lable at ScienceDirect

Marine Environmental Research 103 (2015) 36e45

Contents lists avai

Marine Environmental Research

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

Response of intertidal sandy-beach macrofauna to human trampling:An urban vs. natural beach system approach

Ma Jos�e Reyes-Martínez a, *, Ma Carmen Ruíz-Delgado a, Juan Emilio S�anchez-Moyano b,Francisco Jos�e García-García a

a Departamento de Sistemas Físicos, Químicos y Naturales, Universidad Pablo de Olavide, Ctra. Utrera Km 1, 41013 Sevilla, Spainb Departamento de Zoología, Universidad de Sevilla, Av. Reina Mercedes 6, 41012 Sevilla, Spain

a r t i c l e i n f o

Article history:Received 11 August 2014Received in revised form27 October 2014Accepted 1 November 2014Available online 4 November 2014

Keywords:Sandy beachesMacrofaunaBioindicatorHuman tramplingTourismDisturbance

* Corresponding author.E-mail address: mjreymar@upo.es (M.J. Reyes-Ma

http://dx.doi.org/10.1016/j.marenvres.2014.11.0050141-1136/© 2014 Elsevier Ltd. All rights reserved.

a b s t r a c t

Sandy beaches are subjected to intense stressors, which are mainly derived from the increasing patternof beach urbanization. These ecosystems are also a magnet for tourists, who prefer these locations asleisure and holiday destinations, and such activity further increases the factors that have an adverseeffect on beaches. In the study reported here the effect of human trampling on macrofauna assemblagesthat inhabit intertidal areas of sandy beaches was assessed using a BACI design. For this purpose, threecontrasting sectors of the same beach were investigated: an urban area with a high level of visitors, aprotected sector with a low density of users, and a transitional area with a high level of human occu-pancy. The physical variables were constant over time in each sector, whereas differences were found inthe intensity of human use between sectors. Density variations and changes in the taxonomic structureof the macrofauna with time were shown by PERMANOVA analysis in the urban and transitional loca-tions whereas the protected sector remained constant throughout the study period. The amphipodBathyporeia pelagica appears sensitive to human trampling pressure and the use of this species as abioindicator for these types of impact is recommended.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Ecosystems across theworld are being damaged due to the rapidexpansion of the human population (Defeo et al., 2009). Coastalareas are particularly vulnerable to this phenomenon, especiallygiven that 41% of the global population lives within the coastallimits (Martínez et al., 2007).

In addition to residential uses, coastal areas e and sandy bea-ches in particular e have long been a magnet for tourists (Jennings,2004), who prefer these locations for recreational activities andholiday destinations. Beach ecosystems are therefore subjected tointense stressors as a result of increasing coastal infrastructure, thedevelopment of shoreline armoring, beach nourishment, resourceexploitation, pollution, and grooming (Schlacher et al., 2007). Theseactivities are mainly the result of the increasing pattern of urban-ization of beaches and the improvements of tourist facilities. Thistrend, in which economic sustainability is preferred over biologicalsustainability, leads to substantial environmental costs (Davenport

rtínez).

and Davenport, 2006) that threaten the ecological integrity ofcoastal systems (Lucrezi et al., 2009).

Tourism warrants particular attention since it is the economicengine of many countries (Davenport and Davenport, 2006) andinvolves large numbers of visitors to beaches, especially in thesummer season. The high level of human occupation can disruptcoastal ecosystems through a wide range of activities, such ascamping (Hockings and Twyford, 1997), the use of off-road vehicles(Schlacher and Thompson, 2008), and other recreational pursuits(Fanini et al., 2014). These actions can modify the natural physicalcharacteristics of beaches and have a direct effect on macrofaunacommunities and their distribution patterns, which can in turnresult in a significant loss of biodiversity (Defeo et al., 2009). Adirect effect of the various activities carried out on beaches is hu-man trampling. The effect of trampling on faunal communities is animportant topic that has been addressed for different ecosystems,such as rocky shores (Ferreira and Rosso, 2009), coral reefs (Rodgersand Cox, 2003), and mudflats (Rossi et al., 2007). On sandy beachesthis issue has been considered from different perspectives; forexample, at the population level the effect of human trampling hasbeen well analyzed for supralittoral species of talitrid amphipods(Ugolini et al., 2008; Veloso et al., 2008, 2009; Weslawski et al.,

M.J. Reyes-Martínez et al. / Marine Environmental Research 103 (2015) 36e45 37

2000) or ocypodid decapods (Barros, 2001; Lucrezi et al., 2009). Onthe other hand, at the community level the impact of humantrampling has been addressed both in controlled experiments(Moffet et al., 1998) and by field observations involving comparisonof highly trampled areas with control zones (Jaramillo et al., 1996;Veloso et al., 2006). The results of these studies have shown adecrease in the abundance of macrofaunawithin the trampled area.However, this pattern cannot normally be directly attributed totrampling itself, since the highly trampled areas correspond tohighly urbanized zones and the response of species may thus bedue to a set of influential factors inherent to coastal development or‘compound threats’ (Schlacher et al., 2014) rather than to the iso-lated effect of trampling. To our knowledge, only Schlacher andThompson (2012) have evaluated the isolated effect of tramplingby comparing trampled (access point) and control areas on a beachunmodified by human action. However, the temporal scale was notconsidered in that study.

When the effect of an impact is analyzed, it is recommendedthat the experimental designs consider samplings on differenttime-scales, both before and after a proposed development thatmay have an impact, and on different spatial-scales (Underwood,1994). The information obtained in this way can be used todistinguish between natural changes and those that are attribut-able to impacts, and it also allows themagnitude of the impact to bemeasured (Underwood, 1992).

Before/After/Control/Impact (BACI) design enables the explora-tion of a wide range of responses, such as changes in abundance,diversity, richness, biomass, or body condition (Torres et al., 2011).BACI is therefore a robust design to detect human impacts (Aguado-Gim�enez et al., 2012).

