at SciVerse ScienceDirect

Marine Environmental Research 77 (2012) 60e70

Contents lists available

Marine Environmental Research

journal homepage: www.elsevier .com/locate /marenvrev

Patterns of variation of intertidal species of commercial interest in the ParqueLitoral Norte (north Portugal) MPA: Comparison with three reference shores

Iacopo Bertocci a,*, Rula Dominguez a, Cristiano Freitas a, Isabel Sousa-Pinto a,b

aCIIMAR/CIMAR, Centro Interdisciplinar de Investigação Marinha e Ambiental, Rua dos Bragas, 289, 4050-123 Porto, PortugalbDepartment of Biology, Faculty of Sciences, University of Porto, Rua do Campo Alegre s/n, 4169-007 Porto, Portugal

a r t i c l e i n f o

Article history:Received 20 December 2011Received in revised form1 February 2012Accepted 5 February 2012

Keywords:MPAsRocky intertidalSpatial and temporal heterogeneityParacentrotus lividusMytilus galloprovincialisManagementPortugal

* Corresponding author. Tel.: þ351 223401818; faxE-mail address: ibertocci@ciimar.up.pt (I. Bertocci

0141-1136/$ e see front matter � 2012 Elsevier Ltd.doi:10.1016/j.marenvres.2012.02.003

a b s t r a c t

Marine Protected Areas (MPAs) are world-wide established with the aim of conserving biodiversity andpreventing overexploitation of marine organisms. Evaluating the effectiveness of MPAs is needed inorder to support and implement their management, but it is complicated by the large natural variabilityin space and time of distribution and abundance of natural populations. Here, we tested the hypothesisthat patterns of total abundance and size-frequency distribution of two intensively harvested intertidalspecies (the sea urchin Paracentrotus lividus and the mussel Mytilus galloprovincialis) differed betweena protected and three reference shores along the rocky coast of north Portugal. Response variables werein terms of mean values and measures of variance at different spatial scales (from centimetres to metres)and over time (along a period of about 12 months). A further comparison involved the estimation of thereproductive potential of sea urchins, quantified as variations of Gonad Index (GI ¼ gonad dry weight/body dry weight � 100) at the scale of shore. Results did not generally support a predictable direct effectof protection, as the total abundance and the abundance of larger individuals of both species and GI didnot differ between the MPA and reference shores. However, a considerable temporal and spatial vari-ability at smaller scales was detected for several response variables. Such findings have implications formanagement of MPAs, highlighting the need for sampling designs properly replicated in space and time,in order to examine their effectiveness, and for considering spatial and temporal heterogeneity of targetpopulations and driving processes as a criterion for their implementation and design.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

The impacts of human activities in coastal areas are of increasingconcern for ecologists, policy-makers and the general public(Jackson et al., 2001; Lotze et al., 2006). Anthropogenic distur-bances can modify the physical environment and directly andindirectly affect patterns of distribution, abundance and diversity ofpopulations and assemblages on a global scale (Vitousek et al.,1997). Removing target and non-target species and habitats canalso indirectly alter food webs and biological interactions (Fogartyand Murawski, 1998; Sala et al., 1998a; Guidetti, 2006). Asa consequence, there is an increasing need for effective manage-ment measures aimed at conserving the diversity of species andhabitats and ensuring a sustainable use of living resources(Sherman, 1994). The implementation of marine protected areas(MPAs) is a common tool aimed at achieving these goals (Agardy,1994; Allison et al., 1998; Roberts et al., 2001; Rodrigues et al.,

: þ351 223390608.).

All rights reserved.

2004), through the preservation of stocks of commercial species,the conservation of natural habitats and the protection of speciesconsidered of particular importance due to their functional role(e.g. King and Beazley, 2005 and references therein) or just theircultural or aesthetic value (Palumbi, 2001).

The effectiveness of marine reserves has been widely assessedmostly by comparing patterns of density or size of exploited fishspecies among protected and reference areas (Dayton et al., 2000;García-Charton et al., 2000). A number of previous studies havedocumented increases in abundance and/or size of target speciesinside MPAs compared with unprotected areas (for reviews, seeHalpern, 2003; Pelletier et al., 2005; García-Charton et al., 2008). Inaddition, indirect effects, such as trophic cascades, within or aroundMPAs have been hypothesized and in some cases supported byempirical studies (Lindberg et al., 1998; Castilla, 1999; Pinnegaret al., 2000; Sala et al., 1998a; Shears and Babcock, 2002; Micheliet al., 2005). Such effects depend in complex ways on life-traits,such as dispersal abilities, of species (McClanahan and Mangi,2000) and features of the reserve, including size, age and design(e.g. Botsford et al., 2003; Claudet et al., 2008).

I. Bertocci et al. / Marine Environmental Research 77 (2012) 60e70 61

Intertidal benthic assemblages on rocky shores are particularlyexposed to human activities, such as harvesting gastropods andalgae for commercial purposes (Hockey and Bosman, 1986; Castillaand Bustamante, 1989; Lasiak, 1998; Castilla, 1999), removingspecies for bait (Kingsford et al., 1991) and collecting shellsaesthetically appealing (Underwood, 1993a), and to the effects oftrampling associated to them (e.g. Brosnan and Crumrine, 1994).Analogously to findings from subtidal assemblages, it has beendocumented that the removal of predators can drastically affectintertidal interaction webs and thus deeply alter patterns ofbiodiversity of rocky shores (Hockey and Branch, 1984; Lindberget al., 1998; Castilla, 1999). In other cases, however, no significanteffects of MPAs were indicated on intertidal organisms (e.g. Lasiak,1998). Actually, the effects of MPAs are generally variable indirection and scale, so that examining such heterogeneity and itsbasic mechanisms is key to improve our knowledge of the effec-tiveness of existing MPAs and to implement the design of presentand planned ones (Halpern and Warner, 2002).

Comparing the mean abundance of populations between pro-tected and reference areas is complicated by the large variability inpatterns of distribution and abundance of natural populations atdifferent spatial and temporal scales, rising a number of issuesconcerning tests of hypotheses about protection (Schneider, 1994;Underwood and Chapman,1996; Benedetti-Cecchi et al., 2003). Thecentral point involves the use of sampling design able to separatethe real effects of protection from other uncontrolled sources ofvariation (e.g. García-Charton and Pérez-Ruzafa, 1999; García-Charton et al., 2000; Fraschetti et al., 2005). In addition, effectivemanagement in terms of number, size and design of MPAs shouldspecifically take into account the spatial and temporal variability ofspecies, in order to guarantee the protection of a representativesample of species and assemblages of a region and themaintenanceof the ecological processes that are responsible for such patterns(Horne and Schneider, 1995; Underwood and Chapman, 1996;Schwartz, 1999; Thrush et al., 2000; Benedetti-Cecchi et al., 2003).This requires comparing MPAs and reference areas not only interms of mean density and/or size of populations, but also in termsof patterns of variance at different scales, an approach rarely foundin the ecological literature (but see Benedetti-Cecchi et al., 2003).

The issues above were addressed in the present study on pop-ulations of two species of commercial interest, the sea urchin Par-acentrotus lividus and the mussel Mytilus galloprovincialis, sampledin an MPA and three reference shores in north Portugal. Bothspecies are among the most exploited animals from rocky coastsand shallow subtidal habitats for commercial and recreationalpurposes. P. lividus is subject to intense harvesting particularly inthe Mediterranean Sea (Guidetti et al., 2004; Ceccherelli et al.,2009, 2011), on Atlantic coasts of France and Iberian peninsula(Barnes and Crook, 2001) and in Ireland (Byrne, 1990) due to itsgonads, that represent an appreciated delicacy of high marketvalue. Analogously, there is a long history of harvesting M. gallo-provincialis for food and bait (e.g. de Lumley, 1975) and this speciesis currently severely exploited, particularly in some periods, incountries such as Italy (Airoldi et al., 2005), Spain (Rius and Zabala,2008) and Portugal (Rius and Cabral, 2004). Effects of harvestrestrictions on populations of P. lividus (Guidetti et al., 2005;Gianguzza et al., 2006; Guidetti, 2007; Pais et al., 2007; Ceccherelliet al., 2009, 2011) and M. galloprovincialis (Rius and Zabala, 2008)have been examined only in recent years, providing contrastingfindings. These ranged from increases in total abundance(Gianguzza et al., 2006) or in the abundance of large individuals(Rius and Cabral, 2004; Ceccherelli et al., 2011) of the exploitedspecies in protected compared to reference conditions, to theopposite pattern (Pais et al., 2007), to complex indirect effectsmediated by biological and abiotic factors. These included

predators (Sala and Zabala, 1996; Sala, 1997; Guidetti et al., 2005),wave exposure (Micheli et al., 2005), the design of the reserve (Riusand Zabala, 2008) and interactions between biological and physicaldrivers (Hereu et al., 2005). Therefore, examining the effects ofprotection on populations of such species in unstudied areas is ofoverwhelming commercial and ecological importance.

