Vulnerability of parrotfish functional diversity and reef health in transitioning island socio-ecosystems

Katherine R Rice Corresp. 1

1 Department of Environmental Science, Policy, and Management, University of California, Berkeley, Berkeley, California, United States

Corresponding Author: Katherine R Rice Email address: [email protected]

Mo’orea’s reefs have rebounded from environmental disturbance throughout the years largely due to herbivorous fish that deter damaging algal blooms. This resilience suggests herbivorous fishes act as a in the ecosystem, and the greater island community of Mo’orea. Parrotfish support reef health and stability, and reefs support the development of the local economy by way of tourism and access to medicine, nourishment, and protection. Because island communities rely heavily on coral reef ecosystems, identifying the impact of fishing on the morphology and ecosystem function of parrotfish in a time of marine management and demographic transition can increase our knowledge of the vulnerability and resilience of these complex socio-ecosystems. The 2016 study reported here seeks to understand to what extent changes in fisheries management and off-take rates have affected the historically sustainable relationship between Mo’orea’s fishing population, the lagoon’s supply of larger-sized parrotfish, and the ecological stability of the greater coral reef ecosystem. Specifically, this study measured average parrotfish size at various fishing zones and paired Marine Protected Areas (MPAs) around the island, and then used participatory surveys to quantify fishermen observation of changes in parrotfish size since they started fishing. Both field data and participatory survey data show a decrease in parrotfish size since the establishment of MPAs. Island-wide, parrotfish also appear to be smaller in fished sites than in MPAs. Results suggest that the joint effect of zoning, catch-size enforcements and increased fishing pressure have caused a size-selection of parrotfish in the fishing zones of studied lagoons. These findings highlight the vulnerability of Mo’orea’s coral reef ecosystem to transitions in marine management strategy and size-selective fishing.

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 1 Vulnerability of parrotfish functional diversity and coral 2 reef health in transitioning island socio-ecosystems 3 4 Katherine Rose Rice1 5 6 1 Department of Environmental Science, Policy, and Management, University of California, Berkeley, 7 Berkeley, CA, USA 8 9 [email protected] 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 50 Introduction 51 52 Coral reef ecosystems are fundamentally important to many Pacific Island countries and 53 inhabitants, including those in the South Pacific (Ferraris and Cayré, 2003; Kronen et al., 2010; 54 Moberg and Folke, 1999; Dalzell et al., 2006). Coral reefs support the development of local and 55 national economies by providing goods and services to island communities via fisheries (Moberg 56 and Folke, 1999). Reefs, however, are becoming critically threatened by overfishing as a result 57 of overexploitation of commercial fish (Hughes et al., 2003, 2007; Pandolfi, 2003; Bellwood et 58 al., 2006; Dalzell et al., 2006; Jackson et al., 2001). Unsustainable tampering with the reef’s 59 balance of biological diversity has serious consequences for the goods and services that humans 60 derive from coral reef ecosystems. Consequences include species extinction, reduced ecosystem 61 resilience (i.e. the capacity of an ecosystem to tolerate disturbance without collapsing into a 62 qualitatively different state), and phase shifts from coral to algal dominance (Bellwood et al., 63 2005; Norström et al., 2009; Wilson et al., 2008; Hughes et al., 2010; Jennings and Polunin, 64 1996; Jennings and Lock, 1996; Jennings and Kaiser, 1998; Wilder, 2003). 65 66 The declining health of coral reef ecosystems worldwide has serious implications for their 67 capacity to persist in an era of rapid global change. Reefs have lost their capacity to endure 68 recurrent natural disturbances, and some have undergone long-term phase shifts to degraded 69 ecosystems dominated by fleshy seaweed or other macroalgae (Bellwood et al., 2005; Norström 70 et al., 2009; Wilson et al., 2008; Hughes et al., 2010). Interestingly, however, coral reefs 71 surrounding Mo’orea have historically returned to coral dominance following major 72 environmental disturbances without shifting to a macroalgal-dominated state (Adam et al., 2011). 73 Adam and colleagues (2011) found that the increased algal growth associated with coral loss 74 prompts the response of herbivorous fishes, particularly parrotfish, to graze the reef and 75 consequently help coral reefs recover. Wilder (2003) similarly concluded that if stocks 76 were reduced to low levels through fishing pressure, reefs could be overgrown quickly by dense 77 on a short time scale. Adam (2011) highlighted the importance of parrotfish ecosystem 78 function in maintaining reef resilience in the face of disturbance, and Wilder (2003) revealed the 79 vulnerability of the reef ecosystem to declines in herbivore density. In addition, previous studies 80 have identified herbivorous fish as important factors in regulating algal biomass, cover, and 81 composition on coral reefs (Jennings and Polunin, 1996; Jennings and Lock, 1996; Jennings and 82 Kaiser, 1998). 83 84 The functional diversity within parrotfish populations has been suggested to be a result of 85 differing morphologic characteristics, such as body size (Lokrantz et al., 2008; Bruggemann et 86 al., 1994; Littler et al., 1989; Roff et al., 2011). Populations of large parrotfish are characterized 87 by intense grazing behavior and can harvest the surface of each square meter of reef every 18 88 days, removing up to 40 kg of algae from each square meter per year (Hoey and Bellwood, 89 2008). In addition to grazing capacity, parrotfish size has also been found to be representative of 90 reproductive capacity (Barba, 2010; Choat and Bellwood, 1998; Thresher, 1984). Terminal phase 91 males dominate reproductive activity through a harem-based social system and can be either 92 primary, i.e., born male, or secondary, i.e., females that have undergone sex change (Choat and 93 Bellwood, 1998; Thresher, 1984). Terminal males are usually the largest of the population 94 (Choat and Bellwood, 1998; Thresher, 1984). Larger parrotfish are thus especially critical in 95 sustaining population size and coral reef ecosystem balance.

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 96 97 Fishing pressure in the Pacific Islands is strongly tied to human population density (Jennings and 98 Kaiser, 1998; Russ and Alcala, 1989). In addition, increases in fishing pressure result in changes 99 in target populations (Jennings and Lock, 1996) and in the island fishing community (Jennings 100 and Polunin, 1996). Population census figures for Mo’orea between 1971 and 2007 show an 101 average annual population growth rate of 2.39%, which is higher than the rate for French 102 Polynesia as a whole (1.57%) (Leenhardt et al., 2016). This increase in population implies a 103 respective increase in demand of natural resources, particularly fish from the lagoon, as this is 104 where recreational and subsistence fishing activity is concentrated (Leenhardt et al., 2016; Lison 105 de Loma, 2005). In fact, transition in the demography of French Polynesia has led to a decrease 106 in the density and biomass of harvested fish yields in Mo’orea and Tahiti (which are more 107 heavily fished) (Lison de Loma, 2005). These data suggest that increased fishing pressure may 108 be generally depleting fish stocks, and specifically taking larger individuals thereby impairing 109 the ability of the population to reproduce. Thus, the transitioning socio-ecosystem of Mo’orea 110 presents a model for studying the vulnerability of fish size and abundance to increasing fishing 111 pressure. A favored commercial catch and a dominant herbivore of the lagoon, parrotfish, family 112 Scaridae, represent a fish important to both the sociology and ecology of Mo’orea’s island socio- 113 ecosystem. Parrotfish are thus an applicable bio-indicator for understanding the sociological 114 impact of changes in fishing pressure, and the ecological response of reef stability and resilience. 115 116 Effects of overfishing in Mo’orea were first noticed in 1991 when local citizens and foreign 117 scientists observed that the sizes of fish and overall catches in Mo’orea’s lagoon were declining 118 over time (Walker and Robinson, 2009). As a small-scale fishing community dependent on 119 fishing for income and subsistence (Leenhardt et al., 2016; Walker, 2001), and as a scientific 120 model of an unspoiled island ecosystem for many researchers, concerns about overfishing from 121 local communities, territorial services (Fisheries, Environment, Urbanism), scientific research 122 institutions, and local politicians led to the implementation of a comprehensive marine 123 management plan (Salvat and Aubanel, 2002). The Plan de Gestion de l’Espace Maritime 124 (PGEM) for Mo’orea was established in 2004 and encompassed the entire lagoon and all waters 125 beyond the reef crest out to the 70-m isobath on the outer reef slope (PGEM, 2005). The PGEM 126 established a network of eight “no-take” zones referred to as Marine Protected Areas (MPAs) 127 covering approximately 20% of the lagoon, and enforced size and catch limits on commercial 128 fish species throughout the fishing zones of the lagoon (Walker and Robinson, 2009). 129 130 The implications of dividing the lagoon have both spatial and temporal components (Francour, 131 2000). The spatial components include differences between protected and unprotected zones, 132 such as fishing pressure and the presence of size and catch limits (Russ and Alcala, 1989; 133 Harmelin-Vivien et al., 1995; Francour, 2000). The temporal components include differing 134 ecological responses between the protected no-take MPA reserves and the unprotected fishing 135 zones of the lagoon (Francour, 2000). While no-take MPA reserves establish points of reference 136 to assess human and other impacts on adjacent marine environments, this study suggests that the 137 MPA reserve management strategy may induce additional unanticipated spatial and temporal 138 components for small-scale island fisheries such as Mo’orea’s lagoon. As an island only 61 km 139 in circumference and heavily dependent on fishing, Mo’orea’s fishery is already limited. No-take 140 reserves redistribute fishing pressure to spatially concentrated fishing zones. In addition to 141 increases in fishing pressure as a result of zoning, catch-size regulations on commercial fish

