The Gulf of California is a source of carbon dioxide to the atmosphere El golfo de California es una fuente de bióxido de carbono hacia la atmósfera

Water exchange between the Gulf of California and the Pacific Ocean has a significant vertical component (VCWE). Surface (0–200 m) gulf water flows out into the Pacific Ocean and deep (200–600 m) water flows into the gulf. This is a mechanism that allows for the net input to the gulf of dissolved constituents whose concentrations increase with depth, including dissolved inorganic carbon (DIC). Two scenarios were used to estimate the net input of DIC from the Pacific into the gulf (DICNET INPUT) and to compare this net input with new phytoplankton production in the whole gulf (PNEW) in order to infer if the gulf is a sink or source of CO2. The average annual values of VCWE were 0.67 ± 0.10 Sv in the first scenario and 0.23 ± 0.02 Sv in the second scenario (1 Sv = 106 m3 s–1). After comparing DICNET INPUT with PNEW the result is that the gulf is a source of CO2 to the atmosphere in both scenarios, with an annual average out-gassing of (18.16 ± 6.14) × 1012 and (7.66 ± 2.65) × 1012 grams of carbon in the form of CO2 in the first and second scenarios, respectively. These values are equivalent to an average of 123.5 ± 41.8 and 52.1 ± 18.0 g m–2 yr–1, respectively. The value for the first scenario is higher than the highest value for the eastern equatorial Pacific as reported in the literature (~108 g m–2 yr–1), which leads us to conclude that the value for the second scenario is closer to reality.


INTRODUCTION
The oceans have been considered to be a major sink for CO 2 .Hence, the improved knowledge of the net transport flux across the air-sea interface is important for understanding the fate of this greenhouse gas emitted into the earth's atmosphere (Callendar 1938, Revelle and Suess 1957, Brewer 1978, Siegenthaler and Sarmiento 1993).On the basis of the global distribution of ΔpCO 2 values (ΔpCO 2 = surface water CO 2 partial pressure minus air CO 2 partial pressure: pCO 2w -pCO 2air ), a global net ocean uptake flux for anthropogenic CO 2 emissions of 2.0 ± 1.0 PgC yr -1 was estimated in a reference year, 2000 (one petagram of carbon is 10 15 grams of carbon in the form of CO 2 ) (Takahashi et al. 2009).However, the coastal ocean has been largely ignored in global carbon budgeting efforts, even if the related flows of carbon and nutrients are disproportionately high in comparison with its surface area (Chen et al. 2003).A synthesis of worldwide measurements of pCO 2w indicates that most open shelves in the temperate and high-latitude regions are undersaturated with respect to pCO 2air during all seasons, although the low-latitude shelves seem to be oversaturated (Chen and Borges 2009).These latter authors indicated that continental shelves absorb atmospheric CO 2 (from 0.33 to 0.36 PgC yr -1 .) The Gulf of California has been recognized as a highly productive marginal sea of the Pacific Ocean (Álvarez-Borrego and Lara-Lara 1991).Upwelling occurs off the eastern coast with northwesterly winds ("winter" conditions from December through May) and off the Baja California coast with southeasterly winds ("summer" conditions from July through October), with June and November as transition periods (Roden 1964).Coastal upwelling areas are known to show oversaturation of CO 2 with respect to atmospheric equilibrium because of the input of CO 2 -rich deep waters into the mixed layer (Borges 2005).
The northern gulf exhibits spectacular tidal phenomena, with a range of >7 m during spring tides in the uppermost gulf and >4 m in the Midriff Islands region.In spite of relatively strong stratification during summer, tidal mixing in the Midriff Islands region produces a vigorous stirring of the water column down to >500 m depth, with the net effect of carrying colder, nutrient-rich water to the surface (Simpson et al. 1994) and creating an ecological situation similar to constant upwelling (Álvarez-Borrego 2002).This also has the effect of making the areas around the Midriff Islands a strong source of CO 2 to the atmosphere (Zirino et al. 1997, Hidalgo-González et al. 1997).
