Introduction
Vertical movements have been linked to foraging behaviour and exploration of favourable environmental conditions in several pelagic (Klimley et al., Reference Klimley, Castillo Geniz and Cabrera Mancilla1993; Musyl et al., Reference Musyl, Brill, Boggs, Curran, Kazama and Seki2003; Williams et al., Reference Valenzuela-Quiñonez, Arreguín-Sánchez, Salas-Márquez, García-De León, Garza, Román-Rodríguez and De-Anda-Montañez2015; Thygesen et al., Reference Thorstad, Rikardsen, Alp and Okland2016) and demersal fish (Seitz et al., Reference Rubio-Rodríguez, Villalobos and Nevarez-Martinez2003; Peklova et al., Reference Peklova, Hussey, Hedges, Treble and Fisk2012, Reference Peklova, Hussey, Hedges, Treble and Fisk2014). Vertical movements frequently respond to the distribution of prey, with fish usually moving towards the surface at dusk and sinking at dawn, or vice versa. On the other hand, vertical displacement can also be related to properties that vary across the water column, such as temperature or oxygen concentration, thus limiting their distribution range. As a result, fish susceptibility to a particular fishing gear varies according to water column structure and time of the day (Bone & Moore, Reference Bone and Moore2008; Sims et al., 2008; Abascal et al., Reference Abascal, Mejuto, Quintans and Ramos-Cartelle2010; Chiang et al., Reference Chiang, Musyl, Sun, Chen, Chen, Liu, Su, Yeh, Fu and Huang2011; Afonso et al., Reference Afonso, McGinty, Graca, Fontes, Inácio, Totland and Menezes2014).
Biotelemetry has increased the understanding of the patterns and drivers of these movements, through the use of electronic devices such as pop-up satellite archival tags (PSAT). These are attached externally to fish and record data on depth, temperature and light over a specified period of time. After release, as soon as the tag pops up to the sea surface, it relays data to an Argos satellite. The information recorded is useful to gain a deeper understanding that contributes to better protection and management of fish in both freshwater and marine ecosystems (Thorstad et al., Reference Soto-Mardones, Marinone and Parés-Sierra2013).
Sciaenids are commonly called drums or croakers because of the sound produced with the swim bladder (Ramcharitar et al., Reference Ramcharitar, Gannon and Popper2006). Many species are economically important because their swim bladder is deemed a valued oriental delicacy (Chao, Reference Chao and Fischer1978). The totoaba, Totoaba macdonaldi Gilbert (1890), is a large sciaenid fish (up to 2 m) that is endemic to the Gulf of California, Mexico, characterized by late sexual maturity (6–7 years) and a single reproduction event per year. The distribution of the totoaba in the eastern Gulf of California stretches from the Colorado River mouth to the Fuerte River mouth, and along the west coast from the Colorado River mouth to Conception Bay (Arvizu & Chávez, Reference Arvizu and Chávez1972; Flanagan & Hendrickson, Reference Flanagan and Hendrickson1976).
The totoaba fishery was banned in 1975 (SEMARNAT, Reference Seitz, Wilson, Norcross and Nielsen2002). This species was listed as Critically Endangered by the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES) in 1976 and by the International Union for Conservation of Nature (IUCN) in 1996 (IUCN, 2019). However, despite its current status as critically endangered, conservation policies have been ineffective (Bobadilla et al., Reference Bobadilla, Alvarez-Borrego, Avila-Foucat, Lara-Valencia and Espejel2011) to prevent illegal fishing during its reproductive season, which is fuelled by the high price of its swim bladder, with a market value of up to US$5000 kg−1 on the local black market (De Anda-Montañez et al., Reference De Anda-Montañez, García-De León, Zenteno-Savin, Balart-Paez, Méndez-Rodríguez, Bocanegra-Castillo, MArtínez-Aguilar, Campos-Dávila, Román-Rodríguez, Valenzuela-Quiñonez, Rodríguez-Jaramillo, Meza-Chávez, Ramírez-Rosas, Saldaña-Hernández, Olguín-Monroy and Martínez-Delgado2013; Valenzuela-Quiñonez et al., Reference Turnure, Able and Grothues2015; Brusca et al., Reference Brusca, Álvarez-Borrego, Hastings and Findley2017).