Beach fauna plays a major role in the functioning of beachecosystems (McLachlan and Brown, 2006). Benthos are involved innutrient regeneration (Cisneros et al., 2011), they are trophic linksbetweenmarine and terrestrial systems (Dugan, 1999; Lercari et al.,2010) and are stranded material decomposers (Dugan et al., 2003;Lastra et al., 2008). The identification of factors that cause distur-bance is therefore a crucial task in maintaining the continuity ofsandy beach ecosystems. If one primarily considers human tram-pling, supralittoral species have traditionally been viewed as highlyvulnerable (McLachlan and Brown, 2006) although the swashbeach area, which is inhabited by the greatest diversity of macro-fauna, is most commonly used by people (Schlacher and Thompson,2012). Studies aimed at determining the effects of pedestrian ac-tivity with an emphasis on intertidal species are scarce, despitetheir potential as a tool in the design of management plans andconservation policies in these ecosystems (Jaramillo et al., 1996).The objective of the study reported here was to quantify andevaluate the effect of human trampling onmacrofauna assemblagesthat inhabit the intertidal area of sandy beaches in a gradient ofhuman pressure. The study was carried out using a BACI design. Inthis context, the trajectory of density, richness, diversity index, andcommunity taxonomic structure were evaluated before and afteran episode of high tourist occupancy. In addition, the mostvulnerable species that can be considered as indicators of thesetypes of impact were explored.

2. Material and methods

2.1. Study area

The study was carried out in three sectors of a sandy beach withan anthropogenic pressure gradient. The beach is located in C�adizBay in the southwestern region of the Iberian Peninsula (Fig. 1).C�adiz Bay is a shallow (maximum depth of 20 m) mesotidal basin(maximum tide 3.7 m) with a meanwave height of 1 m (Benavente

et al., 2002). This coastal area has a subtropical climatewith a meanannual temperature of 19 �C and the prevailing winds blow fromthe West and East (Del Río et al., 2013).

The urban sector of Valdelagrana (36�3401300N; 6�1302900W) has ahigh level of urban development (housing and hotels) and highhuman occupancy during the summer season. The backshore isoccupied by constructions and tourism infrastructure (e.g., parkingspaces, streets, boardwalks), which have destroyed the vegetationcover and the dunes system (personal observation). Moreover, thissector is subject to daily mechanical grooming of beach sand toremove debris. In contrast, Levante (36�3205300N; 6�1303400W) is apristine sector that belongs to a protected area (Los Toru~nosMetropolitan Park). In this area the salt-marsh system in thebackshore area is preserved (Veloso et al., 2008) and there is awell-developed dune system that reaches 2 m in height and 50 m inwidth, with natural vegetation cover that is a key area for nestingand shelter for marine birds species (Buitrago and Anfuso, 2011).This area can only be reached on foot. The intermediate sector(36�3303800N; 6�1302600W) is located in the transitional area be-tween Valdelagrana and Levante. This area is not urbanized and islocated within Los Toru~nos Metropolitan Park. The backshore in-cludes a dune system with vegetation cover interrupted by an ac-cess path. Visitors also have other facilities and a tourist traintransports people from the park entrance to this sector. The pro-tected and intermediate sectors are manually groomed (daily) toremove human debris selectively.

2.2. Sampling procedures

The largest tourist influx in Spain occurs during the summermonths (June to August). As a consequence, six sampling cam-paigns were conducted in each sector (urban, intermediate, andprotected) during spring tides; three in each sector before thetourist season (March, April, May 2011) and three in each sectorafter (September, October, November 2011).

At each site six equidistant and across-shore transects wereplaced in a 100 m long-shore area. Each transect comprised 10equidistant points, from the high tidewatermark to the swash zoneto cover the entire intertidal area. At each sampling level, faunasamples were collected with a 25-cm diameter plastic core to adepth of 20 cm. Samples were sieved on site through a 1-mmmeshsieve, preserved in 70% ethanol, and stained with Rose Bengal.Sediment samples were also collected at each sampling level with aplastic tube (3.5-cm diameter) buried at a depth of 20 cm. Thebeach-face slope was estimated by the height difference accordingto Emery (1961).

Themacrofaunawere quantified and identified in the laboratoryand the sediment characteristics (mean grain size, sorting coeffi-cient, sand moisture, and organic matter content) were deter-mined. The mean grain size was determined by sieving drysediment through a graded series of sieves (5, 2, 1, 0.5, 0.25, 0.125,and 0.063 mm) according to the method described by Guitian andCarballas (1976). Sand moisture was measured by the weight lossafter drying the sediment at 90 �C. The organic matter content wasestimated as the difference between dry sediment weight andsediment weight after calcination at 500 �C.

The number of users observed at each sector was used as a proxyto quantify the human trampling intensity. A total of six humancensuses were conducted; three censuses were performed (1census per month at each sector) at the spring tide during theperiod of the greatest inflow of visitors (June, July and August,2011); and three censuses were conducted before impact. Thecounts were performed every 30 min for a 6 h period (until hightide) and were conducted in the same zone as the macrofaunasampling in an area of 50 m along the shore � beach width. In

Fig. 1. Study area showing C�adiz Bay and locations of the three studied sectors: Urban sector Valdelagrana (V); Protected sector Levante (L) and Intermediate sector (I).

M.J. Reyes-Martínez et al. / Marine Environmental Research 103 (2015) 36e4538

addition to the number of beach visitors, the activities undertakenby them were recorded.

2.3. Data analysis

The potential impact of human trampling on the macrofaunaassemblages was analyzed using a modified BACI method thatcontrasts data from urban, intermediate, and protected locationsbefore and after the impact. Here, urban and protected zonesoperate as impacted and control locations, respectively. The nullhypothesis that significant differences did not exist in the benthicassemblages and univariate descriptors (density, richness, andShannon's diversity index) before and after the impact period wastested separately for each sector.

The design for the analyses included three factors: Beach (Be,three levels: urban, intermediate, and protected, fixed), time (Ti,two levels: before and after, fixed), and sampling period (Sp, sixlevels, random and nested in Ti). According to this approach theeffect of human trampling is shown by a statistically significant‘beach � time’ interaction.

The variation over time in the multivariate structure of macro-fauna assemblages and univariate variables was tested by permu-tational multivariate analyses of variance (PERMANOVA)

(Anderson, 2001, 2005) using 9999 permutations. An additional p-value obtained by the Monte Carlo test was used when the numberof permutations was not sufficient (<150). Abiotic variables andhuman trampling (number of people as a proxy) were subjected tothe same design in order to detect changes in the physical char-acteristics and number of users between sectors.

Multivariate patterns were based on BrayeCurtis dissimilaritiesand univariate, abiotic, and human trampling analysis on Euclideandistance similarity matrices, on fourth-root transformed data forbiotic measures. When the interaction of interest was significant,post hoc pair-wise comparisons were performed to identify thesources of these significant differences. The homogeneity ofdispersionwas tested using the PERMDISP routine (Anderson et al.,2008).

