailing effects of different local environmental conditions.

To unravel the two mechanisms, four boulders selected at random in each area were protected by 40 × 15x15 cm plastic mesh cages (hole size 1 × 1 cm) which excluded possible macro-grazers (i.e. fish and sea-urchins), while the remaining four were left uncaged as controls. Because all transplants (both caged and non-caged) in Marotta were lost within 48 h, the experiment was repeated using 1 × 1 mm mesh cages, to exclude both macro and mesograzers, while control plots were left uncaged. We did not include a partial caging control as we did not know the nature of eventual grazers (see 'Discussion'). However, to minimize possible environmental alterations by the cages (e.g. sedimentation or wave action), we conducted the experiment under calm sea conditions. In this experiment, the transplanted units were marble plates (10 × 10 × 2 cm) densely covered with C. barbata juveniles. The plates had been placed at La Vela in March 2009, at the start of the reproductive season, to measure patterns of recruitment and had not been manipulated in any way before this experiment. The density and cover of juveniles were measured for each plate prior to transplantation, and subsequent changes were monitored 4 and 8 days after transplantation. This short interval was sufficient to detect a clear response while limiting possible longer-term artefacts related to the use of fine mesh cages. Differences between treatments were analysed by permutational anova (using the statistical package permanova as illustrated previously) on data from day 8. The model included the factors: Biotic pressures (caged vs. un-caged, fixed factor), Local Environment (Natural bedrock vs. artificial sandy, fixed factor) and Site (random factor nested within Local Environment).

Recruitment experiments

As the feasibility of successfully rearing C. barbata on coastal infrastructures will ultimately depend on its ability to proliferate and recruit onto the artificial substrata following active transplantation, we analysed the effects of small scale complexity on settlement of C. barbata using clay settlement plates (10 × 10 × 2 cm). Six plates were prepared for each of three levels of complexity: low (smooth surface), medium (surface with crevices 1–2 mm deep) and high (surface with crevices c. 5 mm deep), and set randomly at La Vela in March 2009. Complexity was imprinted onto the moist clay using pieces of natural rock, to mimic natural features. Plates were attached to natural substratum close to adult fronds of C. barbata at La Vela during March 2009, at about 3 m depth, using epoxy putty. Recruits of C. barbata were counted at the end of June and August 2009. Differences between levels of complexity (fixed factor) were tested separately for each time by permutational anova (Anderson, Gorley & Clarke 2008).

We also analysed the effects of different materials often used for the construction of marine infrastructures, that is, limestone (marble), concrete and clay. Six replicated plates (10 × 10 × 2 cm) of each material were set randomly at La Vela in March 2009. Recruits of C. barbata were counted at the end of June 2009. No further sampling was possible as these plates were lost during a storm. Effect of material (fixed factor) was tested by permutational anova (Anderson, Gorley & Clarke 2008).


Transplantation experiments

Juveniles of C. barbata transplanted onto both natural bedrocks and artificial habitats had significantly greater survival relative to those transplanted back to their native (source) habitat (Fig. 2 and Table 1a, contrast Native vs. Other Habitats). While virtually no transplants survived after October 2008 in the native habitat (because of boulder overturning and disturbance), many transplants in the other habitats survived until February 2009. Survival was highest in landward artificial habitats, with average survival >30%, in comparison with c. 20% in the natural bedrock habitats and 9% in the seaward artificial habitats (Fig. 2). Nonetheless, differences among these other habitats were not significant (Table 1). Variability among individual replicated plots was high, and some plots had 100% survival throughout the experiment, while others had no surviving transplants. There were no consistent detectable effects in relation to position in any of the three other habitats where it was tested (Fig. 2 and Table 1, effects of Position within Other Habitats contrasts).

Figure 2.