Based on pH and alkalinity values, Hidalgo-González et al. (1997) sampled in the summer to generate a set of pCO 2w and sea-air CO 2 flux estimates for the Midriff Islands region.The calculated CO 2 flux was toward the atmosphere and it was greatest with post-spring tides (up to 23 mmol m -2 d -1 ).Hidalgo-González et al. (1997) indicated that winter CO 2 flux values toward the atmosphere should be much higher than those for summer because of less stratified waters and much higher pCO 2w in winter than during summer.Maxima pCO 2w calculated for the Midriff Islands region have been 560 atm for October 1985 (Zirino et al. 1997), 446 atm for July 1990 (Hidalgo-González et al. 1997), 560 atm for September 1996 (Hernández-Ayón et al. 2007a), and 1200 atm for March 2002 (Hernández-Ayón et al. 2007b).This indicates that the Midriff Islands region is an area that acts as a cuasi-permanent source of CO 2 to the atmosphere throughout the whole year.antropogénico por el océano de 2.0 ± 1.0 PgC año -1 (un petagramo de carbono equivale a 10 15 gramos de carbono en forma de CO 2 ) (Takahashi et al. 2009).No obstante, el océano costero ha sido mayormente ignorado en los esfuerzos globales en torno al presupuesto de carbono, a pesar de que los flujos relacionados de carbono y nutrientes son desproporcionadamente altos en comparación con su área superficial (Chen et al. 2003).Una síntesis de mediciones globales de pCO 2agua indica que la mayoría de las plataformas continentales abiertas en regiones templadas y de alta latitud están subsaturadas con respecto a pCO 2aire durante todas las estaciones del año, aunque las plataformas de baja latitud parecen estar sobresaturadas (Chen y Borges 2009).Estos últimos autores indican que las plataformas continentales absorben el CO 2 atmosférico (entre 0.33 y 0.36 PgC año -1 ).
Con base en valores de pH y alkalinidad, Hidalgo-González et al. (1997) realizaron muestreos en el verano para generar datos de pCO 2agua y el flujo de CO 2 entre el mar y la atmósfera para la región de las islas grandes.El flujo de CO 2 calculado fue hacia la atmósfera y fue mayor durante las mareas posvivas (hasta 23 mmol m -2 d -1 ).Según Hidalgo-González et al. (1997), los flujos de CO 2 hacia la atmósfera en el invierno deberían ser mayores que los del verano debido a una menor estratificación del agua y a la mayor pCO 2agua en invierno que en verano.Se han calculado los siguientes valores máximos de pCO 2agua para la región de las islas grandes: 560 atm para octubre de 1985 (Zirino et al. 1997), 446 atm para julio de 1990 (Hidalgo-González et al. 1997), 560 atm para septiembre de 1996(Hernández-Ayón et al. 2007a) y 1200 atm para marzo de 2002 (Hernández-Ayón The Gulf of California gains heat from the atmosphere (Bray 1988, Lavín andOrganista 1988).This heat has to be exported to the Pacific Ocean somehow, otherwise the gulf's temperature would be increasing (Lavín et al. 1997).Water exchange between the Gulf of California and the Pacific has a vertical component that consists of less dense, warmer, and saltier surface and near-surface waters poor in nutrients (Álvarez-Borrego 2012) and dissolved inorganic carbon (DIC) flowing out from the gulf into the Pacific Ocean, and to balance this water flow, relatively deep, denser, colder, fresher, and nutrient-and DIC-rich waters flow into the gulf.These flows, at the surface and at depth, are present throughout most of the gulf (Bray 1988, Marinone 2003).
The vertical component of water exchange (VCWE) between the Gulf of California and the Pacific Ocean is the laterally integrated transport at the entrance to the gulf.The VCWE does not consist of a vertical component of advection.The horizontal component of water exchange is eliminated when integrating velocity across the mouth of the gulf for each depth (Álvarez-Borrego 2012).A restriction for this kind of estimate is that the water flux out has to be equal to the flux into the gulf, at the entrance to the gulf, because of the conservation of mass principle.Nevertheless, the transport of dissolved constituents, like nutrients and DIC, does not balance at the entrance to the gulf because their concentrations are higher at depth than at the surface.The DIC and nutrients that enter the gulf from the Pacific Ocean at 200-600 m depth are transported throughout the gulf, and through mixing and upwelling, they are input to the euphotic zone to be consumed by new phytoplankton production.Once at the mixed layer, DIC also participates in gas exchange with the atmosphere (fig.1).
There is no published VCWE estimate by physical methods for the entrance to the Gulf of California.Álvarez-Borrego (2012) applied a biogeochemical method to produce a VCWE estimate for the entrance to the gulf of 0.67 ± 0.10 Sv (the number after ± is one standard error = s n -0.5 ), going out to the Pacific at 0-200 m depth and into the gulf at 200-600 m depth.This VCWE is only ~7% of the whole water exchange that has a large horizontal component (e.g., Roden 1972).Álvarez-Borrego's (2012) method is based on assuming steady state for the concentration of nutrients in the gulf.He used the net average annual input of nitrate needed to support new phytoplankton production for the whole gulf (P NEW , kg C yr -1 ) to estimate VCWE.New production is the fraction of total phytoplankton production supported by the input of nitrate from outside the euphotic zone (Dugdale and Goering 1967).
On the other hand, Álvarez-Borrego and Giles-Guzmán (2012) used the average net annual input of dissolved Si needed to support the production of biogenic silica particles preserved in the sediments of the Gulf of California to produce a VCWE independent estimate of 0.23 ± 0.02 Sv for the entrance to the gulf.Hidalgo-González and Álvarez-Borrego ( 2004) estimated new phytoplankton production in different et al. 2007b).Estos valores indican que la región de las islas grandes es un área que actúa como una fuente casi permanente de CO 2 hacia la atmósfera a lo largo del año.
