Vol. 501: 191–206, 2014 MARINE ECOLOGY PROGRESS SERIES Published March 31 doi: 10.3354/meps10711 Mar Ecol Prog Ser

Residency and seasonal movements in Lutjanus argentiventris and rosacea at Los Islotes Reserve, Gulf of California

Thomas TinHan1,*, Brad Erisman2, Octavio Aburto-Oropeza2, Amy Weaver3, Daniel Vázquez-Arce3, Christopher G. Lowe1

1Department of Biological Sciences, California State University Long Beach, Long Beach, California 90815, USA 2Marine Biology Research Division, Scripps Institution of Oceanography, University of California at San Diego, La Jolla, California 92093-0202, USA 3Sociedad de Historia Natural Niparajá A.C., 23020 La Paz, Baja California Sur, Mexico

ABSTRACT: A detailed understanding of inter- and intraspecific movement patterns is required to understand how marine species interact with surrounding ecological communities, their suscepti- bility to anthropogenic disturbance (e.g. fishing pressure), or the feasibility of management strate- gies. Between August 2010 and September 2012, we used acoustic telemetry to continuously mon- itor movements of 31 Lutjanus argentiventris (yellow snapper) and 25 Mycteroperca rosacea (leopard ) at Los Islotes, a small no-take reserve and reported spawning site for both spe- cies in the SW Gulf of California. Though the majority of fish from both species exhibited moder- ate levels of site fidelity to Los Islotes (snapper: present 49 ± 30% of days since tagging, grouper: 64 ± 30%), cluster analyses revealed multiple patterns of site fidelity within species. Approxi- mately 30% of snapper exhibited decreases in site fidelity during the spawning season, and snap- per did not at the reserve during the study. Grouper spawning aggregations at Los Islotes were visually observed in 2011 and 2012, though the abundance of fish and the intensity of courtship behaviors were reduced in comparison with reported aggregations elsewhere in the Gulf. Three snapper and 2 grouper made repeated movements across pelagic waters between Los Islotes and Marisla Seamount, another documented aggregation site in the SW Gulf. The demon- strated variation in movements of these species over multiple temporal and spatial scales warrants consideration of movement patterns in assessments of reserve performance, as well as the combi- nation of traditional fisheries regulations (e.g. size limits) with marine reserves throughout the Gulf.

KEY WORDS: Reef fishes · Yellow snapper · Leopard grouper · Marine Protected Areas · MPAs · Acoustic monitoring · Fish spawning aggregations

Resale or republication not permitted without written consent of the publisher

INTRODUCTION influence the susceptibility of predators to exploita- tion or anthropogenic disturbance (Fahrig 2007), and Members of the snapper-grouper complex (Lut- may also alter distributions of prey items and com- janidae; Epinephelidae) play important roles in the petitors (Holt 1984). Unfortunately, patterns of move- organization of coastal ecosystems via top-down con- ment remain poorly understood for a large number of trol as predators. To better understand their interac- species, many of which are exploited by commercial tions with surrounding marine communities, we need and recreational fisheries worldwide. Population information regarding patterns of space use and declines have been reported for a number of these movement in these species. These characteristics may species (e.g. Lutjanus cyanopterus, Epine phelus stri-

*Corresponding author: [email protected] © Inter-Research 2014 · www.int-res.com 192 Mar Ecol Prog Ser 501: 191–206, 2014

atus; Huntsman 1996, Sala et al. 2001), and manage- range from southern California to Peru, whereas ment efforts have been impeded by the lack of avail- leopard grouper are found primarily in the Gulf, able information regarding the movements of these ranging from the southwest coast of Baja California species. to Jalisco, Mexico (Allen 1985, Heemstra & Randall In the southern Gulf of California, approximately 1993). Both species aggregate seasonally to spawn half of the fish species targeted by coastal fisheries (snapper: May to September; grouper: April to May; comprise snappers and , with landings of Sala et al. 2003, Aburto-Oropeza et al. 2009, Piñon et these species generating nearly 78% of annual ex- al. 2009), during which time they are heavily ex - vessel revenue (i.e. price paid for landings prior to ploited by coastal fisheries (Sala et al. 2003, Erisman onshore handling) (Erisman et al. 2010). Because et al. 2010). There have been no previous studies of many reef-associated species in the Gulf, including reef fish movements in the Gulf, and it is not known snappers and groupers, form seasonal spawning how these species utilize coastal habitats, particu- aggregations at predictable times and locations (Sala larly with respect to spawning. et al. 2003), commercial fishing activities often occur The Los Islotes Marine Reserve is a small rocky in a coincident manner (Sadovy de Mitcheson & Eris- islet in the southwestern Gulf of California, and 1 of man 2012). However, current management strategies 3 no-take reserves (Marine Protected Area, MPA) do not address the seasonal nature of these fisheries, in the Espíritu Santo Archipelago National Park and instead focus on establishing networks of marine (ESANP). The ESANP was established in 2007 with reserves — areas partially or entirely protected from the objectives of sustaining natural resources, con- fishing (Rife et al. 2012). Furthermore, enforcement tributing to regional coastal fisheries, and promo - of reserve areas is sparse, and many of these areas ting sustainable tourist activities within the park are subject to substantial illegal fishing pressure. (CONANP 2011). Though closed to fishing activities, Though such reserves are expected to increase bio- Los Islotes is presumed to be subject to similar levels mass, organism size and diversity within their bound- of illegal fishing as other small protected areas in the aries (Halpern 2003), it is unclear how effective they region. It is also recognized by regional managers as are for species that may migrate to seasonally tar- an aggregation site for several species, including yel- geted aggregation sites. low snapper and leopard grouper. Thus, the Los Species that form spawning aggregations typically Islotes Marine Reserve represents an ideal location in exhibit resident or transient patterns of behavior, in which to quantify the site fidelity, movement patterns which individuals may remain near their home range and space use of these 2 commercially important to spawn in small groups throughout the year (resi- aggregative spawners. A static array of acoustic re - dent), or migrate over varying distances to form large ceivers was used to monitor the long-term move- aggregations in a brief, annual event (transient) ments of yellow snapper and leopard grouper in rela- (Domeier & Colin 1997). Though resident fish may tion to Los Islotes, and to address the following benefit most from small marine reserves, transient in- 4 questions: (1) What is the site fidelity of these spe- dividuals may be afforded little or no protection if ag- cies to the reserve? (2) How do these species differ gregations form in unprotected areas. However, spe- with regard to their temporal movement patterns cies may also adopt a diversity of behaviors that range within the Los Islotes MPA? (3) Do individuals of between the extremes of resident and transient ag- these species exhibit patterns of site fidelity indica- gregation, and multiple patterns of aggregation be- tive of resident or transient spawning? (4) What habi- havior may exist within a species or population (Egli & tat types or depth features do fish associate with at Babcock 2004, Jarvis et al. 2010, Semmens et al. 2010, Los Islotes? Sagarese & Frisk 2011). In order to address such vari- ability in population management strategies, one needs to determine whether multiple aggregation be- MATERIALS AND METHODS haviors are present within a population, and the pro- portion of individuals exhibiting each behavior Study site (Nemeth 2012). Yellow snapper Lutjanus argentiventris and leop- Los Islotes Marine Reserve is positioned around a ard grouper Mycteroperca rosacea are among the small rocky islet at the northern tip of Isla Espiritu most abundant predatory fishes on rocky reefs Santo, Baja California Sur, Mexico (24° 35’ 54’’ N, throughout the Gulf of California (Ramírez & 110° 24’ 6’’ W) (Fig. 1). It is 1 of 3 zones in the southern Rodríguez 1990, Erisman et al. 2010). Yellow snapper Gulf fully protected from fishing, and it covers an TinHan et al.: Snapper and grouper movement patterns 193

