Oceanographic campaigns

Oceanographic campaigns were carried out off the Mexican coast between September 2016 and July 2017. In September 2016, 2 samplings were done aboard the R/V Altaír, which belongs to the Secretaría de Marina Armada de México (SEMAR); the first was done in the marine-coastal zone of Navachiste (off the north of Sinaloa) (NAV2016) and the second in the coastal zone off Guaymas (Sonora) (GUA2016). In March 2017, samplings were carried out in the marine-coastal zone of Navachiste (NAV2017) on board a boat with an outboard motor. In June 2017, 3 areas were sampled aboard the R/V Río Tecolutla of the SEMAR: off Mulegé (MUL2017), in Concepción Bay (BC2017), and off Mazatlán (MAZ2017) (Fig. 1).

Figure 1 Study area in the Gulf of California. Sampled subregions are shown in black squares: Navachiste (a), Guaymas (b), Concepción Bay (c), Mulegé (d), Mazatlán (e), Midriff Islands (f). 

Temperature and salinity were measured using a Sea-Bird 19 Plus CTD, equipped with a 5-L Niskin bottle rosette. Discrete samples of surface water were collected to analyze pH, DIC, and total alkalinity (TA) in 350-mL sodium borosilicate bottles with 100 µL of a saturated solution of mercuric chloride used as a preservative to stop biological activity. Samples were taken at 8 stations at NAV2016, 5 at GUA2016, 9 at NAV2017, 4 at BC2017, 9 at MUL2017, and 7 at MAZ2017. The bottles were sealed with Apiezón grease to prevent evaporation and contact with the atmosphere.

Laboratory analysis

We used an Apollo SciTech AS-C3 instrument to analyze the DIC. We used 6% phosphoric acid to convert all the DIC to CO₂, and we quantified this with a Li-COR 7000 high-precision infrared analyzer (^{Dickson and Goyet 1994}). TA was analyzed with an Apollo SciTech AS-ALK2 titration system and 0.1 M hydrochloric acid solution (the 3 methods are detailed in ^{Hernández-Ayón 1995}). We verified the precision and accuracy of the methods and took care not to exceed 5 µmol·kg^‒1 (0.25% error) in the difference of DIC and TA with respect to reference values. The reference substandards were developed at the Instituto de Investigaciones Oceanológicas of the Universidad Autónoma de Baja California (Mexico).

The calculation of pCO_2W was performed with the CO2SYS program (^{Lewis and Wallace 1998}) using surface data of DIC, TA, temperature, salinity, and the dissociation constants of ^{Lueker et al. (2000)}. We estimated fCO₂ using the ^{Liss and Merlivat (1986)} equation: F = K_wK₀ (∆pCO₂), where K_w is the CO₂ transfer coefficient as a function of wind speed at 10 m above sea level and K₀ is the solubility coefficient of CO₂ as a function of temperature and salinity. The gas transfer coefficient of ^{Ho et al. (2006)} and the CO₂ solubility coefficient of ^{Weiss (1974)} were used. The partial pressure of CO₂ in the atmosphere (pCO_2A) was obtained from the internet portal of the Earth System Research Laboratories (ESRL) of the National Oceanic and Atmospheric Administration (NOAA, USA) for each sampling day, and the values corresponded to the Mauna Loa station, Hawaii (^{NOAA 2017}). Due to failures in the ship’s anemometer, the wind data for the different areas were obtained from the MeteoEarth electronic portal (http://www.meteoearth.com), from the model of the European Center for Medium-Range Weather Forecasts; this model was used because the spectral space is larger than the physical parameterizations in a reduced grid (^{ECMWF 2017}).

