Biological Conservation 166 (2013) 124–131

Contents lists available at SciVerse ScienceDirect

Biological Conservation

journal homepage: www.elsevier .com/locate /b iocon

Short communication

Recruitment hotspots boost the effectiveness of no-take marine reserves

0006-3207/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.biocon.2013.06.017

⇑ Corresponding author. Tel.: +61 7 4781 5569; fax: +61 7 4725 1570.E-mail address: colin.wen@my.jcu.edu.au (C.K.C. Wen).

Colin K.C. Wen a,⇑, Glenn R. Almany b, David H. Williamson b, Morgan S. Pratchett b,Thomas D. Mannering a, Richard D. Evans c,d, Jeffrey M. Leis e, Maya Srinivasan a, Geoffrey P. Jones a,b

a School of Marine and Tropical Biology, James Cook University, Townsville, QLD 4811, Australiab ARC Centre of Excellence for Coral Reef Studies, James Cook University, Townsville, QLD 4811, Australiac Department of Parks and Wildlife, 17 Dick Perry Ave., Kensington, Perth, WA 6151, Australiad Oceans Institute, School of Plant Biology, University of Western Australia, Crawley, WA 6009, Australiae Ichthyology, Australian Museum, Sydney, NSW 2010, Australia and Institute for Marine and Antarctic Studies, University of Tasmania, Hobart, TAS 7001, Australia

a r t i c l e i n f o

Article history:Received 6 February 2013Received in revised form 5 June 2013Accepted 15 June 2013

Keywords:Marine reserveRecruitmentPlectropomusCoral troutLutjanusSnapperGreat Barrier ReefN-mixture model

a b s t r a c t

No-take marine reserves are widely advocated as a means to conserve biodiversity and sustain fisheries.Within adequately protected reserves, the abundance, mean size and age of targeted fish and invertebratespecies are often increased. However, the magnitude and rate of recovery within reserves varies amonglocations and species, and increases in abundance in reserves are not universal. One potential contribut-ing factor to variability in reserve performance is spatial variability in the supply of recruits. Many reeffish species exhibit disproportionate levels of recruitment in relatively small areas (so-called ‘‘recruit-ment hotspots’’). Here we tested the hypothesis that the presence of recruitment hotspots enhancesthe effectiveness of reserves for two important fishery species, coral trout (Plectropomus maculatus)and stripey snapper (Lutjanus carponotatus), on coral reefs of the Keppel Island group, Great Barrier Reef.To do this we compared fish densities in areas that did or did not contain previously identified recruit-ment hotspots, both inside and outside reserves. For both species, the mean densities of adult fish abovethe minimum legal size were approximately three times as high in reserves with hotspots than inreserves without hotspots. Furthermore, the mean densities in reserves without hotspots were similarto those in nearby areas that were open to fishing. In contrast, the densities of sub-legal size individualsof both species were primarily explained by the presence of recruitment hotspots, irrespective of reserveprotection. Our results suggest that identifying and incorporating recruitment hotspots into the selectionof sites for reserves could enhance both conservation and fisheries objectives.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Networks of no-take marine reserves are being implementedaround the globe, with the primary objectives of protecting biodi-versity and/or enhancing fishery sustainability (Gell and Roberts,2003; McCook et al., 2010; Fenberg et al., 2012). Numerous studieshave demonstrated that the abundance, mean body size and age ofexploited fish species can increase dramatically within the bound-aries of adequately protected reserves (Halpern and Warner, 2002;Lester et al., 2009; McLean et al., 2011). However, there is consid-erable variation among studies, both in the magnitude of abun-dance increases in reserves relative to fished areas (20–490%more fish inside reserves; Williamson et al., 2004; Abesamiset al., 2006; Harmelin-Vivien et al., 2008) and in the rates of in-crease after reserves are established (Russ et al., 2008; Russ andAlcala, 2010). A number of studies have also demonstrated littleor no effect of reserves on the abundance of target fishery species

(i.e. Mapstone et al., 2008; McLean et al., 2011). The mechanismsby which some fish populations rapidly increase in abundance inreserves, and others do not, are poorly understood.

A range of factors may explain the variable success of reserves,including historic fishing pressure (Claudet et al., 2010), levels ofcompliance (Guidetti et al., 2008; Claudet and Guidetti, 2010; Poll-nac et al., 2010) and the life history and mobility of target species(Halpern and Warner, 2003; Nardi et al., 2004; Claudet et al., 2010).Assuming that mortality from fishing is largely eliminated insidereserves, the reserve population can persist through time only ifrecruitment levels equal natural mortality, and for abundances toincrease, recruitment must exceed natural mortality. In some casesabundance may increase in the reserve through immigration fromnearby fishing areas, but this effect is likely to be significantly lessimportant than recruitment over the long term (Stockhausen et al.,2000; Gerber et al., 2005). As a result, a key factor that contributesto the success of a reserve is whether or not it contains areas thatconsistently receive high levels of recruitment (Freeman et al.,2012). Despite the obvious importance of recruitment to reserveeffectiveness, patterns of recruitment inside and outside reserves,

C.K.C. Wen et al. / Biological Conservation 166 (2013) 124–131 125

and their relationship to the adult abundance of exploited fishes,have not been examined. Although the selection of sites for re-serves must inevitably take into account a range of social and eco-logical factors (Fernandes et al., 2005; Beger et al., 2010; Edwardset al., 2010; Adams et al., 2011; Hansen et al., 2011), the inclusionof adequate recruitment and nursery habitats inside reserves maybe an important ingredient for their long-term success (Kinin-month et al., 2011).