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 142 increase pressure on larger bodied fish. For parrotfish, the PGEM’s catch-size regulations 143 encourage catch of the more functionally and socially important individuals (>25 cm), thus 144 failing to consider the ecosystem function and complex socio-sexual system of large-bodied 145 individuals (PGEM, 2005; Lokrantz et al., 2008; Choat and Bellwood, 1998; Thresher, 1984). 146 While the Centre de Recherches Insulaires et Observatoire de l’Environnement (CRIOBE) has 147 monitored the biological effects of the MPAs since 2004 (Lison de Loma et al., 2008), 148 CRIOBE’s monitoring plan doesn’t address the PGEM’s effect on fishing pressure and the 149 vulnerability of the collective reef ecosystem to human-derived transitions in management. 150 151 Size-selective fishing pressure from catch-size regulations and an increasing fishing industry 152 mark two notable deviations from Mo’orea’s historically synchronized socio-ecosystem (Adam 153 et al., 2011). Given the social and ecological components of Mo’orea's transitioning lagoon 154 fisheries, understanding their dual dynamics requires integrated methods that consider both 155 systems simultaneously (Jennings and Kaiser, 1998). Pairing ecological field surveys with 156 participatory monitoring techniques allows for comparison of how fishermen understand and 157 interact with their lagoon, and how the lagoon responds to their understanding of its supply of 158 resources. Filling these gaps in our knowledge will enhance the development of marine resource 159 management initiatives that seek long-term sustainability of reef fisheries and foster ecosystem 160 resilience. To promote an understanding of the impact of size-selective fishing pressure on the 161 vulnerability of a small-scale fishery’s coral reef ecosystem, this study compares parrotfish size 162 and abundance in exploited fishing zones and MPAs around the island. To better evaluate the 163 sustainability of the existing fishery, this study also includes a participatory monitoring of 50 164 local fishermen from around the island. Specifically, this study seeks to quantify the increasing 165 population of Mo’orea’s fishing community, and assesses how increase in lagoon resource 166 demand, in conjunction with the PGEM’s zoning and catch-size enforcements, is leading to size- 167 selection of parrotfish in Mo’orea’s lagoon. 168 169 Methods 170 171 Study sites 172 173 Fieldwork was conducted in lagoons on the island of Mo’orea, French Polynesia from October to 174 November, 2016 (Fig. 1). Mo’orea is a volcanic island with 61 km of coastline, encircled by a 175 barrier reef, which forms a 30-km2 lagoon ranging from 500 to 1500 m in width (Galzin, 1985). 176 To calculate the extent to which size-selective fishing pressure impacts parrotfish size in fishing 177 zones of the lagoon, this study conducted transects along a gradient of MPA’s and non-protected 178 fishing zones. Five MPAs (Temae (P1), Pihaena (P2), Tetaiuo (P3), Tiahura Motu (P4), and 179 Afareaitu (P5)) were chosen based on accessibility and proximity to corresponding paired fished 180 sites (F1, F2, F3, and F5) (Fig. 1). The paired sites have close proximity and similar 181 geomorphologies. However, MPA Tiahura Motu and MPA Tetaiuo are geomorphologically 182 similar so a single fishing site in between these two MPAs was chosen (F3) (Fig. 1). The nine 183 survey sites represent geographic diversity and are distributed along the northwestern, northern, 184 and northeastern coasts of the island. Sites were selected to encompass a range of human 185 activity, human development, fishing pressure, and fish density. Other parameters used to choose 186 sites include depth and sea floor composition. 187

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 188 Field Sampling Methods 189 190 Transects (described below) were conducted on two occasions, each in a period of ten days 191 around two principal lunar phases: one during the full moon and one during the new moon to 192 account for potential lunar variation (Galzin, 1987). Observations were always made between 193 0900 and 1630 hours (local time), to minimize any heterogeneity caused by diel variation in fish 194 behavior (Galzin, 1987). 195 196 Five locations around the island were sampled during each lunar occasion. Pairs of sites (MPA 197 and fished) were sampled within each location on the same day (with exception of Tetaiuo and 198 Tiahura Motu, which shared a fished site and were all sampled on the same day). Two distinct 199 reef habitats were sampled at each site: the fringing reef close to shore and the back reef towards 200 the reef crest. Three stations were surveyed per reef habitat, yielding a total of 54 stations: nine 201 sites (five MPAs, four fished sites) x two habitats x three stations per habitat, and 108 surveys 202 (54 stations x two lunar occasions). 203 204 Transect lines were 30 m long and aligned parallel to shore. When surveying at motus, transects 205 were aligned parallel to the motu. In absence of fixed markers, each station was located using a 206 handheld global positioning system (GPS) receiver. Situational characteristics including presence 207 of fishermen, boat traffic, swimmers, current strength, or intense weather conditions were also 208 recorded at each site. 209 210 Preliminary surveys revealed high diversity of parrotfish species (>12 species) at each station, so 211 for simplicity parrotfishes were identified at the family level (Scaridae), counted, and body 212 length estimated to the nearest centimeter using a ruler for scale (counts and average body 213 lengths were recorded for schools of fishes). Transect lines were surveyed continuously by 214 swimming at a slow steady speed (10m/min) to observe fishes in an undisturbed state. 215 216 Fishing Population Surveys 217 218 To indirectly estimate change in average parrotfish size throughout time, this study used a 219 participatory monitoring survey method, drawing upon 50 local fishermen of all ages, genders, 220 and villages around the island as ecological monitors. The survey asked participants to quantify 221 any changes in parrotfish size throughout their lifetime as a fisherman. Size estimates from 1956 222 to 2013 were derived and compared to the average of size estimates from 2016. Individual 223 estimates were plotted on a graph to reveal the island’s collective understanding of parrotfish 224 size over a timescale of 60 years. Estimates of parrotfish body size from 1956 to 2013 were also 225 compared to changes in human population from 1956 to 2013 to look for potential correlation 226 and provide additional evidence for the effect of increases in human population on reef 227 vulnerability. 228 229 Data Analyses 230 231 This study’s aim is to estimate how decreases in parrotfish body size and abundance (functional 232 diversity) may vary between MPA and fished sites. In addition to site, this study considers how 233 functional diversity is affected by location, habitat, and moon phase. From field observations, the

