Seasonal variation of geostrophic velocity and ... - Wiley Online Library

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 109, C07008, doi:10.1029/2003JC002124, 2004

Seasonal variation of geostrophic velocity and heat flux at the entrance to the Gulf of California, Mexico A. S. Mascarenhas Jr.,1 Ruben Castro,2 Curtis A. Collins,3 and Reginaldo Durazo2 Received 11 September 2003; revised 1 April 2004; accepted 5 May 2004; published 3 July 2004.

[1] The divergence of the surface heat flux in the Gulf of California is driven by the

exchange of waters with the Pacific Ocean [Castro et al., 1994]. To estimate these exchanges, geostrophic velocities and heat fluxes were computed from nine sections of closely spaced conductivity-temperature-depth stations across the entrance to the Gulf of California between 1992 and 1999. The mean geostrophic velocity was composed of two alternating cores of inflow and outflow. The two cores that were adjacent to either coast were broader and contained the highest inflow (0.40 m/s) and outflow (0.25 m/s) velocities, supporting the general idea of inflow along Sinaloa and outflow along Baja California (BC). During winter and spring the baroclinic outflow usually occurred near BC, and the baroclinic inflow occurred either through the center of the section and/or along the Sinaloa coast. Both inflow and outflow cores were 45 km wide and extended deeper than 700 dbar. Summer and fall showed a more complex pattern, with alternating cores of inflow and outflow but with inflow along Sinaloa on all cruises. During May the inflow was mainly in the center of the section, while outflow was concentrated along BC. The heat transport associated with the geostrophic flow was calculated and compared with estimates derived from surface heat budgets for the gulf [Castro et al., 1994]. Both the geostrophic flow and heat transport exhibited (for the first empirical orthagonal function mode) a strong seasonal signal with the maximum amplitude in May and the minimum amplitude in late October. A seasonal fit to the net heat transport had an amplitude of 50 1012 W and phase of 133 days, in good agreement with other authors, further validating the geostrophic velocity estimates. The heat added to (subtracted from) the Pacific Ocean accelerates (decelerates) alongshore currents and appears to INDEX TERMS: 4536 Oceanography: Physical: propagate 10 westward as a Rossby wave. Hydrography; 4512 Oceanography: Physical: Currents; 4243 Oceanography: General: Marginal and semienclosed seas; 4223 Oceanography: General: Descriptive and regional oceanography; KEYWORDS: geostrophic heat flux, Gulf of California, geostrophic current Citation: Mascarenhas, A. S., Jr., R. Castro, C. A. Collins, and R. Durazo (2004), Seasonal variation of geostrophic velocity and heat flux at the entrance to the Gulf of California, Mexico, J. Geophys. Res., 109, C07008, doi:10.1029/2003JC002124.

1. Introduction [2] Lying between the arid peninsula of Baja California (BC) to the west and the equally arid states of Sonora and Sinaloa to the east, the Gulf of California comprises a large evaporative basin, the only such marginal sea in the Pacific Ocean. Although its evaporation rate, 0.61 m/yr [Beron-Vera and Ripa, 2000], is comparable to that of the Mediterranean and Red Seas, the gulf differs from these seas because it actually gains heat at an annual rate of 1 Instituto de Investigaciones Oceanolo´gicas, Universidad Autonoma de Baja California, Ensenada, Baja California, Mexico. 2 Facultad de Ciencias Marinas, Universidad Autonoma de Baja California, Ensenada, Baja California, Mexico. 3 Department of Oceanography, Naval Postgraduate School, Monterey, California, USA.

Copyright 2004 by the American Geophysical Union. 0148-0227/04/2003JC002124

118 W/m2 [Castro et al., 1994]. This water loss and heat gain result in modification of water properties, creation of unique water masses, and strong exchange with the Pacific Ocean. These exchanges take place in Pescadero Basin, which lies at the entrance to the gulf (Figure 1). Here depths exceed 2500 m, and the distance between landmasses is 200 km, allowing for free exchange between the waters of the Pacific and the gulf. Roden [1972] was the first to make high-resolution hydrographic observations across the entrance to the Gulf of California. His stations were spaced 9 km apart and revealed a series of geostrophic velocity jets that were 30 km wide and extended to 700 m depth. While such highly resolved sections are not necessary for resolving geostrophic transports if the positions of dynamic troughs and ridges are known, quantities that depend upon the velocity field, such as transport of energy or fresh water, require resolution of the density field at the scale of the internal Rossby radius. Here we present the results for a hydro-

C07008

1 of 9

C07008

MASCARENHAS ET AL.: VELOCITY AND HEAT FLUX VARIATION

C07008

exchanges on the circulation in the Pacific Ocean is discussed.