To better understand the variation of the sea surface temperature (SST) and chlorophyll concentration (Chl_sat) in the GC during the different cruises, 5-day satellite averages of the Aqua-Modis sensor were used for the approximate dates of each campaign. For SST, daytime data were used with radiance of 11 µm wavelength and pixel size of 1.1 × 1.1 km². For Chl_sat, images with processing level 3 and the same pixel size as those of the SST were used. In addition, daily sea surface height anomaly (SSHa) images were obtained, with spatial resolution of 0.25° and L4 level data, produced by AVISO and CMEMS. Geostrophic velocity was calculated according to ^{Pond and Pickard (2013)}. All images were processed in Matlab.

Statistical analysis and data visualization

Principal component analysis (PCA) was performed to determine which factors influenced the variability of pCO_2W in the different areas of the GC. Beforehand, the corresponding analysis of variance was performed to determine the normality of the data; when they were not normal, the PCA was carried out. Data of the SST, salinity, DIC, TA, pCO_2W, pCO_2A, and Chl_sat were normalized; the Broken-stick criterion was used according to ^{Peres-Neto et al. (2003)}, and the correlation and significance of all variables were obtained. All of the above was done with the help of the Statistica software.

Maps of the salinity, DIC, pCO_2W, and fCO₂ were made with the free distribution Ocean Data View software v.5.0 (^{Schlitzer 2016}). Map interpolations were done using the data-interpolating variational analysis (DIVA) of the Ocean Data View to ensure the conservation of structures and better visualize results.

Oceanographic conditions inferred from satellite images

Satellite composites of SSHa, SST, and Chl_sat showed contrasting conditions in September 2016 and March 2017, and transition conditions in June 2017 (Fig. 2). In September 2016 anticyclonic eddies were recorded along the GC. In NAV2016, geostrophic flow (~16 cm·s^-1) towards the south was observed; and in GUA2016, geostrophic flow (~18 cm·s^-1) towards the coast was observed as part of a small anticyclonic eddy (~90 km) with a SSHa of ~24 cm (Fig. 2a). Negative anomalies of the sea level were observed in March 2017 in most of the GC, with the exception of the zone adjacent to La Paz Bay, and an anticyclonic eddy was recorded at the entrance of the gulf (~17 cm·s^-1 and height of 15 cm). Part of a cyclonic eddy (-10 cm) was observed in NAV2017 (Fig. 2b). In June 2017, an intense water inflow occurred in the eastern region of the GC and an intensification of the current with a high sea level (>25 cm·s^-1 and height of 25 cm) was observed in Guaymas, which extended to the region of the Midriff Islands (>30 cm·s^-1 and height of 22-36 cm). In MUL2017, we observed the influence of a cyclonic eddy with flow towards the coast in the northern part (~12 cm·s^-1). In MAZ2017, we recorded the entry of water into the GC with a speed of ~25 cm·s^-1 and a cyclonic circulation at the mouth of the GC (Fig. 2c).

Figure 2 Sea surface height anomalies (SSHa) and geostrophic flow (cm·s^-1) (a-c), sea surface temperature (SST, d-f), and chlorophyll (Chl_sat, g-i) in September 2016 (squares: red, NAV2016; blue, GUA2016), March 2017 (red square, NAV2017), and June 2017 (squares: red, MAZ2017; blue, MUL2017). 

In September 2016, in practically the entire GC, SST was >29 °C (Fig. 2d) and Chl_sat values were <0.5 mg·m^-3. In NAV2016, a coastal strip with Chl_sat values >1.5 mg·m^-3 was observed; in GUA2016, lower values (~0.8 mg·m^-3) were observed (Fig. 2g). In March 2017, SST values were lower than in September throughout the gulf, with differences greater than 8 °C, and Chl_sat was up to ˃1 mg·m^-3 higher than in September, with different mesoscale structures in concordance with SSHa data. In NAV2017, SST values were <22 °C and Chl_sat was >1.5 mg·m^-3 (Fig. 2e, h). Due to the patterns recorded in June, this month was considered a transition period, with intermediate values of SST and Chl_sat between September and March. MUL2017 and MAZ2017 showed SST values <29 °C and Chl_sat values >0.5 mg·m^-3 (Fig. 2f, i).