El componente vertical del intercambio de agua (CVIA) entre el golfo de California y el océano Pacífico es el transporte integrado lateralmente en la entrada al golfo.El CVIA no consiste de un componente vertical de advección.El componente horizontal del intercambio de agua es eliminado al integrar la velocidad a través de la boca del golfo para cada profundidad (Álvarez-Borrego 2012).Una restricción para este tipo de estimación es que el flujo de agua hacia afuera tiene que ser igual al flujo de agua hacia dentro del golfo, en la boca, debido al principio de conservación de la masa.No obstante, el transporte de constituyentes disueltos, tales como nutrientes y CID, no es balanceado a la entrada del golfo porque sus concentraciones son mayores a profundidad que en la superficie.El CID y los nutrientes que entran al golfo del océano Pacífico entre los 200 y 600 m de profundidad son transportados a lo largo del golfo, y mediante surgencias y mezcla son acarreados a la zona eufótica donde se consumen por la producción fitoplanctónica nueva.El CID, cuando ya se encuentra en la capa de mezcla, también participa en el intercambio de gases con la atmósfera (fig.1).
Álvarez-Borrego and Giles-Guzmán (2012) used their VCWE value to estimate the average net annual input of nitrate from the Pacific Ocean to the Gulf of California, and with Redfield's ratio they transformed it to P NEW for the whole gulf ((9.26 ± 3.18) × 10 9 kg C yr -1 ).In a similar manner, the average net annual input of DIC from the Pacific into the gulf (DIC NET INPUT ) may be estimated.Comparing DIC NET INPUT with the annual average P NEW for the whole gulf, an inference may be done about the gulf acting as a sink or source of CO 2 .If DIC NET INPUT is larger than the annual P NEW , the excess CO 2 has to flow from the gulf's water to the atmosphere and vice versa (fig.1).This is a new method to estimate water-air CO 2 fluxes throughout the gulf.It is an alternative to the traditional method that uses pCO 2w and pCO 2air values to estimate the CO 2 fluxes.Data on the dissolved inorganic carbon system of the Gulf of California are very scarce.To generate proper pCO 2w values for the whole gulf would be very expensive and time-consuming, Por otro lado, Álvarez-Borrego y Giles-Guzmán (2012) usaron el aporte neto promedio anual del Si disuelto requerido para sostener la producción de partículas de sílice biogénico preservadas en los sedimentos del golfo de California para producir una estimación independiente del CVIA de 0.23 ± 0.02 Sv para la entrada del golfo.Hidalgo-González y Álvarez-Borrego (2004) estimaron la producción fitoplanctónica nueva en diferentes regiones del golfo mediante imágenes de satélite y modelos, y sus resultados fueron usados por Álvarez-Borrego (2012) para deducir un promedio anual de P NUEVA igual a (31.04 ± 1.58) × 10 9 kg C para todo el golfo.Según Álvarez-Borrego y Giles-Guzmán (2012), la diferencia entre su valor de CVIA y el de Álvarez-Borrego (2012) sugiere que Hidalgo-González y Álvarez-Borrego ( 2004) posiblemente sobrestimaron la producción fitoplanctónica nueva.
El objetivo de este trabajo es cuantificar el CID APORTE NETO y estimar la cantidad promedio anual de CO 2 que fluye a través de la interfase aire-mar en todo el golfo para dos escenarios, cada uno con un valor diferente del CVIA.Nuestros objetivos son: (a) proporcionar una respuesta a la pregunta de si el golfo es sumidero o fuente de CO 2 ; y (b) proporcionar una primera aproximación del valor promedio anual del flujo agua-aire de CO 2 para todo el golfo, para los dos escenarios.Finalmente, se comparan nuestros valores con valores del flujo agua-aire de CO 2 documentadas en la literatura para regiones oceánicas para determinar cuál de los dos escenarios es el más cercano a la realidad.Figura 1. Diagrama simplificado que muestra el aporte neto de carbono inorgánico disuelto (CID APORTE NETO ) y cómo se puede comparar con la producción fitoplanctónica nueva (P NUEVA ) para inferir si el golfo es un sumidero o una fuente de CO 2 .CID APORTE possibly requiring a large number of cruises to obtain average representative values for the whole year.

La entrada al golfo de
The purpose of this work is to quantify DIC NET INPUT and to estimate the average annual amount of CO 2 that is flowing across the air-seawater boundary throughout the gulf for two scenarios, each with a different VCWE value.Our objectives are: (a) to provide an answer to the question of whether the gulf is a sink or source of CO 2 ; and (b) to provide first approximation estimates of the mean annual value of this water-air CO 2 flux for the whole gulf, for the two scenarios.Finally, water-air CO 2 flux values reported in the literature for oceanic regions were compared to our results to determine which of our two scenarios is closest to reality.