For most demersal marine fishes, recruitment occurs after indi-viduals have survived the pelagic larvae stage, settled into benthichabitat and survived the first few weeks post-settlement to jointhe local population (Armsworth, 2002). Coral reef fishes typicallyexhibit predictable spatial patterns in recruitment, both within andamong reefs (Valles et al., 2008). Reef locations that consistentlyreceive above average levels of recruitment have been referred toas ‘‘recruitment hotspots’’ (Booth et al., 2000; Eagle et al., 2012;Wen et al., 2013a). Recruitment hotspots may be explained by avariety of processes and factors, including larval supply (Meekanet al., 1993) and the quality of juvenile habitat (Feary et al.,2007; Wen et al., 2013b). Many exploited fish species that are ex-pected to increase in reserves also exhibit predictable spatial vari-ation in patterns of recruitment (Leis and Carson-Ewart, 1999;Kingsford, 2009; Quéré and Leis, 2010). However, the role ofrecruitment in explaining the effectiveness of coral reef reserveshas received little attention.

The aim of this study was to evaluate the hypothesis that well-connected reserve networks encompassing recruitment hotspotswill be more effective than those established in areas that receivelower levels of recruitment. To test this, we sampled an existingnetwork of reserves with a structured sampling design to comparethe density of adults and sub-adults of two exploited fishes insideand outside reserves, in areas with and without recruitment hot-spots. We focused on two commercially and recreationally impor-tant fishes, coral trout (Plectropomus maculatus) and stripeysnapper (Lutjanus carponotatus). As these species are subject tominimum legal size limits in Australian waters, we also examinedthe effects of the 4 combinations of 2 factors (reserves and hot-spots) on the size-frequency distribution of fishes.

2. Methods

2.1. Study location and species

This study was conducted during February 2010 on fringing cor-al reefs of the Keppel Island group (23�100S, 150�570E), an archipel-ago in the southern region of the Great Barrier Reef Marine Park. Itfocused on three no-take marine reserves that form part of a net-work of six no-take marine reserves in the Keppel Island group(Fig. 1). Two reef fish species were examined, bar-cheeked coraltrout (Plectropomus maculatus; Serranidae) and stripey snapper(Lutjanus carponotatus; Lutjanidae), both important species in rec-reational and commercial fisheries on the Great Barrier Reef. Sig-nificant increases in the abundance and/or mean size of thesespecies have been documented within no-take reserves of the Kep-pel Islands since 2004 (Russ et al., 2008; McCook et al., 2010).

2.2. Defining recruitment hotspots

We defined ‘‘recruitment hotspots’’ as areas that consistentlyreceive higher than average recruitment (sensu Booth et al.,2000; Eagle et al., 2012). Suitable habitat for newly-settled juve-niles (0–1 year-old) of P. maculatus and L. carponotatus have previ-ously been found in nearshore, back reef habitats with abundanttabular and corymbose Acropora corals over sandy substrates(Kingsford, 2009; Wen et al., 2013a). Previous underwater visual

census (UVC) data from reefs in the Keppel Islands has shown thatjuvenile densities in these specific habitats were much higher thanin all other reef habitat types (Wen et al., 2013a).

To make an apriori selection of specific hotspot and non-hot-spot sites for this study, we selected potential sites in both re-serve and non-reserve areas. We compared catch per unit effort(CPUE) data based on diver collections of the two species carriedout during two 20-day field trips in May 2008 and February 2009(Table 1). Twelve divers collected juveniles of both species withinan approximately 200 m-radius area from three boats, whichwere separated by at least 500 m to enable independence andto effectively cover the majority of the reef habitat area at eachsampled location (Fig. 1). Data were standardized to catch perunit effort (number of juvenile fish collected per person per dive).Sites with distinctly high CPUE were defined as recruitment hot-spots (Table 1). The thresholds of CPUE between hotspots andnon-hotspots were >1.5 for P. maculatus and >2 for L. carponota-tus. The reefs around Halfway Island and Clam Bay, which aresplit-zoning, were identified as recruit hotspots (see Table 1).These reefs have large areas of suitable recruitment and juvenilenursery habitat for these species. On the other hand, the reefssurrounding Middle Island, Humpy Island, Miall Island and Mon-key Bay had similar suitable habitat, but less of it, and were de-fined as non-recruitment hotspots. The mean CPUE (±SE) acrossthe four non-hotspot sites was 0.53 ± 0.11 and 0.91 ± 0.15 for P.maculatus and L. carponotatus respectively

2.3. Sampling design and survey methods

An orthogonal sampling design was employed to comparedensities of fish in reserves with and without recruitment hot-spots, and non-reserves with and without recruitment hotspots.This sampling design permitted analyses using two factors; (1)Zone (Reserve or Non-reserve), and (2) Recruitment hotspot(presence or absence). A total of 8 sites were surveyed usingunderwater visual census (UVC), 4 sites within no-take reservesand 4 sites in areas that were open to fishing. All the samplingsites had similar reef environments and 2 sites from both re-serve and fished areas contained recruitment hotspots (Fig. 1).The reef area within each location was measured from satelliteimagery using the software program ImageJ (v. 1.4.6; Rasband,1997). The specific areas were (1) Eastern Clam bay (reservewith recruitment hotspot): 30.8 ha, (2) Western Clam bay (fishedarea with recruitment hotspot): 4.5 ha, (3) Western Halfway Is-land (fished area with recruitment hotspot): 13.2 ha, (4) EasternHalfway Island (reserve with recruitment hotspot): 10.1 ha, (5)Humpy Island (fished area without recruitment hotspot):33.5 ha, (6) Monkey bay (reserve without recruitment hotspot):9.5 ha, (7) Middle Island (reserve without recruitment hotspot):17.9 ha and (8) Miall Island (fished area without recruitmenthotspot): 7.4 ha. The proximity of reserve and non-reserve hot-spots may be considered a potential confounding issue for highlymobile species. However, biannual surveys over the past decadesuggest strong consistent differentiation in mean densities be-tween reserves and fished areas of the Keppel Islands (Russet al., 2008; Evans and Russ, 2004). Furthermore, Plectropomusspp. and L. carponotatus are known to have relatively restrictedhome ranges (Zeller, 2002) and low rates of movement betweenreserves and fished areas are unlikely to significantly change thedemographics of populations within those areas.