An evaluation of movement patterns based on historical fishery data indicates that adults migrate to the Upper Gulf of California in the winter to spawn, mainly from February to April (Arvizu & Chávez, Reference Arvizu and Chávez1972; Flanagan & Hendrickson, Reference Flanagan and Hendrickson1976; Cisneros-Mata et al., Reference Cisneros-Mata, Montemayor-López and Román-Rodríguez1995; De Anda-Montañez et al., Reference De Anda-Montañez, García-De León, Zenteno-Savin, Balart-Paez, Méndez-Rodríguez, Bocanegra-Castillo, MArtínez-Aguilar, Campos-Dávila, Román-Rodríguez, Valenzuela-Quiñonez, Rodríguez-Jaramillo, Meza-Chávez, Ramírez-Rosas, Saldaña-Hernández, Olguín-Monroy and Martínez-Delgado2013; Valenzuela-Quiñonez et al., Reference Turnure, Able and Grothues2015). It has been suggested that after spawning, from late spring to early autumn, the totoaba moves to the vicinity of Angel de la Guarda Island, where juveniles and adults apparently feed on sardines including the Pacific anchoveta (Cetengraulis mysticetus) as their main food type (Román-Rodríguez, 1990; De Anda-Montañez et al., Reference De Anda-Montañez, García-De León, Zenteno-Savin, Balart-Paez, Méndez-Rodríguez, Bocanegra-Castillo, MArtínez-Aguilar, Campos-Dávila, Román-Rodríguez, Valenzuela-Quiñonez, Rodríguez-Jaramillo, Meza-Chávez, Ramírez-Rosas, Saldaña-Hernández, Olguín-Monroy and Martínez-Delgado2013). Following the migration of sardine (Hammann et al., Reference Hammann, Baumgartner and Badan-Dangon1988), totoaba moves in late autumn and early winter to the Sonoran coast, probably another feeding ground for juveniles and adults. Subsequently, they migrate back to the Upper Gulf of California in late winter to early spring, heading towards the spawning ground (Figure 1). The general migration pattern of the totoaba is seemingly linked to seasonal environmental variability and food availability (Flanagan & Hendrickson, Reference Flanagan and Hendrickson1976; Cisneros-Mata et al., Reference Cisneros-Mata, Montemayor-López and Román-Rodríguez1995).
Fig. 1. (a) Theoretical seasonal migration route of the totoaba (Totoaba macdonaldi). Abbreviations indicate distinct zones in the life history of totoaba. J1 = first-time migration of juveniles (2 years of age); PA = post-spawning adults; J2 = pre-adults; FA = adults during their autumn migration; SA = pre-spawning adults. Modified from Cisneros-Mata et al. (Reference Cisneros-Mata, Montemayor-López and Román-Rodríguez1995). (b) Catching and tagging locations of totoaba (Totoaba macdonaldi) (stars), geographic location of satellite tags recovered (dark circles), and Upper Gulf of California and Colorado River Delta Biosphere Reserve (shaded area).
Although the horizontal movement of the totoaba has been previously addressed, there is a lack of knowledge about the underlying factors, particularly in regards to its daily vertical movements and habitat preferences across the water column. This study explored whether the daily vertical movements of the totoaba respond to environmental variability driven by temperature in the water column of the Upper Gulf of California.
Materials and methods
Study area
The Gulf of California is a semi-enclosed sea spanning about 1400 km long and 150–200 km wide. It is bordered by the Baja California Peninsula to the west and by the Mexican mainland to the east. The Upper Gulf is separated from the Lower Gulf by two large islands (Ángel de la Guarda and Tiburón) that, together with several smaller islands, are collectively known as the Midriff Islands. The Upper Gulf is shallow (<200 m), except for the basins surrounding Angel de la Guarda Island (AGI). Depths exceed 1600 m in the Ballenas-Salsipuedes Channel west of AGI. The gulf deepens to the south; the Guaymas Basin is about 1000 m deep, and depths exceed 3000 m towards the mouth (Lavin & Marinone, Reference Lavin, Marinone, Velasco-Fuentes, Sheinbaum and Ochoa2003). The Gulf is a highly dynamic and productive sea; it is fertilized by wind-induced upwelling, tidal mixing and exchange between the Gulf and the Pacific Ocean. In the eastern Gulf, upwelling events occur from December to May, driven by north-westerly winds; in the west coast, these take place from July through October and are driven by south-easterly winds. The tidal range in the Upper Gulf exceeds 7 m during spring and attains about 4 m in the Midriff Islands region. Tidal mixing induces the stirring of the water column down to 500 m, carrying cold, nutrient-rich water to the surface (Alvarez-Borrego, Reference Alvarez-Borrego and Brusca2010).