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 234 vulnerability of parrotfish to size-selective fishing pressure is estimated on varying spatial and 235 temporal scales. This study used R Studio software (R Core Team, 2013) for all statistical 236 analyses. 237 238 Fish size and abundance were analyzed across replica transects for 108 stations. Size data fulfill 239 the assumptions of parametric statistics (normal distribution, similar sample sizes, and equal 240 variances) (Fig. 2); however, abundance data are not normally distributed. To test for differences 241 in mean total parrotfish size between sites (grouped among MPAs or fished), reef habitats 242 (grouped among fringing or back), location (1, 2, 3, 4, or 5), and lunar occasion (full moon or 243 new moon), a four-way analysis of variance (ANOVA) followed by a Tukey HSD post-hoc test 244 were run using these four variables as predictor variables and fish size as the response variable 245 (results shown in Table 1). To analyze fishing pressure’s effect on parrotfish abundance, a 246 Kruskal Wallace test was run for the variable location, and Wilcoxon rank sum test with 247 continuity correction was run for site, reef habitat, and lunar occasion. 248 249 Results 250 251 Field sampling 252 253 Site 254 Island-wide, average body size is greater in MPAs (26 cm ± 4 cm) than in fished sites (24 cm ± 5 255 cm) (p<0.001) (Fig. 3). Post-hoc comparisons using the Tukey HSD test, however, indicated that 256 the average body size is only larger in MPA sites than in fished sites for Locations 1 and 5 257 (p<0.05, p<0.05). In addition to larger average body size in MPAs, parrotfish are more abundant 258 in MPAs (31 ± 61 fish) than in fished sites (12 ± 16 fish) (p<0.05) (Fig. 4). 259 260 Location 261 Regardless of site, however, average body size also differs between locations (Fig. 3 & 5). 262 Average body size in Location 1 (22 cm) is smaller than average body size in Location 3 (26 263 cm), 4 (29 cm), and 5 (26 cm) (p<0.001, p<0.001, p<0.001). In addition, average body size in 264 Location 2 (24 cm) is smaller than average body size in Locations 4 and 5 (p<0.001, p<0.05). 265 Average body size in Location 4 is larger than average body size in Locations 3 and 5 (p<0.05, 266 p<0.001). Contrary to size, parrotfish abundance does not significantly differ among locations 267 (p=0.10) (Fig. 4 & 6). 268 269 Habitat 270 Island-wide, parrotfish are larger in back reef habitats in both MPA (Back: 26 cm, Fringing: 24 271 cm) and fished sites (Back: 26 cm. Fringing: 22 cm) (p=0, p<0.05) (Fig. 3). Post-hoc 272 comparisons using the Tukey HSD test, however, reveal average body size isn’t larger in back 273 reef habitats for all locations. Body sizes are only larger in back reef habitats for Locations 1 274 (p=0), 2 (p=0), and 5 (p=0). Parrotfish are also larger in MPA sites (24 cm) than fished sites (22 275 cm) in all fringing reef habitats (p=0) (Fig. 7). Contrary to size, parrotfish abundance does not 276 significantly differ between habitat types (p=0.22) (Fig. 4). 277 278 Lunar Occasion

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 279 Island-wide, average body size differs according to lunar occasion sampled (Fig. 3 & 5). 280 Average parrotfish are larger when surveyed on the full moon (26 cm) compared to surveys on 281 the new moon (25 cm) (p<0.001). Post-hoc comparisons using the Tukey HSD test indicate 282 average size is larger during the full moon than the new moon for Locations 1 and 5 (p=0, 283 p<0.05), but smaller during the full moon for Location 2 (p<0.05). Larger parrotfish in both 284 fished and MPA sites (p=0.01, p<0.005) are observed during the full moon, in comparison with 285 the new moon. During the new moon, parrotfish are also always larger in MPAs than fished sites 286 (p<0.05); however, during the full moon, there is no difference in observed sizes between sites. 287 Contrary to size, parrotfish abundance does not significantly differ between lunar occasions 288 (p=0.88) (Fig. 4 & 6). 289 290 When no fish were observed in a single transect, fish abundance and size were not recorded, so 291 some habitats only represent one or two stations out of the total three stations surveyed. When a 292 site didn’t physically have one of the two habitat types due to island geomorphology, only the 293 habitat type present was sampled for fish (Fig. 7). 294 295 Fishing Population Surveys 296 297 Results are derived from participatory survey answers provided by 50 fishermen. Surveyed 298 fishermen range from 18 to 81 years of age, include males and females, include residents from 299 every village around the island, and were all surveyed independently to avoid preconceptions. 300 301 Average parrotfish size was noted to decrease throughout the lifetimes of 39 out of the 50 302 surveyed fishermen (Fig. 8). Noted decreases in size range according to the year fishermen 303 started fishing, varying from 5 to 30 cm. Despite the differing timescales of observations, 304 however, the average change in parrotfish size as perceived by fishermen is 9 cm. The average 305 parrotfish size today is 23 cm ± 8 cm (p<0.001), compared to an average size estimate of 32 cm 306 ± 10 cm by fishers who started their fishing career between 1956 and 2002 (p<0.001). 307 Subsequently, while almost every fisherman estimated parrotfish were 20 cm or larger at the start 308 of their fishing career, only 32% of fishermen estimated that parrotfish were 20 cm or larger in 309 2016 (p<0.001) (Fig. 9). 310 311 Discussion 312 313 Previous studies have observed extreme transition over the past two to three decades in 314 Mo’orea’s coral reefs from hard-coral-dominated communities to communities now dominated 315 by fleshy algae (Galzin et al., 2016; Lamy et al. 2015). This type of phase shift has been 316 catalyzed by frequent disturbances, both natural and human derived, and reefs have been forced 317 to adapt quickly, exhibiting cycles of decline and recovery (Hughes et al., 2005; Adam et al., 318 2011; Lamy et al. 2015). Researchers have suggested that phase shifts might be caused by 319 decreases in herbivore density and correspondingly reduced grazing (Galzin et al., 2016; Hughes, 320 1994; Wilder, 2003). Decreases in herbivore density have increased the susceptibility of the 321 lagoon to phase shifts, have led to a less resilient reef (Galzin et al., 2016; Adam et al., 2011; 322 Lison de Loma, 2005), and have consequently made the greater ecosystem exceptionally 323 vulnerable to disturbance. While environmental disturbances cannot be predicted, this study 324 obtained an awareness of the various human-derived stresses affecting the Mo’orea’s reef