We recorded the number and estimated the total length (TL) ofall sighted individuals of both species along 30-min timed-swimUVC transects. The length of each transect was measured using adiver-towed GPS system. The mean transect length was112 m ± 17 m (SE). At each of the 8 sites, 6 replicate transects wereconducted at randomly selected sites along reef flats and reef

Fig. 1. Map of the study area in the Great Keppel Island group, southern Great Barrier Reef. Crosshatch shaded areas with dashed outlines represent the fringing coral reefs.Dark grey areas with dashed outline represent no-take reserves (Clam bay, Monkey bay and Middle Island). Recruitment hotspots were defined according to catch per uniteffort data from diver collections (Table 1). Orthogonal sampling was designed with two factors: recruitment hotspots and reserves. Four sites were in marine reserves (whitecircle): 2 with recruitment hotspots (circle with cross) and 2 without hotspots. Four sites were in areas that were open to fishing (dark circle): 2 with recruitment hotspots(circle with cross) and 2 without hotspots.

Table 1The catch per unit effort (CPUE) or average number of Plectropomus maculatus and Lutjanus carponotatus collected per person per dive, for the hotspot and non-hotspot sites in thisstudy. The arbitrary threshold for defining a location as a recruitment hotspot was a >1.5 CPUE in P. maculatus and >2 CPUE in L. carponotatus.

Site Plectropomus maculatus Lutjanus carponotatus Recruitment level Management status

Western clam bay (Great Keppel Island) 1.54 ± 0.38 2.21 ± 0.04 Hotspot ReserveEastern clam bay (Great Keppel Island) 1.54 ± 0.38 2.21 ± 0.04 Hotspot Non-reserveWestern Halfway Island 3.50 ± 0.08 2.08 ± 0.17 Hotspot Non-reserveEastern Halfway Island 3.50 ± 0.08 2.08 ± 0.17 Hotspot ReserveHumpy Island 0.67 ± 0.50 0.54 ± 0.38 Non-hotspot Non-reserveMiddle Island 0.57 ± 0.10 1.03 ± 0.37 Non-hotspot ReserveMiall Island 0.25 ± 0.08 0.86 ± 0.69 Non-hotspot Non-reserveMonkey bay (Great Keppel Island) 0.42 ± 0.25 0.67 ± 0.33 Non-hotspot Reserve

126 C.K.C. Wen et al. / Biological Conservation 166 (2013) 124–131

slopes, oriented parallel to the reef crest. Reef flat and reef slopeswere selected to representatively sample the adult and sub-adultpopulations. Although these two species undergo ontogeneticshifts in microhabitat utilization and depth preferences (Wenet al., 2013a), sub-adults and adults are most abundant on reefslopes and the survey data from this habitat are representative ofoverall abundance (Williamson et al., 2004). Recruitment hotspotsin this study were classified on the basis of a priori information onjuvenile densities and were simply defined as areas with signifi-cantly greater abundance of juveniles over several years (sensu Ea-gle et al., 2012).

2.4. Data analysis

All recorded individuals of both species were assigned to one oftwo groups based on their estimated length either below or abovethe species-specific Queensland legal minimum size limits (MSL).Fish below the MSL were not exposed to fishing mortality, whereasfish above the MSL were only exposed to fishing mortality in fishedareas (non-reserves) and not within reserves. The MSL in Queens-land waters for P. maculatus is 38 cm TL and L. carponotatus is25 cm TL (QLD Fisheries Act, 2004) and they are set to allow a win-dow for reproduction prior to being exposed to fishing mortality.For the purposes of this study, we classified individuals above

MSL as ‘‘adults’’, and those individuals with lengths betweenrecruits and adults as ‘‘sub-adults.’’ Recruits were defined as youngof the year (0 + years). The defined maximum body length for re-cruit P. maculatus was 15 cm TL, and for L. carponotatus it was12 cm TL.

We used four N-mixture models to evaluate the strongestgoodness of fit (using a combination of factors, 1. zone: reservevs non-reserves and 2. recruitment hotspots: present vs absent)to explain the densities of adult and sub-adult fish (e.g. Josephet al., 2009). The four models were Poisson (P), negative bino-mial (NB), zero-inflated Poisson (ZIP) and zero-inflated negativebinomial (ZINB). Poisson and negative binomial distributionmodels were chosen because of the potential for high variabilityin our data, while the zero-inflated models were included to ac-count for possible over-dispersion of the data (e.g. high propor-tion of zeros; Adams et al., 2011). Factors were examined withinthe framework of the four models using maximum likelihood toestimate parameters and a log-link function to account for thenon-normal distribution of the data. This approach was takenas no prior assumption of homogeneity was necessary. The nullmodel (no factor) and alternative models with a combination ofthe two factors (reserves and hotspots) were compared withAkaike’s information criterion (AIC; Symonds and Moussalli,2011). AIC corrected (AICc) was used in this study due to the

C.K.C. Wen et al. / Biological Conservation 166 (2013) 124–131 127

small sample size. The model with the greatest AICc weight wasselected over other potential models. All statistical analyses wereconducted in the software program R (R 3.0.0, R DevelopmentCore Team, 2011).

Size-frequency distributions of fishes in the 4 treatments werevisually compared in relation to the MSLs for the two fish species.Size distribution histograms were generated using 14 size classesfor P. maculatus and 8 size classes for L. carponotatus in 5 cm incre-ments according to the size range of these two species. Each sizeclass includes size greater (>) than the previous scale and less thanor equal to (6) next scale. The size frequency histograms amongthe four treatments were compared using non-parametric Kol-mogorov–Smirnov tests (Heppell et al., 2012) in the statisticalanalyses software PAST (Hammer et al., 2001). The null hypothesis(H0) is that the two samples are taken from populations with equaldistributions. In addition to the significance value p, the statistic Dis the maximum deviation between the two empirical cumulativedistribution functions.