2. Methods

Figure 1. Map of Pescadero Basin at the entrance to the Gulf of California. The line indicates the location of hydrographic measurements. Isobaths are contoured at 200 m intervals.

graphic section at the entrance to the gulf that was occupied nine times during 1992 – 1999 with closely spaced hydrographic (conductivity-temperature-depth (CTD)) stations (Figure 1). These sections resolve geostrophic flow motions within half the radius of deformation, of order 24 km, which is also the order of magnitude of the radius of the mesoscale gyres that occurs in the entrance of the gulf in which we are interested. The purpose of this paper is to show the resulting distribution of geostrophic velocity and the transport of internal energy. The distribution of geostrophic velocity is relevant because the Gulf of California is one of most productive regions in the world [Hammann et al., 1998]. Eggs, larvae, and nutrients are carried across the entrance of the gulf by currents which are in approximate geostrophic equilibrium. On the other hand, the thermodynamics of the circulation as well as the variation of the physical properties of gulf waters are closely related to the heat fluxes [Ripa, 1997; Beron-Vera and Ripa, 2002; Beier, 1997]. Furthermore, both the velocity and the heat flux across the entrance to the gulf need to be known as boundary conditions for numerical models of the circulation of the gulf. [3] The paper is organized as follows: First, the geostrophic velocity field is described using individual sections. Means and the standard error of the mean for these sections are determined, and anomalies from these means are also shown. The variability of the geostrophic velocity is described using empirical orthogonal function analysis. Mass transports are calculated. Next, energy transports are computed, and their seasonal variability is compared with those obtained from observations of heating within the gulf and from model calculations. Finally, the possible impact of these energy and freshwater

[4] The nine cruises are listed in Table 1. CTD stations were collected along the track shown in Figure 1, not always at the same positions but with spacing between the stations of 10– 15 km (about one fourth the typical internal Rossby radius of deformation for the region). The entrance to the Gulf of California lies in Pescadero Basin, where the Alarcoń Seamount (ASM) is located approximately half of the entrance width into the gulf, at 23.62N and 108.75W. ASM has a horizontal scale of 20.5 km and rises 1000 m from the sea bottom in waters of 2600 m depth. The data collection, calibration, and processing procedures were reported by Rago et al. [1992] and Blanco et al. [1995]. During one of the cruises (October 1994) the CTD measurements were restricted to 1200 m by the length of the hydrographic wire. [5] Objective analysis was used to construct evenly spaced grids for all cruises in the section [Carter and Robinson, 1987]. For each individual cruise the fields of potential temperature, qxp( j), salinity, Sxp( j), and density anomaly, gqxp( j), were objectively mapped onto a 10 km 10 dbar ^xp( j) g grid, yielding estimated fields of ^qxp( j), S^xp( j),r and 2 using a Gaussian covariance function, eð =C Þ , where r2 = Dx2 + (bDp)2, b = 1 km/dbar, C = 25 km, p is pressure, and x is the distance along the section. The grid limits were 4 and 184 km from Cabo Pulmo, BC, and from the surface to 1500 m. These limits do not reach either coast or the bottom but do include most of the upper ocean area where the TS properties show the most variability, which is above the potential density anomaly 26.5 kg/m3, corresponding to the mean depth of 300 m [Castro et al., 1994, 2000]. Geopotential anomaly, F, was computed from the objective maps of temperature and salinity by integrating the specific volume anomaly upward from the bottom. Geostrophic velocity, vg, was then computed as x þ x Fð p; x Þ Fð p; x Þ i iþ1 i iþ1 ; vg p; ¼ fi þfiþ1 2 ðxiþ1 xi Þf

ð1Þ

2

where f is the Coriolis parameter and f is latitude.

3. Heat Flux for the Gulf [6] The high rates of heat gain through the surface in the gulf should be exchanged with the Pacific Ocean since, in Table 1. List of Cruises Carried Out at the Entrance to the Gulf of Californiaa Dates

Research Vessel

Casts

4 – 9 Feb.1994 28 – 29 Feb. 1999 2 – 4 May 1992 28 – 30 May 1998 2 – 4 Aug. 1998 13 – 15 Aug. 1995 31 Oct. to 1 Nov. 1994 19 – 20 Nov.1997 29 – 31 Dec.1992

R/V Francisco de Ulloa R/V Francisco de Ulloa USNS DeSteiguer R/V Francisco de Ulloa R/V Francisco de Ulloa R/V El Puma R/V A. Humboldt R/V Francisco de Ulloa R/V Point Sur

14 22 19 20 20 21 20 23 20

a Cruises are listed sequentially by month and day, and the number of conductivity-temperature-depth casts collected on each transect is given.