MATERIALS AND METHODS
The entrance to the Gulf of California is considered a place where DIC is input from the Pacific Ocean into the gulf and from there it is transported throughout the gulf.Steady state of DIC profiles throughout the gulf is assumed at the scale of annual averages.The method only requires an annual average DIC profile for the entrance to the gulf.The gulf is considered a box open to the Pacific for water and dissolved components exchange, and also open to the atmosphere for gas exchange (fig.1).Once inside the gulf, DIC NET INPUT has to be balanced by consumption by P NEW and water-air CO 2 exchange: If CO 2EXCHANGE is positive there is an excess of DIC NET INPUT after nitrate has been exhausted by P NEW , and CO 2 flows from the water to the atmosphere.If CO 2EXCHANGE is negative there is a deficit of DIC NET INPUT and CO 2 flows from the atmosphere to the water.This occurs regardless of the particular characteristics of DIC profiles in different regions of the gulf.Gas exchange occurs with different intensities in different regions of the gulf according to their particular physical dynamics (mixing and upwelling), but we intend to produce an average CO 2EXCHANGE estimate for the whole gulf.Two possible scenarios were used, one using Álvarez-Borrego's ( 2012 (0-200 and 200-600 m), for the mouth of the gulf, are needed.Arithmetic means do not properly represent AADIC for each layer.These AADIC have to be weighted averages, transportado a lo largo del golfo.Se supone que los perfiles de CID en todo el golfo son constantes en una escala de promedios anuales.Este método sólo requiere un perfil promedio anual de CID para la entrada del golfo.El golfo se considera una caja abierta al Pacífico para el intercambio de agua y componentes disueltos, así como abierta a la atmósfera para el intercambio de gases (fig.1).Una vez adentro del golfo el CID APORTE NETO tiene que ser balanceado por su consumo a través de la P NUEVA y el intercambio aguaaire de CO 2 : Si CO 2INTERCAMBIO es positivo hay un exceso de CID APORTE NETO después de que el nitrato ha sido consumido por P NUEVA , y el CO 2 fluye del agua a la atmósfera.Si CO 2INTERCAMBIO es negativo hay un déficit de CID APORTE NETO y el CO 2 fluye de la atmósfera al agua.Esto sucede a pesar de las características particulares de los perfiles de CID en las diferentes regiones del golfo.El intercambio de gases sucede con intensidades diversas en las diferentes regiones del golfo según su dinámica física particular (mezcla y surgencias), pero nuestro objetivo es producir una estimación promedio de CO 2INTERCAMBIO para todo el golfo.Se usaron dos posibles escenarios, uno con el valor de CVIA de Álvarez-Borrego (2012) y el otro con el de Álvarez-Borrego y Giles-Guzmán (2012).El CID APORTE NETO a través de la boca del golfo de California se calcula como la diferencia entre el transporte hacia dentro del golfo en la capa profunda (200-600 m) (CID INGRESO ) y el transporte hacia afuera del golfo (CID EGRESO ) en la capa superficial (0-200 m) (CID APORTE NETO = CID INGRESO -CID EGRESO ).Se requieren promedios anuales de las concentraciones de CID (PACID) apropiados para cada capa (0-200 y 200-600 m) en la boca del golfo.Las medias aritméticas no representan PACID correctamente.Estos PACID tienen que ser promedios ponderados, donde el factor de ponderación es el transporte de agua integrado horizontalmente en cada profundidad (T INTz , m 2 s -1 ).
where the weighting factor is the horizontally integrated water transport at each depth (T INTz , m 2 s -1 ).
Following Álvarez-Borrego (2012), as a first approximation, relative values representing the shape of an average vertical profile of T INTz may be used as the weighting factor.A similar shape to that of Bray's (1988)  The National Oceanographic Data Center (http:// www.nodc.noaa.gov/)was accessed to obtain pH and total alkalinity (TA) data to generate an annual average DIC profile for the mouth of the gulf and for the 0-600 m depth interval.Unfortunately, only three useful hydrographic stations were retrieved (April 1994 NOAA cruise).Data from two hydrographic stations of the April 1974 Alexander Agassiz cruise (Gaxiola-Castro et al. 1978), from four stations of the November 1985 DeSteiguer cruise (generated by Alberto Zirino and provided by José Martín Hernández-Ayón, pers.comm., IIO-UABC, Ensenada), from three stations of the July 1990 DeSteiguer cruise (our own data), and from one station of the Francisco de Ulloa September 1997 cruise (also provided by JM Hernández-Ayón) made it possible to generate a total of 13 DIC profiles (for sampling locations see figure 3).Five of the stations were occupied during "winter" and eight during "summer".Ideally, these stations should have been on a line connecting Cape San Lucas with Cape Corrientes, defining the entrance to the gulf.