Fig. 1. Map of Los Islotes Reserve, Baja California Sur, Mexico. Inset shows location of Los Islotes in the Gulf of California. Dark green shaded rectangle, area protected from fishing; blue dots, VR2W receivers; light green shaded area, the estimated 50% detection range. Red borders around each receiver indicate the estimated acoustic detection range of individual re- ceivers. Numbers on contour lines represent isobath depths in meters. Green triangle indicates location of acoustic receiver at Marisla Seamount area of 0.61 km2. The reserve boundaries extend 69 kHz coded pulse-train at 110 to 250 s intervals 300 m from shore and encompass a reserve area that (180 s nominal range) for an expected battery life of can be broken down into 3 zones by habitat and 1565 d. With the exception of 1 yellow snapper depth: (1) north Los Islotes: a steep boulder field (29.0 cm total length, TL), all tagged fish were larger extending approximately 100 m from shore and than the reported length-at-maturity for their respec- descending from 10 to 100 m depth; (2) southwest Los tive species (yellow snapper: 32.6 cm TL, leopard Islotes: a narrow, shallow (5 to 20 m), high-relief reef grouper: 34.0 cm TL; Erisman et al. 2008, Piñon et al. crest extending approximately 400 m to the south- 2009). Neither species exhibits external sexual di - west (terminating at the ‘Bajito’, a shallow rocky pin- morphism, thus we were unable to determine the sex nacle immediately outside of the southwest reserve of tagged individuals in this study. Once captured, boundary); and (3) east Los Islotes: a deeper (10 to fish were anesthetized in 100 mg l−1 tricaine me - 30 m) flat of unconsolidated seafloor extending 150 m thanesulfonate (MS-222) for 3 to 5 min. Transmitters southeast rapidly descending beyond a depth of 50 m, were implanted into the peritoneum through a small and a small underwater cave passing in a north- incision, which was then closed by 2 interrupted south direction through the easternmost islet of Los absorbable sutures (Ethicon Chromic-Gut, Johnson & Islotes. The bathymetry of Los Islotes was mapped Johnson). Each fish was also externally tagged in the over a 1 × 1 m grid using multibeam sonar (W. Hey- dorsal musculature with a plastic dart tag (FT-1-94, man unpubl. data). Floy Tag & Mfg) bearing a unique identification number. Fish were then transferred to a fresh seawa- ter bath and allowed to fully recover from anesthesia Tagging before release at the site of their capture.

Between August 2010 and June 2011, 22 yellow snapper and 16 leopard grouper were captured Acoustic monitoring within the reserve using baited hook and line, meas- ured, and surgically implanted with small acoustic To monitor fish movements in relation to the Los transmitters (V13-1L, Vemco; 36 × 13 mm, 147 dB). Islotes Reserve, an array of 7 omni-directional under- An additional 10 fish from each species were simi- water acoustic receivers (VR2W, Vemco) was de - larly captured and implanted with pressure-sensing ployed on subsurface moorings throughout the re - acoustic transmitters (V13P-1L, Vemco; 45 × 13 mm, serve area (Receivers 1 to 7 in Fig. 1). To monitor 150 dB, ±2.5 m accuracy, 0.22 m resolution). Both movements between Los Islotes and other nearby types of transmitters were programmed to emit a rocky reefs, 4 additional receivers were positioned at 194 Mar Ecol Prog Ser 501: 191–206, 2014

prominent locations outside of the reserve. Three Plymouth). Overall site fidelity of fish to Los Islotes receivers were deployed at the northern tip of Espir- was calculated as the proportion of days tagged indi- itu Santo Island, approximately 500 m from the viduals were present within the array since date of reserve boundary (Fig. 1), while the fourth (Receiver tagging. An individual was considered to be present 8 in Fig. 1) was positioned at the Bahito, a rocky pin- within the array if it was detected at least twice in a nacle immediately adjacent to the southwest reserve 24 hr period. Fish were assigned to 3 groups based on boun dary. Fish movements to Marisla Sea mount, a their degree of site fidelity, or the proportion of days deep rocky pinnacle and aggregation site located at liberty that a fish was present in the reserve, with 15 km NE of Los Islotes, were also periodically moni- groups being arbitrarily defined as low (<33%), tored (12 June 2011 to 23 October 2012) by an moderate (33−66%), or high (>66%). To determine acoustic receiver deployed at this site as part of an whether there were any distinct behavioral modes in unrelated study. Through range testing, the mean patterns of presence at Los Islotes, a hierarchical 50% detection range of receivers at Los Islotes was cluster analysis was applied to a Simple Matching estima ted to be approximately 180 m (range: 16 to Similarity matrix generated from the presence-ab- 327 m), providing acoustic coverage for approxi- sence of individual fish on each day of the study mately 60% of the designated reserve area. Though period. Because fish were tagged over the span of detection ranges were highly variable among sites, several months, presence-absence data used in clus- poor surface conditions (Beaufort 4 to 5) during range ter analyses were truncated to begin on the earliest testing suggest these are highly conservative esti- date on which the majority of individuals had been mates of receiver performance. Visual observations tagged (snapper, n = 22; grouper, n = 25). Individuals of both species indicated that fish carrying transmit- tagged after this date were omitted from cluster ters programmed for a 180 s pulse interval were analyses. likely mobile enough to be detected by receivers at To determine whether site fidelity to Los Islotes least twice in a 24 h period (criteria for daily presence). varied with fish size, overall site fidelity was re - When within range of a receiver (<300 m), the gressed against fish total length. Individual patterns transmitter emission is decoded as a unique transmit- of movement in relation to Los Islotes were sporadic ter identification number, and logged with a time and through both spawning and non-spawning periods; date stamp as a detection event. Detections of pres- thus we did not use ‘streaks’ (consecutive days) of sure-sensing transmitters also contained the instan- presence-absence to examine temporal trends in taneous depth of the fish at the time of detection. presence. To identify long-term patterns in the pres- Receiver performance can be influenced by interfer- ence of tagged fish at the population level, the pro- ence from sources of biological or anthropogenic portion of tagged individuals present, or probability noise (e.g. alpheid shrimps, boat traffic) (Heupel et of detecting a randomly selected fish, was calculated al. 2006). For temporal analyses based on frequency for each day and analyzed using the Seasonal Trend of detection, detections of a stationary reference Loess (STL) procedure as described in Cleveland transmitter (V13-IL) were used to calibrate for this et al. (1990). The STL procedure decomposes time effect. A correction factor was calculated for each series data into long-term, seasonal, and remainder hour of the day by dividing the mean detection fre- components. This approach was primarily used to quency of each hour by the overall mean hourly distinguish long-term tag losses from the reserve (via detection frequency of the reference transmitter. emigration, predation or fishing mortality) and po - Adjusted detection frequencies were produced by tentially spawning-related changes in presence to dividing detection frequencies from each hourly bin Los Islotes (seasonal trends). The remainder compo- by their respective correction factor (see Payne et al. nent represented residual variation in daily fish pres- 2010). Only a single reference transmitter was used, ence remaining after the fitting of both seasonal and because it was assumed that receiving conditions long-term trends. The STL cannot be used to exam- were consistent among receivers in the array. ine annual patterns in time series shorter than 2 yr, thus only snapper tagged in 2010 (n = 25) were con- sidered in this analysis. Diel patterns of movement in Data analysis relation to Los Islotes as a whole were examined by pooling detection data for all receivers and all fish Analyses of telemetry data were performed using R into hourly bins (following standardization of detec- v. 2.15.0 (R Development Core Team 2012), Statistica tion frequencies described above), applying a Fast v. 10.0 (Statsoft), and PRIMER v. 6.1.7 (PRIMER-E, Fourier Transform to these frequencies, and plotting TinHan et al.: Snapper and grouper movement patterns 195