Mean sea surface temperature (SST) decreases from the south towards the interior of the gulf, reaching a minimum in the Midriff islands, and then increasing slightly towards the Upper Gulf. From winter to summer, annual SST varies between 10 and 32°C in the Upper Gulf and between 16 and 31°C in the central region. The southern region, which communicates with the tropical Pacific Ocean, has a more complex hydrographic structure due to the confluence of different water masses; besides, SST is also modified by evaporation (Soto-Mardones et al., Reference Sims, Southall, Humphries, Hays, Bradshaw, Pitchford, James, Ahmed, Brierley, Hindell, Morritt, Musyl, Righton, Shepard, Wearmouth, Wilson, Witt and Metcalfe1999).
Fieldwork
Ten totoaba individuals were caught, tagged, and released during the spawning season in the Upper Gulf of California and the Colorado River Delta Biosphere Reserve (31°18′36.30″N and 114°52′1.14″W, Figure 1). These fishes were caught using a 10-inch and 120-m long gillnet (soak times of 1–3 h). Once a fish was captured and landed on board, its eyes were covered with a wet cloth to reduce light stress. A plastic hose connected to an underwater pump was inserted into its mouth to supply a continuous flow of seawater. The total length (TL) of each individual was recorded, and a tag was affixed externally between the two dorsal fins using a stainless-steel needle (12 cm long) attached to a plastic strap (28 cm long) to insert the anchor through the dorsal musculature (Figure 2). The average time taken to tag each fish was 4 min (Table 1). After tagging, each fish was checked to ensure it had no injuries and then released.
Fig. 2. Pictures from the tagging process in the field. Satellite tag (Sea Tag-MOD) placement between the dorsal fins of Totoaba macdonaldi.
Table 1. Summary of data associated with totoaba (Totoaba macdonaldi) caught, tagged with pop-up satellite archival tags, and released in the Upper Gulf of California (N = 10)
Total length (TL in cm).
a Tags recovered with total data.
The satellite tags used were PSAT SeaTag-MODTM (27.5 cm long, 2.5 cm diameter at narrowest point; Desert Star Systems LLC, USA). These include sensors that record ambient water temperature, tag depth, 3-axis acceleration (results not shown), ambient light, and strength of the Earth´s magnetic field for location estimation. Satellite tags were set to record data at 4-minute intervals and transmit a daily summary to the Argos satellite. PSATs were also set to be released gradually throughout the year (Table 1), to obtain information about the movements of fish between two reproductive periods in successive years. All tags were labelled to facilitate being returned if found after detachment.
Some tags detached from fish earlier than planned; these surfaced and transmitted the endpoint geoposition to the Argos satellite system. Field recovery of tags was achieved using an Argos Al-1 PTT locator (Communications specialists, INC. Canada).
The environmental variables included in the analysis were daily SST data and moon phase. For SST retrievals, a MODIS Aqua, thermal-IR SST level 3, 4 km, daily, night-time product was used (https://oceancolor.gsfc.nasa.gov/cgi/l3).
Data analysis
Tagged fishes ranged in size from 107–159 cm TL (134 ± 0.14 cm, mean ± SD). Given the limited and poor-quality records obtained from the light sensor, geolocation along tracking could not be calculated from retrieved tags. Data were edited before the statistical analysis to exclude data recorded during the fish recovery period, defined as the time taken by individual fish to recover from atypical behaviour related to its capture, manipulation and release. The data for the first three days were removed for each tagged fish, following the methodology by Hoolihan et al. (Reference Hoolihan, Luo, Abascal, Campana, De Metrio, Dewar, Domeier, Howey, Lutcavage, Musyl, Neilson, Orbesen, Prince and Rooker2011).