Here, we compared patterns of abundance and distribution ofP. lividus and M. galloprovincialis between the Parque Litoral NorteMPA and reference shores in north Portugal. Although the directionof possible effects of protection was difficult to predict due to therange of contrasting direct and indirect effects of different factorslikely driving such patterns in the studied system, we tested thegeneral hypotheses that patterns of total abundance (number ofindividuals for sea urchins and percentage cover for mussels) and ofdistribution of individuals among size classes of both species at theMPA differed from those at similar rocky shores located in the samegeographical area, but not subject to the same managementmeasures. These hypotheses were examined both in terms of meanvalues of response variables and their temporal and spatial vari-ances (Warwick and Clarke, 1993; Chapman et al., 1995), by meansof a sampling repeated at different times and at a hierarchy ofspatial scales (Underwood, 1992, 1994), from 10 s centimetres(among replicate quadrates) to 10 s metres (between sites in eachshore) at each of the protected and the reference shore (kilometresapart). In addition, we tested the hypothesis that the reproductivepotential of P. lividus differed between the protected and thereference condition at the scale of the shore. This was based on theknowledge that, for instance, reductions of the reproductive outputoccur under large densities, and consequent food limitation, of seaurchins (e.g. Levitan, 1995; Tomas et al., 2005), although fertiliza-tion rates can be positively related with the density of adults (e.g.Whale and Peckham, 1999).

2. Methods

2.1. Study site

The studywas carried out between October 2010 and September2011 in rocky intertidal habitats of a protected and three referenceshores (stretches of coast 100 s m to 1 km long, 5e10 km apart) innorth Portugal. The tidal regime along the Portuguese coast issemidiurnal, with the largest spring tides of 3.5e4 m. The typicalrocky shore is granitic and the coast is exposed to prevailingnorthwest oceanic swells, particularly intense during the winter(Araújo et al., 2005). The protected shore (Mar: 41� 340 5800 N, 08�

480 1900 W) is located within Parque Litoral Norte (PLN), an MPAestablished in 1987 on about 18 km of coast and the adjacent sea,between the city of Apulia in the south to the estuary of the Neivariver in the north. The sampled shore is the largest stretch of rockycoast available within PLN, as the rest of the coast includes just verysmall rocky areas, interspersed among large sandy beaches.Although PLN does not include no-take areas (see http://portal.icnb.pt/ICNPortal/vPT2007-AP-LitoralNorte/HomepageþAreasþProtegidas.htm for detailed information on legislation, organizationandmanagement) all fishing and harvesting activities there requirespecial permits, that are provided only to a small number of locallyresident professionals (to our knowledge, only two licences wereprovided to professional harvesters of sea urchins within PLN andthese are represented by small family units; on the contrary, shoresout of PLN are visited also by professional and recreationalharvesters from nearby areas, such as Galicia in Spain) using, due toboth cultural and law constraints, artisanal techniques. Fishing, inparticular, is mostly based on family activities using trammel nets,longlines and traps as typical gear, whose size and number arelimited by the reduced size of boats and involved people (two

I. Bertocci et al. / Marine Environmental Research 77 (2012) 60e7062

persons per boat in most cases). The use of any tool for harvestingintertidal organisms is prohibited by law, making this activity inprinciple limited to removal by hand during daytime and low tideonly. Enforcement is carried out by a special police sections formarine issues (Polícia Maritima) and by local port authorities(Capitanias do Porto) with recurrent visits at sea and along theshore. The artisanal trait of these activities, however, is associatedto a lack of proper records on collected quantities of organisms, as,for instance, no official devices for recording fished quantities arerequired on board of small boats and most fish and intertidalorganisms are directly sold to consumers (i.e. restaurants andprivate citizens without entering large and better ruled markets).The three reference shores (Amorosa: 41� 380 3000 N, 08� 490 2300 W;Praia Norte: 41� 410 5700 N, 08� 510 1300 W;Moledo: 41� 500 2500 N, 08�

520 2700 W) were randomly selected from those available alongabout 25 km of coast north of PLN. The coast south of PLN is almostexclusively sandy for more than 60 km, preventing to interspersereference shores between both geographic sides of the MPA.Reference shores outside theMPAwere selected as being analogousto the protected one in terms of a number of macroscopic physicaltraits, including spatial extent (i.e.100 sm long), type of substratum(i.e. mostly granite), slope (i.e. almost flat), exposure (i.e. at allshores, the coastline was oriented almost exactly form north tosouth, being exposed to the same prevailing westerly swell andwinds) and accessibility (i.e. all shores were located at comparabledistance from the main road and easily accessible walking fromnear parkings). Both the protected and the reference shores aretypically dominated by mussels (M. galloprovincialis Lamarck) atmid intertidal level and by diverse algal and invertebrate assem-blages lower on the shore, where most sea urchins (P. lividusLamarck) can be found (personal observation). In Portugal, P. lividushas been commercially harvested mostly in the last 20 years asa consequence of the reduction of stocks and increased demandfrom nearby markets (FAO, 2004). Harvesting occurs at low tidemostly in autumn and spring months, when the gonads attain theconsistency, texture and colour suitable for marketing. Professionalharvesting is mostly regulated in terms of number of licences, withloose limitations concerning allowed quantities (i.e.50 kg � person�1 � day�1, to be collected by hand only) andminimum size (i.e. 50 mm in test diameter, established only in2011). Recreational activities are generally not subject to specificlaws. Harvesting of M. galloprovincialis is generally less intense andlimited to family activities for food or collection by sports fish-ermen for bait (Rius and Cabral, 2004). However, it is not specifi-cally regulated at all, being just subject to the general laws rulingrecreational fishing.

2.2. Sampling design and collection of data

A hierarchical sampling design was used to compare patterns ofabundance and distribution of P. lividus and M. galloprovincialisbetween the protected and the reference shores over time and atnested spatial scales. In October 2010, two sites (stretches of shore3e5m long, 10 s m apart) were randomly selected at each protectedand reference shores at low (between 0.1 m and 0.5 m above ChartDatum: P. lividus sampling) or mid (between 1.5 m and 2.0 m aboveChart Datum: M. galloprovincialis sampling) intertidal height, corre-sponding to the shore levels were sea urchins and mussels were,respectively, more abundant. Five quadrates (50 � 50 cm or25 � 25 cm for sea urchins and mussels, respectively) 10 s cm to1e2m apart were randomly sampled at each site, in barren areas, i.e.areas encompassing refuges for sea urchins and dominated byencrusting algae, likely due to grazing (e.g. Bulleri et al., 2002), oremergent rock, respectively. According to the size of sites, the chosennumber of replicate quadrates was considered appropriate to have

a representative sampling of variability at the scale of site. The size ofquadrates was considered appropriate depending on the size oforganisms and were in the range of spatial resolution of experi-mental studies on sea urchins (e.g. Bulleri et al., 2002) and sessileintertidal organisms, including mussels (e.g. Airoldi et al., 2005). Inthe field, all individuals of P. lividus in each quadrate were countedand the test diameter, excluding spines, of each individual wasmeasured with a plastic caliper to the nearest mm. Such measureswere used to assign individuals to each of six size classes (Class1 <10 mm, 10 � Class 2 < 20, 20 � Class 3 < 30, 30 � Class 4 < 40,40 � Class 5 < 50 and Class 6 � 50 mm).

The percentage cover of M. galloprovincialis in each quadratewas estimated in the field by means of a quadrate divided into 25sub-quadrates of 5 � 5 cm each, by assigning to each sub-quadratea score from 0 (mussels absent) to 4 (whole sub-quadrate covered)and then adding up the 25 estimates (Dethier et al., 1993). Logisticdifficulties prevented to measure mussels individually in the field,thus the mussel bed from each quadrate was removed manuallywith a paint scraper, put in separate bags, preserved in ice andtransported to the laboratory for subsequent measures. Measuringeach mussel individual from each replicate was practically impos-sible due to the large numbers occurring in each quadrate. There-fore, a sub-sampling procedure was adopted in order to allowcomparisons of protected and reference shores in terms of size classdistribution of mussels. Mussels from each quadrate were defrozenat ambient temperature and superimposed on a quadrate of thesame size divided into 25 sub-quadrates (5 � 5 cm each). Then, allthe individuals partially or completely covering each of fiverandomly selected sub-quadrates were carefully removed withtweezers, the length of their shell was measured (�0.1 mm) witha caliper and each individual was assigned to each of seven sizeclasses of shell length (Class 1 < 3 mm, 3 � Class 2 < 10, 10 � Class3 < 20, 20 � Class 4 < 30, 30 � Class 5 < 40, 40 � Class 6 < 50 andClass 7 � 50 mm; Class 1 corresponded to the smallest size ofmussels that was possible to handle with a reasonable effort,depending on tools and time needed to individually measuremussels smaller than 3 mm). Data from the five sub-quadrates ineach quadrate were summed, separately for each size class. Weassumed that such a procedure, although not able to build a fullfrequency distribution of sizes, was suitable for the purpose oftesting the proposed hypothesis on differences in the size classdistribution of M. galloprovincialis in the studied system. Anysampling error associated to the procedure, in fact, would havebeen equally present in all samples, likely not affecting, in partic-ular, the relative differences between protected and referenceshores.