Precise and reliable pH data are usually reported because the methodology is relatively simple.But TA data may be unreliable because the use of hydrochloric acid solutions demands very good standardizations.Therefore, surface specific alkalinity (SA = TA/chlorinity = 1.80655 × TA/S, where S is salinity) was assumed to be constant and equal to 119 mol of charge kg -1 , which is the mean of Culberson's (1972) value (120 mol kg -1 ) from the Pacific Ocean off the entrance to the gulf and that of Gaxiola-Castro et al. (1978) (118 mol kg -1 ).This surface specific alkalinity value (119 mol kg -1 ) also results from surface TA values calculated with Lee et al.'s (2006) global relationships (the expression for their oceanic region 1).Specific alkalinity was assumed to change with depth at the same rate as that reported by Gaxiola-Castro (1978) (i.e., 122 μmol kg -1 for 500-600 m depth).Salinity values for each hydro-station were used to calculate the TA profiles (TA = S × SA/ 1.80655).Skirrow's (1965) expression was used to calculate DIC, using Edmond and Gieskes' (1970) apparent dissociation constant of boric acid, and Mehrbach et al.'s (1973) two Se consultó el National Oceanographic Data Center, operado por la National Oceanic and Atmospheric Administration (NOAA) de los Estados Unidos (http:// www.nodc.noaa.gov/),para obtener datos de pH y alcalinidad total (AT) y generar un perfil promedio anual de CID para la boca del golfo y para 0-600 m de profundidad.Desafortunadamente, sólo se encontraron tres estaciones hidrográficas útiles (crucero de la NOAA de abril de 1994).Fue posible generar un total de 13 perfiles de CID a partir de datos de dos estaciones hidrográficas del crucero del B/I Alexander Agassiz de abril de 1974(Gaxiola-Castro et al. 1978), de cuatro estaciones del crucero del B/I DeSteiguer de noviembre de 1985 (generados por Alberto Zirino y proporcionados por José Martín Hernández-Ayón, com.pers., IIO-UABC, Ensenada), de tres estaciones del crucero del B/I DeSteiguer de julio de 1990 (nuestros datos), y de una estación del B/I Francisco de Ulloa de septiembre de 1997 (proporcionados por JM Hernández-Ayón) (la ubicación de las estaciones de muestreo se muestra en la figura 3).Cinco de las estaciones fueron ocupadas en "invierno" y ocho en "verano".Idealmente, estas estaciones deberían de haber estado localizadas en una línea conectando Cabo San Lucas con Cabo Corrientes, que define la entrada al golfo.
El perfil promedio anual de T INT(Z) (fig.2) y el de CID (fig.4) fueron combinados para generar los promedios  (4) and DIC NET INPUT was compared to P NEW to infer if the gulf acts as a source or sink of CO 2 .
In order to explore different possibilities for the air-sea exchange of CO 2 in the gulf, two scenarios were used: in the first one the VCWE was equal to 0.67 ± 0.10 Sv and P NEW was equal to (31.04 ± 1.58) × 10 9 kg C yr -1 (Álvarez-Borrego 2012); in the second scenario the VCWE was equal to 0.23 ± 0.02 Sv and P NEW was equal to (9.26 ± 3.18) × 10 9 kg C yr -1 (Álvarez-Borrego and Giles-Guzmán 2012).The annual average vertical profiles of T INT(Z) (relative values) and DIC were the same for both scenarios.Thus, the values of AADIC (0-200) and AADIC  were also the same for both scenarios.
uncertainty of the VCWE value (± 0.10 Sv in one case and ± 0.02 Sv in the other) should not be taken into account.For the same reason, the uncertainty of P NEW , as calculated by Álvarez-Borrego and Giles-Guzmán ( 2012), depends only on the uncertainty of the weighted average values of NO 3 for each layer and not on the uncertainty of VCWE.Thus, when subtracting P NEW from DIC NET INPUT to infer if there is an excess of DIC or vice versa, instead of using Álvarez-Borrego and Giles-Guzmán's (2012) P NEW value ((9.26 ± 3.18) × 10 9 kg C yr -1 ), the recalculated value (9.26 ± 1.09) × 10 9 kg C yr -1 was used.

DISCUSSION
In the Peruvian and Chilean coastal upwelling systems, known to be among the most productive oceanic areas worldwide, oversaturation of CO 2 with respect to the atmosphere has been reported with pCO 2w values of up to 1200 atm (Borges 2005, Torres et al. 2011).The Gulf of California is a (2012) ((9.26 ± 3.18) × 10 9 kg C año -1 ), se usó el valor recalculado de (9.26 ± 1.09) × 10 9 kg C año -1 .