Vertical movement data were sorted into daytime and night-time periods. Daytime is defined as daylight hours between sunrise (6:00) and sunset (18:00) (as reported in the webpage https://www.esrl.noaa.gov/gmd/grad/solcalc/).
A generalized linear model (GLM) was constructed to determine whether environmental variables influenced vertical movements. Mean daily fish depth was related to mean daily water temperature as recorded by tags (Tsensor), sea surface temperature (SST), moon phases (Moon) and month of the year. The explanatory continuous variables (Tsensor and SST) showed no collinearity (r = 0.2). Before the analysis, we applied a low-pass filter to remove high-frequency (diurnal) signals, using a Goding Filter (Goding, Reference Goding1972) with a half-power of one hour. This was the residual signal used in the GLM.
The GLM selected was:
$${\rm log}\,{\rm \lpar M}v\rpar \,= \,{\rm \beta }_0\,+ \,{\rm \beta }_1{\rm SST}\,+ \,{\rm \beta }_2\,{\rm Month}\,+ \,{\rm \beta }_3\,{\rm Tsensor\;}$$where Mv is the response variable, which is dependent on the intercept (β0), SST, Tsensor and Month. The ‘gamma’ probability distribution was selected for the response variable, as data were continuous and positive (Zuur et al., Reference Williams, Allain, Nicol, Evans, Hoyle, Dupoux, Vourey and Dubosc2009). The modelling was conducted with the lme4 package (Bates et al., Reference Bates, Maechler, Bolker and Walker2015) in R (Core Team, Reference De Anda-Montañez, García-De León, Zenteno-Savin, Balart-Paez, Méndez-Rodríguez, Bocanegra-Castillo, MArtínez-Aguilar, Campos-Dávila, Román-Rodríguez, Valenzuela-Quiñonez, Rodríguez-Jaramillo, Meza-Chávez, Ramírez-Rosas, Saldaña-Hernández, Olguín-Monroy and Martínez-Delgado2018).
Models were constructed using a forward procedure, which consisted of adding explanatory variables to the null model. The model was selected with the Akaike Information Criterion and Bayesian Information Criterion (AIC and BIC, respectively) in R (Core Team, Reference De Anda-Montañez, García-De León, Zenteno-Savin, Balart-Paez, Méndez-Rodríguez, Bocanegra-Castillo, MArtínez-Aguilar, Campos-Dávila, Román-Rodríguez, Valenzuela-Quiñonez, Rodríguez-Jaramillo, Meza-Chávez, Ramírez-Rosas, Saldaña-Hernández, Olguín-Monroy and Martínez-Delgado2018). Both AIC and BIC measure the goodness of fit and complexity of the model. The likelihood value was used for measuring the goodness of fit; the number of parameters, for measuring complexity (Zuur et al., Reference Williams, Allain, Nicol, Evans, Hoyle, Dupoux, Vourey and Dubosc2009).
Results
All PSATs were shed before the programmed time and only three were physically recovered. One tag was released at a linear distance of 90 km and five tags at 223 km from the site where fishes were initially tagged; all fishes migrated southward to the vicinity of Angel de la Guarda Island. Tagged totoabas travelled ~3 km day−1 (Figure 2). One tag was returned by a fisherman and the remaining three tags were lost (Table 1). The three tags recovered recorded information over 46, 79 and 4 days; the latter was excluded from the analysis because the data recorded were few. The fate of specimen A3 with only four days of recording is unclear. Tag detachment after four days could be linked to tag related issues (which is unlikely, since the release system was not activated) or to the death of the individual (very likely). The death of the fish could be due to the capture, handling and release process or to the illegal capture of the specimen. Hereafter, Totoaba1 will refer to a fish with 43 days recorded from May to mid-June; and Totoaba2, to an individual with 75 days recorded (May to mid-July); the days of the recovery period were excluded in both cases.