A non-metric multidimensional scaling ordination (nMDS) of‘beach � time’ interaction centroids was performed to display dif-ferences in community structure. The SIMPER routine wasemployed to detect most species that contribute to the dissimilarityin cases where significant differences in the PERMANOVA analysiswere identified. To detect whether the variation shown in theSimper analysis was natural or induced by human impact, thetrajectory of species density over time was tested by PERMANOVAdesign analysis and this was compared between sectors.

Table 2PERMANOVA result testing for differences in human trampling impact (using thenumber of users as a proxy) between sectors before and during impact and pair-wisecomparison of term ‘Be � Ti’ for pairs of levels of factor: (a) ‘Beach’ and (b) ‘Time’.Urb ¼ Urban sector; Int ¼ Intermediate sector, and Protec ¼ Protected sector.Bef ¼ before impact and Dur ¼ During impact.

Source df MS Pseudo-F P(perm)

Be 2 20.52 19.07 0.001Ti 1 228.05 479.50 0.104Sp(Ti) 4 0.47 0.62 0.639BexTi 2 43.93 40.83 0.0001Bex Sp(Ti) 8 1.07 1.41 0.190Res 252 0.76Total 269

a) Pair-wise test Groups t P(MC)

Before Urb - Int 7.06 0.12Urb - Protec 11.17 0.40Int - Protec 9.65 0.28

During Urb - Int 7.07 0.017Urb - Protec 11.17 0.008Int - Protec 9.65 0.011

b) Pair-wise test Groups t P(MC)

Urban bef- dur 34.57 0.0001Intermediate bef- dur 29.76 0.0001Protected bef- dur 0.72 0.507

M.J. Reyes-Martínez et al. / Marine Environmental Research 103 (2015) 36e45 39

All univariate and multivariate analyses were performed withPRIMER-E v6.1 and PERMANOVAþ (PRIMER-E ltd) (Anderson et al.,2008; Clarke and Warwick, 2006).

Pearson's correlations were used to determine the relationshipbetween changes in the macrobenthos density and human tram-pling intensity (number of users as a proxy). This analysis wasconducted with the software PASW Statistics 18.

3. Results

3.1. Physical environment

Abiotic variables were constant over time in each sector andsignificant variations were not detected from the period prior toimpact to that after impact within each sector (p (perm) > 0.05) orbetween the beach sectors (p(perm) > 0.05 for all variables;Table 1). The urban sector had fine sediment (mean grain size of230 ± 18 mm before and 240 ± 56 mm after), a moderate meansorting coefficient (1.54 ± 0.15 before; 1.46 ± 0.16 after), and ameansediment moisture content of 17 ± 4% before impact and 16.5 ± 3%after. The organic matter content increased slightly after impactcompared to that determined before impact (1.3 ± 0.78% and0.92 ± 0.24%, respectively), but this difference was not statisticallysignificant. The intermediate and protected sectors had a fine me-dian grain size in both periods (180 ± 17 mm and 186 ± 15 mmbefore; 201 ± 52 mm and 212 ± 60 mm after, respectively). The meansorting coefficient was moderate in both sectors (1.53 ± 0.23 and1.48 ± 0.19 before; 1.58 ± 0.21 and 1.61 ± 0.24 after). Themean sandmoisture content was the same in both areas before impact(17 ± 3%) and after impact (18 ± 2%). The organic matter content inthe intermediate and protected sectors varied slightly from before(0.94 ± 0.14%; 1.02 ± 0.28%, respectively) to after (1.02 ± 0.29%;1.06 ± 0.22%, respectively). The beach profile and slope did notdiffer substantially during the study in any sector and the sloperemained constant at 2 ± 0.5%.

3.2. Human use

The human trampling (number of visitors as proxy) registeredsignificantly different trajectories over time (‘beach � time’ inter-action: p (perm) ¼ 0.0001; Table 2). The pair-wise test for thissignificant interaction showed that during impact the number ofusers was significantly higher on the urban and the intermediatesectors (p(MC) < 0.05; Table 2a) whereas before impact differenceswere not detected between sectors (p (MC) > 0.05; Table 2b).Furthermore, within sectors both the urban and intermediateshowed significant differences from before to during impact (p(MC) ¼ 0.001, Table 2b) while differences were not detected at theprotected sector (p (MC) ¼ 0.507, Table 2b).

The numbers of visitors in the sampling area over a diurnal timeperiod, both before and during impact (summer season), in eachsector are shown in Fig. 2. During impact, the urban and

Table 1Asymmetrical analyses of variance (PERMANOVA) testing differences in physical variableafter). Sampling period (Sp) was considered as a random variable.

Source df Median grain size Sorting

MS F P (perm) MS F P(p

Be 2 0.09 1.78 0.22 0.02 0.42 0.6Ti 1 0.03 0.63 0.50 0.01 0.26 0.7Sp(Ti) 4 0.05 1.75 0.13 0.06 1.47 0.2Be � Ti 2 0.00 0.09 0.91 0.06 1.10 0.3Be � Sp(Ti) 8 0.05 1.78 0.07 0.06 1.50 0.1Res 54 0.03 0.04

intermediate sectors showed a similar evolution, with a peak influxbetween 12:00 and 14:00 h, after which the number of beach usersdecreased steadily during the afternoon. In the protected sector thenumber of users was constant over time. By contrast, before impactthe tree sector presented the same lower flow of visitors and amaximum of 15 visitors was reached in the urban sector.

The activities performed by users in the three sectors alsodiffered. In the urban and intermediate sectors, around 80% of theactivities involved relaxation, sunbathing, picnics, ballgames, andbuilding sandcastles, whereas in the protected sector 100% of theusers surveyed were walking or angling.

3.3. Community composition and univariate descriptors

In total, 26 species were found during the study period. Crus-taceans were the most diverse taxa (14 species), followed bypolychaetes (six species), molluscs (four species), nemertea andechinodermata (a single species each). The contributions of themajor taxonomic groups in the community in each sector over timeare shown in Fig. 3. Before impact the dominant taxon in all areaswas crustaceans. After impact, however, crustacean contributionsdecreased by 16% in the protected area and in the intermediate andurban zones this decrease was 68% and 60%, respectively. Amphi-poda and Cumacea were the orders that decreased most markedly.In the protected sector there was an increase of 24% in the contri-bution of the polychaete population after impact, whereas in the

s between sectors (Be; urban, intermediate and protected) and time (Ti; before and

Sand moisture Organic matter content

erm) MS F P (perm) MS F P(perm)

6 40.12 2.30 0.17 40.45 2.27 0.161 101.95 6.98 0.10 102.66 6.66 0.102 14.60 1.47 0.22 15.42 1.53 0.207 31.16 1.79 0.22 31.60 1.77 0.238 17.44 1.76 0.09 17.84 1.77 0.09

9.90 10.09

Fig. 2. Number of beach visitors counted (mean ± SD) per patch (50 m along shore � beach width) and per hour in each sector.