2 of 9


C07008

the mean, the gulf’s waters are not being warmed up and the horizontal advection of heat is carried out mainly by geostrophic currents [Ripa, 1997]. The heat flux for the gulf was formulated in the following manner: The volume, V, of the water in the Gulf of California has a free surface with the area S and a connection to the Pacific Ocean of area A. The precipitation and the river runoff are very small compared to the evaporation rate and are neglected. Precipitation should be considered only when there is exchange of mass through the free surface, as in the case of salt flux calculations. In the case of exchange of heat the effect of precipitation is minimal and restricted to the very surface layer. [7] The amount of heat in volume V will change because of the flow of heat across the boundaries of the volume V. The change of heat is represented by the following divergence of heat flux, HF (the amount of heat by unit time and unit volume): HF ¼ r ðrcP uT Þ;

ð2Þ

where r is the in situ density and cp is the specific heat of seawater. Neglecting any possible heat flux through the bottom of the gulf, the boundaries that allow exchange are the free surface, S, with the atmosphere and the entrance to the gulf, A, with the Pacific Ocean. The theorem of divergence is ZZZ

ZZ ðrcP uT Þ nds ¼

r ðrcP uT ÞdV ;

ZZ ðrcP uT Þ ndA þ

A

ZZZ ðrcP uT Þ ndS ¼

r ðrcP uT ÞdV ; V

S

ð4Þ

where n is the unity outward normal vector to the surface. The second integral term has been estimated using surface meteorological observations for the Gulf of California by many authors. By using the geostrophic velocity and temperature from the cruises listed in Table 1, the first integral term in equation (4) is estimated below. Since for the timescales considered here the gulf is neither warming nor cooling, the right-hand side of equation (4) is small or nearly zero. [9] The geostrophic velocity is constrained by ZZ A

rug ndA ¼ 0:

Z

L

Z

0

FH ¼

Cp qrVdA; 0

ð6Þ

H

where q is the potential temperature referenced to the surface, L is the length of the section, and H is the deepest common depth for two contiguous stations. V is a velocity that will bring the mass flux to zero, and Vij = vij + bi, where vij is the relative geostrophic velocity at level j in the ith station pair and bi is the unknown velocity at the reference level. By minimizing a quadratic measure, bi is obtained, requiring mass conservation through the entire section (the same approach used by Marinone and Ripa [1988]). Empirical orthogonal function (EOF) analysis was used to determine the dominant patterns of residual variance in a sequence of geostrophic velocity, V^ Gxp( j), and heat flux, _ F Hxp( j), sections across the entrance to the Gulf of California.

4. Results

where ds is the area around the volume, V, through which exchange occurs. If the entrance is closed and temporal changes are restricted to seasonal and interannual timescales, the time rate of change of heat in the gulf will only be due to heat exchange through the free surface, S. In other words, for seasonal variability the free surface will move vertically because of the amount of heat received from the atmosphere during the summer and lost to the atmosphere during the winter, but in the mean it will be zero. So the assumption that the vertical velocity, w, of the free surface is zero is not very restrictive, since exchange of mass is not considered. [8] When the exchange of heat with the Pacific Ocean is included, equation (3) becomes ZZ

This constraint requires that the mass flux divergence in the gulf be zero, which from the above arguments is not very restrictive. The computation of the heat flux using oceanographic observations may provide a better estimate of heat flux than the one using surface meteorological observations because of the lack of meteorological observations along the gulf. [10] The energy transport was computed as in equation (4) and following Bacon and Fofonoff [1996] is

ð3Þ

V

ds

C07008

ð5Þ

4.1. Geostrophic Velocity Field [11] Figure 2 shows the geostrophic velocity for each cruise referenced to 1500 dbar. The velocity field was highly complex and variable, with cores of inflow and outflow suggesting in some cases pairs of cyclonic and anticyclonic eddies. The structure was also marked by horizontal and vertical shear, with occurrence of jet-like flows in the surface layer (May 1992 and 1998, August 1995, November 1997, and December 1992). These jets were typically 10– 20 km wide and 200 – 300 m thick. [12] During winter and spring (an exception was February 1994) most of the baroclinic outflow (indicated in Figure 2 by blue) was found near BC, with a mean core 45 km wide which extended deeper than 700 dbar (1400 m in May 1998), and baroclinic inflow occurred either through the center of the section and/or along the Sinaloa coast. Summer and fall had a rather complex pattern with alternating cores of inflow and outflow but with an inflow along Sinaloa observed during all cruises. The maximum outflow velocities were found in spring (May 1998, 0.6– 0.7 m/s), with velocities of 0.2 m/s reaching to 800 dbar. During the 1992 El Ninõ– Southern Oscillation (ENSO) event the May outflow was adjacent to BC and extended to 70 km from the BC coast, and inflow was observed along the rest of the section to the Sinaloa shelf. The geostrophic velocity for November 1997, an ENSO year, also had high outflow along BC (up to 1 m/s at the surface) but was constrained to the upper 500 dbar. The velocity field was not as coherent in November 1997 as during May 1992. [13] The narrow and deep current velocity cores (0.1 – 0.2 m/s to 1400 dbar) suggest that the choice of a shallow reference level will underestimate the volume flux. Deep