3. Results

3.1. Adults

Densities of adult P. maculatus differed between reserves andnon-reserves, and between sites with and without recruitment

(a)

(b)

Fig. 2. Mean density of (a) adults and (b) sub-adults for the two fishery species in reservareas open to fishing with recruitment hotspots (N = 2) and areas open to fishing witho

Table 2Best N-mixture model results testing the influence of reserve status (reserve and non-reservof two predatory fishes (Plectropomus maculatus and Lutjanus carponotatus). AICc (Akaikegoodness-of-fit. Best model with goodness-of-fit was chosen from four N-mixture modenegative binomial (ZINB). Details of the model selected for each species and size group ar

Species Size group Best model for goodness of fit

Plectropomus maculatus Adult NBSubadult NB

Lutjanus carponotatus Adult ZIPSubadult NB

hotspots. The highest densities were recorded in reserves withrecruitment hotspots, where densities were nearly three times ashigh as in the other three treatments (reserves without hotspots,non-reserves areas with hotspots, non-reserves without hotspots;Fig. 2a). The statistical model (NB) with the greatest support con-tained both the zone and recruitment hotspot terms, but therewas no evidence of a statistical interaction between the two factors(Table 2). Similarly, densities of adult L. carponotatus were greatestin reserves that contained recruitment hotspots, and were morethan twice as high as in the other three treatments (Fig. 2a). How-ever, the best-fit statistical model for L. carponotatus was the zero-inflated Poisson (ZIP) with zone and recruitment-hotspot terms(Table 2).

3.2. Sub-adults

In contrast to adults, densities of sub-adults were greatest atrecruitment hotspots, irrespective of zone (Fig. 2b). For both studyspecies, the best-fit model contained only the term for recruitmenthotspots (NB model, Table 2). Densities of sub-adult P. maculatuswere 3–5 times greater at sites with recruitment hotspots thanat those without hotspots (Fig. 2b). Similar patterns were observedfor L. carponotatus, although the magnitude of density differencesbetween sites with and without hotspots was not as large as thatobserved for P. maculatus. High spatial variability of sub-adult L.

es with recruitment hotspots (N = 2), reserves without recruitment hotspots (N = 2),ut recruitment hotspots (N = 2), within the Keppel Islands. Error bars are +/�1 SE.

e) and recruitment hotspots (present and absent) on densities of adults and sub-adultsInformation Criterion, corrected) and Akaike weight were used to identify the best

ls; Poisson (P), negative binomial (NB), zero-inflated Poisson (ZIP) and zero-inflatede provided in supplementary material.

Formula df AICc Akaike weight

Abundance�zone + hotspot 4 267.46 0.156Abundance�hotspot 3 217.74 0.428

Abundance�zone + hotspot 4 230.20 0.227Abundance�hotspot 3 216.50 0.486

Table 3Result of paired Kolmogorov–Smirnov testing the effects of two factors – reserves andrecruitment hotspots – on the length-frequency distribution of (a) Plectropomusmaculatus and (b) Lutjanus carponotatus. Right upper corner of each table is thesignificance value p, Left Lower corner of each table is that statistic distance D, whichis the maximum deviation distance between two cumulative factions from twotreatments.

Res + HS Res � HS nRes + HS nRes � HS

(a)Res + HS – 0.255 0.001a 0.026a

Res � HS 0.1238 – 0.001a 0.398nRes + HS 0.3314 0.3884 – 0.001a

nRes � HS 0.1863 0.1409 0.3919 –

(b)Res + HS – 0.001a 0.001a 0.552Res � HS 0.237 – 0.02a 0.009a

nRes + HS 0.2707 0.1882 – 0.002a

nRes � HS 0.1131 0.2518 0.2795 –

a Denotes a significant difference (a < 0.05).

128 C.K.C. Wen et al. / Biological Conservation 166 (2013) 124–131

carponotatus densities was observed in non-reserves with recruit-ment hotspots.

3.3. Size frequency distributions, recruitment and the minimum legalsize limit

The size-frequency distributions of P. maculatus differed amongthe four treatments (Fig. 3). Areas with recruitment hotspots notonly had higher numbers of recruits (<15 cm TL) than areas with-out hotspots, but also had higher numbers of sub-adult fish belowMSL (38 cm TL). At the other extreme, there was decreased abun-dance of fish larger than the MSL of 38 cm TL on fished reefs, bothwith and without hotspots. Size-frequency distributions of P. mac-ulatus were significantly different (p < 0.05) in four of the six K-Stests, suggesting the two factors (reserves and recruitment hot-spots) affected the size-frequency distribution (Table 3a). How-ever, the test between Res-HS vs Res + HS was not significant,which indicates that the presence of recruitment hotspots didnot alter the size-frequency distribution of P. maculatus in reserves(Table 3a). Additionally, size-frequency distributions were not sig-nificantly different between reserves with hotspots and non-re-serves with hotspots (Res-HS vs nRes-HS, Table 3a).

The size-frequency distributions of L. carponotatus also differedamong the 4 treatments (Fig. 4). There was a higher density of re-cruits in reserves with recruitment hotspots, but this pattern wasless clear for fished reefs (Fig. 4). There was a truncation in the sizefrequency distribution of L. carponotatus at the MSL (25 cm TL) innon-reserves, but not in reserves. K-S test results for L. carponota-tus revealed significant differences in size-frequency distributionsamong five of the six treatments, the one exception was the

Non-reserves

Rec

ruitm

ent h

otsp

ots

Non

recr

uitm

ent h

otsp

ots

Siz

0

8

16

24

32

40

48

56

64

Freq

uenc

y

0

8

16

24

32

40

48

56

64

Freq

uenc

y

0 10 20 30 40 50 60 705 15 25 35 45 55 65 75

70

70

Fig. 3. Body length – frequency histograms of Plectropomus maculatus in the

comparison between reserves with hotspots and non-reserveswithout hotspots (Res + HS vs nRes-HS; Table 3b).