An additional destructive sampling was carried out with the aimof comparing the reproductive potential of P. lividus individualsmostly targeted by human harvesting at the scale of the protectedvs. reference shores. This included haphazardly collecting twentyindividuals out of the largest ones found at each shore. These werepreserved in ice and transported to the laboratory. Here, eachindividual was defrozen, opened with a scalpel and the gonadscarefully separated from the test with tweezers. Both the test(body) and the gonads of each individual were dried at 50 �C for48 h, then their weights measured to the nearest mg. A GonadIndex (GI) aimed at quantifying the relative investment of eachindividual to reproductive compared to somatic structures wasthen calculated as: GI ¼ gonad dry weight/body dry weight � 100(e.g. Byrne, 1990).

Each sampling was repeated with the same procedures at eachof six times (in addition to October 2010, January, February, April,June and September 2011) randomly selected to capture the naturalpatterns of variation of the examined response variables overa period of twelve months.

I. Bertocci et al. / Marine Environmental Research 77 (2012) 60e70 63

2.3. Analysis of data

Differences in response variables between the protected andreference shores were examined with asymmetrical analyses ofvariance (ANOVAs), following a procedure analogous to thatproposed for the analysis of environmental impacts (Underwood,1992, 1993b, 1994; Glasby, 1997; see also Terlizzi et al., 2005). Athree-way model was used to examine patterns in the total numberof individuals of P. lividus, the percentage cover ofM. galloprovincialisand the abundance of individuals of both species in each size class(with the exclusion of P. lividus Class 6 andM. galloprovincialis Class 6and 7, that were generally represented by very few or no individualsand were not formally analyzed). This included the factors Time (sixlevels, random), Shore (four levels, random, crossed with Time) andSite (two levels, random,nestedwithin the interactionTime� Shore),with five replicates in each combination of levels of such factors. TheShore termwas partitioned into the one degree of freedom contrastMPA vs. Reference shores (note that this was the only fixed source ofvariability in the design) and the variability among Reference shores(following the logic described by Underwood, 1992). TheTime � Shore interaction was partitioned into a Time � MPA anda Time� Reference shores interaction. The total variability of the Siteterm was partitioned into Site (Time � MPA) and Site (Referenceshores). Finally, the overall residual variability was partitioned intothe variability among replicate quadrateswithin theMPA [Quadrates(MPA)] and the variability among replicate quadrates within Refer-ence shores [Quadrates (Reference shores)].

F-tests for ANOVAs were built by choosing the appropriatedenominators from expected mean squares (Underwood, 1997; seealso Terlizzi et al., 2005 for a detailed description on how to chooseproper F denominators for asymmetrical designs). Sources of vari-ation individually involving MPA or reference shores were testedover the natural denominator for that term. For example, the site(Time � MPA) term was tested over the quadrates (MPA) term,instead of the overall residual. The overall residual variation wasnot used as denominator for this test as it would have assumed thatthe variance among quadrates at the protected and at the referenceshores was the same, which could not occur if protection affectssmall-scale patterns of variability of organisms.

The analysis on GI was based on a two-waymodel, including thecrossed factors Time and Shore, with the same number of levels and

Table 1Summary of ANOVAs and two-tailed F-tests for the abundance of P. lividus (number of indand reference (Amorosa, Praia Norte and Moledo) shores over six times of sampling. *P <

indicates variances not significantly different.

Source of variation df P. lividus

MS

Time 5 971.24Shore 3 1329.26MPA vs Reference shores 1 2494.97Reference shores 2 746.41

Time � Shore 15 944.84Time � [MPA vs Reference shores] 5 1224.79Time � Reference shores 10 804.87

Site (Time � Shore) 24 329.71Site (Time � MPA) 6 337.28Site (Time � Reference shores) 18 327.18

Residual 192 189.81Quadrates (MPA) 48 214.55Quadrates (Reference shores) 144 181.56

Cochran’s test C ¼ 0.107, nsTransformation None

Variance components and two-tailed F-test: MPA

Quadrates 214.55 ¼Sites 24.55 ¼Time � Shore 88.75 ¼

nature previously described, and their interaction. The Shore term,the Time � Shore interaction and the overall residual variationwere partitioned analogously to the previous analysis. The twentyindividuals sampled at each shore at each time provided thereplicates for this analysis.

The assumption of homogeneity of variances was checkedbefore each ANOVA using Cochran’s C test. In some cases, datawereln(x þ 1) transformed to remove the heterogeneity of variances.When this was not possible, untransformed data were analyzedand results were considered robust if not significant (at P < 0.05),or, to compensate for the increased probability of type I error,significant at P < 0.01 (Underwood, 1997).

Two-tailed F-tests were used to compare the spatial variabilityat the scale of quadrates, sites and Time � Shore interactionbetween the MPA and the three reference shores (Underwood,1992). These tests were based on variance components calculatedfrom ANOVA by equating the empirical and the expected meansquares (Winer et al., 1991; Underwood, 1997). When negativeestimates of variance were obtained, these were interpreted assample underestimates of very small or null variances (Searle et al.,1992) and set to zero (Underwood, 1997). Variance componentswere calculated from untransformed data (e.g. Terlizzi et al., 2005).

3. Results

3.1. P. lividus

The total abundance of P. lividus did not differ between the MPAand reference shores, consistently with the time of sampling, whilesignificant differences were documented among the referenceshores and between sites sampled within them (Table 1, Fig. 1).Patterns of variation at both scales of quadrates and sites andtemporal patterns at the scale of shores for this response variablewere also comparable between the protected and referencecondition (Table 1).

The protected and reference shores differed, consistentlythrough time, for the abundance of individuals of size Class 4(Table 2). This class generally included the relative largest number ofindividuals and at each time of sampling it was more numerous atthe MPA than, on average, at the reference shores (Fig. 2). Both sizeclasses including the smallest individuals were characterized by

ividuals) andM. galloprovincialis (percentage cover) estimated at the protected (Mar)0.05, **P < 0.01, ***P < 0.001, ns: not significant (at P > 0.05). Two-tailed F tests: “¼”

M. galloprovincialis Denominator for F

F MS F

1.03 392.82 0.15 Time � Shore1.41 2756.05 1.07 Time � Shore2.64 2904.05 1.13 Time � Shore0.93 2682.05 1.07 Time � Ref. shores2.87* 2572.42 2.57* Site (Time � Shore)1.52 2683.05 2.68 Time � Ref. shores2.44* 2517.11 2.76 * Site (Time � Shore)1.74* 1000.31 3.59*** Residual1.57 1266.18 4.20** Quadrates (MPA)1.80* 911.68 3.37*** Quadrates (Ref. shores)

278.50301.61270.79

C ¼ 0.093, nsNone

Ref. MPA Ref.

181.56 301.61 ¼ 270.7929.12 192.92 ¼ 128.1847.77 141.69 ¼ 160.54

Site

1

Num

ber

of in

divi

dual

s x

2500

cm

-2

0

10

20

30

40

50

60

70

80

0

10

20

30

40

50

60

70

80

0

10

20

30

40

50

60

70

80

Site

2

Num

ber

of in

divi

dual

s x

2500

cm

-2

0

10

20

30

40

50

60

70

80

0

10

20

30

40

50

60

70

80

0

10

20

30

40

50

60

70

80

Site

1Si

te 2

Site

1Si

te 2

Site

1Si

te 2

Site

1Si

te 2

Site

1Si

te 2

Site

1Si

te 2

Site

1Si

te 2

Site

1Si

te 2

Site

1Si

te 2

Site

1Si

te 2

Site

1Si

te 2

Mar Amorosa P. Norte Moledo Mar Amorosa P. Norte Moledo Mar Amorosa P. Norte Moledo

1102yraurbeF1102yraunaJ0102rebotcO

1102rebmetpeS1102enuJ1102lirpA

Fig. 1. Mean (þSE) number of individuals of P. lividus at the protected (Mar) and three reference (Amorosa, PraiaNorte andMoledo) shores. Data averaged over five replicate quadrates.