a spectrogram. Dominant peaks emerging in the the total abundance of yellow snapper and leopard power spectra indicated the interval of cyclical pat- grouper during their respective spawning periods. terns in presence at Los Islotes. Time-series of detec- Dives were made in the late afternoon and dusk tions were used to identify the movements of all hours, and divers recorded observations of direct tagged fish among receivers at the study site. A hori- (gamete release) or indirect (courtship or chasing zontal movement was defined as any detection of a behaviors, color change, swollen abdomens) indica- single fish at a different receiver than that of the pre- tions of spawning (detailed in Colin et al. 2003). vious detection. Depth data from individuals carrying Dives were made to depths between 5 and 20 m, and depth transmitters were examined to identify vertical visibility ranged from 3 to 15 m. movements, based on the changes in depth between subsequent detections of individuals. Mean hourly rates of movement (ROM) were calculated by indi- RESULTS vidual for both the horizontal and vertical axes. Hor- izontal ROM was calculated as the number of move- Site fidelity ments per detection in hourly bins, while vertical ROM was determined by taking the mean of the A total of 32 yellow snapper and 26 leopard absolute differences in depth between consecutive grouper were tagged between August 2010 and June depth detections in each hour. 2011. In 2010, 25 snapper were tagged, though only Site attachment was estimated by calculating the 3 grouper were tagged at this time due to the diffi- evenness of detections among receivers for each culty of capturing individuals of this species outside individual, using a diversity index of space use of the peak spawning months of April and May. The adapted from Pielou’s evenness index (Pielou 1966): remaining 7 snapper and 23 grouper were tagged in April and May 2011 (Fig. 2a,b). Two snapper were R omitted from the data set as they were presumed to –ln⎡ρρ⎤ ∑ ⎣ ii()⎦ have died within the array shortly after tagging. A Spatial evenness = i=1 ln()R third snapper (Larg22) was caught by fishers at the Bajito in March 2011 and was thus omitted from ρ where R is the number of receivers in the array and i analyses of site fidelity or long-term temporal trends. is the relative number of detections at the i th receiver. Grouper exhibited the greatest overall site fidelity to Spatial evenness (se) values can range from zero to Los Islotes, being detected on 64 ± 30% of days at one, with values approaching one indicating even liberty (mean ± SD). Thirteen grouper (50% of indi- use of all areas. Spatial evenness for each species viduals) exhibited high site fidelity, 8 (30%) showed was calculated by taking the mean of the spatial moderate site fidelity and 5 (19%) showed low site evenness values calculated for each tagged individ- fidelity. Overall, snapper displayed moderate site ual. The relationship between spatial evenness and fidelity (49 ± 30% of days at liberty). Eight snapper fish size was examined in a linear regression of spa- (25% of individuals) exhibited high site fidelity, 12 tial evenness values calculated for individual fish and (38%) showed moderate site fidelity and 9 (28%) fish total length. Overall patterns in space use within showed low site fidelity. Los Islotes Reserve were compared by determining Cluster analysis of patterns in presence-absence of the proportion of detections occurring at each re - fish at Los Islotes revealed 3 significant groups for ceiver for individual fish, then calculating the mean yellow snapper, and 2 groups for leopard grou per for each receiver. (Fig. 2c,d). These distinct groups were considered to represent individuals exhibiting resident, transient or mixed patterns of presence to the reserve. For yel- Spawning observations low snapper, the largest group of individuals (~30%) exhibited moderate to high levels of site fidelity dur- Diver observations were made from 21 to 23 April ing non-spawning months, but sustained periods of and 28 to 30 September 2012, during the peak absence during the months of spawning. Approxi- months of spawning for yellow snapper and leopard mately 80% of grouper displayed continuous pat- grouper, respectively, to determine the presence and terns of residence at the reserve, and though the extent of spawning or aggregation behavior at Los remaining individuals showed low overall site fide- Islotes. Teams of divers swam the circumference of lity, these individuals also exhibited high site fidelity the reserve area to monitor spawning and estimate over short time scales (i.e. weeks). A significant rela- 196 Mar Ecol Prog Ser 501: 191–206, 2014

Fig. 2. Lutjanus argentiventris, Mycteroperca rosacea. (a,b) Detection plots of individual daily presence-absence at Los Islotes Reserve and (c,d) non-metric multi-dimensional scaling (nMDS) plots of individual patterns of residence at the reserve for (a,c) yellow snapper and (b,d) leopard grouper. Short, vertical coloured lines in detection plots indicate days on which fish were de- tected at Los Islotes; black triangles represent detections at Marisla Seamount. Shaded areas (a,b) indicate spawning seasons for the species shown; rows are arranged in order of similarity between individuals. Short, vertical gray lines in (a) indicate in- dividuals not included in nMDS analyses. Overlaid contour lines (in c,d) from hierarchical cluster analysis represent groups showing different residency patterns (corresponding group membership represented by red, blue, and green lines in detection plots) tionship was found between site fidelity and body and seasonal fluctuations corresponding with the 2 size for grouper (F1,23 = 8.637, r = 0.273, p = 0.007; spawning season (May to September) (Fig. 4). The Fig. 3). However, the significance and strength of this proportion of tagged snapper present at Los Islotes 2 relationship were greatly improved (F1,22 = 21.71, r = declined steadily over a 2 yr period at a mean rate of 0.49, p = 0.0001) when grouper Mros17 was omitted 35.5% per year (Fig. 4c). Cyclical changes in snapper from analysis as an outlier (Fig. 3). No such relation- presence to Los Islotes were also identified, with the ship between size and site fidelity was observed for probability of detecting a tagged fish decreasing by 2 snapper (F1,27 = 1.934, r = 0.06, p = 0.175). approximately 30% during the months of peak spawning. In each year, the proba bility of detection increased rapidly between the months of November Temporal patterns and January, shortly after the end of the reported spawning season (Fig. 4b). However, there remained STL decomposition of the daily proportion of substantial daily fluctuations in snapper presence tagged yellow snapper present over the course of the (approx. ±10%) that could not be explained by the study revealed long-term declines in fish presence identified seasonal or long-term trends (Fig. 4d). As TinHan et al.: Snapper and grouper movement patterns 197