DISCUSIÓN
Para los sistemas de surgencia de Perú y Chile, una de las áreas oceánicas más productivas del mundo, se ha registrado sobresaturación de CO 2 con respecto a la atmósfera, con valores de pCO 2agua de hasta 1200 atm (Borges 2005, Torres et al. 2011).El golfo de California es un sistema de surgencia costera y actúa como una fuente de CO 2 a la atmósfera en ambos escenarios.Esto se debe a que la pendiente de la relación CID-nitrato es mayor que la razón de Redfield en aguas subsuperficiales y profundas del golfo (fig.5).Cuando aguas subsuperficiales y relativamente profundas son acarreadas a la zona eufótica por surgencias y/o mezcla, después de que todo el nitrato es consumido por la producción fitoplanctónica nueva, CID queda como un exceso, como ya ha coastal upwelling system and it behaves as a source of CO 2 to the atmosphere in both scenarios.This is because the slope of the DIC-nitrate relationship is greater than Redfield's ratio in subsurface and deep waters of the gulf (fig.5).When subsurface and relatively deep water are carried to the euphotic zone by upwelling and/or mixing, after all the nitrate is consumed by new phytoplankton production, DIC is left as an excess, as indicated by Borges (2011).The Gulf of California is a source of CO 2 to the atmosphere because of the DIC-nitrate relationship, regardless of the VCWE value.
At depth, respiration increases DIC and NO 3 presumably following Redfield's ratio, but respiration is not the only process affecting DIC and NO 3 .The DIC excess with relation to NO 3 is due to the dissolution of calcium carbonate skeletons at depth (Park 1965), to denitrification processes associated with the oxygen minimum zone in the eastern Pacific Ocean (Thomas 1966), and to differences of preformed DIC (Park 1965).The processes of calcium carbonate dissolution and denitrification occur along the trajectory of the water masses from their origin at high latitudes and not only in the gulf.Calcium carbonate dissolution occurs in waters deeper than ~200 m because of undersaturation with respect to both aragonite and calcite in the Gulf of California (Gaxiola-Castro et al. 1978) and in the whole northeastern Pacific Ocean (Park 1968).Denitrification occurs at 100-800 m depth in the eastern Pacific Ocean because of nitrate reduction by bacteria sido indicado por Borges (2011).El golfo de California es una fuente de CO 2 a la atmósfera debido a la relación CIDnitrato, sin importar el valor de CVIA.
when dissolved oxygen concentration is very low (Thomas 1966).Since the deeper the water masses the lower their temperature, deep gulf waters had a larger solubility of gases (including CO 2 ) at their latitude of origin, when they were in contact with the atmosphere (Culberson and Pytkowicz 1970), and hence greater preformed DIC than those of shallow depths in the gulf.The DIC required by the average annual P NEW for the whole Gulf of California has to be compensated by an export from the gulf to the Pacific Ocean, and inside the gulf from the water column to the sediments.The majority of organic carbon produced by P NEW is exported from the gulf to the Pacific Ocean in the form of dissolved organic carbon (Álvarez-Borrego 2012).Based on estimates by Thunell et al. (1993), export of particulate organic carbon to the sediments is only ~3% of the carbon required by P NEW .Thus the "biological pump" is taking every year an average of between ~280 × 10 9 and 930 × 10 9 grams of particulate organic carbon to the gulf's bottom.
A sensitivity analysis was performed to assess the effect of changing the specific alkalinity profile and the VCWE value on DIC NET INPUT .Also, an exercise was run to see the effect of changing the average DIC profile on DIC NET INPUT , equilibrating the mixed layer waters with the 2013 NOAA atmospheric average pCO 2 value of 396 ppm.If the surface specific alkalinity is changed to 118 mol kg -1 instead of 119, and its rate of change with depth is maintained the same as in the previous calculations, the weighted averages of DIC   Since our data are from years in the period 1974-1997, they have the effect of a large fraction of the anthropogenic CO 2 that has been absorbed by this region of the ocean.At the entrance to the gulf, waters deeper than ~33 m (average mixed layer depth) have pCO 2w higher than 396 ppm (not illustrated).Thus, as an average, the atmospheric 2013 pCO 2 value only affects pCO 2w from the surface to ~33 m.If the mixed layer pCO 2w is equilibrated with the atmospheric 2013 value, surface pH decreases 0.02 units as an average and AADIC (0-200) is 2.104 ± 0.012 mol m -3 , and the respective weighted average for 200-600 m remains the same as the original value.Under these conditions, in both the first and second scenarios, DIC NET INPUT values are not significantly different from the original values.Thus, up to 2013, la alcalinidad específica superficial a 120 mol kg -1 , PACID (0-200) y PACID (200-600) cambian a valores mayores que los originales.Una vez más, los resultados en ambos escenarios para CID APORTE NETO son básicamente los mismos que los valores originales.Por ende, el valor de CID APORTE NETO no varía significativamente con los cambios de los perfiles de alcalinidad.La relación entre el valor de CID APORTE NETO y CVIA es directo y lineal.Si el CVIA cambia por un cierto porcentaje, CID APORTE NETO también cambia por el mismo porcentaje y en la misma dirección.