M.J. Reyes-Martínez et al. / Marine Environmental Research 103 (2015) 36e4540

urban and intermediate sector the increases were 60% and 85%,respectively. These increases were primarily due to an increase inindividuals of the order Spionida.

For community descriptors, PERMANOVA showed variationsover time for density only, with a significant ‘beach � time’ inter-action (p (perm) ¼ 0.03). The pair-wise comparison of this inter-action showed differences from before to after impact in the urbanand intermediate sectors (p (MC) < 0.05), but differences were notfound in the protected area (Table 3). The density in the protectedsector increased over time (212.2 ± 28.6 ind/m2 before and240.8 ± 48.6 ind/m2 after impact), whereas at the other locationsthe opposite pattern was observed. In the urban sector the density

Fig. 3. Pie charts representing the proportion of taxa in the community, in each sector,and before and after impact.

varied from 158.4 ± 17.4 ind/m2 before impact to 82 ± 21.8 ind/m2

after impact, while in the intermediate site the density decreasedfrom 331.5 ± 39 ind/m2 before impact to 91.8 ± 10.8 ind/m2 afterimpact (Fig. 4).

Significant time differences were not found in the richness anddiversity index (p (perm) > 0.05). Nonetheless, the communitydescriptors showed a more stable response than in the other areasalthough a decrease in these variables was observed in the pro-tected sector.

A global significant and negative correlationwas found betweenmacrobenthos density and the number of users (r ¼ 0.36,p ¼ 0.003). A Person's correlation between these two factors wasalso performed in each sector. In the urban and intermediate sec-tors a significant and negative correlation was found (r ¼ �0.21,p ¼ 0.01; r ¼ �0.42, p ¼ 0.001, respectively), while in the protectedsector the correlation was not significant (r ¼ �0.01, p ¼ 0.84),despite the fact that these factors were negatively correlated.

3.4. Multivariate analysis

Macrofauna assemblages changed from before to after impact,with a significant ‘beach � time’ interaction (p (perm) ¼ 0.0008).Pair-wise comparisons indicate that the taxonomic structure of themacrofauna at the impacted site changed statistically from beforeto after impact (p (MC) ¼ 0.0001). The same trend was observed inthe intermediate sector, but in the protected sector differenceswere not detected. The PERMANOVA test also showed a significanteffect on the beach factor (p (perm) < 0.01) (Table 4).

The differences in the structures of the community can beobserved in the nMDS plot (Fig. 5), where the direction of changeover time was different for the urban and intermediate sectorscompared with the protected sector. At each sector, heterogeneitywas not found in the multivariate dispersion over time (PERMDISP;Urban F1,142 ¼ 2.93, p(perm) ¼ 0.13; Intermediate F1,142 ¼ 4.19,p(perm) ¼ 0.06; Protected F1,142 ¼ 2.48, p(perm) ¼ 0.14).

The SIMPER test showed a high level of dissimilarity in thecommunities between the time before and after impact, both in theurbanized (92.42%) and intermediate (90.22%) sectors (Table 5). Inboth areas, the amphipod Bathyporeia pelagica, the polychaeteScolelepis squamata, the mollusc Donax trunculus, and the cumaceaCumopsis fageiwere the taxa that contributed most to the temporaldifferences, accounting for 56% of the total dissimilarity betweensampling periods in the urban sector and 46% in the intermediatesector. Moreover, the polychaete Paraonis fulgens and the amphipod

Table 3Results of three-way PERMANOVA and pair-wise comparisons testing for differences in univariate measures. Only taxa showing a significant ‘beach � time’ interaction areshown.

Source df Richness Diversity index Density Bathyporeia pelagica

MS F P MS F P MS F P MS F P

Be 2 1.60 4.90 0.0396 3.18 15.002 0.0028 14.06 6.69 0.0213 9.97 15.16 0.012Ti 1 11.49 12.96 0.1028 15.34 26.47 0.1019 88.60 7.54 0.0987 1139.5 8.06 0.046Sp(Ti) 4 0.88 4.77 0.0014 0.57 3.46 0.0084 11.74 6.93 0.0001 17.31 11.63 0.0001BexTi 2 0.57 1.75 0.2344 1.24 5.88 0.295 12.13 5.77 0.0318 14.83 22.61 0.0007BexSp(Ti) 8 0.33 1.76 0.0878 0.21 1.26 0.2517 2.10 1.24 0.2665 0.66 0.44 0.89Res 414 0.18 0.16 1.69 1.49Total 431

Pair-wise test Groups Density B. pelagica

t P (MC) t P (MC)

Urban bef, after 3.11 0.0359 4.56 0.0096Intermediate bef, after 2.79 0.048 3.41 0.0292Protected bef, after 0.93 0.4024 0.868 0.4403

M.J. Reyes-Martínez et al. / Marine Environmental Research 103 (2015) 36e45 41

Pontocrates arenarius also contributed greatly to the differencesbetween periods in the intermediate sector. A complete list ofspecies that contributed to the differences between times in eachsector is shown in Table 5.

Of all the species identified in the SIMPER analysis, only Bath-yporeia pelagica showed a significant ‘beach � time’ interaction (p(perm) < 0.05) (Table 3). In the protected sector, Bathyporeiapelagica decreased in density after impact (27.6 2 ± 49.7 ind/m2)

Fig. 4. Temporal variation (mean ± SE) in each sector of (a) richness, (b) density (ind/m2) animpact.

compared to before impact (59.1 ± 17.8 ind/m2), but not as mark-edly as in the other two sectors. In the intermediate sector thedensity decreased from 90.6 ± 19.6 ind/m2 before impact to2.4 ± 7 ind/m2 after impact, while in the urban sector individualswere not found after impact (from 36.2 ± 8.2 ind/m2 to 0 ind/m2).Furthermore, changes in the density of three species were recor-ded. The densities of Eurydice affinis and Haustorius arenariusincreased after impact in the protected area, while in the other

d (c) diversity index. Black bars represent before impact and white bars represent after

Table 4PERMANOVA result testing for differences in macrofauna assemblages betweensectors and pair-wise of term 'BexTi' interaction.