activity, also peaked during the crepuscular hours for both species, although these increases did not corre- spond to any particular area of the reserve. Snapper activity levels became more variable during the night, though there was no clear overall increase in activity during this period (Fig. 6a). Grouper main- tained an increased level of activity throughout the day before becoming more quiescent at night (Fig. 6b). Examination of vertical activity patterns for individuals tagged with depth-sensing transmitters (yellow snapper: n = 9; leopard grouper: n = 10) revealed temporal patterns similar to those in the horizontal axis, with both species increasing in activ- ity during the crepuscular hours. However, snapper sustained their levels of vertical activity throughout the night, while grouper showed an overall pattern of increased crepuscular and diurnal vertical activity (Fig. 6c,d). Between June 2011 and October 2012, 3 yellow snapper (Larg12, Larg20, Larg26) and 2 leopard grouper (Mros07, Mros11) were detected by acoustic receivers at the Marisla Seamount, approximately 15 km to the northeast of Los Islotes (Fig. 1). Marisla Seamount comprises 3 submerged rocky pinnacles Fig. 3. Lutjanus argentiventris, Mycteroperca rosacea. that rise from a depth of 1000 m to within 15 m of the Natural-log relationship between fish length and site fidelity of (a) yellow snapper and (b) leopard grouper at the Los Is- surface. All 3 snapper were first detected at Mar - lotes Reserve. Reported regression statistics for leopard isla Seamount between July and September — the grouper omit an outlier (Mros17) shown as ‘×’ in lower right months of peak spawning for yellow snapper. Two corner of plot (b) snapper (Larg12, Larg20) remained at this site for extended periods (>6 mo), while the third (Larg26) data for leopard grouper were limited to a period of was only detected on 9 non-consecutive days over 1 yr, STL analysis could not be used to identify the subsequent 8 mo. Of these 3, only 1 snapper annual cycles for this species. However, similar long- (Larg12) returned to Los Islotes from Marisla Sea - term declines (72%) in fish presence to Los Islotes mount. In contrast, leopard grouper movements to were also identified for this species (data not shown). Marisla Seamount did not coincide with the spawn- No size difference was found between yellow snap- ing season and tended to be brief, with all fish travel- per that disappeared from the array and those re - ing to Marisla Seamount and back to Los Islotes maining at Los Islotes at the end of the study period within 24 h. Transit times between the 2 sites for yel- (September 2012) (t-test, t = −0.243, p = 0.980). How- low snapper ranged from 3 h to 8 d, while the shortest ever, 18 grouper were no longer detected at Los transit time for leopard grouper was 102 min at an Islotes at the end of the study, and these were signif- estimated swimming speed of 2.5 m s−1 or 2.9 body icantly smaller than those that remained in the array lengths s−1. (t-test, t = −4.377, p = 0.0002). Of the 18 ind. that dis- appeared from the array, 7 (36%) showed declines in site fidelity in the weeks prior to their disappearance Spawning observations (Fig. 2b; e.g. Mros08). Spectral analysis of adjusted hourly detection fre- We confirmed the presence of at least 1, and pos - quencies revealed 2 dominant peaks at periods of 12 sibly 2, small spawning aggregations of leopard and 24 h for yellow snapper and a single large peak grouper at Los Islotes during underwater surveys at 24 h for leopard grouper (Fig. 5). Detection fre- conducted in the afternoon (13:00 to 20:00 h) on 21 to quency did not vary significantly with tide height, 23 April 2012. Throughout this period, we observed a tidal phase, or lunar phase. The frequency of move- group of approx. 80 adult leopard groupers loosely ments made among receivers, or level of horizontal aggregated off the east point of the island. Fish were 198 Mar Ecol Prog Ser 501: 191–206, 2014

Fig. 4. Lutjanus argentiventris. Time series of proportion of tagged yellow snapper present at Los Islotes (a) over the duration of the study, and decomposed into (b) seasonal, (c) trend and (d) remainder components. Gray shading represents spawning seasons. The sign of y-values in (b,d) represents changes in relation to the trend component in (c)

present both in the water column above the thermocline (depth range = 5 to 10 m) or on the reef itself (depth range = 10 to 20 m). Indirect evidence of spawning via the presence of nu me - rous females with enlar ged (swollen) abdomens was observed on all dives. We also observed direct evidence of spawning during visual surveys con- ducted the hour before sunset (i.e. 18:45 to 19:45 h) on 22 and 23 April at the easternmost point of the island, on both the north and south side of the point. During these dives, we re - corded nearly the entire repertoire of behaviors associated with courtship and spawning in leopard grouper: Fig. 5. Lutjanus argentiventris, Mycteroperca rosacea. Spectral analyses of courtship color change, following, Fast Fourier Transformed (FFT) hourly detection frequencies at Los Islotes courtship chase, mobbing, female for (a) yellow snapper and (b) leopard grouper darting, rubbing, bumping, lateral dis- TinHan et al.: Snapper and grouper movement patterns 199

domens. However, court ship or spawning was not ob served. Several small groups of 10 to 20 adult leop- ard grouper were documented on reefs on the north side of the island, indicating the presence of ap - prox. 180 to 200 adult leopard grouper at Los Islotes during the April spawning period. Total abundance of yellow snapper at Los Islotes was estimated to be no more than 100 ind. in September 2012. We observed an aggregation of 50 to 100 adult yellow snapper within the cave at Los Islotes during all sur- vey months; however, we found no direct evidence that this aggregation represented a spaw ning aggre- gation. Though small groups of snapper (<5 ind.) were observed loosely swimming together, there were no definitive signs of aggregation, and no ob - servations were made that indicated spawning activ- ity in yellow snapper at this site.

Space use

Yellow snapper exhibited significantly lower spa- tial evenness index values (0.60 ± 0.17, mean ± SD) than leopard grouper (0.71 ± 0.14) (Kruskal-Wallis χ2 = 5.13, df = 1, p = 0.024). Though grouper were found to make greater use of the available reserve area, both species were most frequently detected by receivers along the east point and northern side of the islet — locations composed primarily of steep boulder fields and wall habitat fringed by sand bot- tom habitat (Table 1). There was no relationship between spatial evenness and body length for either species (snapper: r2 = 0.009, p = 0.394; grouper: r2 = 0.063, p = 0.119). However, both species were found to occupy a similar distribution of depths (F = 3.40, Fig. 6. Lutjanus argentiventris, Mycteroperca rosacea. Ac- df = 1, p = 0.085). Though no correlation was found tivity levels by time of day (hourly): rate of (a,b) horizontal between body length and mean depth for snapper movement and (c,d) vertical movement for yellow snapper (r2 = 0.06, p = 0.247), a significant positive relation- and leopard grouper. Grey shaded areas indicate nighttime. Note differences in scale along vertical axes Table 1. Lutjanus argentiventris, Mycteroperca rosacea. Re - la tive proportion of detections calculated across individuals play, gaping, burst rise, head shake, and spawning at each acoustic receiver at Los Islotes Reserve rush (Erisman & Allen 2006, Erisman et al. 2007). Courtship occurred within subgroups of 3 to 6 ind. Station Yellow snapper Leopard grouper that, led by a gravid female with a swollen abdomen, Mean ± SD Mean ± SD would engage in bouts of chasing behavior and mob- bing of the female, followed by rapid ascents off the 1 0.205 ± 0.212 0.160 ± 0.147 2 0.087 ± 0.153 0.143 ± 0.144 reef up into the water column. While rushing behav- 3 0.311 ± 0.215 0.174 ± 0.183 ior of subgroups was observed, ga mete release could 4 0.125 ± 0.153 0.145 ± 0.145 not be confirmed due to the poor visibility and strong 5 0.201 ± 0.209 0.090 ± 0.139 currents. A second aggregation of approx. 60 leopard 6 0.025 ± 0.027 0.058 ± 0.074 grouper was observed at the west point of the island, 7 0.014 ± 0.030 0.120 ± 0.167 8 0.029 ± 0.087 0.105 ± 0.186 which included gravid fe males with enlarged ab- 200 Mar Ecol Prog Ser 501: 191–206, 2014