El aporte neto de nutrientes y CID no es transportado a la zona eufótica de manera homogénea a lo largo del golfo ya que hay diferencias regionales de su dinámica física.Como lo menciona Álvarez-Borrego (2012), las surgencias en la corrections to the estimates of DIC NET INPUT due to the fact that surface waters of the gulf tend to be equilibrated with an increasing atmospheric pCO 2 may be considered negligible given our large standard errors.On the other hand, it is reasonable to assume that anthropogenic CO 2 stored in the Gulf of California is practically the same as that of the adjacent Pacific Ocean (~15 mol CO 2 m -2 , Sabine et al. 2004), so that it would not make any appreciable difference in the exchange between the two.This storage has accumulated for the last one and a half century and a large fraction of it must be part of the DIC profile that we are using.
The choice of two layers, 0-200 and 200-600 m, is not the only option.Marinone (2003) used a three dimensional model to predict the circulation of the Gulf of California.When integrating the average annual circulation predicted by this model across the mouth of the gulf, the VCWE between the gulf and the Pacific Ocean results in four layers: 0-200, 200-600, 600-1200, and 1200-2600 m.However, there is no physical known mechanism that transports nutrients and DIC from very deep waters, such as those below 600 m, to the euphotic zone to be used by phytoplankton.Besides, in the Midriff Islands region and the central gulf the depth of sills between basins is not greater than ~500 m.
The net nutrient and DIC input to the gulf is not transported to the euphotic zone homogeneously throughout it because there are regional differences in its physical dynamics.As Álvarez-Borrego (2012) indicated, upwelling along most of the eastern gulf during "winter" conditions, cyclonic eddies in different parts of the gulf, and strong mixing off the Midriff Islands throughout the year (mainly with spring tides and during "winter") are mechanisms that transport deep nutrient-and DIC-rich waters to the euphotic zone.The Midriff Islands region is the area within the gulf with the highest CO 2 water-to-air fluxes throughout the year; it is the area with the largest pCO 2w values, as mentioned above (e.g., Hernández-Ayón et al. 2007b).The "winter" upwelling region off the eastern coast might be the area with the second highest CO 2 water-to-air fluxes in the gulf.
These estimates of DIC NET INPUT as annual averages are first approximations to reality, and there are opportunities for future works on its time variability such as seasonal changes and those caused by the incidence of El Niño events, as it was indicated by Álvarez-Borrego (2012) for the net input of nitrate from the Pacific Ocean to the gulf.
It would be of interest to compare the results obtained here, comparing DIC NET INPUT with P NEW , with those obtained with the traditional method that uses pCO 2w and pCO 2air values to estimate water-to-air CO 2 fluxes; however, as mentioned above, data on the DIC system of the whole Gulf of California are very scarce.To generate pCO 2w values for the whole gulf in enough quantities, and to obtain average representative values for the different regions and for the whole year, will be very expensive and time consuming, possibly requiring a large number of cruises.This is a task for future work.costa este del golfo durante condiciones de "invierno", los remolinos ciclónicos en diferentes partes del golfo y la mezcla intensa en la región de las islas grandes durante todo el año (principalmente con mareas vivas y durante el "invierno") son mecanismos que transportan aguas profundas ricas en nutrientres y CID a la zona eufótica.La región de las islas grandes es la zona del golfo con los mayores flujos de CO 2 del agua a la atmósfera durante todo el año; como ya se mencionó, es la zona con los mayores valores de pCO 2agua (e.g., Hernández-Ayón et al. 2007b).La región de surgencias de "invierno" frente a la costa oriental podría ser la zona del golfo con los segundos flujos más altos de CO 2 del agua a la atmósfera.
Transforming the results of the gulf's water-to-air CO 2 fluxes for each scenario into grams per square meter per year, in the first scenario the average value is 123.5 ± 41.8 g m -2 yr -1 and in the second scenario it is 52.1 ± 18.0 g m -2 yr -1 .The maxima water-to-air annual average CO 2 fluxes of the world's ocean, as reported by Takahashi et al. (2009), are between 24 and 108 g m -2 yr -1 , in places like the eastern equatorial Pacific Ocean, which has continuous upwelling.The Gulf of California is almost at equilibrium with the atmosphere during "summer" conditions, with exception of the Midriff Islands region, and during "winter", upwelling occurs mostly on the eastern side.Thus, an annual average CO 2 flux per unit area for the whole gulf cannot be larger than the maximum for places like the eastern equatorial Pacific.This indicates that the second scenario is more acceptable with an average CO 2 output to the atmosphere of (7.66 ± 2.65) × 10 12 g C yr -1 for the whole gulf, and that the VCWE value of (0.23 ± 0.02) Sv is closer to reality than (0.67 ± 0.10) Sv.This CO 2 input from the gulf to the atmosphere is only ~1.7% of the annual CO 2 output to the atmosphere of the whole eastern equatorial Pacific (0.48 Pg C yr -1 , Takahashi et al. 2009), which has a very large area compared to that of the gulf.But, when adding up all the coastal areas of the world's oceans, the figure may be a very significant one (i.e., Chen and Borges 2009).