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 325 ecosystem. This study observed two major transitions in Mo’orea’s societal relationship with the 326 lagoon: increase in resource demand, and transition in lagoon management. 327 328 First, this study found an annual population growth rate of 3.79 from 1956 to 2012 on Mo’orea 329 (Fig. 10) (ISPF, 2015), implying increased demand for fish around the lagoon. Roadside fish 330 stands and personal communication with 50 surveyed fishermen revealed the high demand for 331 parrotfish in particular (Rice, 2016, Personal observation). In addition, both the fishing 332 demographic and the demand for parrotfish have increased since previous accounts (Aubanel, 333 1993; Brenier, 2009; Madi Moussa, 2010; Vieux, 2002; Yonger, 2002) The extent of this 334 increase, however, is difficult to measure, as supply and demand for fish meet along the roadside 335 rather than in the markets (Madi Moussa, 2010). Increases in fishing pressure are also unapparent 336 to the average fisherman because fish stocks are non-concentrated and fishing activity is spread 337 around the lagoon (Leenhardt et al., 2016). The combination of these circumstances suggests that 338 fishermen are unaware of their collective footprint on the lagoon ecosystem, and consequently, 339 demand for lagoon resources may be exceeding supply. 340 341 In addition to increase in resource demand by a growing island population, foreign stakeholders 342 have infiltrated management of the lagoon. Disregarding the complexities and uncertainties of 343 Mo’orea’s spatially and temporally dispersed fisheries as described above, recent management 344 practices have limited and regulated areas of the lagoon previously relied upon by fishermen as 345 sources of subsistence (Walker, 2001). Since the establishment of MPAs, protected areas of the 346 lagoon have seen increases in fish biomass (Lison de Loma et al., 2008); however, the effect of 347 concentrated activity in the fishing zones has been overlooked. Fishing zones have only been 348 used in research as ‘controls,’ assuming that MPAs are the only zones with a changing 349 ecosystem (Lison de Loma et al., 2008; Lamy et al., 2015). This study, however, reveals that 350 transitions in lagoon management, including zoning and catch-size regulations, affect the entire 351 reef ecosystem. 352 353 Despite conservational aims, MPAs concentrate fishing pressure into limited zones of the lagoon, 354 and catch-size regulations encourage catch of larger-bodied commercial fish stocks in those 355 spatially limited fishing zones. Observed decrease in parrotfish size and abundance inside fishing 356 zones compared to MPAs reveal the vulnerability of fishing zones to transitions in lagoon 357 management, as well as the ensuing size-selective fishing pressure of functionally critical 358 parrotfishes. 359 360 Spatial Dynamics 361 362 Site 363 Results indicate that MPAs support a higher biomass of parrotfish than adjacent fished sites, in 364 terms of average body size and abundance of fishes (Fig. 3 & 4). The smaller body size of 365 parrotfish in fished sites suggests that parrotfish are indeed sensitive to fishing pressure. Despite 366 differences in size between sites, however, it is important to note that the difference was minimal 367 (1.3 cm ± 4 cm). This suggests that parrotfish island-wide may be responding to influences other 368 than fishing pressure. Human-derived factors such as tourism, boat traffic, and pollution, as well 369 as environmental factors such as ocean current, wave action, or natural disturbance are potential 370 additional factors influencing size. While MPA and fished pairs were chosen based on proximity

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 371 and similarity of geomorphological conditions, some MPAs were exposed to a much stronger 372 current than their fished pairs (such was the case at Temae, P1 & F1), and other MPAs were 373 exposed to heavier boat traffic and tourist activity than their fished pairs (such was the case at 374 Tiahura Motu and its fished pair, P4 & F3). 375 376 The minimal difference in average body size between sites may also be due to a lack of data on 377 parrotfish range in Mo’orea. Parrotfish range is extremely variable, differing according to life 378 history stage, species, depth, latitude, and lunar phase (Howard, et al., 2013). While the spatial 379 proximity between fished and MPA pairs were determined based on the home range of parrotfish 380 reported in Howard’s (2013) study, it is possible that individual home ranges ‘spilled over’ 381 between pairs. In addition, terminal phase individuals (large bodied) have larger home ranges 382 than initial phase individuals (Howard, et al., 2013); therefore, the same large-bodied parrotfish 383 may have been observed in both the MPA site and its fished pair. If this is the case, it can be 384 suggested that parrotfish are vulnerable to size-selective fishing regardless of which site they 385 were observed in during this study. Subsequently, this proposes that MPAs in Mo’orea may not 386 have as great of a “reserve effect” on parrotfish abundance and size as they have been found to 387 have with other fish species in different locations of the world (Polunin and Roberts, 1993; 388 Walker and Robinson, 2009; Starr et al., 2015). While managers may use MPAs as tools to 389 regulate fishing pressure and maintain biodiversity, this study confirms that MPAs are 390 redistributing fishing pressure into concentrated zones of the lagoon. Overexploitation of large- 391 bodied parrotfish in those zones is inadvertently leading to populations dominated by smaller- 392 bodied parrotfish throughout the lagoon. 393 394 The short timescale of this study is an additional factor to consider when comparing the effects 395 of the PGEM on parrotfish size between sites. For instance, MPA reserve benefits might be slow 396 to accumulate given the relatively recent establishment of MPAs. Previous studies suggest that 397 20 years or more may be needed to detect significant changes in response variables that are due 398 to the establishment of MPAs (Starr et al., 2015). Factors such as short-term environmental 399 variability and the high spatial and temporal variability of fish recruitment patterns could 400 influence the impression of how MPAs are working, making short-term ecosystem monitoring 401 inconclusive and unrepresentative of greater ecological patterns (Starr et al., 2015). Given the 402 high frequency of environmental disturbances in Mo’orea (Lamy et al., 2015; Adam et al., 2011), 403 long-term monitoring is needed to identify greater patterns of ecosystem responses to human- 404 derived disturbances such as size-selective fishing. 405 406 The negligible difference between average parrotfish size in MPA and fished sites has important 407 consequences for the future trajectory of Mo’orea’s reef ecosystem. The decline in grazing 408 capacity of fish in both managed areas and fished areas suggests reduced resilience of reefs 409 throughout the lagoon, including reefs exposed to fishing pressure and reefs absent of fishing 410 pressure. Consequences of an island-wide decrease in parrotfish size may include a decreased 411 response rate and functional aptitude of populations to graze the reef following natural 412 disturbance-mediated shifts to macroalgal dominance (Barba, 2010). This, in turn, may entail 413 unrecoverable phase shifts of the lagoon ecosystem to algal domination (Roff and Mumby, 414 2012; Bonaldo and Bellwood, 2008; Hoey and Bellwood, 2008; Roff et al., 2011; Wilder, 2003; 415 Hay, 1984). Island-wide declines in functional diversity may also denote a decreased gene pool 416 of large-bodied parrotfish (Pope and Macer, 1996). Implications of this for future parrotfish