3 of 9

C07008


C07008

Figure 2. Geostrophic velocity for a section across Pescadero Basin between Baja California and Sinaloa. Panels correspond to individual cruises, which are identified by month and year. Blue indicates flow out of the Gulf of California. The zero isotach is indicated by a gray line. currents with velocities of 0.1 m/s, measured in situ, were previously observed by Collins et al. [1997]. Above ASM some sections display pairs of cyclonic and anticyclonic eddies (August 1995, October 1994, and December 1992). Since similar flow patterns have been produced above seamounts in laboratory experiments [Zhang and Boyer, 1993] and numerical experiments [Chapman and Haidvogel, 1993], these may be due to occurrence of vortex shedding by ASM. Unfortunately, there are not enough oceanographic observations to verify the occurrence of vortex shedding in the region. [14] Figure 3 shows the mean geostrophic velocity and the standard error of the mean. The mean field was composed of two alternating cores of ingoing and outgoing flow. The two broadest cores had inflow of 0.15 m/s next to Sinaloa and outflow of 0.15 m/s next to BC. Narrower cores of outflow (inflow) are found at the surface over the western flank (summit) of Alarcoń seamount, but these broaden below 800 dbar. The standard error of the mean was largest at the surface, where it ranged from 0.05 m/s near the Sinaloa coast to 0.15 m/s over Alarcoń seamount. In the upper 200 dbar the standard error was of the same order of magnitude as the mean velocities and indicated the high level of variability in the geostrophic velocity field. [15] Figure 4 shows the anomaly of the geostrophic velocities, i.e., the fields displayed in Figure 2 minus the

mean field displayed in Figure 3. The largest anomalies were observed during May 1992 and November 1997 El Ninõ cruises. During May the observed inflow was larger than the mean inflow along the Sinaloa side, while on the BC side the observed outflow was larger than the mean outflow. This did not happen in November, when only the higher inflow on the Sinaloa side was observed. This suggests that even though the observations were made during El Ninõ conditions at the entrance to the gulf, the pattern of inflow/outflow clearly changed. The multivariate ENSO index for the seven strongest El Ninõ events indicated that the May 1992 and November 1997 observations were made at different phases of El Ninõ. The May 1992 cruise was performed during the initial phase of El Ninõ, while the November 1997 cruise was performed during the decaying phase. The difference between the two flow regimes could represent the changes of circulation associated with the different phases of El Ninõ. 4.2. Geostrophic Velocity and Heat Flux EOFS [16] To further examine the seasonal space and time variability of the geostrophic velocity field and heat exchanges, EOF analysis was performed. The spatial patterns associated with the first geostrophic velocity mode and the temporal variability of the amplitudes for the nine sections across the entrance are depicted in Figure 5. The

4 of 9

C07008


Figure 3. (left) Mean geostrophic velocity. Blue indicates flow out of the Gulf of California. Contour interval is 0.05 m/s. (right) Standard error of the mean geostrophic velocity. Contour interval is 0.02 m/s.

C07008

first EOF mode accounted for 37% of the total variance and exhibited a strong seasonal signal. The maximum (minimum) amplitude occurred in May (October). The first mode indicated a maximum change of inflow (outflow) during May (October) in the center of the gulf, with changes of outflow (inflow) concentrated along BC as well along Sinaloa. [17] The first mode for the EOF decomposition of the heat flux represented 34.7% of the variance. As expected, the flux of heat into the gulf is modulated by the geostrophic velocity field. The first mode amplitude varies seasonally (Figure 5c), with heat exported from the gulf during winter and spring (except in February 1994) and imported during summer and fall (except in November 1997). The spatial structure (Figure 5b) shows more horizontal variability than geostrophic velocity (Figure 5a), with numerous changes in sign near the surface between BC and Sinaloa, probably due to the commingling surface waters of different thermohaline properties [Castro et al., 2000]. Changes of heat outflow (inflow) were largest near BC in May 1992 and 1998 (August 1995 and 1998 and October 1994). Similar changes occurred with smaller amplitude near the Sinaloa coast. A secondary outflow core occurs next to the Sinaloa side. During these periods the heat inflow (outflow) occurred mainly in the center of the section. The latter is probably