Variability of salinity, dissolved inorganic carbon, pCO_2W, and fCO₂

Salinity distribution showed that NAV2016 (Fig. 3a) had lower values (34.70 ± 0.21, mean and standard deviation, respectively, here and henceforth) than NAV2017 (35.29 ± 0.29) (Fig. 3c). There was little variation (35.14 ± 0.01) in GUA2016 (Fig. 3b). Homogeneous values were also recorded (35.94 ± 0.13) in BC2017 (Fig. 3d). Values were 35.28 ± 0.13 in MUL2017 (Fig. 3e), and the central region had the maximum values. MAZ2017 also had little variation in salinity (35.02 ± 0.05) (Fig. 3f).

Figure 3 Horizontal distribution of salinity and dissolved inorganic carbon (DIC) for the different subregions: NAV2016 (a, g), GUA2016 (b, h), NAV2017 (c, i), BC2017 (d, j), MUL2017 (e, k), and MAZ2017 (f, l). The figures were confectioned using Ocean Data View. 

DIC and pCO_2W values showed contrasting conditions in NAV2016 (1,994 ± 3.52 µmol·kg^-1 and 428 ± 12.21 µatm, respectively; Figs. 3g, 4a) and NAV2017 (2,076 ± 25.08 µmol·kg^-1 and 363 ± 46.94 µatm, respectively, Figs. 3i, 4c). In NAV2017, variations of DIC and pCO_2W were similar; values of both properties were relatively low at the northernmost station (~2,070 µmol·kg^-1 and ~315 µatm, respectively), increased in the central area (~2,120 µmol·kg^-1 and ~420 µatm), and decreased towards the south (~2,080 µmol·kg^-1 and ~330 µatm). GUA2016 showed little variation of DIC and pCO_2W, although with higher values towards the south of the area (2,015 ± 16.22 µmol·kg^-1 and 434 ± 29.06 µatm, respectively) (Figs. 3h, 4b). In BC2017, little variation of DIC and slightly more marked variation of pCO_2W were recorded (2,017 ± 14.69 µmol·kg^-1 and 377 ± 29.45 µatm, respectively) (Figs. 3j, 4d). In MUL2017 there was greater variability; DIC had low values towards the central zone in the open sea and high values towards the edges (2,058 ± 15.05 µmol·kg^-1), and pCO_2W had higher values to the north of the study area (414 ± 37.79 µatm) (Figs. 3k, 4e). DIC and pCO_2W had the lowest values (1,934 ± 5.94 µmol·kg^-1 and 305 ± 12.64 µatm, respectively) in MAZ2017, although pCO_2W increased in the center of the area (Figs. 3l, 4f). In general, DIC and pCO_2W had the highest average values in NAV2017 and GUA2016, respectively, whereas both variables had the lowest values in MAZ2017. All the above information is summarized in Table 1.

Figure 4 Horizontal distribution of seawater CO₂ partial pressure (pCO_2W) and CO₂ flux variability (fCO₂) for the different subregions: NAV2016 (a, g), GUA2016 (b, h), NAV2017 (c, i), BC2017 (d, j), MUL2017 (e, k), and MAZ2017 (f, l). The figures were confectioned using Ocean Data View. 