Percentage (mean ± 1SE, n = 8, i.e. four plots for each of two areas) of recruits that survived out of those transplanted (five per plot) in July 2008 into four habitats (Natural bedrock, Seaward artificial, Landward artificial and Native boulder) at three positions [horizontal with surrounding adults (■), horizontal without surrounding adults (□), vertical (▲)] over the study period. Native habitat (△) had no comparable positions (see text for further explanation). Natural habitat data have been presented in part in Perkol-Finkel & Airoldi (2010).

Table 1. Asymmetrical analysis of the effects of habitat type and position on (a) percentage survival and (b) size of transplanted Cystoseira barbata recruits in October 2008. Factors are the following: habitat type (Native boulder was compared with three Other habitats: Natural bedrocks, Artificial seaward, Artificial landward, all fixed factors), area (two random areas, nested in habitat type) and position [horizontal surrounded with adults (HA), horizontal without adults (HW), and vertical (V), fixed factor] orthogonal to the Other habitats only. The analysis consists of two parts, one (upper) contrasting Native vs. Other habitats and the other (lower) examining survival within Other habitats in relation to the different positions. We used the statistical package permanova to partition the variability and obtain F-statistics on a matrix of Euclidean distances from the original raw data, and calculated all P-values using 9999 random permutations of the appropriate exchangeable units (Anderson, Gorley & Clarke 2008)
Source of variationd.f.MS F
  1. a

    < 0·05.

Native vs. other habitats16796·357·77a
Area (habitat)4874·350·73
Among other habitats
Other habitats2190·39350·20
Area (other habitats)3961·63290·84
Position × other habitats41496·84560·59
Position × area (other Ha)62547·67332·22
Native vs. other habitats11·15440·007
Area (habitat)4161·1914·34
Among other habitats
Other habitats272·720·55
Area (other habitats)3190·3118·78a
Position × other habitats448·3153·4807
Position × area (other Ha)613·9211·373

No significant differences in the size of transplanted juveniles were found between native and other habitats (Fig. 3 and Table 1b, contrast Native vs. Other Habitats). However, all transplanted juveniles that survived were larger on average than naturally unmanipulated juveniles in the study region. These differences were significant (t-test, < 0·01) at the last monitoring date (February 2009) when transplanted juveniles had an average size of 10·79 ± 6·08 cm (mean ± 1 SD, n = 83), while natural unmanipulated juveniles were only 8·27 ± 3·88 cm (mean ± 1 SD, n = 49). Moreover, some transplanted thalli that survived to the following spring (2009) both in natural bedrocks and in artificial habitats were observed to have grown to adult size and hold reproductive structures. In fact, during the summer, first generation recruits were observed in close proximity to these transplants. While at the natural sites, it is possible that these new recruits originated from other adults in the area, and this was unlikely at the artificial sites where very few adults occurred naturally.

Figure 3.

Size of transplanted recruits at four habitats (Natural bedrock, Seaward artificial, Landward artificial and Native boulders) and three positions [horizontal with surrounding adults (■), horizontal without surrounding adults (□), vertical (▲)]. Native habitat (△) had no comparable positions (see text for further explanation). Size of natural, un-manipulated recruits in the study areas is also plotted. Data points show the mean (±1SE, n = 8, i.e. four plots for each of two areas) size of survived thalli of transplanted recruits within each plot (no measures were available for February 2009 for control habitats because no transplants remained). Data for natural un-manipulated recruits (*) are the mean (±1ES) of 80 recruits from three different boulders sampled from natural habitats at each time point. Natural habitat data have been presented in part in Perkol-Finkel & Airoldi (2010).

None of the juveniles transplanted onto breakwaters at Punta Marina, Lido Adriano or Marotta survived longer than 2–3 days. During these experiments, the sea was calm, leading to exclude a possible dislodgment by waves, and there were no signs of vandalism.