Source df MS Pseudo-F P(perm)

Be 2 23377 9.10 0.0002Ti 1 95410 18.22 0.099Sp(Ti) 4 5234.5 2.34 0.0003BexTi 2 12944 5.04 0.0008Bex Sp(Ti) 8 2568 1.15 0.2277Res 414 2230.5Total 431

Pair-wise test Groups t P(MC)

Urban bef, aft 4.33 0.0001Intermediate bef, aft 3.55 0.0001Protected bef, aft 1.55 0.0714

Table 5SIMPER analysis to evaluate the contributions of taxa to dissimilarities from beforeto after impact in urban and intermediate sectors.

Groups: Urban before & urban after; Average dissimilarity: 92:42%

Species urban sector Before After Av.Diss Diss/SD Contrib% Cum.%

Av.Abund Av.Abund

Bathyporeia pelagica 1.46 0 15.67 0.88 16.96 16.96Scolelepis squamata 0.51 1.12 14.94 0.69 16.17 33.13Cumopsis fagei 1.34 0.03 11.21 0.89 12.13 45.26Donax trunculus 0.66 0.65 10.46 0.65 11.32 56.57Pontocrates arenarius 0.71 0.08 7.73 0.59 8.36 64.93Mactra stoultorum 0.59 0 5.04 0.44 5.46 70.39Eurydice affinis 0.3 0.04 4.41 0.33 4.78 75.17Nepthys hombergii 0.28 0.18 3.55 0.44 3.84 79.01

M.J. Reyes-Martínez et al. / Marine Environmental Research 103 (2015) 36e4542

sectors this parameter decreased. The Pontocrates arenarius den-sities followed the same pattern of decline in all sectors afterimpact, but this change was less pronounced in the protectedsector. However, these differences were not detected in the PER-MANOVA analysis (Fig. 6).

4. Discussion

In this study, the response of macrofauna assemblages thatinhabit sandy beaches to human trampling, which mainly occurs inthe summer season, was analyzed. For this purpose, three con-trasting sectors of the same beach were investigated: an urban areawith a high level of visitors, a protected sector belonging to a nat-ural park with a low density of users, and an intermediate zone,which was also within the natural park but had high level of humanoccupancy.

The density of macrofauna and community composition showeddifferent trajectories in each sector. The urban and intermediatesectors followed the same pattern, i.e., a dramatic reduction inspecies density and a significant change in the structure of thecommunity from before to after impact. However, the protectedsector showed greater stability throughout the study period andsignificant changes were not observed in the community de-scriptors and community structure. It is well known that macro-fauna vary within a beach in the along-shore directions dependingon the susceptibility of each species to environmental factors. As aconsequence, changes in sand particle size, swash climate, andmorphodynamics, among other factors, can explain these

Fig. 5. Non-metric multidimensional scaling ordination (nMDS) based on theBrayeCurtis dissimilarity measure of centroids for each sector and for after and beforeimpact. Triangles represent the urban sector, squares the intermediate, and circles theprotected sector. Black figures indicate before impact and white figures after impact.

variations (Defeo and McLachlan, 2005). The results obtained inthis study showed that the physical variables remained constantover time in each sector and between sectors. It therefore appearsthat this is not themain factor that causes biotic variation. Althoughseasonal variations may also affect macrofauna communities(Harris et al., 2011), our study was carried out on a spatial scale thatis not sufficiently large for biotic differences are caused by thisphenomenon. Human activity is also considered to be an additionalsource of variability (Defeo and McLachlan, 2005). Since the num-ber of beach users differed statistically between sectors and wasnegatively correlated with the species density, the biotic variationcan be tentatively attributed to human trampling activity.

In many cases, it is difficult to disentangle the effects of tram-pling from those generated by other impacts inherent in coastaldevelopment (see Schlacher and Thompson, 2012). The factors thatare most valued by visitors to a beach have been identified as beachcleanliness, comfort and safety, good access, parking areas, andgood facilities (such as restaurants, bars, boulevard, access to thebeach, litter bins, and shower facilities) (Roca and Villares, 2008;Rolfe and Gregg, 2012). Thus, in an effort to promote and supporttourism, beach managers initiate infrastructure improvements thattransform the beaches into increasingly urbanized areas and suchimprovements become increasing stressors on these ecosystems.Although tourism has undoubted economic benefits, it is usuallyassociated with substantial environmental costs (Davenport andDavenport, 2006). Different studies concerning nourishment(Leewis et al., 2012; Peterson et al., 2014; Schlacher et al., 2012),beach cleaning (Dugan and Hubbard, 2010; Gilburn, 2012), andcoastal armoring (Dugan et al., 2008; Hubbard et al., 2013) have

Corbula gibba 0.26 0.2 3.22 0.46 3.49 82.5Dispio uncinata 0.29 0.13 3.09 0.38 3.35 85.84Paraonis fulgens 0.31 0.06 2.97 0.41 3.22 89.06Glycera tridactyla 0.23 0.14 2.65 0.38 2.87 91.93

Groups: Intermediate Before & Intermediate After; Average dissimilarity:90.22%

Species Intermediatesector

Before After Av.Diss Diss/SD

Contrib%

Cum.%

Av.Abund Av.Abund

Cumopsis fagei 2.18 0.12 13.87 1.23 15.38 15.38Bathyporeia pelagica 1.79 0.24 12.88 0.89 14.28 29.65Scolelepis squamata 0.26 0.93 7.68 0.58 8.51 38.17Donax trunculus 0.95 0.65 7.54 0.75 8.36 46.52Paraonis fulgens 0.95 0.25 6.18 0.74 6.85 53.38Pontocrates arenarius 0.78 0.42 6.14 0.71 6.81 60.18Gastrosaccus sanctus 0.86 0 4.96 0.63 5.5 65.68Corbula gibba 0.67 0.11 4.49 0.6 4.98 70.66Haustorius arenarius 0.36 0.4 4.47 0.5 4.95 75.62Glycera tridactyla 0.32 0.21 3.04 0.46 3.37 78.98Nepthys hombergii 0.2 0.24 2.88 0.4 3.19 82.17Dispio uncinata 0.26 0.27 2.66 0.47 2.95 85.13Eurydice affinis 0.21 0.19 2.62 0.36 2.9 88.03Mactra stoultorum 0.29 0.08 2.36 0.31 2.62 90.65

Fig. 6. Mean density (±SE) of (a) Bathyporeia pelagica, (b) Eurydice affinis, (c) Haustorius arenarius, and (d) Pontocrates arenarius.