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 417 populations could include: reduced growth, shorter life span, earlier maturation, or earlier sex 418 reversal, as found in a similar study analyzing the effects of fishing pressure on two parrotfish 419 species (C. sordidus and S. psittacus) (Barba, 2010). Loss of genetic diversity (Smith, Francis & 420 McVeagh, 1991) and change in assemblage structure (Russ and Alcala, 1989) are further 421 consequences of increased fishing pressure. 422 423 Location 424 Parrotfish size varied with location. Average size was different between each location, being 425 greatest at Tiahura (29 cm) and smallest at Pihaena (24 cm). Tiahura was unusual because its 426 paired MPA and fishing site differed greatly in terms of human presence. The MPA site was 427 situated next to Tiahura Motu, a motu frequently visited by tourists and in close proximity to 428 picnic tables and a restaurant (Fig. 1). The fished site was barren of tourists and fishermen on 429 both occasions surveyed; however, despite this, fish were still large and plenty in that site. Fish 430 size is therefore naturally larger in this location, possibly due to the increased availability of food 431 from nearby picnicking activity. The smaller size of parrotfish in Pihaena may be explained by 432 both sites’ proximity to channel markers and corresponding boat traffic. Regardless of these 433 anthropogenic influences, however, Mo’orea’s lagoon has been observed to be naturally 434 heterogeneous throughout (Lison de Loma et al., 2008; Lamy et al., 2015; Adam et al., 2011). 435 Similarly to this study, Lamy (2015) found a large variation in the functional responses of 436 herbivorous fish across the western and northeastern reefs of Mo’orea, despite homogenization 437 of the entire lagoon habitat toward reduced coral cover and complexity. This spatial 438 heterogeneity of coral-reef fish morphology and functional diversity provides further evidence 439 that species might differ with respect to other factors that influence their responses, such as 440 environmental characteristics. This could explain the variation in parrotfish size between 441 locations around Mo’orea, especially as ocean current, wave exposure, boat traffic, and tourism 442 activities vary island-wide. 443 444 Habitat 445 In addition to site and location, this study also found differences in parrotfish size in two 446 different reef habitats within the lagoon. In almost every site, parrotfish size and abundance were 447 greater in the back reef than in the fringing reef (Fig. 7). Body size was especially larger in back- 448 reef habitats in fished sites compared to that in fringing reef habitats in fished sites (Fig. 7). 449 Combined, these data suggest that the fringing reef may be more vulnerable to size-selective 450 fishing than the back reef. Interactions with local fishermen support this claim, as the majority of 451 fishermen are recreational fishermen, fish before 7 am and after 7 pm, and use spearfishing as 452 their main method (Fig. 9) (Rice, 2016, Personal observation). Given that most recreational 453 fishermen do not use boats (Leenhardt et al., 2016), and the preferred times to fish are in periods 454 of limited sunlight, responses suggest that most fishermen fish just off shore in the fringing reef, 455 rather than the distant back reef. Furthermore, the reserve effect of MPAs has been found to be 456 greatest towards the back reef of the lagoon, and smallest within the fringing reef (Lison de 457 Loma et al., 2008). Together with my results, this implies that, regardless of site, parrotfish in the 458 fringing reef are more vulnerable to size-selective fishing pressure that parrotfish in the back 459 reef. 460 461 The present study found variation in parrotfish size on many spatial scales. Such variation 462 suggests the differing vulnerabilities of parrotfish to fishing pressure according to location and

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 463 reef habitat. While Lamy (2015) asserts that differing spatial responses of herbivorous fishes to 464 natural disturbance may provide Mo’orea’s reef ecosystem with greater resilience, differing 465 spatial responses as a result of human-derived disturbances such as fishing may not provide the 466 same ecosystem benefits. Unlike environmental disturbances (such as cyclones and bleaching 467 events), anthropogenic disturbances are not part of the regime under which coral-reef ecosystems 468 have evolved. Thus, differing spatial responses of parrotfish size to fishing pressure in a non- 469 environmentally disturbed state may be either a reflection of the variability of fishing pressure at 470 each location, or the variability of vulnerabilities of parrotfish throughout different parts of the 471 lagoon. Regardless, we cannot assume Mo’orea’s reefs will remain resilient in the face of 472 disturbance as they have in the past, especially when considering the mounting consequences of 473 an ever-increasing human population on ecosystems such as Mo’orea’s lagoon. 474 475 Temporal Dynamics 476 477 Lunar occasion 478 Over the short temporal scale of this study, fish abundance at any one site and time is highly 479 variable and unreliable as an absolute measure of herbivore pressure (Starr et al., 2015). The 480 current study observed greater average body sizes during the full moon in comparison with the 481 new moon, and no differences in fish abundance between lunar occasions. Previous studies have 482 found that fish abundance and behavior differ according to moon phase (Vinson, 2014); 483 however, finding a difference in average fish-body size between lunar occasions was an 484 unexpected result. This finding suggests that either different parrotfish were observed at each site 485 during the two survey occasions, or that smaller parrotfish were simply not observed due to 486 cryptic behavior during the full moon (Galzin, 1987). 487 488 Mo’orea’s fishing population 489 490 Mo’orea’s population has increased at an unsustainable rate in recent decades. From 2007 to 491 2012 alone, Mo’orea’s population increased 6.33% (ISPF, 2015) (Fig. 10). In addition, estimates 492 of fishing density in 2007 revealed 77 fishermen per km2 (Brenier, 2009). If this calculated 493 fishing pressure is accurate, it is quite high considering that five fishermen per km2 is the upper 494 limit at which coral reef resources can be safely exploited (McClanahan et al., 2002). Assuming 495 population and fishing density are proportional, this indicates that fishing density in 2012 was 82 496 fishers per km2. Furthermore, assuming population growth rate hasn’t increased since 2012, the 497 current fishing density in 2016 may be as high as 86 fishers per km2; however, as population 498 growth appears to be increasing exponentially (Fig. 10), fishing density today is most likely even 499 higher. 500 501 The implications of this unprecedented increase in population are noteworthy, to say the least. 502 While consequences are apparent on land (i.e. pollution, hillside construction, crowded public 503 spaces), the most threatening costs are hidden beneath the surface of the lagoon. Given that 504 fishermen attain higher economic gains from larger fish and understanding the effect of the 505 PGEM’s regulation on minimum catch size, an increased population means increased targeting 506 pressure on larger parrotfish. For example, roadside fish size evaluations from Madi Moussa’s 507 (2010) study in 2007 found average parrotfish catch size was 27 cm ± 2 cm, compared to a 508 minimum legal catch size of 25 cm. In addition, as lagoon fishing exploits only a few species,

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 509 with only five to six species representing more than 80% of total sales (Madi Moussa, 2010), 510 parrotfish in the lagoon are especially vulnerable to size-selective fishing pressure. 511 512 To attain an understanding of responses of the lagoon ecosystem to increases in demand for 513 larger parrotfish, this study took a socio-ecological approach using a survey technique referred to 514 as “participatory monitoring” (Leenhardt et al., 2016) and included the local population in the 515 quest for ecological information, both qualitative and quantitative. In the context of ecological 516 awareness, such anecdotal monitoring has its value, but also its limits. In all cases, participatory 517 monitoring presupposes that research provides itself with the means of its supervision, betting in 518 some way on its instructive and administrative efficiency. Although the information collected 519 was quantitative, it involved substantial uncertainty because it relied on the long-term memory of 520 the person interviewed and his or her ability to convert an image or a memory into a physical 521 size (Gilbert, 2006). Responses from surveys are thus considered qualitative. 522 523 Estimates of parrotfish body size from the beginning of a fisherman’s career revealed the 524 vulnerability of parrotfish to varying levels of demand over a timescale of 60 years (Fig. 8). 525 Given that each interview was independent from the next, and observations included a wide 526 demographic of fishermen, the mutual decline in parrotfish size from 1956 to 2013 reputes 527 fishermen knowledge, and also reinforces the perceived decline in parrotfish size over time. 528 Additionally, when comparing size estimates from 1956 to 2013 to population size over the same 529 time scale, a weak inverse relationship between estimated fish size and human population size is 530 evident. Thus, estimating observable ecological variables through participatory monitoring— 531 especially in a rural community deeply connected with their natural environment (Budd-Falen, 532 1995; Jentoft, 1999)— can contribute to the collective understanding of ecological change over 533 time, especially in the absence of scientific data from earlier time periods. 534 535 Application to future marine management strategy 536 537 Civic engagement ensures environmental and economic sustainability in rural communities 538 (Budd-Falen, 1995; Jentoft, 1999), and the intergenerational environmental knowledge of local 539 resource users is comprehensive and relevant to modern conservation objectives. This case study 540 in Mo’orea raises questions about the assumed connections between local control, public 541 participation, and successful conservation results. Especially in Mo’orea’s lagoon where 542 underwater resources and ecological processes are not sedentary, visible, or easily quantified, 543 there was considerable debate between differing stakeholders including locals, the state, and 544 scientists over indicators of lagoon health, patterns of and larval transport, the 545 dynamics of land-based pollution effluents, and the location and importance of different lagoon 546 uses and meanings (Walker, 2001). While the government and biologists cited “scientific” 547 studies and spatial data to support the creation of the PGEM, Mo’orea’s stakeholders likewise 548 asserted their own knowledge of the lagoon by describing traditional lagoon management and 549 fishing laws, reciting Maohi legends about Mo’orea and its lagoon, and explicating their life- 550 long, daily interactions with the fish, coral reefs, sharks, and other organisms of Mo’orea’s 551 lagoon (Walker, 2001). Walker (2001) found that many of Mo’orea’s fishermen keep detailed 552 diaries of fishing information, which have been passed down for generations. These diaries 553 include daily explanations of where different species of fish are found in the lagoon, based on a 554 variety of indicators such as currents, winds, lunar cycle, and seasons. Older fishermen, in