I. Bertocci et al. / Marine Environmental Research 77 (2012) 60e7064

temporally variable differences between the MPA and referenceshores, as indicated by the significant interaction Time � [MPA vsReference shores] (Table 2). In particular, Class 1 was absent at theMPA and present, although with very low abundances, at thereference shores at all sampling times between October 2010 andJune 2011, while it was completely absent at all shores in September2011 (Fig. 2). Class 2 was less represented at the MPA than, onaverage, at the reference shores in October 2010, January 2011 and,

Table 2Summary of ANOVAs and two-tailed F-tests for the number of individuals of P. lividus in ea4 < 40 and 40 � Class 5 < 50) estimated at the protected (Mar) and reference (Amoros***P < 0.001, ns ¼ not significant (at P > 0.05). Two-tailed F tests: “¼” as in Table 1; no sywere set to zero. Note that when the full dataset was transformed, the same transformpartitioning of variance, independently of whether variances of this reduced dataset wereMPA vs Reference shores term for Class 1, Class 2 and Class 3, which was tested over the

Source of variation df Class 1 Class 2

MS F MS F

Time 5 3.14 1.64 3.32 3.25Shore 3 12.25 6.38** 14.97 14.70MPA vs Referenceshores

1 2.45 0.14 4.87 0.24

Reference shores 2 17.15 6.70* 20.03 19.83Time � Shore 15 1.92 0.94 1.02 1.18Time � [MPA vsReference shores]

5 0.63 0.25 1.04 1.03

Time � Referenceshores

10 2.56 1.25 1.01 1.17

Site (Time � Shore) 24 2.04 1.83* 0.86 2.27Site (Time � MPA) 6 0.08 0.67 0.14 0.33Site (Time � Referenceshores)

18 2.69 1.86* 1.10 3.03

Residual 192 1.12 0.38Quadrates (MPA) 48 0.13 0.43Quadrates (Referenceshores)

144 1.45 0.36

Cochran’s test C ¼ 0.619, P < 0.01 C ¼ 0.064, nsTransformation None Ln(x þ 1)

Variance components and two-tailed F-test: MPA Ref. MPA Re

Quadrates 0.13 <1.45*** 6.06 <18.9Sites 0.00 0.25 0.00 7.5Time � Shore 0.06 0.00 2.77 >0.1

less evidently, February and April 2011, while it included compa-rable numbers of individuals under both conditions at the last twotimes of sampling (Fig. 2). Significant differences between the MPAand reference shores at the scale of sitewere documented for Class 3and Class 4 (Table 2), with a larger variability between protectedcompared to reference sites that was particularly evident in October2010, January 2011 and April 2011. The variance at the smallest scaleexamined (among quadrates) was smaller within the MPA than

ch of five size classes (Class 1<10mm,10� Class 2< 20, 20� Class 3< 30, 30� Classa, Praia Norte and Moledo) shores over six times of sampling. *P < 0.05, **P < 0.01,mbols reported in all cases where no test was done as negative estimates of varianceation was applied to the dataset of reference shores only used to do the indicatedhomogenous or not. Denominators for F tests as in Table 1, with the exception of theReference shores term as this was significant.

Class 3 Class 4 Class 5

MS F MS F MS F

* 4.34 5.23** 2.54 2.95* 1.20 0.41*** 5.24 6.31** 6.86 7.98** 1.78 0.61

0.11 0.01 19.61 22.80** 0.02 0.01

*** 7.80 7.65** 0.49 0.55 2.67 1.060.83 0.78 0.86 1.16 2.94 2.12*0.45 0.44 0.80 0.90 3.80 1.51

1.02 0.96 0.89 1.20 2.51 1.81

** 1.06 2.37*** 0.74 1.76* 1.39 2.36***2.19 3.46** 1.16 3.54** 0.92 1.64

*** 0.68 1.77* 0.60 1.33 1.54 2.59***

0.45 0.42 0.590.63 0.33 0.560.39 0.45 0.60

C ¼ 0.089, ns C ¼ 0.101, ns C ¼ 0.047, nsLn(x þ 1) Ln(x þ 1) Ln(x þ 1)

f. MPA Ref. MPA Ref. MPA Ref.

7*** 169.21 >33.00*** 155.07 >78.14** 11.36 <19.44 *5 30.08 >3.56*** 43.50 >3.70*** 3.99 ¼ 3.937*** 0.00 4.92 0.00 7.51 7.56 ¼ 3.77

Fig. 2. Mean (þSE) number of individuals of P. lividus in each of six size classes (from 1 to 6: Class 1 < 10 mm, 10 � Class 2 < 20, 20 � Class 3 < 30, 30 � Class 4 < 40, 40 � Class5 < 50 and Class 6 � 50 mm) at the protected (Mar: ) and three reference (Amorosa: , Praia Norte: and Moledo: ) shores. S1: site 1; S2: Site 2. Data averaged over fivereplicate quadrates.

I. Bertocci et al. / Marine Environmental Research 77 (2012) 60e70 65

within the reference shores for Class 1, Class 2 and Class 5, while theopposite pattern was shown by Class 3 and Class 4 (Table 2).

The MPA did not differ from reference shores for the GonadIndex (GI), while significant differences, not consistent throughtime, were detected among reference shores. This was indicated bythe significant interaction Time � Reference shores (Table 3).Nevertheless, it was graphically observable a tendency towardsrelatively larger GI values at the MPA in October 2010, April 2011,September 2011 and, less evident, January 2011,and comparablevalues in February and June 2011 (Fig. 3).

3.2. M. galloprovincialis

Similarly to the abundance of P. lividus, no main or temporallydependent differences between theMPA and reference shores werefound for the percentage cover of M. galloprovincialis (Table 1,

Table 3ANOVA for the Gonad Index (GI) of P. lividus estimated at the protected (Mar) andreference (Amorosa, Praia Norte and Moledo) shores over six times of sampling.*P < 0.05, **P < 0.01, ***P < 0.001, ns ¼ not significant (at P > 0.05).

Source of variation df MS F Denom. for F

Time 5 102.42 0.69 Time � ShoreShore 3 546.72 3.67* Time � ShoreMPA vs Referenceshores

1 887.49 82.10*** Time � Shore

Reference shores 2 376.33 2.07 Timex � Ref. sh.Time � Shore 15 148.80 10.81*** ResidualTime � [MPA vsReference shores]

5 83.43 4.76 Indiv.(MPA)

Time � Referenceshores

10 181.48 14.51*** Indiv.(Ref. sh.)

Residual 456 13.77Individuals (MPA) 114 17.54Individuals(Reference shores)

342 12.51

Cochran’s test C ¼ 0.348, P < 0.01Transformation None

Fig. 4). Effects of protection on this response variable, however,were documented at the scale of sites, with measures of spatialvariance at this scale tending to be larger at the protected comparedto reference shores (Table 1).

The protected shore differed from the reference shores for theabundance of individuals belonging to size Class 2 only. The overallabundance of this class was larger at the reference than at theprotected shores, although this pattern was difficult to visualize,mainly due to the large variability between sites existing at boththe MPA and the reference shores and characterizing the otherclasses too (Table 4, Fig. 5). Patterns of spatial variance in thedistribution of individuals among size classes differed between theMPA and reference shores at both the scales of quadrates and sitesfor all classes including individuals smaller than 20 mm in shelllength (i.e. Classes 1e3), although, at each scale, Class 1 showedlower variance at the MPA compared to reference shores, whileClass 2 and Class 3 showed the opposite pattern (Table 4). Thespatial variance of Class 4 was affected by protection at the scale ofsites only, in the same direction of Classes 2 and 3 (Table 4), whilethe abundance and patterns of variance of Class 5 were generallycomparable between the MPA and reference shores (Table 4, Fig. 5).