ship was found for grouper (r2 = 0.47, p = 0.017) Oropeza & A. Weaver unpubl. data; pers. comm. with (Fig. 7), with the largest individuals being detected at park guards and CONANP officials), and it is there- mean depths of 16 to 18 m. fore unknown whether declines in the presence of tagged fish at Los Islotes are the result of long-term emigration, natural mortality by predation, or fishing DISCUSSION mortality. Another possibility for fish disappearance is tag failure, though unpublished data from other Site fidelity studies indicate extremely low failure rates (<3%) for similar tags produced by the manufacturer (e.g. M. Overall, yellow snapper and leopard grouper A. Dance et al. unpubl., B. W. Wolfe et al. unpubl.). tagged in this study exhibited site fidelity to the Los Nevertheless, estimates of site fidelity for fish at this Islotes Reserve. However, species differed in their location should be viewed as conservative. patterns of presence to the reserve, as well as their Though the majority of grouper exhibited moder- overall degrees of site fidelity. The majority of leop- ate to high levels of site fidelity to Los Islotes, periods ard grouper (80%) were detected on at least 50% of of absence were often prolonged, and many individ- days, whereas less than half (48%) of yellow snapper uals did not return to the reserve prior to completion were detected on at least 50% of days (Fig. 3). Exam- of the study (Fig. 2b). There was considerable varia- ination of long-term changes in site fidelity for both tion in site fidelity among individuals, and a small species revealed an overall decline in the presence of percentage (16%) exhibited low site fidelity to the tagged fish at Los Islotes over the course of the study. reserve. It was found that much of this variation was Over a 2 yr period (789 d), approximately 70% of explained by body size, with smaller individuals snapper were no longer detected at Los Islotes, while exhibiting lower site fidelity to the reserve than a similar proportion of leopard grouper were unde- larger grouper (Fig. 3b). At the end of the study, tected in the reserve after 515 d. Similar patterns the majority of tagged grouper had disappeared have been reported in previous studies of reef fish from Los Islotes, and these fish were also found to movements, although residence times are variable be significantly smaller than those remaining in the among studies. Topping & Szedlmayer (2011) ob- reserve. Ontogenetic habitat shifts have been de- served comparable declines in red snapper Lutjanus scribed in at least one other species of Mycterop- campechanus presence on reefs in the Gulf of Mex- erca (gag grouper; M. microlepis), where individuals ico, with 20 to 25% of tagged fish remaining at study move from shallow to deeper reef habitats with sites after a period of ~750 d, while a study of snapper maturity, presumably to exploit larger or more abun- and grouper in the Keys reported residence dant prey items found in habitats presenting a times of <1 yr (Lindholm et al. 2005). While Los greater predation risk to smaller individuals (Bullock Islotes is a no-take marine reserve, enforcement is & Smith 1991, Dahlgren & Eggleston 2000). Shapiro sporadic due to the re mote location of the reserve in et al. (1994) also observed an increase in home range relation to the nearest coastal city (La Paz; distance = size with body size for red hind gutta- 50 km). Illegal fishing at night by hookah divers is tus, as well as the disappearance of approximately known to occur at this site (B. Erisman, O. Aburto- 44% of tagged individuals from a small fore reef in Puerto Rico over a 5 mo period. However, if Los Islotes was abandoned by maturing individuals in favor of habitat more suitable to larger fish, this does not explain why large grouper are encountered in any substantial number at Los Islotes. In fact, fish tagged in the present study were at least as large (53 cm mean SL) as mature leopard grouper captured in spawning aggregations near Loreto, Mexico (47 cm mean SL) (Erisman et al. 2007). Illegal fishing has been reported to occur at Los Islotes at night, when enforcement levels are lowest, and the abrupt disap- pearance of some individuals (Fig 2b; e.g. Mros03) from the array may have resulted from their capture Fig. 7. Mycteroperca rosacea. Relationship between fish to- tal length and mean depth for leopard grouper at Los Islotes by fishers within the reserve. However, some individ- Reserve (r2 = 0.47, p = 0.017) uals lost from the array showed declines in site TinHan et al.: Snapper and grouper movement patterns 201

fidelity in the weeks prior to their disappearance to transit this area undetected while making directed (Fig. 2b; e.g. Mros08), and it is unlikely that these movements. It is also plausible that fish venturing particular losses resulted from illegal fishing within away from Los Islotes simply did not orient to the the Los Islotes Reserve. nearest available habitat, and instead traveled along Though the fate of leopard grouper after leaving routes outside of the detection range of these the reserve remains unknown, these declines in site receivers until reaching more preferable sites further fidelity may represent periods in which grouper along Espiritu Santo. began exploring new habitats before eventually emi- Site fidelity in yellow snapper was also variable grating from Los Islotes. We saw no indication that among individuals, but size could not explain the individuals departed from, or returned to, the array variation in site fidelity in this species (Fig. 3a). Clus- via any particular part of the reserve, but we suspect ter analysis of patterns in presence-absence also pro- that fish might have been capable of traversing the duced groups indicating fish from both species range of one or more receivers between transmitter exhibited differential patterns in movement and site pulses, obscuring any directional pattern of emigra- fidelity in relation to Los Islotes. Such patterns have tion. If indeed these disappearances were the result also been reported for other species inhabiting coas - of emigration, there are 2 potential explanations for tal waters, and it is presumed that multiple patterns the size difference seen. Firstly, smaller grouper may in movement or space use may provide individuals have been competitively excluded from Los Islotes with a means of minimizing intraspecific competition by larger individuals and were thus forced to visit for resources such as food or space. For example, a sites elsewhere in order to satisfy their energetic study of movements of juvenile Atlantic sharpnose requirements. Secondly, the abundance of sea lions shark Rhizoprionodon terraenovae identified multi- Zalophus californianus at Los Islotes (~400 ind.; ple peaks in the distribution of site fidelity and home CONANP 2011) might have posed enough of a dis- range values (Carlson et al. 2008). Similar results turbance or predation risk to smaller grouper to merit were obtained from another study of sparid snapper emigration to habitats with fewer potential predators. Pagrus auratus movements in a New Zealand MPA, On several occasions, divers in the present study where 2 distinct groups of fish were identified as observed sea lions harassing both yellow snapper exhibiting either low or high degrees of site fidelity and leopard grouper, and interrupting spawning (Egli & Babcock 2004). Two possible interpretations behaviors in the latter (B. Erisman & A. Weaver pers. of the apparent differences in site fidelity seen obs.), though studies of sea lion diet have not explic- among individuals are that (1) snapper with low site itly identified these species as prey items (e.g. Gar- fidelity to Los Islotes exhibit high site fidelity to cía-Rodríguez & Aurioles-Gamboa 2004). another area, or (2) snapper with low site fidelity to Snapper were detected consistently over the Los Islotes are simply more transient than those dis- course of the study but were frequently absent for playing high degrees of site fidelity. Snapper move- short periods (Fig. 2a). The brief yet common nature ments among different activity spaces in the Gulf of absences for snapper suggests individuals utilized remain unclear, and future studies using active other nearby areas in addition to habitat at Los tracking may be successful in identifying the scale Islotes. Though it is possible that these absences and frequency of these movements, as well as the resulted from fish moving into areas of the reserve influence of environmental factors. outside of the detection range of receivers, it is unlikely fish would remain in these locations for the multi-day periods observed. Previous studies of Temporal patterns snapper movement have found fish to make repeated forays among multiple areas and habitat types sepa- A moderate seasonal component was also identi- rated by small or intermediate distances (Meyer et al. fied in patterns of presence of tagged yellow snapper 2007, Luo et al. 2009). Los Islotes is separated from at Los Islotes. The greatest proportion of tagged fish the nearest rocky reef (northern Isla Espiritu Santo) was detected between February and April, followed by waters ranging in depth from 80 to 100 m, and a by a steady decline through the summer and early distance of approx. 700 m. However, on only one fall (Fig. 4b). Yellow snapper in the Gulf of California occasion was a snapper detected by the array at have been found to spawn throughout the summer northern Espiritu Santo. Because of the relatively months (May to Sept) (Sala et al. 2003, Aburto- long pulse interval of the transmitters used in this Oropeza et al. 2009), though local fishers report that study (110 to 250 s), it is possible that fish were able spawning activity and aggregative behaviors peak in 202 Mar Ecol Prog Ser 501: 191–206, 2014