Four of the 5 regions studied behaved as CO₂ sinks in March and June 2017: NAV2017, BC2017, MUL2017, and MAZ2017. NAV2016 behaved as a light source of CO₂ into the atmosphere; the average fCO₂ there was 0.38 ± 0.52 mmol C·m^-2·d^-1, and values ranged from practically in equilibrium with the atmosphere (0.03 mmol C·m^-2·d^-1) to 1.57 mmol C·m^-2·d^-1 (Fig. 4g). In GUA2016, the average fCO₂ was 0.56 ± 0.46 mmol C·m^-2·d^-1; the spatial pattern showed equilibrium values at the most coastal station and maximum values of 1.14 mmol C·m^-2·d^-1 at the most oceanic station (Fig. 4h). The positive values for September 2016 indicate a slight flow from the sea to the atmosphere. NAV2017 behaved differently; 78% of the area had negative fCO₂ values, with an average of -0.40 ± 0.39 mmol C·m^-2·d^-1 and a range of values from -1.03 to 0.39 mmol C·m^-2·d^-1 (Fig. 4i). BC2017 trended to larger negative values towards the inner bay, with an average CO₂ flux of -0.90 ± 0.87 mmol C·m^-2·d^-1 (Fig. 4j). MUL2017 showed positive values to the north and negative values to the south, with an average of -0.05 ± 0.74 mmol C·m^-2·d^-1 and a range of values from -1.01 to 1.37 mmol C·m^-2·d^-1 (Fig. 4k). The entire area of MAZ2017 showed negative values, with an average of -2.26 ± 1.85 mmol C·m^-2·d^-1 and a range of values from -4.99 to -0.42 mmol C·m^-2·d^-1 (Fig. 4l).

Statistics

The PCA showed an explained variance of 34% for the first component and 30% for the second (Fig. 5). In the first component, the variables with the greatest weight were DIC, TA, and SST (0.95, 0.80, and -0.74, respectively), and the one with the least contribution was Chl_sat (0.57). In the second component, the variables with the greatest weight were pCO_2W and pCO_2A (0.95 and -0.74, respectively). The zones that were centered throughout the first component were NAV2016, GUA2016, and some MUL2017 stations. NAV2017 and MAZ2017 showed an opposite behavior (Fig. 5). Due to the amount of data recorded per zone, it was difficult to indicate that the weight of the variables suggested that in the primary processes of the carbonate system in NAV2016 and GUA2017 thermal effects predominate and that biological ones do so to a lesser degree. The PCA suggested that biological effects were stronger than physical effects at NAV2017, whereas at MUL2017, BC2017, and MAZ2017, both effects may have contributed similarly.

Figure 5 Principal component analysis (PCA) for Navachiste (NAV2016, NAV2017), Guaymas (GUA2016), Concepción Bay (BC2017), Mulege (MUL2017), and Mazatlán (MAZ2017). 

Oceanographic conditions

The circulation pattern in the GC is determined by the presence of eddies at the mouth. In September 2016 (summer, Fig. 2a), we recorded a cyclonic eddy at the mouth, the inflow of water from the western part of the gulf, the presence of anticyclonic eddies, and the outflow of water in the Sinaloa coast. The aforementioned inflow and outflow of water from the gulf coincides with that reported by ^{Portela et al. (2016)}, but for spring; for summer, they indicated that the water entered through the eastern coast (Sinaloa) and outflowed through the western coast (Baja California). ^{Beier (1997)} reported a cyclonic circulation in the GC in summer, which does not coincide with the pattern recorded in this study. The anticyclonic eddies recorded in Guaymas and Sinaloa coincide with that reported by ^{Pegau et al. (2002)} from a time series from 1997 to 2001. In March 2017 (winter, Fig. 2b), an anticyclonic eddy appeared at the mouth of the GC, with negative sea level anomalies in practically the entire GC that, despite the geostrophic velocities, were small; in addition, we observed a flow towards the south of the gulf on the peninsular coast, but the influence of the anticyclonic eddy that crossed towards Sinaloa suggests a limited circulation towards the inner gulf. ^{Beier (1997)} and ^{Castro et al. (2017)} reported an anticyclonic circulation in winter, whereas ^{Castro et al. (2000)} reported, in their 1992-1998 study, a cyclonic circulation between winter and spring in the central part of the gulf. This difference in circulation patterns has been reported by ^{Castro et al. (2017)}, who concluded that mesoscale eddies can modify the seasonal exchange of water at the entrance of the GC. In June 2017 (Fig. 2c) a geostrophic flow towards the inner gulf was recorded off the coast of Sinaloa, indicative of a cyclonic circulation, which coincides with ^{Beier (1997)} and ^{Portela et al. (2016)}. Therefore, this study suggests that the circulation pattern was anticyclonic in the GC in September 2016 (summer) and cyclonic in June 2017 (transition).