4. Discussion

This study did not indicate differences between the MPA andthree reference shores in mean values and measures of spatial andspatial � temporal variance of the abundance of sea urchins,P. lividus, while the protected area was characterized by a largerabundance of individuals between 30 and 40mm in diameter (Class4). The other classes showed idiosyncratic responses. These rangedfrom patterns spatially and/or temporally variable to a lack ofsignificant differences between the MPA and reference shores, suchas the two classes including the larger individuals, which were veryreduced or virtually absent at all shores. At the smallest spatial scaleexamined (among quadrates), the variance of all the size classes

Gon

ad I

ndex

0

2

4

6

8

10

12

14

16

18

0

2

4

6

8

10

12

14

16

18

0

2

4

6

8

10

12

14

16

18G

onad

Ind

ex

0

2

4

6

8

10

12

14

16

18

0

2

4

6

8

10

12

14

16

18

0

2

4

6

8

10

12

14

16

18

Mar Amorosa P. norte Moledo Mar Amorosa P. norte Moledo Mar Amorosa P. norte Moledo

1102yraurbeF1102yraunaJ0102rebotcO

1102rebmetpeS1102enuJ1102lirpA

Fig. 3. Mean (þSE) value of Gonad Index (GI ¼ gonads dry weight/whole animal dry weight � 100) of P. lividus at the protected (Mar) and three reference (Amorosa, Praia Norte andMoledo) shores. Data averaged over twenty replicate individuals.

I. Bertocci et al. / Marine Environmental Research 77 (2012) 60e7066

analyzed differed between the MPA and reference shores, althoughwith contrasting patterns between Class 1, Class 2 and Class 5 vs.Class 3 and Class 4. The reproductive potential of P. lividus esti-mated at the scale of shores was also temporally variable, although,when differences were evident, these were always in the directionof larger values of calculated GI at the MPA compared to referenceshores. The percentage cover of M. galloprovincialis was alsocomparable between the protected and reference shores, but thespatial variance of mussels at the scale of sites was generally largerat the former compared to the latter. The distribution of this speciesat the MPA, however, was characterized by a reduced abundance of

Site

1

% c

over

x 6

25 c

m-2

0

10

20

30

40

50

60

70

80

90

100

0

10

20

30

40

50

60

70

80

90

100

Site

2

% c

over

x 6

25 c

m-2

0

10

20

30

40

50

60

70

80

90

100

0

10

20

30

40

50

60

70

80

90

100

Site

1Si

te 2

Site

1Si

te 2

Site

1Si

te 2

Site

1Si

te 2

Site

1

Mar Amorosa P. norte Moledo Mar Amor

naJ0102rebotcO

uJ1102lirpA

Fig. 4. Mean (þSE) percentage cover of M. galloprovincialis at the protected (Mar) and threequadrates.

Class 2 individuals compared to reference shores. Finally, the spatialvariance among quadrates and between sites of the distribution ofthe smallest (Class 1) individuals ofM. galloprovincialiswas reducedat the MPA compared to reference shores, an opposite pattern wasdisplayed by individuals up to 20 mm in shell length (Classes 2 and3) and by individuals between 20 mm and 30 mm (Class 4), but atthe scale of sites only, while patterns of variance of the larger sizeswere comparable between the two conditions. Globally, thesefindings were inconsistent with a model assuming a pervasivedirect effect of protection on sea urchins and mussels populationsin the Parque Litoral Norte area, while other concomitant processes

0

10

20

30

40

50

60

70

80

90

100

0

10

20

30

40

50

60

70

80

90

100

Site

2

Site

1Si

te 2

Site

1Si

te 2

Site

1Si

te 2

Site

1Si

te 2

Site

1Si

te 2

Site

1Si

te 2

osa P. norte Moledo Mar Amorosa P. norte Moledo

1102yraurbeF1102yrau

1102rebmetpeS1102en

reference (Amorosa, Praia Norte and Moledo) shores. Data averaged over five replicate

Table 4Summary of ANOVAs and two-tailed F-tests for the number of individuals ofM. galloprovincialis in each of five size classes Class 1 < 3 mm, 3 � Class 2 < 10, 10 � Class 3 < 20,20 � Class 4 < 30 and 30 � Class 5 < 40 estimated at the protected (Mar) and reference (Amorosa, Praia Norte and Moledo) shores over six times of sampling. *P < 0.05,**P < 0.01, ***P < 0.001, ns ¼ not significant (at P > 0.05). Denominators for F tests (with the exception of the MPA vs Reference shores term for Class 3, which was tested overthe Reference shores term as this was significant) and symbols for two-tailed F tests as in Table 1.

Source ofvariation

Class 1 Class 2 Class 3 Class 4 Class 5

df MS F MS F MS F MS F MS F

Time 5 19.07 3.31 * 7.46 2.95 * 3.86 2.63 2.17 1.89 14.97 1.27Shore 3 13.60 2.36 10.78 4.26 * 9.47 6.46** 0.98 0.85 7.72 0.65MPA vs Referenceshores

1 19.05 3.30 14.43 5.70 * 4.08 0.32 0.71 0.62 3.76 0.32

Reference shores 2 10.88 1.34 8.96 3.00 12.16 6.27* 1.12 0.90 9.71 0.66Time � Shore 15 5.77 4.21*** 2.53 0.95 1.47 0.57 1.15 1.53 11.81 1.44Time � [MPA vsReference shores]

5 1.07 0.13 1.62 0.54 0.52 0.27 0.97 0.78 6.12 0.42

Time � Referenceshores

10 8.11 5.92*** 2.98 1.12 1.94 0.75 1.24 1.65 14.65 1.79

Site (Time � Shore) 24 1.37 2.34*** 2.65 3.86*** 2.59 6.09*** 0.75 2.86*** 8.18 1.65*Site (Time � MPA) 6 1.09 1.42 4.23 5.78*** 3.45 8.37*** 1.21 4.30** 9.45 1.56Site (Time � Referenceshores)

18 1.46 2.78*** 2.13 3.16*** 2.30 5.36*** 0.60 2.33** 7.75 1.70*

Residual 192 0.59 0.69 0.42 0.26 4.94Quadrates (MPA) 48 0.77 0.73 0.41 0.28 6.08Quadrates (Referenceshores)

144 0.53 0.67 0.43 0.26 4.56

Cochran’s test C ¼ 0.088, ns C ¼ 0.058, ns C ¼ 0.091, ns C ¼ 0.065, ns C ¼ 0.108, nsTransformation Ln(x þ 1) Ln(x þ 1) Ln(x þ 1) Ln(x þ 1) None

Variance components and two-tailed F-test: MPA Ref. MPA Ref. MPA Ref. MPA Ref. MPA Ref.

Quadrates 66.68 <153.86*** 1128.87 >391.31*** 190.03 >73.89*** 24.34 ¼16.25 6.08 ¼4.56Sites 7.25 <255.18*** 723.67 >88.15*** 273.71 >83.81* 26.73 >3.74*** 0.68 ¼0.64Time � Shore 39.70 ¼39.80 0.00 6.43 0.00 0.00 0.00 3.44 0.00 ¼0.69

I. Bertocci et al. / Marine Environmental Research 77 (2012) 60e70 67

could provide a further, and even greater, contribution to drivetheir documented patterns of spatial and temporal heterogeneity.

A direct effect of protection from human harvesting shouldresult in larger abundances of target species at protected than atcomparable, but unmanaged, areas, as it has been documented byprevious studies involving comparisons between un- vs. intensively

October 2010

Num

ber

of in

divi

dual

s

0

20

40

60

80

100

120

140

160Jan

0

20

40

60

80

100

120

140

160

April 2011

Num

ber

of in

divi

dual

s

0

20

40

60

80

100

120

140

160Ju

0

20

40

60

80

100

120

140

160

S1 S2S1 S2S1 S2 S1 S2

1234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

2

1234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

2

S1S2S1S

Fig. 5. Mean (þSE) number of individuals of M. galloprovincialis in each of seven size classe30 � Class 5 < 40, 40 � Class 6 < 50 and Class 7 � 50 mm) at the protected (Mar: ) and th2. Data averaged over five replicate quadrates (after sub-sampling, see text for details).

exploited populations of both P. lividus (Sala et al., 1998b; Guidettiet al., 2005; Gianguzza et al., 2006; Pais et al., 2007) andM. galloprovincialis (Rius and Cabral, 2004; Robinson et al., 2007;Rius and Zabala, 2008). Moreover, since humans usually removepreferably the larger individuals (e.g. Guidetti, 2006), such activityis likely to affect the structure rather than, or in addition to, the

uary 2011 February 2011

0

20

40

60

80

100

120

140

160

ne 2011 September 2011

0

20

40

60

80

100

120

140

160

34567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

2345

67

34567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

234567 1

2345

67

S1 S2S1 S2 2S1S2 S1 S2S1 S2 S1 S2

s (from 1 to 7: Class 1 < 3 mm, 3 � Class 2 < 10, 10 � Class 3 < 20, 20 � Class 4 < 30,ree reference (Amorosa: , Praia Norte: and Moledo: ) shores. S1: site 1; S2: Site