Figure 4. Geostrophic velocity anomaly. Panels correspond to individual cruises, which are identified by month and year. Blue indicates negative anomalies. The zero isotach is indicated by a gray line. 5 of 9

C07008


C07008

Figure 5. First mode of the empirical orthogonal function (EOF). (a) Geostrophic velocity. Contour interval is 0.02 m/s. (b) Heat flux. Contour interval is 0.02 TW. (c) Temporal variability of the first amplitude of the first EOF for geostrophic velocity (solid line) and heat flux (dotted line). due to the seasonality of the wind stress field [Pare´s-Sierra et al., 2003] that blows strongly from the NW during fall and winter, pushing the water out of the gulf. 4.3. Geostrophic Heat Flux [18] Castro et al. [1994] and Beron-Vera and Ripa [2000] have estimated the seasonal heat flux from the Pacific Ocean into the gulf. They obtained their heat flux estimate by dividing the gulf into a series of boxes that extended across the gulf from BC to Sinaloa or Sonora and used historical data to assume a balance between the stored heat and the heat exchanged across the sea surface. Ripa [1997] used a slightly different approach and divided the gulf into boxes that extended along the gulf. For each box he fit a straight line to temperature and salinity data across the gulf and then estimated a mean along-gulf velocity profile using the across-gulf density gradient. To compute the mean horizontal advection due to the heat flux, Ripa [1997] used a zero mass transport condition, while geostrophic velocity was based on direct observations across the section. [19] The heat flux at the entrance for each cruise is listed in Table 2. An annual harmonic was fit to the heat flux data to extract a seasonal signal. The data for October/ November 1994 were not included because of its shallower sampling. The fit, which explains 80% of the variance, is shown in Figure 6, together with results discussed in the previous paragraph. Maximum heat gain, 27 1012 W, occurred on 13 May, and maximum heat loss, 75.4 1012 W, occurred on 16 November. The phase of the annual cycle and the minimum value agreed well with estimates by Castro et al. [1994] and Beron-Vera and Ripa [2000], but the amplitude and the mean were

about halfway between their estimate and that of Ripa [1997]. The latter results in a heat flux into the gulf, which was about twice that estimated by Castro et al. [1994], and extends (shortens) the period of time during which the Pacific (gulf) supplies heat to the gulf (Pacific) by 1 month. Although the approach used here could be considered more accurate because of the sampling methodology, the fact that there was good agreement with previous work in phase and not in estimates of the magnitude of the winter exchanges [Castro et al., 1994; Beron-Vera and Ripa, 2000] is probably due to the fact that the data used by Castro et al. [1994], Ripa [1997], and Beron-Vera and Ripa [2002] were from the Centro de Investigacioń Cientı´fica y Educacioń Superior de Ensenada historical data bank, which had few observations for winter at the mouth of the gulf (as shown in Figure 2 of Castro et al. [1994] and Figure 5 of Ripa [1997]). On the other hand, the data used in this study were from a different set of oceanographic sections across the entrance,

Table 2. Heat Fluxes Into the Gulf of California Dates

Heat Flux, 1012 W

4 – 9 Feb. 1994 28 – 29 Feb. 1999 2 – 4 May 1992 28 – 30 May 1998 2 – 4 Aug. 1998 13 – 15 Aug. 1995 31 Oct. to 1 Nov. 1994 19 – 20 Nov. 1997 29 – 31 Dec. 1992 Mean

8.5 14.3 23.8 27.0 19.5 11.3 22.1 75.4 36.6 1.61

6 of 9


C07008

Figure 6. Annual variation of the flux of heat at the entrance to the Gulf of California. The thick solid line was derived by a least squares fit for the annual harmonic from heat flux estimates in this paper and is compared to heat flux estimates of other authors. obtained with a CTD in closely spaced stations and vertically averaged to 2 dbar intervals.

5. Discussion 5.1. Geostrophic Velocity [20] Geostrophic velocity is important in the computation of fluxes, since it transports energy and property tracers.