Totoaba1 and Totoaba2 remained 47% of the total time within a depth range of 25–35 m (Figure 3). The depth profiles showed that fishes swam to deeper water throughout the study period; their vertical movements comprised from 10 m to ~50 m for Totoaba1 and to 76 m for Totoaba2 (Figure 4). Totoaba1 and Totoaba2 spent 80% of the total time within a mean temperature between 21 and 23°C (range 17–26 °C) (Figure 3). The temperature profiles showed smaller 15-day variations in early May than in the rest of the tracking period; however, fishes remained at ~21 and 23°C (Figure 5).
Fig. 3. Depth and temperature occupancy of Totoaba1 and Totoaba2 individuals tracked for 43 days and 75 days, respectively.
Fig. 4. Daily depth (mean ± standard deviation) recorded for Totoaba1 and Totoaba2 in (a) May, (b) June and (c) July.
Fig. 5. Daily seawater temperature (mean ± standard deviation) recorded for Totoaba1 and Totoaba2 in (a) May, (b) June, and (c) July.
The diel diving for Totoaba1 and Totoaba2 in May and June exhibited movement patterns that alternated descent in daytime and ascent in night-time. These patterns, however, were not constant all the time; for example, Totoaba2 was recorded at a similar depth in daytime and night-time in July (Figures 6 & 7).
Fig. 6. Depth records for Totoaba1 over 43 days of free displacement in (a) May and (b) June. Dashed vertical lines show a 7-day period in each month (A, B respectively) during which vertical movements are displayed. Circles show daytime (grey) and night-time (dark) transitions.
Fig. 7. Depth records for Totoaba2 covering 75 days of free displacement in (a) May, (b) June, and (c) July. Dashed vertical lines show a 7-day period in each month (A, B, C, respectively) during which vertical movements are displayed. Circles show daytime (grey) and night-time (dark) transitions.
The general linear models had different statistical properties because they included different explanatory variables. The best GLM models based on AIC and BIC criteria include the same explanatory variables except for Moon Phase. Since this variable explains a little fraction of the variance (2.02%), it was not included in the BIC. Therefore, the most parsimonious model explaining the vertical movement of fishes includes SST, Month and Tsensor. All variables showed a significant relationship (P < 0.05, Tables 2 & 3). SST and Month showed the highest individual explained deviance (Tables 2 & 3). The diagnostic plots of residuals showed that the variance is homogeneous for the explanatory variables in the model. In addition, a nearly linear relationship between observed and predicted values was found, confirming a reasonably good fit of the data (see Appendix).
Table 2. Model construction and selection process of a generalized linear model to describe vertical movements of the totoaba (Totoaba macdonaldi) based on the Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC) and per cent deviance explained
Table 3. Model variables in the final fitted GLM
SE, Standard error.
*Significant differences (P < 0.05) from the intercept.
Discussion
This work provided a preliminary insight about daily vertical movements of the totoaba in respond to environmental variability across the water column in the Upper Gulf of California. Three PSATs were retrieved from 10 tags initially deployed; according to the size at first maturity of 124 cm TL described by De Anda-Montañez et al. (Reference De Anda-Montañez, García-De León, Zenteno-Savin, Balart-Paez, Méndez-Rodríguez, Bocanegra-Castillo, MArtínez-Aguilar, Campos-Dávila, Román-Rodríguez, Valenzuela-Quiñonez, Rodríguez-Jaramillo, Meza-Chávez, Ramírez-Rosas, Saldaña-Hernández, Olguín-Monroy and Martínez-Delgado2013), 80% of tagged individuals were adults. Geolocations along the path of the tags retrieved were not calculated due to scarce and poor-quality data recorded with the light sensor. The depth profiles showed that fishes swam to deeper water over time, remained 47% of the time at a depth ranging from 25–35 m, and preferred a mean temperature of 21–23°C. Diel diving exhibited movement patterns that combined descent in daytime and ascent in night-time, although these were not constant throughout the study period. Generalized linear models showed that sea surface temperature (SST) and ambient water temperature recorded by sensors tag (Tsensor) were related to swimming depth.