M.J. Reyes-Martínez et al. / Marine Environmental Research 103 (2015) 36e45 43

shown the negative effects of these actions on the beach fauna,mainly because they cause changes in the habitat, destroy the dunesystems, change the natural physical characteristics of the beaches,eliminate food sources, and reduce habitats and shelter areas,among others. Furthermore, these actions indirectly affect othercomponents of the food chain, such as shorebirds and fish, due to areduction in their food sources (Defeo et al., 2009). The resultsobtained in this study are consistent with this scenario as theyshow that the urban area before impact had the lowest values forcommunity descriptors. In addition, the correlation coefficient be-tween benthos density and number of users was lower than in theintermediate sector, which could suggest that in the urban areaother factors are influencing the density decrease, i.e., coastalarmoring and urbanization.

The effect of trampling can be addressed experimentally, butthe results will probably not reflect natural conditions (Ugoliniet al., 2008) due to the inability to mimic the real impact onboth the temporal and spatial scales. This limitation occursbecause temporally experiments have a fixed period and do notlast as long as the real impact and spatially they are performedwithin limited areas, which might be avoided by the beach faunaby simply moving to undisturbed areas. The transitional zoneselected in this study is a suitable enclave to evaluate the effect oftrampling on macrofauna communities uncoupled from otherfactors. This area has natural characteristics (e.g., without man-made structures and with a backshore with dune systems) but, likethe urban sector, it receives a large tourist influx during thesummer as facilities are provided for human access. Thus, the highcorrelation coefficient found between macrofauna density sug-gests that trampling has a negative effect on the beach fauna,causing a decrease in density and altering the composition of thecommunity.

At the population level, amphipods have commonly beenconsidered as bioindicators, especially supralittoral speciesbelonging to the family Talitridae (Fanini et al., 2005; Ugolini et al.,2008; Veloso et al., 2009; Weslawski et al., 2000). In fact, in 2008Veloso et al. performed a study at the same beach and they founddifferences in the Talitrus saltator density between sectors. Talitridpopulations in the protected and intermediate sites were main-tained throughout the year, while in the urban area they werenonexistent. So, the absence of this species, combined with theresults obtained in this study, show the negative effects that urbanbeaches have on the macrofauna that inhabit them. These negativeinfluences are due to the high number of beach visitors and thesignificant modifications made to the beach.

Beyond the Talitritridae family, species of Haustoridae, Ponto-poreiidae, Oedicerotidae, and also Cirolanidae isopods are consid-ered to be susceptible to the enrichment of organic matter (Chaoutiand Bayed, 2009), although very little is known about the ecologicalimplications of human activity. Haustorius arenarius, Pontocratesarenarius, and E. affinis showed changes in density throughout thestudy and these may be due to pedestrian activity. However, onlychanges in Bathyporeia pelagica were significant. In all sectors thisamphipod density decreased after impact. The decline was moremarked in the intermediate and urban sectors, where the densityreached minimum values to the extent that specimens were notfound at all in some cases. The annual cycle of the Bathyporeiagenus includes two reproductive peaks in spring and autumn (Fishand Preece, 1970; Mettam, 1989) and the declining density trendobserved suggests that these species are highly vulnerable totrampling impact. Theway inwhich activity has a negative effect onbeach communities is probably the result of sediment compaction,which might hinder burrowing and thus reduce the probability ofsurvival (Ugolini et al., 2008) or increase the probability of being

M.J. Reyes-Martínez et al. / Marine Environmental Research 103 (2015) 36e4544

killed by direct crushing (Rossi et al., 2007). In addition to the ef-fects at the population and community levels, human tramplingmay also have adverse consequences at the ecosystem level. In fact,protected beaches are more complex (with higher trophic levels),organized, mature, and active environments than urbanized bea-ches (Reyes-Martínez et al., 2014).

Although the potential for recovery of the beach fauna was notaddressed in this work, since the study area has been subjected tohuman impact for years, the ‘before impact’ state considered herecould be seen as a reflection of subsequent recovery. Thus, althoughtrampling causes a significant decrease in species density, main-tenance of the natural characteristics of the beaches (as in the in-termediate sector) might enable a potentially rapid recovery of thecommunity (see Carr, 2000). However, the consequence of inten-sive use by beach visitors in urbanized areas could be a long-termloss of biodiversity which might become irreversible. Further-more, the stability of the macrofauna communities found withinthe protected area highlights the importance of these areas in theconservation and maintenance of biodiversity.

Given the important role of macrofauna on the beaches(McLachlan and Brown, 2006), as well as the numerous servicesprovided by these ecosystems (Defeo et al., 2009), it is critical thatmanagement policies focus on the protection of these areas and onthe recovery and restoration of those areas that have already beendegraded. Recommendations that consider macrofauna are beingdeveloped so that managers can ensure the suitable use of beaches(McLachlan et al., 2013). However, such steps are still not sufficientbecause they are rarely applied and these ecosystems continue tobe ignored in conservation initiatives (Harris et al., 2014).

In conclusion, human trampling is an important disturbingfactor for the macrobenthos that inhabit sandy beaches. This factoracts by decreasing benthic densities and, consequently, a change inthe community occurs. When this activity occurs in urbanizedareas, a long-term irreversible loss biodiversity could result. Not allspecies respond to impact in a similar way and it seems that theamphipod Bathyporeia pelagica is highly sensitive to human tram-pling pressure. The use of this amphipod as a bioindicator for thisimpact type is therefore recommended. Although areas that retainnatural features may have a high capacity for recovery, futurestudies should be performed in order to test this hypothesis.

Acknowledgements

Special thank to Natural Park “Los Toru~nos” (C�adiz) and its stafffor the facilities provided during sampling. We also thank anony-mous reviewers for their constructive comments on the finalmanuscript and Neil Thompson for English revision. This work wassupported by the incentive program to excellent research projects,financed by the Government of Andalucía (P09-HUM-4717)through a PhD Grant awarded to the first author. M.C. Ruiz-Delgadowas supported by the Spanish Ministry of Education via a predoc-toral grant (FPU) (AP-2009-3906).

References

Aguado-Gim�enez, Piedecausa, M.A., Guti�errez, J.M., García-Charton, J.A.,Belmonte, A., 2012. Benthic recovery after fish farming cessation: a “beyond-BACI” approach. Mar. Pollut. Bull. 64, 729e738.

Anderson, M.J., 2001. A new method for non-parametric multivariate analysisofvariance. Austral. Ecol. 26, 32e46.

Anderson, M.J., 2005. Permanova: a FORTRAN Computer Program for PermutationalMultivariate Analysis of Variance. Department of Statistics, University ofAuckland, New Zealand, Auckland.

Anderson, M.J., Gorley, R.N., Clarke, K.R., 2008. PERMANOVA for PRIMER: Guide toSoftware and Statistical Methods. PRIMEReE Ltd., Plymouth, United Kingdom.