C07008

The across-gulf structure of the geostrophic velocity was complex on any given cruise, with stronger flow near the surface (Figure 2). Although stronger currents were observed near the surface, cores with velocities of 0.1 m/s were observed much deeper, reaching 1400 dbar, which included Pacific Intermediate Water in all cruises. Rather than being organized in depth layers, the inflows and outflows extend from surface to bottom and are organized across the entrance. These deep cores are so robust that they appear in the mean flow section of Figure 3. Mean geostrophic currents depicted the outflow taking place along BC, inflow taking place over the central portion of the section, and variable flow taking place next to Sinaloa. In the region over Alarcoń seamount, baroclinic structures often displayed a cyclonic or anticyclonic eddy as deep as 600 m with diameters between 25 and 40 km, likely related to the shear on either side of the gulf or to interaction of the deeper flow with ASM [Zhang and Boyer, 1993; Chapman and Haidvogel, 1993; Boyer and Davies, 2000]. In the upper 300 m the cores of higher inflow (outflow) often comprise Tropical Surface Water (TSW) and California Current Water (CCW) along Sinaloa, while the outflow is made up of Gulf of California Water. Subsurface Subtropical Water was found as inflow and outflow all along the section [Wyrtki, 1967; Griffiths, 1968; A´lvarez-Sańchez et al., 1978; Castro et al., 2001]. The mean geostrophic field smoothed out many of the features that were seen in individual sections. Many processes contributed to the variability reflected in the high values of standard error. In addition to the processes discussed above, frontal instabilities [Griffiths, 1968; Collins et al., 1997], internal waves, and shelf waves [Merrifield and Winant, 1989; Zamudio et

Figure 7. Annual harmonic fit to sea level height anomaly along the eastern boundary of North America from coastal crossings of TOPEX/Poseidon, 1992 –2002 (observations during the El Ninõ period from July 1997 to June 1998 were not included). (left) Amplitude as a function of the distance from the southernmost data point. (middle) Phase (the time at which sea level is maximum) as a function of the distance from the southernmost data point. (right) Location of TOPEX/Poseidon data points. 7 of 9

C07008


al., 2002] have been observed in the region. The cruises performed during El Ninõ showed different velocity and heat flux structure; it has been suggested that during these years, TSW intrudes well into the gulf, and CCW is absent or feeble. Therefore we may not expect a unique geostrophic velocity or heat flux structure. [21] The first EOF of the geostrophic velocity and the heat flux varied annually. A seasonal variation in temperature is expected because of the annual heating. However, at the entrance to the gulf, seasonal salinity changes were not well defined. Salinity changes did not appear to be related to the evaporation-precipitation balance but were mostly due to northward incursions of the Costa Rica current (September to December), inflow of California Current Waters during February to June, and southward flow of surface and subsurface gulf waters [Castro et al., 2000]. 5.2. Heat Flux [22] How important are the heat fluxes? What is their effect on the eastern Pacific circulation? Bograd et al. [2001] computed the net flux of heat into a control volume within the southern California Current System (CCS) using 55 California Cooperative Oceanic Fisheries Investigations cruises over a 14 year period. They found a net flux of heat out of the box equivalent to 5.4 108 W. This estimate can be used as a reference to quantitatively evaluate the importance of the delivery of heat by the Gulf of California to that region of the Pacific. Bograd et al. [2001] estimate contrasts with an import of 35 1012 W of heat in late May and an export of 78 1012 W in November for the gulf. The gulf values are 3 orders of magnitude higher than the ones for the CCS. Thus a considerable amount of heat is added to or subtracted from Pacific coastal waters seasonally. The response should be similar to the Rossby adjustment problem for momentum. Charney [1973] noted that if the ratio of the Rossby radius of deformation, Ro, to the size of the strip of added momentum, LE, was small, almost all kinetic energy would be radiated as gravity waves. This is the case at the entrance to the gulf as Ro/LE = 48 km/200 km = 0.24, where the width of the entrance to the gulf is used for LE. The internal Rossby radius of deformation was computed as by Pedlosky [1987]. The obtained value is of the order of the values determined by Emery et al. [1984] and by Chelton et al. [1998] for the region. [23] Kelvin waves fit this model, and their phase speed is estimated at 20 km/d. However, Rossby waves could also be forced at the entrance to the gulf by heat exchanges with the gulf. Note that if we estimate that the width of the gulf entrance as the distance between Cabo San Lucas and Cabo Corrientes (to the west of Puerto Vallarta), the distance increases to 500 km. According to Chelton and Schlax [1996], Rossby waves are the large-scale dynamical response of the ocean to wind forcing and buoyancy forcing (heating and cooling) at the eastern boundaries and over the ocean interior. The release of a considerable amount of heat from the gulf will perturb the geostrophic state leading to a geostrophic adjustment-type process. Chelton and Schlax [1996] provide evidence of propagating signals along 21 and 32N (entrance to the gulf from 20 to 23N). [24] The variation of the heat flux (Figure 6) exhibited an annual character. This annual signal is expected to be manifested in sea level. A time series of sea level height