The results of the present study indicate the influence of temperature on the vertical movements of totoaba. The temperature range recorded here varied between 17 and 26°C; this is consistent with the range of temperatures reported during the catch of juvenile and adult totoabas (15 and 29°C) in the Upper Gulf of California before the fishing prohibition became effective (True et al., Reference Thygesen, Sommer, Evans and Patterson1997; Hernández-Aguilar et al., Reference Hernández-Aguilar, Zenteno-Savin, De-Anda-Montañez and Méndez-Rodríguez2018). The preference to remain within a temperature range (21–23°C) in both tagged totoabas and the displacement of Totoaba2 to deeper (~70 m) and cooler waters (21.2°C ± 0.99 SD) during July when sea surface temperature rose (28.55°C ± 0.48 SD), suggest that temperature is a key driver of vertical displacement and habitat occupancy in the totoaba. However, further tracking over longer periods and distances is needed to determine how daily vertical migration patterns are affected by temperature and other oceanographic conditions.
Previous studies on the totoaba suggest that after reproduction in the Upper Gulf of California, it migrates southward in search of suitable food-rich cold waters near Angel de la Guarda Island and the Ballenas-Salsipuedes Channel (Flanagan & Hendrickson, Reference Flanagan and Hendrickson1976; Cisneros-Mata et al., Reference Cisneros-Mata, Montemayor-López and Román-Rodríguez1995; Márquez-Farías & Rosales-Juárez, Reference Márquez-Farías and Rosales-Juárez2013; Valenzuela-Quiñonez et al., Reference Turnure, Able and Grothues2015). Premature pop-up locations of 50% of the PSATs initially deployed are consistent with this migratory route, and the southward movement of tagged fishes may seek favourable post-spawning feeding grounds. This area has the lowest recorded temperatures throughout the year within the Gulf of California (Soto-Mardones et al., Reference Sims, Southall, Humphries, Hays, Bradshaw, Pitchford, James, Ahmed, Brierley, Hindell, Morritt, Musyl, Righton, Shepard, Wearmouth, Wilson, Witt and Metcalfe1999; Marinone & Lavín, Reference Marinone, Lavín, Velasco Fuentes, Sheinbaum and Ochoa2003; Lluch-Cota et al., Reference Lluch-Cota, Aragón-Noriega, Arreguín-Sánchez, Aurioles-Gamboa, Jesús Bautista-Romero, Brusca, Cervantes-Duarte, Cortés-Altamirano, Del-Monte-Luna, Esquivel-Herrera, Fernández, Hendrickx, Hernández-Vázquez, Herrera-Cervantes, Kahru, Lavín, Lluch-Belda, Lluch-Cota, López-Martínez, Marinone, Nevárez-Martínez, Ortega-García, Palacios-Castro, Parés-Sierra, Ponce-Díaz, Ramírez-Rodríguez, Salinas-Zavala, Schwartzlose and Sierra-Beltrán2007). It is characterized by strong tidal currents that cause turbulence across the water column, reaching more than 50 m depth, resulting in the resuspension of sediments and carrying nutrients and cold water to the surface (Alvarez-Borrego, Reference Alvarez-Borrego and Brusca2010). These events increase productivity, boosting the abundance of sardines and other small pelagic fish such as the Pacific anchoveta, Cetengraulis mysticetus (Lluch-Belda et al., Reference Lluch-Belda, Magallon and Schwartzlose1986), that are major components of the totoaba diet (Román-Rodríguez, 1990; De Anda-Montañez et al., Reference De Anda-Montañez, García-De León, Zenteno-Savin, Balart-Paez, Méndez-Rodríguez, Bocanegra-Castillo, MArtínez-Aguilar, Campos-Dávila, Román-Rodríguez, Valenzuela-Quiñonez, Rodríguez-Jaramillo, Meza-Chávez, Ramírez-Rosas, Saldaña-Hernández, Olguín-Monroy and Martínez-Delgado2013).