Barros, F., 2001. Ghost crabs as a tool for rapid assessment of human impacts onexposed sandy beaches. Biol. Conserv. 97, 399e404.

Benavente, J., Del Río, L., Anfuso, G., Gracia, F.J., Reyes, J., 2002. Utility of morpho-dynamic characterisation in the prediction of beach damage by storms. J. Coast.Res. 36, 56e64.

Buitrago, N.R., Anfuso, G., 2011. Morphological changes at Levante Beach (C�adiz, SWSpain) associated with storm events during the 2009e2010 winter seasons.J. Coast. Res. 64, 1886e1890.

Carr, M.H., 2000. Marine protected areas: challenges and opportunities for under-standing and conserving coastal marine ecosystems. Environ. Conserv. 27,106e109.

Chaouti, A., Bayed, A., 2009. Categories of importance as a promising approach tovaluate and conserve ecosystem integrity: the case study of Asilah sandy beach(Morocco). In: Bayed, A. (Ed.), Sandy Beaches and Coastal Zone Management,Proceedings of the Fifth International Symposium on Sandy Beaches (Rabat,Morocco), vol. 6. Travaux de l'Institut Scientifique, pp. 107e110.

Cisneros, K.O., Smit, A.J., Laudien, J., Schoeman, D.S., 2011. Complex, dynamiccombination of physical, chemical and nutritional variables controls spatio-temporal variation of sandy beach community structure. PloS One 6, e23724.

Clarke, K.R., Gorley, R.N., 2006. PRIMER v6: User Manual/Tutorial. PRIMER-E,Plymouth.

Davenport, J., Davenport, J.L., 2006. The impact of tourism and personal leisuretransport on coastal environments: a review. Estuar. Coast. Shelf Sci. 67,280e292.

Defeo, O., McLachlan, A., 2005. Patterns, processes and regulatory mechanisms insandy beach macrofauna: a multi-scale an�alisis. Mar. Ecol. Prog. Ser. 296, 1e20.

Defeo, O., McLachlan, A., Schoeman, D.S., Schlacher, T.A., Dugan, J., Jones, A.,Lastra, M., Scapini, F., 2009. Threats to sandy beach ecosystems: a review.Estuar. Coast. Shelf Sci. 81, 1e12.

Del Río, L., Gracia, F.J., Benavente, J., 2013. Morphological and evolutionary classi-fication of sandy beaches in Cadiz coast (SW Spain). In: Conley, D.C.,Masselink, G., Russell, P.E., O'Hare, T.J. (Eds.), Proceedings 12th InternationalCoastal Symposium (Plymouth, England). J. Coast. Res. Special Issue vol. 65,2113e2118.

Dugan, J., 1999. Utilization of Sandy Beaches by Shorebirds: Relationships to Pop-ulation Characteristics of Macrofauna Prey Species and Beach Morphodynamics.Draft Final Technical Report. In: Outer Continental Shelf Study. Minerals Man-agement Service, Caramillo, CA.

Dugan, J.E., Hubbard, D.M., 2010. Loss of coastal Strand habitat in Southern Cali-fornia: the role of Beach grooming. Estuaries Coasts 33, 67e77.

Dugan, J.E., Hubbard, D.M., McCrary, M., Pierson, M., 2003. The response of mac-rofauna communities and shorebirds to macrophyte wrack subsidies onexposed beaches of Southern California. Estuar. Coast. Shelf Sci. 58S, 133e148.

Dugan, J.E., Hubbard, D.M., Rodil, I.F., Revell, D.L., Schroeter, S., 2008. Ecologicaleffects of coastal armoring on sandy beaches. Mar. Ecol. 29, 160e170.

Emery, K.O., 1961. A simple method of measuring beach profiles. Limnol. Oceanogr.6, 90e93.

Fanini, L., Cantarino, C.M., Scapini, F., 2005. Relationship between the dynamics oftwo Talitrus saltator populations and the impacts of activities linked to tourism.Oceanologia 47, 93e112.

Fanini, L., Zampicinini, G., Pafilis, E., 2014. Beach parties: a case study on recrea-tional human use of the beach and its effects on mobile arthropod fauna. Ethol.Ecol. Evol. 26, 69e79.

Ferreira, M.N., Rosso, S., 2009. Effects of human trampling on a rocky shore fauna onthe Sao Paulo Coast, Southeastern Brazil. Braz. J. Biol. 69, 993e999.

Fish, J.D., Preece, G.S., 1970. The annual reproductive patterns of Bathyporeia pilosaandBathyporeia pelagica (Crustacea: Amphipoda). J. Mar. Biological Assoc. U. K.50, 475e488.

Gilburn, A.S., 2012. Mechanical grooming and beach award status are associatedwith low strandline biodiversity in Scotland. Estuar. Coast. Shelf Sci. 107, 81e88.

Guitian, F., Carballas, J., 1976. T�ecnicas de an�alisis de suelos (Pico Sacro. Santiago deCompostela. Espa~na).

Harris, L., Nel, R., Smale, M., Schoeman, D., 2011. Swash away? Storm impacts onsandy beach macrofaunal communities. Estuar. Coast. Shelf Sci. 94, 210e221.

Harris, L., Campbell, E.E., Nel, R., Schoeman, D., 2014. Rich diversity, strong ende-mism, but poor protection: addressing the neglect of sandy beach ecosystemsin coastal conservation planning. Divers. Distributions 1e16.

Hockings, M., Twyford, K., 1997. Assessment and management of beach campingwithin Fraser Island World Heritage Area, South East Queensland. Aust. J. En-viron. Manag. 4, 25e39.

Hubbard, D.M., Dugan, J.E., Schooler, N.K., Viola, S.M., 2013. Local extirpations andregional declines of endemic upper beach invertebrates in southern California.Estuar. Coast. Shelf Sci. http://dx.doi.org/10.1016/j.ecss.2013.06.017.

Jaramillo, E., Contreras, H., Quijon, P., 1996. Macroinfauna and human disturbance ina sandy beach of south-central Chile. Rev. Chil. Hist. Nat. 69, 655e663.

Jennings, S., 2004. Coastal tourism and shoreline management. Ann. Tour. Res. 31,899e922.

Lastra, M., Page, H.M., Dugan, J.E., Hubbard, D.M., Rodil, I.F., 2008. Processing ofallochthonous macrophyte subsidies by sandy beach consumers: estimates offeeding rates and impacts on food resources. Mar. Biol. 154, 163e174.