Figure 1 .
Figure 1.Simplified diagram showing the net input of dissolved inorganic carbon (DIC NET INPUT ) and how it can be compared with new phytoplankton production (P NEW ) to infer if the gulf is a sink or source of CO 2 .DIC NET INPUT = DIC INPUT -DIC OUTPUT .When P NEW > DIC NET INPUT the gulf is a sink of CO 2 ; when P NEW < DIC NET INPUT the gulf is a source of CO 2 .
) VCWE value and the other using that ofÁlvarez-Borrego and Giles-Guzmán (2012).DIC NET INPUT through the entrance to the Gulf of California is calculated as the difference between the transport into the gulf in the deep layer (200-600 m) (DIC INPUT ) and the transport out of the gulf in the surface and subsurface water layer (0-200 m) (DIC OUT- PUT ) (DIC NET INPUT = DIC INPUT -DIC OUTPUT ).Proper annual averages of DIC concentrations (AADIC) for each layer average integrated (across the central gulf) transport profile, and Marinone's (2003) results on heat and salt transport were used to generate a T INTz profile with relative values (T INT(Z) ) for 0-600 m, with zero relative integrated transport at 200 and 600 m (fig.2, taken from Álvarez-Borrego 2012).Depths with zero T INT(Z) (200 and 600 m) are not necessarily without motion; they are depths with equal input and output of water (Álvarez-Borrego 2012).

Figure 2 .
Figure 2. Shape of the annual average of the vertical distribution of water transport integrated across the mouth of the Gulf of California (T INT(Z) , m 2 s -1 , relative values) (similar to the one proposed by Bray 1988 and modified taking into consideration the result obtained by Marinone 2003) (taken from Álvarez-Borrego 2012).
of carbonic acid as modified byPlath et al. (1980).These expressions require pH data in the National Bureau of Standards (NBS) scale.The 1974The  ,  1985The  , and 1990   pH data are in the NBS scale, the 1994 data are in the seawater scale, and the 1997 data are in the total hydrogen ion scale.Because of the relative unimportance of the fluoride ion, the total and seawater scales differ only very slightly(Zeebe and Wolf-Gladrow 2001).Thus, the 1994 and 1997 pH data were transformed to the NBS scale followingMillero et al. (1988), as if both sets of data were in the total hydrogen ion scale.Since DIC does not change with temperature in a closed reservoir, the measured pH was used with the laboratory temperature values to calculate DIC.The annual average T INT(Z) profile (fig.2) and the one for DIC (fig.4) were combined to generate weighted averages for DIC for each layer, 0-200 and 200-600 m: AADIC (0-200) = Σ(DIC (Z) × T INT(Z) )/Σ(T INT(Z) ), with Z changing from 0 to 200 m, and similarly for 200-600 m.The average output of DIC from the gulf to the Pacific Ocean in the 0-200 m layer (DIC OUTPUT ) was calculated multiplying AADIC (0-200) (mol m -3 ) by the water transport (10 6 × VCWE m 3 s -1 ), and similarly for the average input of DIC from the Pacific Ocean into the gulf in the 200-600 m layer (DIC INPUT ).Each of the two results was transformed into an annual DIC flux: DIC OUTPUT mol yr -1 = (AADIC (0-200) mol m -3 )(10 6 × VCWE m 3 s -1 )(86,400 s d -1 )(365 d yr -1 ) ( 2 ) DIC INPUT mol yr -1 = (AADIC (200-600) mol m -3 )(10 6 × VCWE m 3 s -1 )(86,400 s d -1 )(365 d yr -1 )

Figure 3 .
Figure 3. Location of hydrographic stations.The insert in the upper right side shows the whole Gulf of California with the location of the Midriff Islands.

Figure 4 .
Figure 4. Annual average of the vertical distribution of dissolved inorganic carbon (DIC, mol L -1 ) for the mouth of the Gulf of California.Horizontal bars represent ± one standard error (s n -0.5 ).

Figure 5 .
Figure 5. Relationship between dissolved inorganic carbon (DIC) and NO 3 for the 0-600 m depth interval at the entrance to the Gulf of California.
and AADIC (200-600) ) change to lower values than the original ones.The results for DIC NET INPUT in both scenarios are basically the same as the original values.If the surface specific alkalinity is changed to 120 mol kg -1 , AADIC (0-200) and AADIC (200-600) change to higher values than the original ones.Again, the results for DIC NET INPUT in both scenarios are basically the same as the original values.Thus, the DIC NET INPUT value does not vary significantly with changes of the alkalinity profiles.The relation between the VCWE value and DIC NET INPUT is direct and linear.If VCWE is changed by a certain percentage, DIC NET INPUT does it by the same percentage and in the same direction.