August and September. However, many fish that study made migrations over a distance of 15 km to were less likely to be detected during spawning peri- Marisla Seamount, these movements did not corre- ods did not simply leave Los Islotes for extended spond with spawning periods and were thus not be - periods, but instead made multiple trips away from lieved to be spawning migrations. We observed the the reserve throughout the spawning season (Fig 2a). synchronous disappearance of 7 grouper from Los Yellow snapper are serial spawners, and these re - Islotes at the start of the 2012 spawning season peated short-term migrations may therefore repre- (27 April to 13 May 2012) (Fig. 2b), and though the sent movements to other spawning sites in the Gulf. possibility of fishing mortality could not be elimi- In this regard, yellow snapper appear to exhibit sim- nated (landings peak between March and May; Eris- ilar spawning patterns to those identified in earlier man et al. 2010), we speculate these movements studies of lutjanid reproduction, where species tend were related to spawning. Dive surveys also pro- to spawn throughout an ex tended spawning season vided evidence of leopard grouper spawning at Los containing multiple peaks in spawning activity (Jo - Islotes during the months of April and May, suggest- hannes 1981, Grimes 1987, Collins et al. 1996). ing grouper at this site may instead represent a sub- Though the repeated seasonal declines in presence population of both resident and transient spawners. to a home range seen in the present study are charac- There is evidence that aggregating species may teristic of transient aggregation behavior, it is impor- exhibit partial migration, where a single population tant to note that the proportion of tagged fish present contains both resident and transient spawners (Jarvis at Los Islotes fluctuated by ~30% between spawning et al. 2010, Semmens et al. 2010). Partial migration in and non-spawning periods (Fig. 4b). It is unknown fish may represent a suite of strategies within a pop- what proportion of the adult population spawns each ulation whereby individuals may maximize their fit- year, so the possibility remains that some individuals ness under a combination of environmental and simply did not engage in spawning during this study. genetic factors (e.g. food availability, sex, maturation Local divers and tour operators have reported seeing or growth rates) (Jonsson & Jonsson 1993). Previous yellow snapper spawn at Los Islotes, though dive sur- studies have shown that the distribution and avail- veys performed in the present study provided little ability of resources, such as food or habitat, can evidence of spawning or aggregation behavior at this impact social structures and reproductive behaviors site. On several occasions, snapper were seen to form (e.g. Travis & Slobodchikoff 1993), and it is possible small groups and engage in chasing behaviors that that differences in habitat quality are influencing the are typically followed by courtship displays and differences in aggregation behavior seen between spawning, but these groups were often interrupted Los Islotes and sites sampled near Loreto. For exam- by rushes from sea lions that inhabit Los Islotes, and ple, Marsh et al. (2000) report decreased aggregation no direct observations of gamete release or spawning behavior and increased site fidelity in male tungara rushes were made. However, yellow snapper have frogs as the distance between ponds (breeding habi- also been observed to form large spawning aggrega- tat) increased. In contrast to coastal areas like Loreto, tions at sites elsewhere in the Gulf, such as Isla Las where fish may be able to move easily among aggre- Animas, Cabo Pulmo National Park and Marisla Sea - gation sites via contiguous coastal habitat, Los Islotes mount (B. Erisman pers. obs.). Interestingly, 3 tagged may instead represent one of several insular, patchily snapper (Larg12, Larg20, Larg26) made a synchro- distributed aggregation sites, where it is adaptive for nized migration from Los Islotes at the end of July individuals to form resident spawning aggregations. 2011 and were detected shortly afterwards on acous - Spectral analyses revealed a diel pattern in the tic receivers deployed at Marisla Seamount. Because hourly detection frequencies of a stationary refer- these migrations coincided with the beginning of ence transmitter, indicating that diel rhythms in what are thought to be the months of peak spawning background noise interfered, to some degree, with for yellow snapper, we speculate that they may have detection efficiencies of receivers. These results are been related to reproduction, though further study is broadly consistent with the findings of Payne et al. needed to understand the nature and degree of con- (2010), in which predictable diel fluctuations in nectivity between these 2 sites. receiver performance were identified and attributed Erisman et al. (2007) reported seasonal shifts in to interference from external noise sources (e.g. leopard grouper abundance among a number of sites soniferous fishes or invertebrates). However, in con- near Loreto, Mexico, and suggested that individuals trast to the opposing diel patterns shown in Payne et migrate over significant distances to specific sites in al. (2010), the similarity of patterns seen between cor- order to spawn. Though 2 grouper in the present rected and uncorrected detection frequencies at Los TinHan et al.: Snapper and grouper movement patterns 203

Islotes suggest environmental noise did not funda- Vázquez et al. 2008). Hobson (1965) observed that mentally alter observations of fish movement at this leopard grouper typically utilized sand bottom habi- site. Spectrograms of hourly detection frequencies at tats to ambush schooling baitfish near the surface, Los Islotes revealed a dominant 24 h period in activ- and larger grouper in the present study were de - ity, indicating clear patterns of diel movement for tected most commonly at depths >15 m, coinciding both species. A secondary peak in activity occurring with sand bottom/rock edge habitat surrounding Los at 12 h intervals was observed in snapper, but pat- Islotes. Larger grouper might simply occupy deeper terns of fish presence to the reserve were not found waters irrespective of habitat type, but another possi- to correspond with tidal height or phase, so this peak bility is that larger, more piscivorous individuals are most likely represents the 12 h interval between cre- seeking out deeper sand bottom habitats to better puscular periods of activity in this species (Fig. 5a). prey on small schooling fish as described in Hobson Examination of levels of horizontal and vertical (1965). movement in snapper also revealed peaks in activity Yellow snapper exhibited spatial evenness values during the crepuscular periods (Fig. 6a,c). This is lower than those calculated for leopard grouper, indi- consistent with previous observations of yellow snap- cating a greater degree of site attachment to specific per movements in the Gulf of California made by areas of the reserve. Diver observations indicate that Hobson (1965), who found snapper to be most active large numbers of snapper would congregate in a when foraging at dusk and at night, and least active small underwater cave passing in a north-south when seeking refuge in rock crevices during the day. direction through the easternmost islet of Los Islotes. Because of the low probability of detecting fish when A large proportion of detections at eastern Los Islotes sheltered in rock crevices, it is likely that the 24 h may be attributed to the aggregation of tagged snap- interval between periods of presence and absence to per within this feature, as 2 out of the 3 receivers hav- the reserve observed in spectral analyses for both ing the greatest number of yellow snapper detections species are the result of diurnal refuging behavior, had an unimpeded ‘line-of-sight’ into this cave. rather than diurnal movements away from the re - Detections of yellow snapper along the western point serve area. of Los Islotes were infrequent. Though range tests While no substantive secondary peak was ob - indicated receiver performance to be poor in this served in spectral analysis of leopard grouper move- area, detection frequencies of grouper at this site ments in relation to Los Islotes (Fig. 5b), small in - were comparable to those in other parts of the re - creases in activity within the reserve were observed serve, suggesting receiver performance might have during crepuscular periods (Fig. 6b,d). This indicates been adequate for detection of fish movements that while grouper become more active during the through this area. Leopard grouper were detected crepuscular hours, they show no shift in their overall more evenly across all receivers in the array, but sim- presence at the reserve during these periods. These ilar to snapper, detection frequencies were greatest periods of activity are most likely foraging-related as along the easternmost point and north side of Los leopard grouper are thought to be primarily crepus- Islotes. However, grouper were not observed to con- cular foragers (Hobson 1965). gregate in the eastern islet cave, indicating detec- tions at these receivers were of fish moving along deeper boulder and rocky wall habitats present in Space use this area. These results are consistent with diver observations of fish habitat associations by Aburto- Crepuscular and nocturnal foraging behaviors are Oropeza & Balart (2001), in which the highest densi- common among predatory fishes, where low light ties of both species were found in association with levels probably allow predators to increase their like- the wall habitats along the eastern and western lihood of capturing small, fast-moving prey items points of the islets. (Hobson et al. 1981, Helfman 1986, Parrish 1992), such as schools of flat-iron herring Harengula thris- sina that frequent Los Islotes. Flat-iron herring are a Implications for management principal prey item of adult yellow snapper and leop- ard grouper, but both species also show ontogenetic The objectives of the establishment of Los Islotes shifts in diet: Juveniles primarily consume benthic Marine Reserve as a part of the larger Espiritu Santo invertebrates (and fish eggs, for snapper), while Archipelago National Park (ESANP) were primarily adults are increasingly piscivorous (Hobson 1965, to conserve biodiversity, ecosystem function, and 204 Mar Ecol Prog Ser 501: 191–206, 2014