Leewis, L., Van Bodegom, P.M., Rozema, J., Janssen, G.M., 2012. Does beach nour-ishment have long-term effects on intertidal macroinvertebrate species abun-dance? Estuar. Coast. Shelf Sci. 113, 172e181.

Lercari, D., Bergamino, L., Defeo, O., 2010. Trophic models in sandy beaches withcontrasting morphodynamics: comparing ecosystem structure and biomassflow. Ecol. Model. 221, 2751e2759.

M.J. Reyes-Martínez et al. / Marine Environmental Research 103 (2015) 36e45 45

Lucrezi, S., Schlacher, T.A., Robinson, W., 2009. Human disturbance as a cause of biasin ecological indicators for sandy beaches: experimental evidence for the ef-fects of human trampling on ghost crabs (Ocypode spp.). Ecol. Indic. 9, 913e921.

Martínez, M.L., Intralawan, A., V�azquez, G., P�erez-Maqueo, O., Sutton, P.,Landgrave, R., 2007. The coasts of our world: ecological, economic and socialimportance. Ecol. Econ. 63, 254e272.

McLachlan, A., Brown, A.C., 2006. The Ecology of Sandy Shores. Academic Press,Burlington, Massachusetts.

McLachlan, A., Defeo, O., Jaramillo, E., Short, A.D., 2013. Sandy beach conservationand recreation: guidelines for optimising management strategies for multi-purpose use. Ocean Conservation 71, 256e268.

Mettam, C., 1989. The life cycle of Bathyporeia pilosa Lindstr€om (Amphipoda) in astressful, low salinity environment. Sci. Mar. 53, 543e550.

Moffet, M.D., McLachlan, A., Winter, P.E.D., De Ruyck, A.M.C., 1998. Impact oftrampling on sandy beach macrofauna. J. Coast. Conservation 4, 87e90.

Peterson, C.H., Bishop, M.J., D'Anna, L.M., Johnson, G.A., 2014. Multi-year persistenceof beach habitat degradation from nourishment using coarse shelly sediments.Sci. Total Environ. 487, 481e492.

Reyes-Martínez, M.J., Lercari, D., Ruíz-Delgado, M.C., S�anchez-Moyano, J.E., Jim�enez-Rodríguez, A.J., P�erez-Hurtado, A., García-García, F.J., 2014. Human pressure onsandy beaches: implications for trophic functioning. Estuaries Coasts. http://dx.doi.org/10.1007/s12237-014-9910-6.

Roca, E., Villares, M., 2008. Public perceptions for evaluating beach quality in urbanand semi-natural environments. Ocean Coast. Manag. 51, 314e329.

Rodgers, K.S., Cox, E.F., 2003. The effects of trampling on Hawaiian corals along agradient of human use. Biol. Conserv. 112, 383e389.

Rolfe, J., Gregg, D., 2012. Valuing beach recreation across a regional area: the GreatBarrier Reef in Australia. Ocean Coast. Manag. 69, 282e290.

Rossi, F., Forster, R.M., Montserrat, F., Ponti, M., Terlizzi, A., Ysebaert, T.,Middelburg, J.J., 2007. Human trampling as short-term disturbance on intertidalmudflats: effects on macrofauna biodiversity and population dynamics of bi-valves. Mar. Biol. 151, 2077e2090.

Schlacher, T.A., Thompson, L.M.C., 2008. Physical impacts caused by off-road vehi-cles (ORVs) to sandy beaches: spatial quantification of car tracks on anAustralian barrier island. J. Coast. Res. 24, 234e242.

Schlacher, T.A., Thompson, L.M.C., 2012. Beach recreation impacts benthic in-vertebrates on ocean-exposed sandy shores. Biol. Conserv. 147, 123e132.

Schlacher, T.A., Dugan, J., Schoeman, D.S., Lastra, M., Jones, A., Scapini, F.,McLachlan, A., Defeo, O., 2007. Sandy beaches at the brink. Divers. Distributions13, 556e560.

Schlacher, T.A., Noriega, R., Jones, A., Dye, T., 2012. The effects of beach nourishmenton benthic invertebrates in eastern Australia: Impacts and variable recovery.Sci. Total Environ. 435e436, 411e417.

Schlacher, T.A., Schoeman, D.S., Jones, A.R., Dugan, J.E., Hubbard, D.M., Defeo, O.,Peterson, C.H., Weston, M.A., Maslo, B., Olds, A.D., Scapini, F., Nel, R., Harris, L.R.,Lucrezi, S., Lastra, M., Huijbers, C.M., Connolly, R.M., 2014. Metrics to assessecological condition, change, and impacts in sandy beach ecosystems.J. Environ. Manag. 144, 322e335.

Torres, A., Palacín, C., Seoane, J., Alonso, J.C., 2011. Assessing the effects of a highwayon a threatened species using BeforeeDuringeAfter and BeforeeDuringeAfter-ControleImpact designs. Biol. Conserv. 144, 2223e2232.

Ugolini, A., Ungherese, G., Somigli, S., Galanti, G., Baroni, D., Borghini, F., Cipriani, N.,Nebbiai, M., Passaponti, M., Focardi, S., 2008. The amphipod Talitrus saltator as abioindicator of human trampling on sandy beaches. Mar. Environ. Res. 65,349e357.

Underwood, A.J., 1992. Beyond BACI: the detection of environmental impactsonpopulations in the real, but variable, world. J. Exp. Mar. Biol. Ecol. 161,145e178.

Underwood, A.J., 1994. On beyond BACI: sampling designs that might Reliablydetect environmental disturbances. Ecol. Appl. 4, 3e15.

Veloso, V.G., Silva, E.S., Caetano, C.H.S., Cardoso, R.S., 2006. Comparison between themacroinfauna of urbanized and protected beaches in Rio de Janeiro State, Brazil.Biol. Conserv. 127, 510e515.

Veloso, V.G., Neves, G., Lozano, M., Perez-Hurtado, A., Gago, C.G., Hortas, F., García-García, F., 2008. Responses of talitrid amphipods to a gradient of recreationalpressure caused by beach urbanization. Mar. Ecol. 29, 126e133.

Veloso, V.G., Sallorenzo, I.A., Ferreira, B.C.A., Souza, G.N., 2009. Atlantorchestoideabrasiliensis (Crustacea: Amphipoda) as an indicator of disturbance caused byurbanization of a beach ecosystem. Braz. J. Oceanogr. 58, 13e21.

Weslawski, J.M., Stanek, A., Siewert, A., Beer, N., 2000. The sandhopper (Talitrussaltator, Montagu 1808) on the Polish Baltic Coast. Is a victim of increasedtourism? Oceanol. Stud. 29, 77e87.