C07008

anomaly was used in order to find a relation between heat exchange at the entrance to the gulf and a possible annual northward propagation of a warm signal. Figure 7 shows the annual harmonic fitting of the sea level height anomaly from TOPEX/Poseidon altimetry data from 1992 – 2002 along the eastern boundary of North America (El Ninõ data from July 1997 to June 1998 were not included in the least squares fit for the annual harmonic). The amplitude of the annual oscillation was 5 cm both north of 38.5N and south of 18.5N; it doubled near the entrance to the gulf and was almost 3 times as large, 15 cm, at 25N within the gulf. The phase changed 2 months across the entrance to the gulf, and the observed phase near the southern tip of BC, 2.5 months, coincided with the phase of the maximum horizontal heat flux out of the gulf shown in Figure 6. The observed propagation speed from the southern tip of BC to the next northern point at 24N was 300 km in 20 days or 15 km/d, but along the rest of the BC coast to the north the phase changed little; for example, the predicted Kelvin wave propagation did not occur. At the annual frequency the alongshore slope of sea level shown in Figure 7 would favor flow toward the mouth of the gulf during February to April and away from the mouth from August to October in agreement with the observations of California Current water at the entrance to the gulf. [25] Acknowledgments. This project has been supported by Consejo Nacional de Ciencia y Tecnologıá (CONACyT), projects T-9201, 26653-T, and 4271P-T9601; the Oceanographer of the U.S. Navy; Universidad Autonoma de Baja California; and Secretaria de Marina de Mexico. S. Larios, R. Blanco, T. Rago, M. Cook, P. Jessen, A. Sanchez Devora, and E. Gil helped collect the data. We also thank John Ryan for the TOPEX/ Poseidon data and Miguel Lavin and two anonymous reviewers for suggestions that improved the manuscript. The strong support of the masters and crew of USNS DeSteiguer, R/V Point Sur, R/V A. Humboldt, R/V Francisco de Ulloa, and R/V El Puma is appreciated.

References ´ lvarez-Sańchez, L. G., M. R. Stevenson, and B. Wyatt (1978), Circulacioń A y masas de agua en la regioń de la Boca del Golfo de California en la Primavera de 1970, Cienc. Mar., 5, 57 – 69. Bacon, S., and N. Fofonoff (1996), Oceanic heat flux calculation, J. Atmos. Oceanic Technol., 13, 1327 – 1329. Beier, E. (1997), A numerical investigation of the annual variability in the Gulf of California, J. Phys. Oceanogr., 27, 615 – 632. Beron-Vera, F. J., and P. Ripa (2000), Three-dimensional aspects of the seasonal heat balance in the Gulf of California, J. Geophys. Res., 105, 11,441 – 11,457. Beron-Vera, F. J., and P. Ripa (2002), Seasonal salinity balance in the Gulf of California, J. Geophys. Res., 107(C8), 3100, doi:10.1029/2000JC000769. Blanco, R., S. Larios, A. S. Devora, E. Gil, A. S. Mascarenhas Jr., and C. Collins (1995), Datos hidrograficos del crucero PESCAR 5 en la Cuenca Pescadero, Octubre 26 – Noviembre 2, 1994, report, 61 pp., Inst. de Invest. Oceanol., Univ. Auton. de Baja Calif., Ensenada, Mex. Bograd, S. J., T. K. Chereskin, and D. Roemmich (2001), Transport of mass, heat, salt, and nutrients in the southern California Current System: Annual cycle and interannual variability, J. Geophys. Res., 106, 9255 – 9275. Boyer, D. L., and P. A. Davies (2000), Laboratory studies of orographic effects in rotating and stratified flows, Annu. Rev. Fluid Mech., 32, 165 – 202. Carter, E. F., and A. R. Robinson (1987), Analysis models for the estimation of oceanic fields, J. Atmos. Oceanic Technol., 4, 49 – 74. Castro, C. G., F. P. Chavez, and C. A. Collins (2001), Role of the California Undercurrent in the export of denitrified waters from the eastern tropical North Pacific, Global Biogeochem. Cycles, 15(4), 819 – 830. Castro, R., M. F. Lavıń, and P. Ripa (1994), Seasonal heat balance in the Gulf of California, J. Geophys. Res., 99, 3249 – 3261. Castro, R., A. S. Mascarenhas, R. Durazo, and C. A. Collins (2000), Seasonal variation of the temperature and salinity at the entrance of the Gulf of California, Mexico, Cienc. Mar., 26, 561 – 583.