genetic diversity among populations, and to promote exhibit similar patterns of site fidelity and seasonal sustainable use of commercial species in the region. movement in relation to these habitats. Since there exist no studies assessing such endpoints Frisk et al. (2013) emphasized the importance of as biodiversity, abundance, genetic diversity or fish- adult movement — in contrast to larval dispersal — in eries contribution prior to Los Islotes’ establishment the connectivity, structuring and recruitment of pop- as a marine reserve, it is difficult to ascertain whether ulations. To assess the success of small marine re - it has succeeded under the original objectives of the serves in meeting their objectives of maintaining ESANP. However, the findings of the present study genetic diversity or fisheries contribution will require may be used to assess the potential contribution of an understanding of the connectivity between puta- future reserve design and management approaches. tive secondary spawning sites such as Los Islotes and The moderate levels of site fidelity seen for the other aggregation sites (e.g. Marisla Seamount). majority of fish suggest that the protection afforded Nevertheless, the intraspecific behavioral differ- to individual fish by Los Islotes is limited over both ences seen for yellow snapper and leopard grouper daily and annual time scales, and reserve benefits at Los Islotes suggest that mixed management strate- might be improved if future reserves are designed to gies affording protection to the full diversity of encompass larger areas than Los Islotes. Buffer zones aggregation and movement behaviors within a popu- around such reserve boundaries may also be instru- lation may be the most appropriate approach for mental in protecting individuals that utilize habitats these species. nearing the boundaries of the reserve area, particu- larly if these may be the largest and oldest individu- Acknowledgements. We thank 3 anonymous reviewers for als as seen for leopard grouper at Los Islotes (Nemeth their help in improving the manuscript. We also thank D. et al. 2007). Furthermore, there currently exist no Hernandez, the CMBC team in La Paz, and most impor- fisheries regulations for these species in the Gulf of tantly, the entire Niparajá team for their logistic and scien- California. The demonstrated high level of vulnera- tific support in the field: Dani, Pollo, Pelón, Tito, and Zorro. We thank the La Paz CONANP office for allowance of this bility of fish at Los Islotes due to their daily and sea- project within a protected area, as well as J. Rivera, F. León, sonal movements (coincident with the times of peak Javier, Jorge, Cachas and A. Garcia and the fishers of SCPP exp loitation) across reserve boundaries indicates that Pescadores del Esterito, Baja Pirata Fleet and Baja Expedi- management efforts should be complemented by tions for assistance capturing fish for tagging. Thank you to J. Ketchum for kindly sharing detection data from acoustic the incorporation of traditional regulations (e.g. receivers at Marisla Seamount. We are grateful to B. Allen catch, size limits, time/area closures). Observations for comments on an early draft of the manuscript. B. Allen, J. of spaw ning at this site did not compare with the Archie and B. Wolfe provided in valuable assistance with magnitude of snapper and grouper aggregations statistical analyses. Thanks to W. Heyman and S. Kobara for reported at other locations in the Gulf, and due to the providing bathymetry maps of Los Islotes. This project was funded by the Walton Family Foundation, the David and lack of spawning observations for yellow snapper, Lucile Packard Foundation, along with support from the this is likely not an important spawning site for this California State University (CSU) Long Beach College of species. Samoilys (1997) and Zeller (1998) found Natural Sciences and Mathematics, CSU COAST, the AAUS coral trout leopardus to aggregate less Kevin Gurr Scholarship, and the SCTC Marine Biology Foundation. All work was carried out in accordance with predictably and in smaller numbers at a number of use protocols of the University of California, San sites on the and suggested these Diego (protocol S09323) and conducted under permit from were secondary spawning sites. Los Islotes may be the National Commissions of Fisheries and Aquaculture an example of one such site in the Gulf of California, (CONAPESCA) (permit #DGOPA-05356-140710-3457). though an alternate possibility is that heavy fishing pressure at this site prior to its closure (or as a result LITERATURE CITED of poaching) has reduced the number of participants and level of spawning activity seen at this site. It is Aburto-Oropeza O, Balart E (2001) Community structure of important that future studies examine the relation- reef fish in several habitats of rocky reef in the Gulf of California. Mar Ecol 22:283−305 ship between habitat quality and aggregation behav- Aburto-Oropeza O, Dominguez-Guerrero I, Cota-Nieto J, ior with the aim of understanding determinant char- Plomozo-Lugo T (2009) Recruitment and ontogenetic acteristics of which sites become primary spawning habitat shifts of the yellow snapper (Lutjanus argentiven- sites. Though boulder, wall and sand bottom/rock tris) in the Gulf of California. Mar Biol 156:2461−2472 Allen GR (1985) FAO species catalogue, Vol 6. Snappers edge habitats appear to be the most important habi- of the world. An annotated and illustrated catalogue of tats at Los Islotes, it is also crucial to determine lutjanid species known to date. FAO Fish Synop No. whether fish encountered in other areas of the Gulf 125(6) TinHan et al.: Snapper and grouper movement patterns 205