4. Discussion

Our results strongly suggest that areas of high recruitment en-hance the build-up of fishery-targeted species within marine re-serves. Adult densities of two species (P. maculatus and L.carponotatus) were 2–3 times higher in reserves with recruitmenthotspots than in reserves without hotspots. In fished areas that

Reserves

e (cm TL)80 0 10 20 30 40 50 60 70 805 15 25 35 45 55 65 75

Keppel islands using two factors (reserves and recruitment hotspots).

C.K.C. Wen et al. / Biological Conservation 166 (2013) 124–131 129

encompassed recruitment hotspots, adult densities of both specieswere similar to the densities observed inside reserves without hot-spots. In contrast, sub-adult densities of both species were highestat sites with recruitment hotspots irrespective of their manage-ment status. Previous work has demonstrated that P. maculatushas stronger habitat selectivity in the early juvenile phase than L.carponotatus (Wen et al., 2013a), and this may have been a key dri-ver of the stronger effects of recruitment hotspots in P. maculatusthan L. carponotatus observed in the present study. Despite thesubtle differences between the study species, the findings pre-sented here suggest that reserves that are established in areas withrecruitment hotspots have a greater potential for sustained in-creases in adult fish numbers.

Spatial variation in recruitment has the potential to explainmuch of the existing variation in the effectiveness of marine re-serves. Reserves that underperform or fail to accumulate higherdensities of target species may do so, in part, because recruitmentto the reserve is insufficient to exceed natural mortality over thelong term. Furthermore, our study suggests one possible mecha-nism for observed differences in the rate of population recovery in-side reserves – rapid increases following protection are more likelyif the reserve contains one or more recruitment hotspots (Halpernand Warner, 2002; Russ et al., 2008; Russ and Alcala, 2010),whereas in reserves that receive lower levels of recruitment, pop-ulation densities are likely to increase more slowly.

It is clear that a range of factors other than recruitment mayalso explain spatial differences in the magnitude of reserve effects.To date, most studies attempting to quantify reserve effects havefocused on comparing areas exposed to high and low (or negligi-ble) fishing effort and mortality (Gerber and Heppell, 2004; Russand Alcala, 2010). Reserve size has been reported as an important

Non-reserves

Rec

ruitm

ent h

otsp

ots

Non

recr

uitm

ent h

otsp

ots

Size0 10 20 30

0

10

20

30

40

50

60

Freq

uenc

y

15

10

20

30

40

50

60

Freq

uenc

y

25 4355

Fig. 4. Body length – frequency histograms of Lutjanus carponotatus in the

factor, where relatively small reserves may be much more effectivefor site-attached fishes than for more mobile species (Nardi et al.,2004; Grüss et al., 2011; Januchowski-Hartley et al., 2013). In addi-tion, spatial variation in enforcement and/or compliance may alsobe implicated (Mapstone et al., 2008; McLean et al., 2011).Although enforcement is applied evenly at the scale of the KeppelIslands, there are some data to suggest that compliance levels mayvary among different parts of the Great Barrier Reef reserve net-work (Davis et al., 2004; Evans and Russ, 2004; McCook et al.,2010). Nevertheless, a full understanding of the population dynam-ics in reserves requires a thorough description of demographicrates, including habitat availability and condition, recruitment,migration between zones, various sources of mortality, and fishingintensity.

The relationship between recruitment levels and adult numbersacross a network of reserves is likely to be complex. Recruitment isknown to be an important demographic parameter that explainsmuch of the variation in adult reef fish numbers (e.g. Booth et al.,2000), and therefore may at least partly explain why some reserveshave higher density of exploited fishes than others. However, it isalso possible that variation in adult densities through other causeshas a reciprocal effect on juvenile densities, either through topdown effects of predation or through increased self-recruitment(e.g. Swearer et al., 2002; Jones et al., 2005). In a previous study,we found no effects of reserve status on juvenile density or variousother aspects of their ecology (Wen et al., 2012). However, geneticstudies have shown that many of the recruits of the two study spe-cies in the Keppel Islands were the offspring of local adult fishes(Harrison et al., 2012). Hence there may be a strong feed-back loopbetween increased adult numbers within reserves and highrecruitment at specific locations. This finding also supports reserve

Reserves

(cm TL)0 40350 10 20 3015 255

Keppel islands using two factors (reserves and recruitment hotspots).

130 C.K.C. Wen et al. / Biological Conservation 166 (2013) 124–131

design theory that emphasizes the importance of maintainingdemographically relevant connectivity between recruitment hot-spots and source populations (Kininmonth et al., 2011).

There was a clear truncation in the size frequency distributionabove the legal minimum size limit (MSL) in fished areas comparedto reserves. This suggests that fishers are having a substantial im-pact on the size structure of P. maculatus and L. carponotatus pop-ulations in fished areas of the Keppel Islands. The presence orabsence of recruitment hotspots in fished areas appeared to havea lesser effect than the presence of reserves on the abundancesof larger size-class fishes. The dramatic decline of adult P. macula-tus in fished areas appeared to also apply to fish that were slightlybelow the MSL. This may indicate a level of fishing pressure on fishslightly under the MSL outside reserves, which has also been notedin other studies (e.g., Powell et al., 2010; McLean et al., 2011; Cookeet al., 2012). The non-significant difference between size-fre-quency distributions of L. carponotatus in reserves with hotspotsand non-reserves without hotspots (Res + HS vs nRes-HS) runscounter to the rest of the results in this study. This may be dueto limited larval supply (no recruitment hotspots) and intensivefishing pressure (fished areas) which resulted in decreased abun-dance of all size classes, but the shapes of the size-frequency distri-butions are similar. Although a K-S test did not detect a differencebetween low-density populations (nRes-HS) and high-density pop-ulations (i.e., Res + HS), differences in relative densities can be ob-served in Fig. 2.