The climatic conditions in the GC were contrasting in September 2016 and March 2017, and transitional in June 2017, which coincides with that reported by ^{Roden (1964)}. September is considered one of the warmest months and March one of the coldest (^{Soto-Mardones et al. 1999}). In September 2016 (Fig. 2d, g), high temperatures (˃30 °C) were associated with low chlorophyll concentrations (˂0.1 mg·m^‒3) near the eastern coast, which coincides with what was reported by ^{Santamaría-del-Angel et al. (1999)}. For March 2017 (Fig. 2e, h), low SSTs (~22 °C) were associated with high Chl_sat concentrations (˃2 mg·m^‒3) in the eastern coast of the gulf, due to the effect of coastal upwelling in winter. In particular, comparisons of NAV2016 with NAV2017 suggested the presence of coastal upwelling in March and its absence in September (Fig. 2d, e, g, h). In June 2017 (Fig. 2f, i), in the central and southern regions of the GC, SSTs were higher by the mainland than by the peninsula, which coincides with ^{Santamaría-del-Ángel et al. (1999)} and ^{Soto-Mardones et al. (1999)}, mainly for summer (~1 °C difference).

Variability of salinity, dissolved inorganic carbon, pCO_2W, and fCO₂

The lowest average salinity was recorded in NAV2016, and the highest salinities were recorded in NAV2017 and BC2017 (Fig. 3a, c, d). In GUA2016 (Fig. 3b), surface salinity values >35 indicated the presence of the Gulf of California Water (GCW), which had already been described by ^{Delgadillo-Hinojosa et al. (2001)}. In NAV2016 (Fig. 3a), values <35 were a result of the presence of transition water and GCW (^{Portela et al. 2016}). The salinity recorded in GUA2016 (Fig. 3b) represents, according to ^{Bray (1988)}, the general circulation of the GC. In NAV2017 (Fig. 3c), in agreement with what was reported by ^{Castro et al. (2000)}, salinity values were a combined product of coastal upwelling water and GCW flowing south into the gulf. Concepción Bay (Fig. 3d) is a semi-enclosed bay where solar radiation in summer could have caused water evaporation and increased salinity. In MAZ2017 (Fig. 3f), salinity was homogeneous throughout the area, which suggests that transitional water entered the gulf (^{Portela et al. 2016}). Therefore, in the sampling months (September 2016, March and June 2017), according to salinity records, GCW and transitional water were located at the surface.