I. Bertocci et al. / Marine Environmental Research 77 (2012) 60e7068

overall density of populations, reducing the abundance of the largerspecimens at the reference compared to the protected areas(Guidetti et al., 2004; Rius and Cabral, 2004; Pais et al., 2007). Thelack of significant differences between our protected and referenceshores in the total abundance of P. lividus andM. galloprovincialis is,however, consistent with previous findings on both species(Ceccherelli et al., 2011; Rius and Zabala, 2008, respectively) andcould be a consequence of interactions between contrastingconcomitant effects of protection, including the direct reduction ofhuman harvesting and the increased removal of individuals bynatural predators whose abundance could increase when fishing isbanned or reduced. Several large fishes, including species of Spar-idae of high commercial interest, are known to be active predatorsof both sea urchins (Sala and Zabala, 1996; Sala, 1997; Guidetti,2004, 2006; Hereu et al., 2005; Guidetti et al., 2005) and mussels(Lloret et al., 2005; Rius and Zabala, 2008).When fishing is reduced,such as within MPAs, the abundance of such fishes can increase,causing an increased predation pressure on populations of theirbenthic preys that could buffer the potential direct positive effectsof reduced human harvesting of such populations. Trophiccascading mechanisms have been documented for subtidal (Salaet al., 1998a; Shears and Babcock, 2002; Guidetti, 2006) andintertidal (Hockey and Branch, 1984; Lindberg et al., 1998; Castilla,1999; Connell and Anderson, 1999; Halpern et al., 2006) systemsand might have applied in the present case too, maintainingcomparable abundance of sea urchins and mussels at the MPA andthe reference shores. Although the present study could not provide,due to resource limitations, direct and quantitative data on theintensity of fishing and harvesting activities, other explanationscould involve the assumption that such activities are negligible orcomparably intense at both the MPA and the reference shores. Thefirst alternative is unlikely according to the known importance ofhuman harvesting of benthic organisms on Portuguese rockyshores (e.g. Rius and Cabral, 2004), while the second model couldbe at least in part supported by not enough strong (i.e. presence ofauthorities in the field not enough frequent to avoid poaching)enforcement of the existing laws and/or to the possible severeeffect of licensed harvesters, determining levels of recreational andprofessional activities comparable across shores independently ofsupposed protection. The virtual absence of large individuals ofP. lividus at all the sampled shores, whose accessibility to humanswas also similar (see also Ceccherelli et al., 2011), might supportsuch model.

Analogously to the total abundance, patterns of distribution ofindividuals among size classes of sea urchins and mussels wereinconsistent with a clear effect of protection, while a large spatialand/or temporal variability was documented on mean values andvariances of most response variables at all shores, in agreementwith previous findings (Ceccherelli et al., 2009, 2011). Natural largespatial and temporal heterogeneity in patterns of distribution andabundance is a documented trait of intertidal sea urchins andmussels populations. Several biological and abiotic processes,including predation (Sala and Zabala, 1996; Guidetti, 2004, 2006;Hereu et al., 2005), competition (Guidetti et al., 2004), recruitment(Tomas et al., 2004; Rius and Zabala, 2008), wave disturbance(Denny,1987) and substratum heterogeneity (Benedetti-Cecchi andCinelli, 1995; Sala et al., 1998a; Barnes and Crook, 2001; Hereu et al.,2005) can be responsible for patterns of distribution of P. lividus andM. galloprovincialis at spatial and temporal scales such as thoseexamined by the present study. Similarly, the reproductive poten-tial of P. lividus could be affected by temporally variable factorsindependent of direct and indirect effects of protection. The avail-ability of food, for instance, is a determinant driver positivelyrelated with the gonad development of sea urchins (Levitan, 1995),so that large densities can determine a reduced development of

reproductive organs and production of gametes by these animals(e.g. Whale and Peckham, 1999). The lack of differences in theabundance of P. lividus between our protected and reference shores,and in particular the drastically reduced abundance of large indi-viduals at all shores, likely prevented to differentiate the twoconditions according to such a mechanism, while natural fluctua-tions of food resources, such as the availability of ephemeral greenalgae, might have occurred in comparable ways at all shores,explaining the present results on Gonad Index values.

Specifically examining the effects of the several factors andprocesses likely responsible for differences in spatial and temporalpatterns of response variables at the protected vs. reference shoreswas beyond the goals of the present study. Nevertheless, our find-ings have importantmanagement implications. First, analogously toassessments of environmental impacts (Bishop et al., 2002), theyhighlight the need for a sampling design includingmultiple scales inorder to detect differences between protected and reference shores,particularly when these are examined in terms of variance ofresponse variables. This means that, in spite of the difficulties toanticipate the specific spatial and temporal scales at which theeffects of protection should occur, evaluations of the effectiveness ofMPAs should be conducted over a range of scales. Second, presentresults contribute to the growing literature demonstrating theimportance of spatial and temporal heterogeneity as a crucialcomponent of management of MPAs. It is evident, in fact, that thedesignation and implementation of an MPA should guaranteea proper representation of the relevant scales of variation of theorganisms that are intended to be protected, and of the relevantabiotic and biological processes (e.g. Benedetti-Cecchi et al., 2003).Finally, and even more importantly on the short term for themanagement of present andotherMPAs, this study suggests that theunequivocal evaluation of patterns of heterogeneity in distributionandabundanceof organisms as a criterion to assess the effectivenessof MPAs can not prescind from an accurate enforcement of restric-tions to human activities. The social and cultural acceptance of thisis, inmost cases, bad (Russ and Zeller, 2003), but it can be critical formaintaining and improving the ecological role of MPAs and forchoosing better scientifically-based management options.

Acknowledgements

This study was funded by a grant (241/10/PC09) to CIIMAR anda fellowship to R.D. from Polis Litoral Norte S.A. through the project“Estudo de caracterização da actividade pesqueira costeira e dosseus impactes nos recursos e nas comunidades marinhas do LitoralNorte”. I.B. was supported by FCT within the Programa Ciência2008 e Fundo Social Europeu. We thank F. Tuya for discussions atthe early steps of the project, M. Rubal, P. Veiga, M. Silva, R. Vieira, F.Arenas and the students J. Ferreira, L. Magalhães, S. Rocha and R.Vasconcelos for their help with field and laboratory work atdifferent stages of the study and M. Matias for valuable commentson the manuscript.

References

Agardy, M.T., 1994. Advances in marine conservation: the role of marine protectedareas. Trends in Ecology and Evolution 9, 267e270.

Airoldi, L., Bacchiocchi, F., Cagliola, C., Bulleri, F., Abbiati, M., 2005. Impact ofrecreational harvesting on assemblages in artificial rocky habitats. MarineEcology Progress Series 299, 55e66.

Allison, G.W., Lubchenco, J., Carr, M.H., 1998. Marine reserves are necessary but notsufficient for marine conservation. Ecological Applications. 8-Supplement,79e92.

Araújo, R., Bárbara, I., Sousa-Pinto, I., Quintino, V., 2005. Spatial variability ofintertidal rocky shore assemblages in the northwest coast of Portugal. Estua-rine, Coastal and Shelf Science 64, 658e670.

I. Bertocci et al. / Marine Environmental Research 77 (2012) 60e70 69

Barnes, D.K.A., Crook, A.C., 2001. Quantifying behavioural determinants of thecoastal European sea-urchin Paracentrotus lividus. Marine Biology 138,1205e1212.

Benedetti-Cecchi, L., Cinelli, F., 1995. Habitat heterogeneity, sea urchin grazing andthe distribution of algae in littoral rock pools on the west coast of Italy (westernMediterranean). Marine Ecology Progress Series 203, 203e212.

Benedetti-Cecchi, L., Bertocci, I., Micheli, F., Maggi, E., Fosella, T., Vaselli, S., 2003.Implications of spatial heterogeneity for management of marine protectedareas (MPAs): examples from assemblages of rocky coasts in the northwestMediterranean. Marine Environmental Research 55, 429e458.

Bishop, M.J., Underwood, A.J., Archambault, P., 2002. Sewage and environmentalimpact on rocky shores: necessity of identifying relevant spatial scales. MarineEcology Progress Series 236, 121e128.

Botsford, L.W., Micheli, F., Hastings, A., 2003. Principles for the design of marinereserves. Ecological Applications 13, S25eS31.

Brosnan,D.M., Crumrine, L.L.,1994. Effects of human tramplingonmarine rocky shorecommunities. Journal of Experimental Marine Biology and Ecology 177, 79e97.

Bulleri, F., Bertocci, I., Micheli, F., 2002. Interplay of encrusting coralline algae andsea urchins in maintaining alternative habitats. Marine Ecology Progress Series243, 101e109.