Bullock LH, Smith GB (1991) Memoirs of the Hourglass reserves work and does reserve size matter? Ecol Appl cruises. Seabasses (Pisces: ). Florida Marine 13: S117−S137 Research Institute, St. Petersburg, FL Heemstra PC, Randall JE (1993) FAO species catalogue, Vol Carlson JK, Heupel MR, Bethea DM, Hollensead LD (2008) 16. Groupers of the world. An annotated and illustrated Coastal habitat use and residency of juvenile Atlantic catalogue of the grouper, rockcod, hind, coral grouper sharpnose sharks (Rhizoprionodon terraenovae). Estuar and lyretail species known to date. FAO Fish Synop no. Coast 31: 931−940 125(16) Cleveland RB, Cleveland WS, McRae JE, Terpenning I Helfman GS (1986) Fish behaviour by day, night and twi- (1990) STL: a seasonal-trend decomposition procedure light. In: Pitcher TJ (ed) The behaviour of teleost fishes. based on loess. J Off Stat 6: 3−73 Croom Helm, London, p 366−387 Colin PL, Sadovy YJ, Domeier ML (2003) Manual for the Heupel MR, Semmens JM, Hobday AJ (2006) Automated study and conservation of reef fish spawning aggrega- acoustic tracking of aquatic : scales, design and tions. SCRFA (Spec Publ), San Diego, CA deployment of listening station arrays. Mar Freshw Res Collins LA, Johnson AG, Keim CP (1996) Spawning and 57: 1−13 annual fecundity of the red snapper (Lutjanus cam- Hobson ES (1965) Diurnal-nocturnal activity of some inshore pechanus) from the northeastern . In: fishes in the Gulf of California. Copeia 1965:291−302 Arreguin-Sánchez F, Munro JL, Balgos MC, Pauly D Hobson ES, McFarland WN, Chess JR (1981) Crepuscular (eds) Biology, fisheries and culture of tropical groupers and nocturnal activities of Californian nearshore fishes, and snappers. Proc 48th Int Cent Living Aquat Res Man- with consideration of their scotopic visual pigments and age, Manila, Phillipines, p 174−188 the photic environment. Fish Bull 79:1−30 CONANP (Comisíon Nacional de Áreas Naturales Protegi- Holt RD (1984) Spatial heterogeneity, indirect interactions, das) (2011) Programa de manejo del Parque Nacional and the coexistence of prey species. Am Nat 124: exclusivamente la zona marina del Archipiélago de 377−406 Espíritu Santo. CONANP, México City Huntsman G (1996) IUCN 2013. IUCN Red List of threat- Dahlgren CP, Eggleston DB (2000) Ecological processes ened species v. 2013.2. www.iucnredlist.org (accessed 04 underlying ontogenetic habitat shifts in a coral reef fish. July 2013) Ecology 81: 2227−2240 Jarvis ET, Linardich C, Valle CF (2010) Spawning-related Domeier ML, Colin PL (1997) Tropical reef fish spawning movements of barred sand bass, Paralabrax nebulifer, in aggregations: defined and reviewed. Bull Mar Sci 60: southern California: interpretations from two decades of 698−726 historical tag and recapture data. Bull South Calif Acad Egli DP, Babcock RC (2004) Ultrasonic tracking reveals mul- Sci 109: 123−143 tiple behavioural modes of snapper (Pagrus auratus) in a Johannes RE (1981) Words of the lagoon: fishing and marine temperate no-take marine reserve. ICES J Mar Sci 61: lore in the Palau district of Micronesia. University of 1137−1143 California Press, Berkeley, CA Erisman BE, Allen LG (2006) Reproductive behavior of a Jonsson B, Jonsson N (1993) Partial migration: niche shift temperate serranid fish, Paralabrax clathratus (Girard), versus sexual maturation in fishes. Rev Fish Biol Fish 3: from Santa Catalina Island, California, USA. J Fish Biol 348−365 68: 157−184 Lindholm J, Kaufman L, Miller S, Wagschal A, Newville M Erisman BE, Buckhorn ML, Hastings PA (2007) Spawning (2005) Movement of yellowtail snapper (Ocyurus chrysu- patterns in the leopard grouper, Mycteroperca rosacea, rus Block 1790) and black grouper (Mycteroperca bonaci in comparison with other aggregating groupers. Mar Biol Poey 1860) in the northern Florida Keys National Marine 151: 1849−1861 Sanctuary as determined by acoustic telemetry. Marine Erisman BE, Rosales-Casián JA, Hastings PA (2008) Evi- Sanctuaries Conservation Series, Silver Spring, MD dence of gonochorism in a grouper, Mycteroperca rosa - Luo J, Serafy JE, Sponaugle S, Teare PB, Kieckbush D (2009) cea, from the Gulf of California, Mexico. Environ Biol Movement of grey snapper Lutjanus griseus among sub- Fishes 82: 23−33 tropical seagrass, mangrove, and coral reef habitats. Mar Erisman B, Mascarenas I, Paredes G, Sadovy de Mitcheson Ecol Prog Ser 380: 255−269 Y, Aburto-Oropeza O, Hastings P (2010) Seasonal, Marsh DM, Stanley Rand A, Ryan MJ (2000) Effects of inter- annual, and long-term trends in commercial fisheries for pond distance on the breeding ecology of tungara frogs. aggregating reef fishes in the Gulf of California, Mexico. Oecologia 122: 505−513 Fish Res 106: 279−288 Meyer CG, Papastamatiou YP, Holland KN (2007) Seasonal, Fahrig L (2007) Non-optimal animal movement in human- diel, and tidal movements of green jobfish (Aprion altered landscapes. Funct Ecol 21:1003−1015 virescens, ) at remote Hawaiian atolls: impli- Frisk MG, Jordaan A, Miller TJ (2013) Moving beyond the cations for marine protected area design. Mar Biol 151: current paradigm in marine population connectivity: are 2133−2143 adults the missing link? Fish Fish, doi: 10.1111/faf.12014 Nemeth RS (2012) Ecosystem aspects of species that aggre- García-Rodríguez FJ, Aurioles-Gamboa D (2004) Spatial gate to spawn. In: Sadovy de Mitcheson Y, Colin PL (eds) and temporal variation in the diet of the California sea Reef fish spawning aggregations: biology, research and lion (Zalophus californianus) in the Gulf of California, management, Book 35. Springer, New York, NY, p 21−55 Mexico. Fish Bull 102: 47−62 Nemeth RS, Blondeau J, Herzlieb S, Kadison E (2007) Spa- Grimes CB (1987) Reproductive biology of the Lutjanidae: a tial and temporal patterns of movement and migration at review. In: Polovina JJ, Ralston S (eds) Tropical snappers spawning aggregations of red hind, Epinephelus gutta- and groupers: biology and fisheries management. West tus, in the U.S. Virgin Islands. Environ Biol Fishes 78: View Press, Boulder, CO, p 239−294 365−381 Halpern BS (2003) The impact of marine reserves: Do Parrish JK (1992) Levels of diurnal predation on a school 206 Mar Ecol Prog Ser 501: 191–206, 2014

of flat-iron herring, Harengula thrissina. Environ Biol Nassau grouper spawning aggregations in Belize: fish- Fishes 34: 257−263 ery management and conservation needs. Fisheries 26: Payne NL, Gillanders BM, Webber DM, Semmens JM (2010) 23−30 Interpreting diel activity patterns from acoustic tele - Sala E, Aburto-Oropeza O, Paredes G, Thompson G (2003) metry: the need for controls. Mar Ecol Prog Ser 419: Spawning aggregations and reproductive behavior of 295−301 reef fishes in the Gulf of California. Bull Mar Sci 72: Pielou EC (1966) The measurement of diversity in different 103−121 types of biological collections. J Theor Biol 13:131−144 Samoilys MA (1997) Periodicity of spawning aggregations of Piñon A, Amezcua F, Duncan N (2009) Reproductive cycle of coral trout Plectropomus leopardus (Pisces: Serranidae) female yellow snapper Lutjanus argentiventris (Pisces, on the northern Great Barrier Reef. Mar Ecol Prog Ser , Lutjanidae) in the SW Gulf of California: 160: 149−159 gonadic stages, spawning seasonality and length at sex- Semmens JM, Buxton CD, Forbes E, Phelan MJ (2010) ual maturity. J Appl Ichthyol 25:18−25 Spatial and temporal use of spawning aggregation sites R Development Core Team (2012) R: a language and envi- by the tropical sciaenid Protonibea diacanthus. Mar Ecol ronment for statistical computing. R Foundation for Sta- Prog Ser 403: 193−203 tistical Computing, Vienna. www.R-project.org Shapiro DY, Garcia-Moliner G, Sadovy Y (1994) Social Ramírez RM, Rodríguez CM (1990) Specific composition of system of an inshore stock of the red hind grouper, Epi- the small scale fishery capture at Isla Cerralvo, BCS, nephelus guttatus (Pisces: Serranidae). Environ Biol Mexico. Investig Mar CICIMAR 5: 137−141 Fishes 41: 415−422 Rife AN, Erisman B, Sanchez A, Aburto-Oropeza O (2012) Topping DT, Szedlmayer ST (2011) Site fidelity, residence When good intentions are not enough … insights on net- time and movements of red snapper Lutjanus campe - works of ‘paper park’ marine protected areas. Conserv chanus estimated from long-term acoustic monitoring. Lett 6:200–212 Mar Ecol Prog Ser 437:183−200 Sadovy de Mitcheson Y, Erisman B (2012) Fishery and bio- Travis SE, Slobodchikoff CN (1993) Effects of food resource logical implications of fishing spawning aggregations, distribution on the social system of Gunnison’s prairie and the social and economic importance of aggregating dog (Cynomys gunnisoni). Can J Zool 71:1186−1192 fishes. In: Sadovy de Mitcheson Y, Colin PL (eds) Reef Vázquez RI, Rodríguez J, Abitia LA, Galván F (2008) Food fish spawning aggregations: biology, research and man- habits of the yellow snapper Lutjanus argentiventris agement, Book 35. Springer, New York, NY, p 225−284 (Peters, 1869) (Percoidei: Lutjanidae) in La Paz Bay, Sagarese SR, Frisk MG (2011) Movement patterns and Mexi co. Rev Bio Mar Oceanogr 43:295−302 residence of adult winter flounder within a Long Island Zeller DC (1998) Spawning aggregations: patterns of move- estuary. Mar Coast Fish Dyn Manag Ecosyst Sci 3: ment of the coral trout Plectropomus leopardus (Ser- 295−306 ranidae) as determined by ultrasonic telemetry. Mar Ecol Sala E, Ballesteros E, Starr RM (2001) Rapid decline of Prog Ser 162: 253−263

Editorial responsibility: Nicholas Tolimieri, Submitted: July 31, 2013; Accepted: January 3, 2014 Seattle, Washington, USA Proofs received from author(s): March 13, 2014