Our results suggest that identifying and incorporating recruit-ment hotspots into the design of marine reserve networks couldenhance reserve performance in achieving both conservation andfisheries objectives. Further experimental work is required toinvestigate causal relationships between recruitment and adultdensities, as well as the role of recruitment in reserve success.Realistically, the selection of sites for reserves must be based ona large suite of biological and social factors (Nicholson et al.,2006; Ban et al., 2011), and the scope for incorporating species-specific criteria such as recruitment hotspots may be limited. How-ever, highly sought after fishery species are the most likely to ben-efit from protection and including key recruitment habitats intoreserve network design appears to be a good way of maximizingthose benefits.

Acknowledgements

Financial support was provided by the Marine and Tropical Sci-ences Research facility (MTSRF), the National Environment Re-search Program (NERP) Tropical Ecosystems Hub, Great BarrierReef Marine Park Authority (GBRMPA) and the Australian ResearchCouncil Centre of Excellence for Coral Reef Studies at James CookUniversity. We are grateful to S. Blowes and K. Nash for adviceon the data analyses and numerous volunteers for fish collectionand fieldwork. We especially appreciate the great help from P. Wil-liams and R. Ramage (Keppel Lodge) through the whole project inGreat Keppel Islands.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.biocon.2013.06.017.

References

Abesamis, R.A., Russ, G.R., Alcala, A.C., 2006. Gradients of abundance of fish acrossno-take marine reserve boundaries: evidence from Philippine coral reefs. Aquat.Conserv.: Mar. Freshwat. Ecosyst. 16, 349–371.

Adams, V.M., Mills, M., Jupiter, S.D., Pressey, R.L., 2011. Improving socialacceptability of marine protected area networks: a method for estimating

opportunity costs to multiple gear types in both fished and currently unfishedareas. Biol. Conserv. 144, 350–361.

Armsworth, P.R., 2002. Recruitment limitation, population regulation, and larvalconnectivity in reef fish metapopulations. Ecology 83, 1092–1104.

Ban, N.C., Adams, V.M., Almany, G.R., Ban, S., Cinner, J.E., McCook, L.J., Mills, M.,Pressey, R.L., White, A., 2011. Designing, implementing and managing marineprotected areas: emerging trends and opportunities for coral reef nations. J. Exp.Mar. Biol. Ecol. 408, 21–31.

Beger, M., Grantham, H.S., Pressey, R.L., Wilson, K.A., Peterson, E.L., Dorfman, D.,Mumby, P.J., Lourival, R., Brumbaugh, D.R., Possingham, H.P., 2010.Conservation planning for connectivity across marine, freshwater, andterrestrial realms. Biol. Conserv. 143, 565–575.

Booth, D.J., Kingsford, M.J., Doherty, P.J., Beretta, G.A., 2000. Recruitment ofdamselfishes in One Tree Island lagoon: persistent interannual spatialpatterns. Mar. Ecol. Prog. Ser. 202, 219–230.

Claudet, J., Guidetti, P., 2010. Fishermen contribute to protection of marine reserves.Nature 464, 673.

Claudet, J., Osenberg, C., Domenici, P., Badalamenti, F., Milazzo, M., Falcón, J.,Bertocci, I., Benedetti-Cecchi, L., García-Charton, J., Goñi, R., 2010. Marinereserves: fish life history and ecological traits matter. Ecol. Appl. 20, 830–839.

Cooke, S.J., Suski, C.D., Arlinghaus, R., Danylchuk, A.J., 2012. Voluntary institutionsand behaviours as alternatives to formal regulations in recreational fisheriesmanagement. Fish Fish., in press, doi: 10.1111/j.1467-2979.2012.00477.x.

Davis, K.L.F., Russ, G.R., Williamson, D.H., Evans, R.D., 2004. Surveillance andpoaching on inshore reefs of the great barrier reef marine park. Coast. Manag.32, 373–387.

Eagle, J.V., Baird, A.H., Jones, G.P., Kingsford, M.J., 2012. Recruitment hotspots:consistent spatial patterns in the relative abundance of coral recruits at OneTree Island Australia. Galaxea 14, 5–22.

Edwards, H.J., Elliott, I.A., Pressey, R.L., Mumby, P.J., 2010. Incorporating ontogeneticdispersal, ecological processes and conservation zoning into reserve design.Biol. Conserv. 143, 457–470.

Evans, R.D., Russ, G.R., 2004. Larger biomass of targeted reef fish in no-take marinereserves on the great barrier reef, Australia. Aquat. Conserv.: Mar. Freshwat.Ecosyst. 14, 505–519.

Feary, D.A., Almany, G.R., McCormick, M.I., Jones, G.P., 2007. Habitat choice,recruitment and the response of coral reef fishes to coral degradation. Oecologia153, 727–737.

Fenberg, P.B., Caselle, J.E., Claudet, J., Clemence, M., Gaines, S.D., Antonio García-Charton, J., Gonçalves, E.J., Grorud-Colvert, K., Guidetti, P., Jenkins, S.R., 2012.The science of European marine reserves: status, efficacy, and future needs.Mar. Policy 36, 1012–1021.