Byrne, M., 1990. Annual reproductive cycles of the commercial sea urchin Para-centrotus lividus from an exposed intertidal and a sheltered habitat on the westcoast of Ireland. Marine Biology 104, 275e289.

Castilla, J.C., 1999. Coastal marine communities: trends and perspectives fromhuman-exclusion experiments. Trends in Ecology and Evolution 14, 280e283.

Castilla, J.C., Bustamante, R.H., 1989. Human exclusion from rocky intertidal of LasCruces, central Chile: effects on Durvillaea antarctica (Phaeophyta, Durvilleales).Marine Ecology Progress Series 50, 203e214.

Ceccherelli, G., Pinna, S., Sechi, N., 2009. Evaluating the effects of protection onParacentrotus lividus distribution in two contrasting habitats. Estuarine, Coastaland Shelf Science 81, 59e64.

Ceccherelli, G., Pais, A., Pinna, S., Sechi, N., Chessa, L.A., 2011. Human impact onParacentrotus lividus: the result of harvest restrictions and accessibility oflocations. Marine Biology 158, 845e852.

Chapman, M.G., Underwood, A.J., Skilleter, G.A., 1995. Variability at different spatialscales between a subtidal assemblage exposed to discharge of sewage and twocontrol sites. Journal of Experimental Marine Biology and Ecology 189,103e122.

Claudet, J., Osenberg, C.W., Benedetti-Cecchi, L., Domenici, P., García-Charton, A.,Pérez-Ruzafa, A., Badalamenti, F., Bayle-Sempere, J., Brito, A., Bulleri, F.,Culioli, J.-M., Dimech, M., Falcón, J.M., Guala, I., Milazzo, M., Sánchez-Meca, J.,Somerfield, P.J., Stobart, B., Vandeperre, F., Planes, S., 2008. Marine reserves: sizeand age do matter. Ecology Letters 11, 481e489.

Connell, S.D., Anderson, M.J., 1999. Predation by fish on assemblages of intertidalepibiota: effects of predator size and patch size. Journal of Experimental MarineBiology and Ecology 241, 15e29.

Dayton, P.K., Sala, E., Tegner, M.J., Thrush, S., 2000. Marine reserves: parks, base-lines, and fishery enhancement. Bulletin of Marine Science 66, 617e634.

de Lumley, H., 1975. Cultural evolution in France in its paleoecological settingduring the Middle Pleistocene. In: Isaac, G.L., Butzer, K.W. (Eds.), After theAustralopithecines: Stratigraphy, Ecology and Culture in the Middle Pleistocene.Mouton, The Hague, pp. 745e808.

Denny, M.W., 1987. Lift as a mechanism of patch initiation in mussel beds. Journal ofExperimental Marine Biology and Ecology 113, 231e245.

Dethier, M.N., Graham, E.S., Cohen, S., Tear, L.M., 1993. Visual versus random-pointpercent cover estimations: ‘Objective’ is not always better. Marine EcologyProgress Series 110, 9e18.

FAO Fishery Information, Data and Statistics Unit (FAO-FIDI), 2004. Collation,Analysis and Dissemination of Global and Regional Fishery Statistics. FI Pro-gramme Websites (UN Food and Agriculture Organization, Rome, 2006); see.www.fao.org/fi/default.asp.

Fogarty,M.J., Murawski, S.A.,1998. Large-scale disturbance and the structure ofmarinesystems: fishery impacts on Georges Bank. Ecological Applications 8, S6eS22.

Fraschetti, S., Terlizzi, A., Bussotti, S., Guarnieri, G., D’Ambrosio, P., Boero, F., 2005.Conservation of Mediterranean seascapes: analyses of existing protectionschemes. Marine Environmental Research 59, 309e332.

García-Charton, J.A., Pérez-Ruzafa, A., 1999. Ecological heterogeneity and the eval-uation of the effects of marine reserves. Fisheries Research 42, 1e20.

García-Charton, J.A., Williams, I.D., Pérez-Ruzafa, A., Milazzo, M., Chemello, R.,Marcos, C., Kitsos, M.-S., Koukouras, A., Riggio, S., 2000. Evaluating theecological effects of Mediterranean marine protected areas: habitat, scale andthe natural variability of ecosystems. Environmental Conservation 27, 159e178.

García-Charton, J.A., Pérez-Ruzafa, A., Marcos, C., Claudet, J., Badalamenti, F., Bene-detti-Cecchi, L., Falcon, J.M., Milazzo, M., Schembri, P.J., Stobart, B., Vandeperre, F.,Brito, A., Chemello, R., Dimech, M., Domenici, P., Guala, I., Le Direach, L., Maggi, E.,Planes, S., 2008. Effectiveness of European Atlanto-MediterraneanMPAs: do theyaccomplish the expected effects on populations, communities and ecosystems?Journal for Nature Conservation 16, 193e221.

Gianguzza, P., Chiantore, M., Bonaviri, C., Cattaneo-Vietti, R., Vielmini, I., Riggio, S.,2006. The effects of recreational Paracentrotus lividus fishing on distributionpatterns of sea urchins at Ustica Island MPA (Western Mediterranean, Italy).Fisheries Research 81, 37e44.

Glasby, T.M., 1997. Analysing data from post-impact studies using asymmetricalanalysis of variance: a case study of epibiota on marinas. Australian Journal ofEcology 22, 448e459.

Guidetti, P., 2004. Consumers of sea urchins, Paracentrotus lividus and Arbacia lixula,in shallow Mediterranean rocky reefs. Helgoland Marine Research 58, 110e116.

Guidetti, P., 2006. Marine reserve re-establish lost predatory interactions andcauses community changes in rocky reefs. Ecological Applications 16, 963e976.

Guidetti, P., 2007. Potential of marine reserves to cause community-wide changesbeyond their boundaries. Conservation Biology 21, 540e545.

Guidetti, P., Terlizzi, A., Boero, F., 2004. Effects of the edible sea urchin, Paracentrotuslividus, fishery along the Apulian rocky coast (SE Italy, Mediterranean Sea).Fisheries Research 66, 287e297.

Guidetti, P., Bussotti, S., Boero, F., 2005. Evaluating the effects of protection on fishpredators and sea urchins in shallow artificial rocky habitats: a case study in thenorthern Adriatic Sea. Marine Environmental Research 59, 333e348.

Halpern, B.S., 2003. The impact of marine reserves: do reserves work and does sizematter? Ecological Applications 13, S117eS137.

Halpern, B.S., Warner, R.R., 2002. Marine reserves have rapid and lasting effects.Ecology Letters 5, 361e366.

Halpern, B.S., Cottenie, K., Broitman, B.R., 2006. Strong top-down control inSouthern California kelp forest ecosystems. Science 312, 1230e1232.

Hereu, B., Zabala, M., Linares, C., Sala, E., 2005. The effects of predator abundanceand structural complexity on survival of juvenile sea urchins. Marine Biology146, 293e299.

Hockey, P.A.R., Bosman, A.L., 1986. Man as an intertidal predator in Transkei:disturbance, community convergence and management of a natural foodresource. Oikos 46, 3e14.

Hockey, P.A.R., Branch, G.M., 1984. Oystercatchers and limpets: impact and impli-cations. A preliminary assessment. Ardea 72, 199e206.

Horne, J.K., Schneider, D.C., 1995. Spatial variance in ecology. Oikos 74, 18e26.Jackson, J.B.C., Kirby, M.X., Berger, W.H., Bjorndal, K.A., Botsford, L.W., Bourque, B.J.,

Bradbury, R.H., Cooke, R., Erlandson, J., Estes, J.A., Hughes, T.P., Kidwell, S.,Lange, C.B., Lenihan, H.S., Pandolfi, J.M., Peterson, C.H., Steneck, R.S., Tegner, M.J.,Warner, R.R., 2001. Historical overfishing and the recent collapse of coastalecosystems. Science 293, 629e637.

King, M.C., Beazley, K.F., 2005. Selecting focal species for marine protected areanetwork planning in the Scotia-Fundy region of Atlantic Canada. AquaticConservation: Marine and Freshwater Ecosystems 15, 367e385.

Kingsford, M.J., Underwood, A.J., Kennelly, S.J., 1991. Humans as predators on rockyreefs in New South Wales, Australia. Marine Ecology Progress Series 72, 1e14.

Lasiak, T., 1998. Multivariate comparisons on rocky intertidal macrofaunal assem-blages from replicate exploited and non-exploited localities on the Transkeicoast of South Africa. Marine Ecology Progress Series 167, 15e23.

Levitan, D.R., 1995. The ecology of fertilization in free-spawning invertebrates. In:McEdward, L.R. (Ed.), Ecology of Marine Invertebrate Larvae. CRC Press, BocaRaton, Florida, pp. 123e156.