8 of 9

C07008


Chapman, D. C., and D. B. Haidvogel (1993), Generation of internal lee waves trapped over a tall isolated seamount, Geophys. Astrophys. Fluid Dyn., 69, 33 – 54. Charney, J. G. (1973), Planetary fluid mechanics, in Dynamical Meteorology, edited by P. A. Morel, pp. 97 – 351, D. Reidel, Norwell, Mass. Chelton, D. B., and M. G. Schlax (1996), Global observations of oceanic Rossby waves, Science, 272, 234 – 238. Chelton, D. B., R. A. DeSzoeke, and M. G. Schlax (1998), Geographical variability of the first baroclinic Rossby raius of deformation, J. Phys. Oceanogr., 28, 433 – 460. Collins, C. A., N. Garfield, A. S. Mascarenhas Jr., M. Spearman, and T. A. Rago (1997), Ocean currents across the entrance to the Gulf of California, J. Geophys. Res., 102, 20,927 – 20,936. Emery, W. J., W. G. Lee, and L. Magaard (1984), Geographical and seasonal distributions of Brunt-Vaïsa¨la¨ frequency and Rossby radii in the North Pacific and North Atlantic, J. Phys. Oceanogr., 14, 294 – 352. Griffiths, R. C. (1968), Physical, chemical, and biological oceanography of the entrance to the Gulf of California, spring 1960, U.S. Fish Wildl. Serv. Spec. Sci. Rep. Fish., 573, 1 – 47. Hammann, M. G., M. O. Nevarez-Martinez, and Y. Green-Ruiz (1998), Spawning habitat of the Pacific sardine (Sardinops sagax) in the Gulf of California: Egg and larval distribution 1956 – 1957 and 1971 – 1991, CalCOFI Rep. 39, pp. 169 – 179, Calif. Coop. Oceanic Fish. Invest., La Jolla, Calif. Marinone, S. G., and P. Ripa (1988), Geostrophic flow in the Guaymas basin, central Gulf of California, Cont. Shelf Res., 8, 159 – 166. Merrifield, M. A., and C. D. Winant (1989), Shelf circulation in the Gulf of California: A description of the variability, J. Geophys. Res., 94, 18,133 – 18,160. Pare´s-Sierra, A., A. Mascarenhas, S. G. Marinone, and R. Castro (2003), Temporal and spatial variation of the winds in the Gulf of California, Geophys. Res. Lett., 30(6), 1312, doi:10.1029/2002GL016716.

C07008

Pedlosky, J. (1987), Geophysical Fluid Dynamics, 710 pp., SpringerVerlag, New York. Rago, T. A., R. Mitchel, L. F. Navarro-Olache, N. A. Garfield, and C. Collins (1992), Hydrographic data from the Pegasus in the Sea of Cortes area cruise (PESCAR-01), Tech. Rep. NPS-OC-92-009, 51 pp., Nav. Postgrad. Sch., Monterey, Calif. Ripa, P. (1997), Toward a physical explanation of the seasonal dynamics and thermodynamics of the Gulf of California, J. Phys. Oceanogr., 27, 597 – 614. Roden, G. I. (1972), Thermohaline structure and baroclinic flow across the Gulf of California entrance and the Revilla Gigedo Islands region, J. Phys. Oceanogr., 2, 177 – 183. Wyrtki, K. (1967), Circulation and water masses in the eastern equatorial Pacific Ocean, Int. J. Oceanol. Limnol., 1, 117 – 147. Zamudio, L., H. E. Hurlburt, E. J. Metzger, and O. M. Smedstad (2002), On the evolution of coastally trapped waves generated by Hurricane Juliette along the Mexican West Coast, Geophys. Res. Lett., 29(23), 2141, doi:10.1029/2002GL014769. Zhang, X., and D. L. Boyer (1993), Laboratory study of rotating, stratified, oscillatory flow over a seamount, J. Phys. Oceanogr., 23, 1122 – 1141.

R. Castro and R. Durazo, Facultad de Ciencias Marinas, Universidad Autonoma de Baja California, Ensenada, BC 22820, Mexico. (rubenc@ uabc.mx; [email protected]) C. A. Collins, Naval Postgraduate School, Department of Oceanography, 833 Dyer Road, Room 328, Monterey, CA 93943, USA. ([email protected]. navy. mil) A. S. Mascarenhas Jr., Instituto de Investigaciones Oceanolo´gicas, Universidad Autonoma de Baja California, Ensenada, BC 22820, Mexico. ([email protected])

9 of 9