Caging experiments

At the natural bedrock sites, caging did not influence the survival or the cover of juveniles, and all transplants remained equally intact both inside and outside of the cages (Fig. 4a,b). At the artificial sites in Marotta, uncaged transplants showed severe decline, with nearly 80% of the coverage lost within 8 days. These losses persisted when large mesh cages were used. Conversely, both survival and cover of transplanted juveniles at the artificial sites significantly increased when fine mesh cages were used (Table 2a,b, permanova, significant Treatment × Habitat interaction, F(d.f. = 1,18) = 5·8739 < 0·05 for cover and F(d.f. = 1,18) = 47·459 < 0·05 for survival). In these treatments, after 8 days, both cover and survival matched the values measured at the natural bedrock sites (Fig. 4a,b).

Figure 4.

(a) Relative cover and (b) percentage survival (in relation to initial cover and/or count, respectively, mean ± 1 SE) of caged (1 mm mesh size cage, full symbols) and un-caged (open symbols) recruits transplanted onto two breakwaters at Marotta [Sandy artificial (■)] vs. two natural bedrock areas in La Vela [natural bedrock (▴)]. Four plates covered with Cystoseira barbata recruits per treatment and site within each habitat. Values are presented as 100% when transplanted (Time 0) and then 4 and 8 days following transplantation (Time 1, Time 2, respectively).

Table 2. Results of tests for (a) Relative cover and (b) Percentage survival (in relation to initial cover and/or count respectively) of caged (1 mm mesh size cage) and uncaged recruits transplanted onto two breakwaters at two Sandy Artificial sites vs. two Natural bedrock sites. Factors are the following: treatment (Cages vs. Un-caged), habitat (Artificial Sandy vs. Natural Bedrock), site (nested in habitat: two breakwaters at Marotta and two areas in La Vela). Four plates covered with Cystoseira barbata recruits per treatment and site within each habitat. The tests were carried out 4 days following transplantation. We used the statistical package permanova to partition the variability and obtain F-statistics on a matrix of Euclidean distances from the original raw data and calculated all P-values using 9999 random permutations of the appropriate exchangeable units (Anderson, Gorley & Clarke 2008)
Source of variationd.f.MS F
  1. a

    < 0·05.

Site (habitat)2812·32·5831
Caging × habitat13327·55·8739a
Caging × site (habitat)2567·621·805
Site (habitat)22057·42·6137
Caging × habitat18670·547·459a
Caging × Site (habitat)2363·510·4618

Recruitment experiments

By the end of the reproductive season (June 2009), all settlement plates had some C. barbata juveniles. The density of recruits was highly variable among individual plates, ranging from 6 to 64 individuals. Initially, complexity appeared to have a significant influence on settlement (permanova: F(d.f. = 2,14) = 3·893, < 0·05), and densities of settlers on plates with medium complexity were on average almost double than those with low and high complexities (Fig. 5). Two months later, average densities was still highest on plates with medium complexity, but differences between complexities were no longer significant (permanova: F(d.f. = 2,14) = 2·72, > 0·05). The density of recruits was lower on concrete than on limestone and clay (Fig. 6), but there was a large variability among plates, and substratum composition did not appear to significantly affect settlement of C. barbata (permanova: F(d.f. = 2,13) = 1·684, > 0·05).

Figure 5.

Number of Cystoseira barbata recruits (mean ± 1SE, n = 6) on clay plates of different complexity levels: L = low, M = medium, H = high. Plates were set on March 2009 and counted in June and August 2009. Superimposed circles represent significant differences by Student–Newman–Keuls (SNK) test (M > H = L,< 0·05).

Figure 6.

Number of Cystoseira barbata recruits (mean ± 1SE, n = 6) on plates of different composition: concrete, clay and limestone. Plates were set on March 2009 and counted in June 2009.