Lindberg, D.R., Estes, J.A., Warheit, K.I., 1998. Human influences on trophic cascadesalong rocky shores. Ecological Applications 8, 880e890.

Lloret, J., Galzin, R., Gil de sola, L., Souplet, A., Demestre, M., 2005. Habitat relateddifferences in lipid reserves of some exploited fish species in the north-westernMediterranean continental shelf. Journal of Fish Biology 67, 51e65.

Lotze, H.K., Lenihan, H.S., Bourque, B.J., Bradbury, R.H., Cooke, R.G., Kay, M.C.,Kidwell, S.M., Kirby, M.X., Peterson, C.H., Jackson, J.B.C., 2006. Depletion,degradation, and recovery potential of estuaries and coastal seas. Science 312,1806e1809.

McClanahan, T.R., Mangi, S., 2000. Spillover of exploitable fishes from a marine parkand its effect on the adjacent fishery. Ecological Applications 10, 1792e1805.

Micheli, F., Benedetti-Cecchi, L., Gambaccini, S., Bertocci, I., Borsini, C., Osio, G.C.,Romano, F., 2005. Cascading human impacts, marine protected areas, and thestructure of Mediterranean reef assemblages. Ecological Monographs 75,81e102.

Pais, A., Chessa, L.A., Serra, S., Ruiu, A., Meloni, G., Donno, Y., 2007. The impact ofcommercial and recreational harvesting for Paracentrotus lividus on shallowrocky reef sea urchin communities in North-western Sardinia, Italy. Estuarine,Coastal and Shelf Science 73, 589e597.

Palumbi, S.R., 2001. The ecology of marine protected areas. In: Bertness, M.D.,Gaines, S.D., Hay, M.E. (Eds.), Marine Community Ecology. Sinauer Associates,Sunderland, MA, pp. 509e530.

Pelletier, D., García-Charton, J.A., Ferraris, J., David, G., Thébaud, O., Letourneur, J.,Claudet, J., Amand, M., Kulbicki, M., Galzin, R., 2005. Designing indications forevaluating the impact of marine protected areas on coral reef ecosystems:a multi-disciplinary standpoint. Aquatic Living Resources 18, 15e33.

Pinnegar, J.K., Polunin, N.V.C., Francour, P., Badalamenti, F., Chemello, R., Harmelin-Vivien, M.-L., Hereu, B., Milazzo, M., Zabala, M., D’Anna, G., Pipitone, C., 2000.Trophic cascades in benthic marine ecosystems: lessons for fisheries andprotected-area management. Environmental Conservation 27, 179e200.

Rius, M., Cabral, H.N., 2004. Human harvesting of Mytilus galloprovincialis Lamarck,1819, on the central coast of Portugal. Scientia Marina 68, 545e551.

Rius, M., Zabala, M., 2008. Are marine protected areas useful for the recovery of theMediterranean mussel populations? Aquatic Conservation: Marine and Fresh-water Ecosystems 18, 527e540.

Roberts, C.M., Bohnsack, J.A., Gell, F.R., Hawkins, J.P., Goodridge, R., 2001. Effects ofmarine reserves on adjacent fisheries. Science 294, 1920e1923.

Robinson, T.B., Branch, G.M., Griffiths, C.L., Govender, A., 2007. Effects of experi-mental harvesting on recruitment of an alien mussel Mytilus galloprovincialis.Journal of Experimental Marine Biology and Ecology 345, 1e11.

Rodrigues, A.S.L., Akçakaya, H.R., Andelman, S.J., Bakarr, M.I., Boitani, L., Brooks, T.M.,Chanson, J.S., Fishpool, L.D.C., Fonseca, G.A.B., Gaston, K.J., Hoffmann, M.,

I. Bertocci et al. / Marine Environmental Research 77 (2012) 60e7070

Marquet, P.A., Pilgrim, J.D., Pressey, R.L., Schipper, J., Sechrest, W., Stuart, S.N.,Underhill, L.G., Waller, R.W., Watts, M.E.J., Yan, X., 2004. Effectiveness of theglobal protected area network in representing species diversity. Nature 428,640e643.

Russ, G.R., Zeller, D., 2003. From mare liberum to mare reservarum. Marine Policy27, 75e78.

Sala, E., 1997. Fish predators and scavengers of the sea urchin Paracentrotus lividusin protected areas of the north-west Mediterranean Sea. Marine Biology 129,531e539.

Sala, E., Zabala, M., 1996. Fish predation and the structure of the sea urchin Para-centrotus lividus populations in the NW Mediterranean. Marine Ecology Prog-ress Series 140, 71e81.

Sala, E., Boudouresque, C.F., Harmelin-Vivien, M., 1998a. Fishing, trophic cascades,and the structure of algal assemblages: evaluation of an old but untestedparadigm. Oikos 82, 425e439.

Sala, E., Ribes, M., Hereu, B., Zabala, M., Alvà, V., Coma, R., Garrabou, J., 1998b.Temporal variability in abundance of the sea urchins Paracentrotus lividus andArbacia lixula in the northwestern Mediterranean: comparison betweena marine reserve and an unprotected area. Marine Ecology Progress Series 168,135e145.

Schneider, D.C., 1994. Quantitative Ecology: Spatial and Temporal Scaling. AcademicPress, San Diego, CA, USA.

Schwartz, M.W., 1999. Choosing the appropriate scale of reserves for conservation.Annual Review of Ecology and Systematics 30, 83e108.

Searle, S.R., Casella, G., McCulloch, C.E., 1992. Variance Components. John Wiley &Sons Inc., New York.

Shears, N.T., Babcock, R.C., 2002. Marine reserves demonstrate top-down control ofcommunity structure on temperate reefs. Oecologia 132, 131e142.

Sherman, K., 1994. Sustainability, biomass yields, and health of coastal ecosystems:an ecological perspective. Marine Ecology Progress Series 112, 277e301.

Terlizzi, A., Benedetti-Cecchi, L., Bevilacqua, S., Fraschetti, S., Guidetti, P.,Anderson, M.J., 2005. Multivariate and univariate asymmetrical analyses inenvironmental impact assessments: a case study of Mediterranean subtidalsessile assemblages. Marine Ecology Progress Series 289, 27e42.

Thrush, S.F., Hewitt, J.E., Cummings, V.J., Green, M.O., Funnell, G.A., Wilkinson, M.R.,2000. The generality of field experiments: interactions between local andbroad-scale processes. Ecology 81, 399e415.

Tomas, F., Romero, J., Turon, X., 2004. Settlement and recruitment of the sea urchinParacentrotus lividus in two contrasting habitats in the Mediterranean. MarineEcology Progress Series 282, 173e184.

Tomas, F., Romero, J., Turon, X., 2005. Experimental evidence that intra-specificcompetition in seagrass meadows reduces reproductive potential in the seaurchin Paracentrotus lividus (Lamarck). Scientia Marina 69, 475e484.

Underwood, A.J., 1992. Beyond BACI: the detection of environmental impact onpopulation in the real, but variable world. Journal of Experimental MarineBiology and Ecology 161, 145e178.

Underwood, A.J., 1993a. Exploitation of species on the rocky coast of New SouthWales (Australia) and options for its management. Ocean and CoastalManagement 20, 41e62.

Underwood, A.J., 1993b. The mechanisms of spatially replicated sampling pro-grammes to detect environmental impacts in a variable world. AustralianJournal of Ecology 18, 145e178.

Underwood, A.J., 1994. On beyond BACI: sampling designs that might reliably detectenvironmental disturbances. Ecological Applications 4, 3e15.

Underwood, A.J., 1997. Experiments in Ecology: Their Logical Design and Interpre-tation Using Analysis of Variance. Cambridge University Press, Cambridge, UK.

Underwood, A.J., Chapman, M.G., 1996. Scales of spatial patterns of distribution ofintertidal invertebrates. Oecologia 107, 212e224.

Vitousek, P.M., Mooney, H.A., Lubchenco, J., Melillo, J.M., 1997. Human dominationof earth’s ecosystems. Science 277, 494e499.

Warwick, R.M., Clarke, K.R., 1993. Increased variability as a symptom of stress inmarine communities. Journal of Experimental Marine Biology and Ecology 172,215e226.

Whale, R.A., Peckham, S.H., 1999. Density-related reproductive trade-offs in thegreen sea urchin, Strongylocentrotus droebachiensis. Marine Biology 134,127e137.

Winer, B.J., Brown, D.R., Michelis, K.M., 1991. Statistical Principles in ExperimentalDesign. McGraw-Hill, USA.