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 555 particular, claim knowledge of their home lagoon areas at the scale of individual coral heads, and 556 they are able to explain precisely where, at what time, and on what day one can go to catch a 557 particular species of fish (Walker, 2001). Despite fishermen’s comprehensive and relevant 558 environmental knowledge of their lagoon, the government alienated locals from the deliberation 559 process through the privileged use of GIS decision-making, a resource not accessible to the 560 majority of Mo’orea’s local population (Walker, 2001). As a result, state-mandated MPAs 561 spurred significant political struggles and prompted resistance among locals unlike any previous 562 resource regulation in French Polynesia (Yousing, 2016; Rey, 2016; Bambridge, 2016; Aubanel 563 et al., 2013; Walker and Robinson, 2009; Walker, 2001). Consequently, and, as this current study 564 suggests, absence of local cooperation might ultimately render MPA management plans 565 unsustainable. The mapping of MPAs in Mo’orea highlighted discrepencies between 566 policymaker recommendations and fishermen ecological knowledge, conflicts over access to 567 lagoon space and resources, and disagreements over evolving forms of lagoon conservation 568 (Yousing, 2016; Rey, 2016; Walker, 2001). Westerners, developers, and tourists were handed 569 control of large zones of the lagoon, promoting tourist activities in parts of the lagoon that 570 residents previously relied on for sustenance (Yousing, 2016; Rey, 2016; Walker, 2001). 571 Accordingly, alienated stakeholders formed politicized local associations to defend their own 572 livelihoods and sovereignty, and display their opposition to government interference into lagoon 573 management, as well as foreign exploitation of the lagoon (Bambridge, 2016; Rey, 2016; 574 Walker, 2001). 575 576 Several of these local associations support a movement to revert to a traditional management 577 system that many islands in French Polynesia previously used, and some islands such as Rapa 578 and Maiao still use (Agence des aires marines protégées, 2012). Referred to as “Rahui,” this 579 traditional style of management accounts for local understanding of, and relationship with, 580 natural resources, embracing holistic and rational modes of enforcement and avoiding 581 overregulation (Bambridge, 2016). The ocean and the reef ecosystem have survived centuries of 582 disturbance without human interference. However, as indicated by this study, human-centered 583 management strategy, despite well-intentioned environmental objectives, shows detrimental 584 effects on the reef ecosystem. Locals that have been fishing on Mo’orea their whole lives 585 understand the consequences of a poorly managed reef (Yousing, 2016). Managing land should 586 be in the best interest of the natural environment and the local stakeholders, not foreign 587 developers. If we want the reef to remain resilient in the face of increasing demand of its 588 resources, we must equally integrate local stakeholders, and consider the ecological dynamics 589 that sustain a reef, such as fish life history, reproductive stage, and especially, ecosystem 590 function, as Rahui considers. 591 592 Conclusions 593 The effects of fishing on marine ecosystem structure and processes are significant and complex. 594 Results from this study highlight the inverse relationship between size-selective fishing 595 intensity and the average size of the herbivorous reef-fish family Scaridae greatly targeted by 596 Mo’orea’s fishing population. Ecological relationships and questions relevant to the marine 597 environment must be studied on many spatial and temporal scales, as the marine environment 598 can be incredibly variable according to location, habitat, and moon phase, as this study revealed. 599 The use of bio-indicators such as herbivorous fish size can enhance our understanding of fishing 600 effects throughout space and time. Assessing both the spatial and temporal variations in

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 601 parrotfish and other herbivorous species’ composition is important to comprehensively explore 602 the effects of size-selective fishing pressure, and thus the vulnerability of the coral-reef 603 ecosystem to reductions in resilience. Enforcements to capture parrotfish characteristic of the 604 highest grazing activity, in a small-scale fishery with limited fish supply and an increasing 605 fishing population may hinder the grazing capacity of parrotfish further, and make phase shifts to 606 a macroalgal-dominated state more likely. Size-selective fishing of a species keystone to the reef 607 ecosystem is a result of population-based marine management strategy, and the resulting 608 vulnerability of the reef ecosystem has only just gained recognition. More evidence is needed to 609 determine a scientific basis for a change from population-based management to ecosystem-based 610 management for the vulnerable marine ecosystem in Mo’orea and elsewhere. Extensive time 611 series data on the responses of various, diverse ecosystems to anthropogenic transitions such as 612 population increase and management strategy will create a record for future managers to consult 613 when strategizing sustainable solutions for preserving the earth’s natural resources. 614 615 Acknowledgements 616 617 I would like to thank all of my Professors, Patrick O’Grady, Jonathon Stillman, Justin Brashares, 618 Cindy Looy, and Ivo D. for all of the help and time that they put into developing my project and 619 editing my paper. I would also like to thank the graduate student instructors Natalie Stauffer- 620 Olsen, Ignacio Escalante, and Eric Armstrong for their endless support and dedication to making 621 this experience one of a lot of learning, as well as a lot of fun. I would also like to thank Alex 622 Yokley and Ryan Mullen for always being willing survey buddies, and my family for their 623 inspiration and support. Additionally, I would like to thank the wonderful Gump Station and 624 Atitia Center staff, Val, Irma, Jaques, Tony, Frank, and Hinano for sharing their bliss and culture 625 with us. Finally, I would like to thank all of my amazing classmates who made this experience 626 truly unforgettable. Maururu roa. #nofermentedfish2k16. 627 628 References 629 630 Adam T, Schmitt R, Holbrook S, Brooks A, Edmunds P, Carpenter R, Bernardi G. 2011. 631 Herbivory, Connectivity, and Ecosystem Resilience: Response of a Coral Reef to a Large-Scale 632 Perturbation. PLoS ONE 6:e23717. 633 634 Agence des aires marines protégées. 2012. Web. http://www.aires-marines.fr/ 635 636 Aubanel A. 1993. Evaluation socio-économique de la pêche en milieu corallien dans l'île de 637 Moorea. Journal de la Société des océanistes 96:49-62. 638 639 Aubanel A., Salvat B, and Féral F. 2013. Analyse du Role des Scientifiques et de la Place des 640 Connaissances Scientifiques dans L'elaboration et la Mise en œuvre du Plan de Gestion de 641 L'espace Maritime (PGEM) de Moorea et de sa Gouvernance. CRIOBE, BEST CORAL 642 Reports, Communication for Policy. 643 644 Bambridge T. 2016. The Rahui: Legal pluralism in Polynesian traditional management of 645 resources and territories. Acton, A.C.T.: ANU Press. 646