Transplanting C. barbata juveniles proved technically feasible on both natural bedrocks and man-made habitats in the area of Monte Conero, indicating the potential of coastal infrastructures to provide a suitable habitat for the growth of this threatened species. Overall, landward, sheltered rocky artificial habitats seemed most successful, regardless the presence of surrounding adults. Furthermore, the chances for survival and growth of transplanted individuals in the study area were greater than those measured in the native habitats, where this species is threatened because of long-term recruitment failure related to increasing instability of the substrata (Perkol-Finkel & Airoldi 2010). Assisted re-introduction or translocation can facilitate recovery of damaged populations (Lotze et al. 2011). Therefore, developing simple techniques to grow C. barbata on suitable habitats, either natural or artificial, could enhance the recovery potential of locally endangered populations of this species.

Transplanted juveniles showed no consistent survival patterns that relate to substratum orientation. This suggests that coastal infrastructures such as seawalls, breakwaters and pilings could provide potentially adequate habitats despite the greater proportion of inclined surfaces compared to natural habitats (Bulleri, Chapman & Underwood 2005; Chapman & Blockley 2009). Moreover, the survival of transplants was not impaired by lack of surrounding adults, suggesting that this would not be a limiting factor when managing assemblages on newly built man-made infrastructures that would obviously lack adult canopies.

Transplantation success was greater on landward, sheltered sides compared to exposed seaward sides of the breakwaters. This is in agreement with findings from Jonsson et al. (2006) who demonstrated that the higher flow speed on seaward compared to landward sides of breakwaters induced greater dislodgment of fucoid macroalgae. Indeed, the different sides of marine infrastructures provide distinct habitats for the growth of a variety of macroalgae and invertebrate species (e.g. Bacchiocchi & Airoldi 2003; Bulleri & Airoldi 2005; Burt et al. 2009). This ecological characteristic of many coastal infrastructures must be considered if we are to design and manage these structures for achieving desired secondary management goals and for enhancing their contribution to local biodiversity and ecosystem functions.

While several transplants survived and reproduced for over a year post-transplantation, we did not establish a substantial self-sustaining population (which was beyond the scope of the current research). Nonetheless, as most of the receiving habitats had relatively high levels of survivals several months after deployment, much of the mortality can be attributed to rough sea conditions and not as an immediate reaction to the transplantation procedure. Moreover, our transplantation efforts were limited in scale and only small sized recruits were transplanted. Future work should explore whether a larger-scale transplantation effort onto sheltered portions of coastal infrastructures would be self-sustaining and whether using larger transplants would increase their survival and thus facilitate establishment of viable populations.

While transplantation of C. barbata proved successful onto coastal infrastructures along a rocky coastline, survival was not as promising when the structures were located along sedimentary coastlines, a typical setting of many coastal defence infrastructures (Airoldi et al. 2005b). The results of the caging experiment suggested that lack of survival of C. barbata transplants along sedimentary coastlines was not related to environmental factors (e.g. reduced water quality or excess sedimentation). Instead biotic disturbance was a determining factor limiting the survival of C. barbata in these habitats. Our tests with cages of different mesh sizes initially suggested that such biotic disturbance could be related mainly to the activity of small organisms (0·1–1 cm). However, preliminary results of ongoing experimental work by our group (aiming at clarifying the nature, distribution and generality of such biotic disturbance with the aid of underwater video cameras) suggest that loss of Cystoseira at some structures is related to a complex of both consumptive and non-consumptive disturbance by a variety of organisms of different sizes, ranging from small crabs to mullets (F. Ferrario, L. Ivesa, S. Perkol-Finkel, A. Jacklin & L. Airoldi, unpublished data). Some of these organisms are also present at natural rocky sites, but at lower densities and they do not show the same degree of interaction with the Cystoseira. Coastal infrastructures set on sedimentary shorelines represent ‘oasis’ of hard bottoms in a soft bottom environment (Airoldi et al. 2005a). As such, they might attract a greater abundance of predators compared to nearby natural habitats, similarly to what is thought to occur in other oasis systems such as seamounts (Rowden et al. 2010a,b). This unexplored aspect of the ecology of marine infrastructures deserves further attention.