In NAV2016, we recorded the highest temperature and Chl_sat ( x- = 29.94 °C and x- = 1.23 mg·m^-3, respectively), and pCO_2W ( x- = 428 μatm) indicated a supersaturation of CO₂ and a positive fCO₂ ( x- = 0.38 mmol C·m^-2·d^-1). In GUA2016, we also recorded a SST of x- = 29.94 °C, but low concentrations of Chl_sat ( x- = 0.23 mg·m^-3), a pCO_2W ( x- = 434 μatm) above the average for the pCO_2A (402 μatm), and a positive fCO₂ of x- = 0.56 mmol C·m^-2·d^-1. In both areas there were summer conditions, anticyclonic eddies, CO₂ supersaturation, and a slight CO₂ release into the atmosphere, with pCO₂ values above equilibrium and those reported by ^{Rodríguez-Ibáñez et al. (2013)} for summer, which indicated that the GC is almost in equilibrium with the atmosphere. The water mass that predominates in summer is GCW (^{Álvarez-Borrego and Schwartzlose 1979}). ^{Portela et al. (2016)} reported, for this same period, the presence of Superficial Tropical Water and GCW, which suggests that both masses of water are present in NAV2016 and GUA2016. The contrast in the concentration of Chl_sat between both zones could be explained by differences in the intensity of grazing by zooplankton; it is important to consider that Chl_sat concentrations only reflect the first optical depth and that, due to the anticyclonic eddy, the subsurface maximum of chlorophyll of GUA2016 could be below this optical depth, and thus cause low concentrations to be observed in satellite images. This behavior was recorded by ^{Espinosa-Carreón et al. (2012)} for the southern region of the California Current. ^{Flores-Trejo et al. (2019)} reported surface DIC values of 2,055 μmol·kg^-1 associated with the GCW for Punta Lobos, Sonora (south of Guaymas), in October 2018, which were slightly higher than those reported in this work for the GUA2016 area (2,015 μmol·kg^-1). The ocean is so dynamic that the fCO₂ can vary over the course of days. ^{Morales-Urbina et al. (2017)} reported data from an oceanographic buoy anchored in the NAV2016 area during the summer of 2016, and their average fCO₂ value for 10 days was 2.70 mmol C·m^‒2·d^‒1, higher than the average fCO₂ for NAV2016 recorded in the present study, which suggests an increase in the contribution of CO₂ to the atmosphere.

The highest DIC concentration was recorded in NAV2017 ( x- = 2,076 μmol·kg^-1), which suggests that upwelling conditions could have occurred. However, based on the values obtained for pH ( x- = 8.08, data not shown) and fCO₂ ( x- = -0.40 mmol C·m^-2·d^-1), said event was in a relaxation phase. The negative fCO₂ indicated that the GC behaved as a sink, and the Chl_sat values ( x- = 1.09 mg·m^-3) indicated that the phytoplankton was increasing and consuming DIC. ^{Coronado-Álvarez et al. (2017)} reported similar conditions for the upwelling area off the northwest of the Baja California Peninsula, and they indicated that when upwelling relaxes, nutrients are used more effectively, which causes DIC and pCO_2W to decrease; changes in these variables were associated with factors ranging from semidiurnal variabilities (sea breezes), upwelling events, to interannual variabilities (El Niño/Southern Oscillation). DIC values found in this study were lower than those reported by ^{Linacre et al. (2010)} (~2,215 μmol·kg^-1) for the upwelling zone off the western coast of Baja California at a depth of ~80 m.

In BC2017, fCO₂ values ranged from -2.15 to -0.16 mmol C·m^‒2·d^‒1 (Fig. 4j). Chl_sat values (data not shown) indicated that much of the area was mesotrophic to eutrophic, with possible high photosynthetic activity in the days prior to sampling, which could have caused the negative fCO₂ values. However, there are spatiotemporal variations between the absorption and contribution of CO₂ in the study area that should be considered in subsequent research to describe the variability of fCO₂.

In the MUL2017 region, the average DIC concentration was 2,058 μmol·kg^-1, lower than that reported by ^{Hernández-Ayón et al. (2013)}, who indicated that there was a relationship between the variables of the carbonate system and the water masses in the region of the Midriff Islands and reported values of ~2,080 μmol·kg^-1 for the summer of 2004. The spatial variability of DIC in this region could be due to the cyclonic eddy found in this area, with differences in the intensity of the geostrophic current (Fig. 2c); this eddy could have influenced the pCO_2W values to be below equilibrium to the south of the area and to be positive to the north of the area, probably due to the shearing of the water, which promotes the release of CO₂ into the atmosphere. Therefore, the fCO₂ value of -0.05 mmol C·m^-2·d^-1 (Fig. 4k) was probably due to combined physical and biological effects.