Cisneros-Mata et al. (Reference Cisneros-Mata, Montemayor-López and Román-Rodríguez1995) related the southward displacement of totoaba in the Gulf of California to the migration of sardines. In addition, schools of small pelagic fish such as sardines and anchovies are abundant in the water column at around 30–40 m depth near Angel de la Guarda Island and the Ballenas-Salsipuedes Channel, where vertical migrations have been described (Rubio-Rodríguez et al., Reference Román-Rodríguez2018). This information is consistent with the depth occupancy and diel vertical movement by the tagged individuals. Therefore, vertical movements of the totoaba probably respond to the displacement of its prey. In other species, the recurrent displacement at mid water suggests that this might be a preferred strategy when seabed depth changes continuously, while a more sustained maximum depth would lead to travelling close to the seabed (Hobson et al., Reference Hobson, Righton, Metcalfe and Hays2009). From the Upper Gulf of California to Angel de la Guarda Island, where satellite tags were released, there is a complex bathymetric scenario characterized by shallow areas of 10 m and basins as deep as ~200–2000 m, together with tidal ridges averaging 8–9 m in vertical relief (Alvarez et al., Reference Alvarez, Suárez-Vidal, Mendoza-Borunda and González-Escobar2009).
Studies of other sciaenids are consistent with movements driven by tides, as in Argyrosomus japonicas (Næsje et al., Reference Næsje, Cowley, Diserud, Childs, Kerwath and Thorstad2012) and Cynoscion regalis (Turnure et al., Reference True, Loera and Castro2015); these species seemingly use tides to guide a feeding strategy that minimizes energy expenditure on movements by consuming prey that follow tidal currents. Although the influence of moon phases is not well understood, the Upper Gulf of California is known as a region where the tidal amplitude can be >7 m, with a current speed exceeding 1 m s−1 (Lavin et al., Reference Lavin, Godínez and Alvarez1998; Alvarez-Borrego, Reference Alvarez-Borrego and Brusca2010). It is possible that movements of the totoaba are also driven by tides, although the limited number of tracking data are insufficient to show any well-defined effect.
Six of the 10 tags placed emitted signals to the Argos system prematurely, while the rest sent no signals to the satellite. Broell et al. (Reference Broell, Taylor, Litvak, Bezanson and Taggart2016), who used the same type of satellite tags in demersal fish together with acoustic tags, indicated that some of the tags placed were released early and sent no information to the satellite, although the potential causes were not mentioned. In our case, the tags that transmitted signals before the scheduled date suggest that the respective fish were recaptured, since the location of prematurely released tags in the Upper Gulf of California is known by locals as an area of illegal totoaba fishing. Five of the six tags were released on different dates at around the same geographic position, a deserted area outside the boundaries of the Upper Gulf of California and the Colorado River Delta Biosphere Reserve; however, further research including a larger number of individuals and a more prolonged time is needed to confirm these assertions.
Despite the premature-release issues mentioned above, the information obtained in this study provided preliminary insights into the vertical displacement of this species in the study area. The results showed that, after reproduction, both individuals migrated to the southern Gulf of California at mid water searching for a particular temperature (21–23°C). This pattern could reflect their occupancy of a particular habitat in the summer; this preliminary information should be supported by studies involving longer tracking throughout the year of more totoaba individuals.
Acknowledgements
We thank Horacio Bervera-León, Oswaldo Rodríguez-García, Gaston Antonio Bazzino Ferreri and Crisalejandra Rivera Pérez for their invaluable assistance in the field. To the Federación de Cooperativas de San Felipe, B.C. particularly to Captain Raúl Bugarín, and the vessel crew. Thanks also to Hector Aguilar, Arturo and Ramón Franco for their support in sampling. Oscar Sosa-Nishizaki of CICESE provided equipment for the recovery of satellite tags. The Secretaria del Medio Ambiente y Recursos Naturales through the Dirección General de Vida Silvestre issued the permit (SGPA/DGVS/00492/16) to conduct scientific collections, and staff from Procuraduría Federal de Protección al Ambiente provided support during fieldwork. C. Hernández gratefully acknowledges scholarship no. 342449 from CONACYT. We also thank the anonymous reviewers and editors for their helpful comments and suggestions, which helped to improve the manuscript. María Elena Sánchez-Salazar edited the English manuscript.
Financial support
This work was supported by the Mexican Consejo Nacional de Ciencia y Tecnología (CONACYT Grant CB-2011-01; 165376).
Appendix
Fig. A1. Diagnostic plots for the vertical movement model: (A) observed values vs residual deviance, (B) predicted values vs residual deviance; and (C) observed vs predicted values.