Substratum composition and complexity have a strong influence on settlement, recruitment and survival of benthic fauna and flora in both natural (Harlin & Lindbergh 1977; Wells, Moll & Bolton 1989; Johnson & Brawley 1998; Guarnieri et al. 2009) and artificial substrata (Spieler, Gilliam & Sherman 2001; Chapman 2003; Burt et al. 2009). Coastal infrastructures such as seawalls, breakwaters and jetties may be constructed of stone, concrete, wood, steel or geotextiles (Dugan et al. 2011) and may be designed to incorporate greater habitat complexity (Moreira, Chapman & Underwood 2007; Chapman & Blockley 2009). For example, subtle change in infrastructure complexity, at small scale (e.g. addition of pits to a seawall as in Martins et al. (2010) or medium scale (e.g. addition of holes to concrete wave energy foundations as in Langhamer & Wilhelmsson (2009) can significantly increase the ability of the infrastructure to sustain greater abundance of organisms. Our tests with concrete and limestone (the most common materials in our study region) and clay (a potentially practical substratum for transplantation) showed similarly high levels of recruitment. Settlement was initially double on surfaces with medium complexity in comparison with simple or highly complex ones, yet this facilitation was apparently transient, probably due to post-settlement processes related to natural self-thinning (Reed 1990; Kendrick & Walker 1995; Johnson & Brawley 1998). Moreover, it is possible that engineering species like Cystoseira (Sala & Ballesteros 2012) modify their immediate environment once established in terms of, for example, hydrodynamic and/or sedimentation patterns, thus masking further effect of complexity. This suggests that the artificial substrata in the study area provide potentially suitable substrata for this canopy-forming alga and that other biological or ecological factors (such as those suggested by the caging experiment) limit its natural recruitment on the infrastructures.

Understanding how man-made habitats function in urbanized seascapes is fundamental if we are to design and manage these habitats in a way that enhances their contribution to marine biodiversity and flow of ecosystem services (Airoldi et al. 2005a; Chapman & Blockley 2009; Bulleri & Chapman 2010). We demonstrate that managing assemblages on marine infrastructures for desirable secondary management goals can be feasible, but requires a good understanding of the different ecology of these artificial systems. This is in agreement with Moschella et al. (2005) concluding that infrastructures can be modified to influence the abundance and species composition of epibiota to achieve desired management goals such as controlling growth of nuisance algae or promoting diversity of habitats and species for recreational activities. This emerging approach complements the evolving paradigm of ecological engineering (Mitsch 1996), aimed at integrating ecological, economic and social needs into the design of man-made ecosystems.

In conclusion, the current study contributes to bridging the gap between growing societal needs of coastal development and the need for conserving the marine environment (Mitsch 1996; Chapman & Blockley 2009; Inger et al. 2009; Bulleri & Chapman 2010). The ability to utilize coastal infrastructures as scaffolds for recovery of threatened species or for enhancement of desirable species has important applications for the conservation of biodiversity in globally expanding coastal urban environments. For example, current restoration or enhancement efforts based on the construction of artificial reefs (Reed et al. 2006; Schmidt et al. 2007; Dupont 2008) could be best replaced by utilizing existing infrastructures. This approach could be more sustainable in the long term, and be efficiently incorporated into marine spatial planning.


The work was supported from the EU projects MarUrbe (EU – FP7-PEOPLE-2007-2-1-IEF-219818) and THESEUS (EU – FP7 – ENV2009-1, grant 244104), and from project AdriaBio (University of Bologna, 2007–2009). Further support to S.P.F. from the Sussman Family Center for the Study of Environmental Sciences at The Weizmann Institute of Science and to L.A. from the exchange programme E.C.C.O between Wesleyan University and Bologna University. We thank the Civil Defence and the Port Autorities of Numana, Ancona and Marotta for continuous support, and Roberto Balzano, Giulio Franzitta, and Giovanni Fontana, for their valuable assistance in the field. We also thank Lisandro-Benedetti Cecchi, Gustavo Martins and one anonymous reviewer who greatly improved the final version of this manuscript.