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 647 Barba J. 2010. Demography of parrotfish: age, size and reproductive variables. Masters 648 (Research) thesis, James Cook University. 649 650 Bellwood DR, Hughes TP, Connolly SR, Tanner J. 2005. Environmental and geometric 651 constraints on Indo-Pacific coral reef biodiversity. Ecol. Lett. 8:643–651. 652 653 Bellwood DR, Hughes TP, Hoey A. S. 2006. Sleeping functional group drives coral reef 654 recovery. Curr. Biol 16:2434–2439. 655 656 Bonaldo RM, Bellwood DR. 2008. Size-dependent variation in the functional role of the 657 parrotfish rivulatus on the , Australia. Mar. Ecol. Prog. 360: 237–244. 658 659 Brenier A. 2009. Pertinence des approches participatives pour le suivi écosystémique des 660 pêcheries récifales. Thèse de Doctorat. Université Paris VIPierre et Marie Curie et UFR des 661 Sciences de la vie, Université de Tuléar-Institut halieutique et des sciences marines. 143. 662 663 Bruggemann JH, Kuyper MWM, Breeman AM. 1994. Comparative analysis of foraging and 664 habitat use by the sympatric Caribbean parrotfish Scarus vetula and 665 viride (Scaridea). Marine Ecology-Progress Series, 112:51–66. 666 667 Budd-Falen K. 1995. Protecting community stability and local economies: opportunities for local 668 government influence in federal decision- and policy-making processes. A Wolf in the Garden 669 Eds P Brick, R Cawley (Rowman & Littlefield, Lanham, MD) 73 – 86. 670 671 Burkepile DE, Hay ME. 2008. Herbivore species richness and feeding complementarity affect 672 community structure and function on a coral reef. Proc. Natl Acad. Sci. USA 105, 16 201–16 673 206. 674 675 Choat H, Bellwood DR. 1998. & Parrotfishes. Pp. 209-210 in W Eschmeyer, J 676 Paxton, eds. Encyclopedia of Fishes – second edition. San Diego, CA: Academic Press. 677 Dalzell P, Adams TJH, Polunin NVC. 2006. Coastal fisheries in the Pacific Islands. Oceanogr. 678 Mar. Biol. Annu. Rev. 34, 395–531. 679 680 Ferraris J, Cayré P. 2003. Les pêcheries récifales dans le Pacifique sud : d’une gestion intuitive 681 vers une gestion écosystémique raisonnée. Océanis, 29(3-4): 397-414. 682 683 Francour P. 2000. Evolution spatio-temporelle à long terme des peuplements de poissons des 684 herbiers à Posidonia oceanica de la réserve naturelle de Scandola (Corse, Méditerranée nord- 685 occidentale). Cybium 24(3): 85–95. 686 687 Galzin R, Lecchini D, Lison de Loma T, Moritz C. Parravicini V, Siu G. 2016. Long term 688 monitoring of coral and fish assemblages (1983–2014) in Tiahura reefs, Moorea, French 689 Polynesia. Cybium 40:31–41. 690 691 Galzin R. 1985. Écologie des poissons rêcifaux de Polynésie française. Cybium 9:403–407. 692

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 693 Galzin R. 1987. Structure of fish communities of French Polynesian coral reefs. 2. Temporal 694 scales. Mar. Ecol. Progr. Ser. 41:137–145. 695 696 Gilbert A. 2006. Prêcision et Biais des Estimations de la Consommation de Poissons par 697 Enquêtes Auprés des Mênages. Cas de la commune de Haapiti (Moorea) Polynêsie française. 698 UR-CoReUs, CRISP. 699 700 Harmelin-Vivien M, Harmelin JG, Chauvet C, Duval C, Galzin R, Lejeune P, et al. 1985. 701 Evaluation visuelle des peuplements et populations de poissons. Méthodes et problémes. Terre 702 vie 40: 467–539. 703 704 Hay ME. 1984. Patterns of fish and urchin grazing on Caribbean coral reefs: are previous results 705 typical? Ecology 65:446–454. 706 707 Hoey AS, Bellwood DR. 2008. Cross-shelf variation in the role of parrotfishes on the Great 708 Barrier Reef. Coral Reefs 27:37–47. 709 710 Howard KG, Claisse JT, Clark TB, Boyle K, Parrish JD. 2013. Home range and movement 711 patterns of the Redlip Parrotfish (Scarus rubroviolaceus) in Hawaii. Marine 712 Biology 160(7):1583-1595. 713 714 Hughes TP. 1994 Catastrophes, phase shifts, and large-scale degradation of a Caribbean coral 715 reef. Science 265:1547–1551. 716 717 Hughes TP, Baird AH, Bellwood DR, Card M, Connolly SR, Folke C. 2003. Climate change, 718 human impacts, and the resilience of coral reefs. Science 301:929–933. 719 720 Hughes TP, et al. 2007. Phase shifts, herbivory and the resilience of coral reefs to climate 721 change. Curr. Biology 17:1–6. 722 723 Hughes TP, Graham NAJ, Jackson JBC, Mumby PJ, Steneck RS. 2010. Rising to the challenge 724 of sustaining coral reef resilience. Trends Ecol. Evol. 25:633–642. 725 726 ISPF. 2015. Populaiton Légale du Descencement de la Populaiton de 2012 en Polynésie 727 Française. Papeete: Institut de la Statistique de la Polynésie Française. 728 729 Jackson JBC, et al. 2001. Historical overfishing and the recent collapse of coastal ecosystems. 730 Science 293:629–638. 731 732 Jennings S, Kaiser MJ. 1998. The effects of fishing on marine ecosystems. Adv. Mar. 733 Biol. 34:201–352. 734 735 Jennings S, Lock JM. 1996. Population and ecosystem effects of fishing. Reef Fisheries 193-218. 736 737 Jennings S, Polunin NVC. 1996. Effects of fishing effort and catch rate upon the structure and 738 biomass of Fijian reef fish communities. Journal of Applied Ecology 33:400-412.

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 739 740 Jentoft S. 1999. Healthy fishing communities: an important component of fish stocks. Fisheries 741 24:28-29. 742 743 Kronen M, Vunisea A, Magron F, McArdle B. 2010. Socio-economic drivers and indicators for 744 artisanal coastal fisheries in Pacific island countries and territories and their use for fisheries 745 management strategies. Mar. Policy 34:1135–1143. 746 747 Lamy T, Legendre P, Chancerelle Y, Siu G, Claudet J. 2015. Understanding the Spatio-Temporal 748 Response of Communities to Natural Disturbances: Insights from Beta-Diversity 749 Decomposition. PLOS ONE 10:e0138696. 750 751 Leenhardt P, Lauer M, Madi Moussa R, Holbrook SJ, Rassweiler A, Schmitt RJ and Claudet J. 752 2016. Complexities and Uncertainties in Transitioning Small-Scale Coral Reef Fisheries. Front. 753 Mar. Sci. 3:70. 754 755 Lison de Loma T. 2005. Liens entre les caracte´ristiques re´cifales des ıˆles, la densite´ humaine 756 et les populations de poissons re´cifaux commerciaux des ıˆles de la Socie´te´. Convention 757 Ministe`re de la Peˆche, de l’Industrie et des Petites et Moyennes Entreprises. 4:95. 758 759 Lison de Loma T, Osenberg C, Shima J, Chancerelle Y, Davies N, Brooks A, Galzin R. 2008. A 760 Framework for Assessing Impacts of Marine Protected Areas in Moorea (French Polynesia). 761 Pacific Science 62:431-441. 762 763 Littler MM, Taylor PR, Littler DS. 1989. Complex interactions in the control of coral zonation 764 on a Caribbean reef flat. Oecologia 80:331-340. 765 766 Lokrantz J, Nyström M, Thyresson M, Johansson C. 2008. The non-linear relationship between 767 body size and function in parrotfishes. Coral Reefs 27: 967–974. 768 769 Madi Moussa R. 2010. Estimation de la taille des poissons lagonaires vendus sous la forme de tui 770 en bord de route sur l’île de Moorea (Polynésie fran- çaise) par analyse de clichés numériques. 771 Cybium 34:73–82. 772 773 McClanahan T, Polunin N, Done T. 2002. Ecological states and the resilience of coral 774 reefs. Conservation Ecology 6:18. 775 776 Moberg F, Folke C. 1999. Ecological goods and services of coral reef ecosystems. Ecol Econ 777 29:215–233. 778 779 Norström A, Nyström M, Lokrantz J, Folke C. 2009. Alternative states on coral reefs: beyond 780 coral–macroalgal phase shifts. Marine Ecology Progress Series 376:295-306. 781 782 Pandolfi J. 2003. Global Trajectories of the Long-Term Decline of Coral Reef Ecosystems. 783 Science 301:955-958. 784