The flow of water into the GC along the eastern coast in MAZ2017 carried properties that resulted in the lowest concentration records of Chl_sat, DIC, pCO_2W, and fCO₂ ( x- = 0.18 mg·m^-3, 1,934 μmol·kg^-1, 305 μatm, -2.26 mmol C·m^-2·d^-1, respectively) (Fig. 2c). ^{Franco et al. (2014)} reported, for the area south of Cabo Corrientes in August 2010, DIC concentrations of ~1,980 μmol·kg^-1 and minimum fCO₂ values of -4.4 mmol C·m^-2·d^-1, values which are higher than those of the present study for MAZ2017. ^{Trucco-Pignata et al. (2019)} reported, for June 2015, transition water at the entrance of the GC that traveled at the surface (0-115 m) with a DIC concentration of ~2,050 μmol·kg^-1 (higher than that reported in the present study). Therefore, in MAZ2017, the advection of water and its productivity determined that the region can act as a source or sink of CO₂ (^{Franco et al. 2014}).

Correlations

The PCA grouped, on the one hand, NAV2016 and GUA2016, where anticyclonic eddies predominated and the variables were similar, except Chl_sat. On the other hand, it grouped NAV2017 and MAZ2017, both with the lowest fCO₂ values; in the first, the negative sea level anomaly predominated, and in the second, the coastal flow on the eastern coast. Therefore, the fCO₂ is determined by the oceanographic conditions of the area, which acts as a source or sink at different times.

It is important to consider that the GC is part of the great oceanographic dynamics of the eastern Pacific. According to ^{Rodríguez-Ibáñez et al. (2013)}, the net contribution of nutrients and DIC is not transported homogeneously throughout the GC, due to regional differences in physical dynamics, which promotes variations in CO₂ fluxes. These authors reported that, in a one-year average, the GC was a light emitter. In the present study, the average of the samplings carried out in 2016 and 2017 in the 5 coastal zones was -0.42 mol C·m^-2·d^-1, which suggests that the GC behaved as a light CO₂ sink. ^{Laruelle et al. (2014)} classified GC as a CO₂ emitter (0-1 mol C·m^-2·y^-1). In regions dominated by coastal upwelling on the eastern edges, when there are changes in coastal currents, and these occur in short time scales, the physical and biological processes are modified, as well as the ocean-atmosphere CO₂ equilibrium conditions (^{Cai et al. 2020}). Concentrations of DIC, pCO_2W, and, therefore, the fCO₂ are associated with the oceanographic dynamics of the area and, among others, with the presence of mesoscale processes, coastal upwelling, advection, and biological processes. Thus, oceanographic conditions in the GC (such as cyclonic eddies, anticyclonic eddies, and coastal upwelling) and its biology, which vary throughout it and in time, influenced changes in pCO_2W and fCO₂. Based on the information obtained in this study in the 5 zones, we describe possible responses to understand the mechanisms by which a positive or negative ocean-atmosphere CO₂ exchange occurs; nevertheless, substantial efforts are needed for more information at the spatial and temporal level to further the understanding of CO₂ flux dynamics in the GC.

^{Coronado-Álvarez et al. (2017)} created time series of SST, salinity, and pCO_2W with high-frequency point data (every 3 h) recorded by a MAPCO₂ system (for more details see ^{Sutton et al. 2014}) anchored to an oceanographic buoy. With the Observatorio del Monitoreo Costero (Coastal Monitoring Observatory), these authors obtained a 7-year time series, off the northwest of Baja California, in an area of intense upwelling. ^{Coronado-Álvarez et al. (2017)} observed significant changes in these properties and in fCO₂ that lasted for periods as short as hours and as long as interannual segments. Their spectral analysis showed significant semidiurnal, diurnal, ~15-d, and ~28-d variance components. Their description is very limited because they only used data for one geographic point, but their results indicate that those of the present study should be taken as preliminary results because they are instantaneous for each cruise. Continued efforts are needed to understand and more appropriately piece together the puzzle of the role GC plays in fCO₂.