Fernandes, L., Day, J., Lewis, A., Slegers, S., Kerrigan, B., Breen, D., Cameron, D., Jago,B., Hall, J., Lowe, D., Innes, J., Tanzer, J., Chadwick, V., Thompson, L., Gorman, K.,Simmons, M., Barnett, B., Sampson, K., De’Ath, G., Mapstone, B., Marsh, H.,Possingham, H., Ball, I., Ward, T., Dobbs, K., Aumend, J., Slater, D., Stapleton, K.,2005. Establishing representative no-take areas in the great barrier reef: large-scale implementation of theory on marine protected areas. Conserv. Biol. 19,1733–1744.

Freeman, D.J., MacDiarmid, A.B., Taylor, R.B., Davidson, R.J., Grace, R.V., Haggitt, T.R.,Kelly, S., Shears, N.T., 2012. Trajectories of spiny lobster Jasus edwardsii recoveryin New Zealand marine reserves: is settlement a driver. Environ. Conserv. 39,295–304.

Gell, F.R., Roberts, C.M., 2003. Benefits beyond boundaries: the fishery effects ofmarine reserves. Trends Ecol. Evol. 18, 448–455.

Gerber, L.R., Heppell, S.S., 2004. The use of demographic sensitivity analysis inmarine species conservation planning. Biol. Conserv. 120, 121–128.

Gerber, L.R., Heppell, S.S., Ballantyne, F., Sala, E., 2005. The role of dispersal anddemography in determining the efficacy of marine reserves. Can. J. Fish. Aquat.Sci. 62, 863–871.

Guidetti, P., Milazzo, M., Bussotti, S., Molinari, A., Murenu, M., Pais, A., Spanò, N.,Balzano, R., Agardy, T., Boero, F., Carrada, G., Cattaneo-Vietti, R., Cau, A.,Chemello, R., Greco, S., Manganaro, A., Notarbartolo di Sciara, G., Russo, G.F.,Tunesi, L., 2008. Italian marine reserve effectiveness: does enforcement matter?Biol. Conserv. 141, 699–709.

Grüss, A., Kaplan, D.M., Guénette, S., Roberts, C.M., Botsford, L.W., 2011.Consequences of adult and juvenile movement for marine protected areas.Biol. Conserv. 144, 692–702.

Halpern, B.S., Warner, R.R., 2002. Marine reserves have rapid and lasting effects.Ecol. Lett. 5, 361–366.

Halpern, B.S., Warner, R.R., 2003. Matching marine reserve design to reserveobjectives. Proc. R. Soc. Lond. Ser. B: Biol. Sci. 270, 1871–1878.

Hammer, Ø., Harper, D., Ryan, P., 2001. PAST: paleontological statistics softwarepackage for education and data analysis. Palaeontol. Electron. 4, 9.

Hansen, G.J.A., Ban, N.C., Jones, M.L., Kaufman, L., Panes, H.M., Yasué, M., Vincent,A.C.J., 2011. Hindsight in marine protected area selection: a comparison ofecological representation arising from opportunistic and systematicapproaches. Biol. Conserv. 144, 1866–1875.

Harmelin-Vivien, M., Le Diréach, L., Bayle-Sempere, J., Charbonnel, E., García-Charton, J.A., Ody, D., Pérez-Ruzafa, A., Reñones, O., Sánchez-Jerez, P., Valle, C.,2008. Gradients of abundance and biomass across reserve boundaries in sixMediterranean marine protected areas: evidence of fish spillover. Biol. Conserv.141, 1829–1839.

Harrison, H.B., Williamson, D.H., Evans, R.D., Almany, G.R., Thorrold, S.R., Russ, G.R.,Feldheim, K.A., van Herwerden, L., Planes, S., Srinivasan, M., Berumen, M.L.,

C.K.C. Wen et al. / Biological Conservation 166 (2013) 124–131 131

Jones, . Larval export from marine reserves and the recruitment benefit for fishand fisheries. Curr. Biol. 22, 1023–1028.

Heppell, S.A., Semmens, B.X., Archer, S.K., Pattengill-Semmens, C.V., Bush, P.G.,McCoy, C.M., Heppell, S.S., Johnson, B.C., 2012. Documenting recovery of aspawning aggregation through size frequency analysis from underwater lasercalipers measurements. Biol. Conserv. 155, 119–127.

Januchowski-Hartley, F.A., Graham, N.A.J., Cinner, J.E., Russ, G.R., 2013. Spillover offish naïveté from marine reserves. Ecol. Lett. 16, 191–197.

Joseph, L.N., Elkin, C., Martin, T.G., Possingham, H.P., 2009. Modeling abundanceusing N-mixture models: the importance of considering ecological mechanisms.Ecol. Appl. 19, 631–642.

Jones, G.P., Planes, S., Thorrold, S.R., 2005. Coral reef fish larvae settle close to home.Curr. Biol. 15, 1314–1318.

Kininmonth, S., Beger, M., Bode, M., Peterson, E., Adams, V.M., Dorfman, D.,Brumbaugh, D.R., Possingham, H.P., 2011. Dispersal connectivity and reserveselection for marine conservation. Ecol. Model. 222, 1272–1282.

Kingsford, M.J., 2009. Contrasting patterns of reef utilization and recruitment ofcoral trout (Plectropomus leopardus) and snapper (Lutjanus carponotatus) at OneTree Island, southern great barrier reef. Coral Reefs 28, 251–264.

Leis, J.M., Carson-Ewart, B.M., 1999. In situ swimming and settlement behaviour oflarvae of an Indo-Pacific coral-reef fish, the coral trout Plectropomus leopardus(Pisces: Serranidae). Mar. Biol. 134, 51–64.

Lester, S.E., Halpern, B.S., Grorud-Colvert, K., Lubchenco, J., Ruttenberg, B.I., Gaines,S.D., Airamé, S., Warner, R.R., 2009. Biological effects within no-take marinereserves: a global synthesis. Mar. Ecol. Prog. Ser. 384, 33–46.