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 785 PGEM 2005. L’essentiel du PGEM (Plan de gestion de l’espace maritime à Moorea). Journal 786 Officiel de la Polynésie française. 419-429. 787 788 Polunin N, Roberts C. 1993. Greater biomass and value of target coral-reef fishes in two small 789 Caribbean marine reserves. Marine Ecology Progress Series 100:167-176. 790 791 Pope J, Macer C. 1996. An evaluation of the stock structure of North Sea cod, haddock, and 792 whiting since 1920, together with a consideration of the impacts of fisheries and predation 793 effects on their biomass and recruitment. ICES Journal of Marine Science 53:1157-1169. 794 795 R Core Team. 2013. R: A language and environment for statistical computing. R Foundation for 796 Statistical Computing, Vienna, Austria. URL http://www.R-project.org/. 797 798 Rey J. 2016. Moorea - Moorea Rahui souhaite imposer le principe d'une gestion lagonaire 799 communautaire. La Depeche de Tahiti. 800 801 Rice, K. 2016. Personal observation. 802 803 Roff G, Ledlie M, Ortiz J, Mumby P. 2011. Spatial Patterns of Parrotfish Corallivory in the 804 Caribbean: The Importance of Coral Taxa, Density and Size. PLoS ONE 6:e29133. 805 806 Roff G, Mumby P. 2012. Global disparity in the resilience of coral reefs. Trends in Ecology & 807 Evolution 27:404-413. 808 809 Russ G, Alcala A. 1989. Effects of intense fishing pressure on an assemblage of coral reef fishes. 810 Marine Ecology Progress Series 56:13-27. 811 812 Salvat B, Aubanel A. 2002. La gestion des récifs coralliens de Polynésie française. Rev. Ecol. 813 (Terre Vie) 57:193- 251. 814 815 Smith P, Francis RM, McVeagh M. 1991. Loss of genetic diversity due to fishing pressure. 816 Fisheries Research 10:309-316. 817 818 Starr RM, Wendt DE, Barnes CL, Marks CI, Malone D, Waltz G, Yochum N. 2015. Variation in 819 Responses of Fishes across Multiple Reserves within a Network of Marine Protected Areas in 820 Temperate Waters. PLoS ONE 10(3):e0118502. 821 822 Thresher R. 1984. Reproduction in reef fishes. British Crown Colony of Hong Kong: T.F.H. 823 Publications. 824 825 Vieux C. 2002. La Peche Lagonaire á Moorea (Polynésie française): Évolution 826 Quantitative et Socio- Économique Depuis 1992. Papeete: CRIOBE. 827 Vinson MR, Angradi TR. 2014. Muskie Lunacy: Does the Lunar Cycle Influence Angler Catch 828 of Muskellunge (Esox masquinongy). PLOS ONE 9(5): e98046. 829

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 830 Walker BLE. 2001. Mapping Moorea's lagoons: conflicts over marine protected areas in French 831 Polynesia. Proceedings of the Inaugural Pacific Regional Meeting of the International 832 Association for the Study of Common Property, Brisbane, Australia. 23 pp. 833 834 Walker BLE, Robinson M. 2009. Economic development, marine protected areas and gendered 835 access to fishing resources in a Polynesian lagoon. Gender Place Cult. 16:467–484. 836 837 Wilder RM 2003. The influence of algal production on the effects of herbivorous fish and sea 838 urchins on coral reef algae. In: Algae-Herbivore Interactions on the Coral Reefs of Moorea, 839 French Polynesia. Dissertation. University of California, Santa Barbara, USA, 108-155. 840 841 Wilson S, Fisher R, Pratchett M, Graham N, Dulvy N, Turner R, Cakacaka A, Polunin Nrushton 842 S. 2008. Exploitation and habitat degradation as agents of change within coral reef fish 843 communities. Global Change Biology 14:2796-2809. 844 845 Yonger M. 2002. Approche de la Pêcherie Récifo-Lagonaire de Moorea (Polynésie française): 846 Évaluation de la Production Halieutique et de la Population de Pêcheurs. Rennes: ENSA. 847 848 Yousing, Tony. 2016. Mo’orea, French Polynesia. Personal communication.

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 Figure 1

Marine Protected Areas (MPAs) defined by the PGEM in Mo’orea’s lagoon (PGEM 2005).

MPAs are represented in black zones. I selected five out of the total eight MPAs defined by the PGEM (labeled P1, P2, P3, P4, and P5), and I selected four paired fishing zones (labeled F1, F2, F3, and F5), represented in white zones.

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 Figure 2

Histogram of response variable (cm) showing normal data.

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 Figure 3

Average parrotfish size (cm) between variables: site (A), location (B), reef habitat (C), and lunar occasion (D).

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 Figure 4

Average parrotfish abundance between variables: site (A), location (B), reef habitat (C), and lunar occasion (D). Only “Site” was statistically significant.

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 Figure 5

Parrotfish size according to geographic location of paired fishing and MPA sites. Error bars represent 95% confidence intervals.

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 Figure 6

Parrotfish abundance according to geographic location of paired fishing and MPA sites. Error bars represent standard error around the mean.

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 Figure 7

Box-plot showing interactions between site (in red and blue), location (z-axis), and habitat type (x-axis) with fish size (cm) (y-axis).

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 Figure 8

Estimates of parrotfish size (cm) over time. Blue dots represent fishermen estimates at the start of their fishing career (1956-2013). Red dot represents averaged fisherman estimate of parrotfish size (cm) today (2016).

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 Figure 9

Participatory survey responses from 50 fishermen. Y-axes represent frequency of response per question.

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 Figure 10

Mo’orea’s population from 1959 to 2012 as recorded by ISPF (ISPF 2015).

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 Table 1(on next page)

Four-way ANOVA with size (cm) as the response variable and site, location, habitat type, and lunar occasion (“phase”) as the predictor variables.

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016 Df Sum Sq Mean Sq F value Pr(>F) Site 1 277 277.1 22.061 3.19e-06 *** Loc 4 1608 402.0 31.998 < 2e-16 *** Hab_Type 1 701 700.6 55.774 2.46e-13 *** Phase 1 442 441.6 35.153 4.81e-09 *** Site:Loc 4 397 99.3 7.902 3.13e-06 *** Site:Hab_Type 1 23 23.3 1.857 0.17336 Loc:Hab_Type 4 729 182.2 14.500 2.23e-11 *** Site:Phase 1 1 0.5 0.041 0.84012 Loc:Phase 4 551 137.6 10.956 1.28e-08 *** Hab_Type:Phase 1 217 216.8 17.258 3.67e-05 *** Site:Loc:Hab_Type 2 242 121.1 9.637 7.45e-05 *** Site:Loc:Phase 4 725 181.2 14.423 2.56e-11 *** Site:Hab_Type:Phase 1 101 100.9 8.028 0.00474 ** Loc:Hab_Type:Phase 4 366 91.4 7.277 9.63e-06 *** Site:Loc:Hab_Type:Phase 1 5 4.9 0.391 0.53192 Residuals 692 8693 12.6 --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.2657v1 | CC BY 4.0 Open Access | rec: 21 Dec 2016, publ: 21 Dec 2016