Mapstone, B.D., Little, L.R., Punt, A.E., Davies, C.R., Smith, A.D.M., Pantus, F.,McDonald, A.D., Williams, A.J., Jones, A., 2008. Management strategy evaluationfor line fishing in the great barrier reef: Balancing conservation and multi-sector fishery objectives. Fish. Res. 94, 315–329.

McCook, L.J., Ayling, T., Cappo, M., Choat, J.H., Evans, R.D., De Freitas, D.M., Heupel,M., Hughes, T.P., Jones, G.P., Mapstone, B., Marsh, H., Mills, M., Molloy, F.J.,Pitcher, C.R., Pressey, R.L., Russ, G.R., Sutton, S., Sweatman, H., Tobin, R.,Wachenfeld, D.R., Williamson, D.H., 2010. Adaptive management of the greatbarrier reef: a globally significant demonstration of the benefits of networks ofmarine reserves. Proc. Natl. Acad. Sci. 107, 18278–18285.

McLean, D.L., Harvey, E.S., Meeuwig, J.J., 2011. Declines in the abundance of coraltrout (Plectropomus leopardus) in areas closed to fishing at the HoutmanAbrolhos Islands, Western Australia. J. Exp. Mar. Biol. Ecol. 406, 71–78.

Meekan, M.G., Milicich, M.J., Doherty, P.J., 1993. Larval production drives temporalpatterns of larval supply and recruitment of a coral reef damselfish. Mar. Ecol.Prog. Ser. 93, 217.

Nardi, K., Jones, G.P., Moran, M.J., Cheng, Y.W., 2004. Contrasting effects of marineprotected areas on the abundance of two exploited reef fishes at the sub-tropical Houtman Abrolhos Islands, Western Australia. Environ. Conserv. 31,160–168.

Nicholson, E., Westphal, M.I., Frank, K., Rochester, W.A., Pressey, R.L., Lindenmayer,D.B., Possingham, H.P., 2006. A new method for conservation planning for thepersistence of multiple species. Ecol. Lett. 9, 1049–1060.

Pollnac, R., Christie, P., Cinner, J.E., Dalton, T., Daw, T.M., Forrester, G.E., Graham,N.A.J., McClanahan, T.R., 2010. Marine reserves as linked social–ecologicalsystems. Proc. Natl. Acad. Sci. 107, 18262–18265.

Powell, E.N., Bochenek, E.A., DePersenaire, J., 2010. Evaluation of bag-and-size-limitoptions in the management of summer flounder Paralichthys dentatus. Fish. Res.105, 215–227.

Queensland Fisheries Act, 2004 Fisheries (Coral Reef Fin Fish) Management Plan2003. The State of Queensland, Queensland Government.

Quéré, G., Leis, J.M., 2010. Settlement behaviour of larvae of the stripey snapper,Lutjanus carponotatus; (Teleostei: Lutjanidae). Environ. Biol. Fishes 88, 227–238.

R Development Core Team, 2011. R: a language and environment for statisticalcomputing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL <http://www.Rproject.org/>.

Rasband, W. S., 1997. ImageJ, US National Institutes of Health, Bethesda, Maryland,USA.

Russ, G.R., Alcala, A.C., 2010. Decadal-scale rebuilding of predator biomass inPhilippine marine reserves. Oecologia 163, 1103–1106.

Russ, G.R., Cheal, A.J., Dolman, A.M., Emslie, M.J., Evans, R.D., Miller, I., Sweatman, H.,Williamson, D.H., 2008. Rapid increase in fish numbers follows creation ofworld’s largest marine reserve network. Curr. Biol. 18, R514–R515.

Stockhausen, W.T., Lipcius, R.N., Hickey, B.M., 2000. Joint effects of larval dispersal,population regulation, marine reserve design, and exploitation on productionand recruitment in the Caribbean spiny lobster. Bull. Mar. Sci. 66, 957–990.

Swearer, S.E., Shima, J.S., Hellberg, M.E., Thorrold, S.R., Jones, G.P., Robertson, D.R.,Morgan, S.G., Selkoe, K.A., Ruiz, G.M., Warner, R.R., 2002. Evidence of self-recruitment in demersal marine populations. Bull. Mar. Sci. 70, 251–271.

Symonds, M.R.E., Moussalli, A., 2011. A brief guide to model selection, multimodelinference and model averaging in behavioural ecology using Akaike’sinformation criterion. Behav. Ecol. Sociobiol. 65, 13–21.

Valles, H., Kramer, D.L., Hunte, W., 2008. Temporal and spatial patterns in therecruitment of coral-reef fishes in Barbados. Mar. Ecol. Prog. Ser. 363, 257–272.

Wen, C.K., Almany, G.R., Williamson, D.H., Pratchett, M.S., Jones, G.P., 2012.Evaluating the effects of marine reserves on diet, prey availability and preyselection by juvenile predatory fishes. Mar. Ecol. Prog. Ser. 469, 133–144.

Wen, C.K.C., Pratchett, M.S., Almany, G.R., Jones, G.P., 2013a. Patterns of recruitmentand microhabitat associations for three predatory coral reef fishes on thesouthern great barrier reef, Australia. Coral Reefs 32, 389–398.

Wen, C.K.C., Pratchett, M.S., Almany, G.R., Jones, G.P., 2013b. Role of prey availabilityin microhabitat preferences of juvenile coral trout (Plectropomus: Serranidae). J.Exp. Mar. Biol. Ecol. 443, 39–45.

Williamson, D.H., Russ, G.R., Ayling, A.M., 2004. No-take marine reserves increaseabundance and biomass of reef fish on inshore fringing reefs of the great barrierreef. Environ. Conserv. 31, 149–159.

Zeller, D.C., 2002. Tidal current orientation of Plectropomus leopardus (Serranidae).Coral Reefs 21, 183–187.