JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 117, C04024, doi:10.

1029/2011JC007517, 2012

Relaxation from upwelling: The effect on dissolved oxygen

on the continental shelf
Uwe Send1 and SungHyun Nam1
Received 29 August 2011; revised 23 February 2012; accepted 24 February 2012; published 14 April 2012.
[1] Continental shelves in upwelling regimes are subject to sequences of upwelling and
relaxation events, each on timescales of order 1 week typically. These episodes have
pronounced impacts on the temperature and density structure on the shelves and also on the
along-shore and cross-shore flow regimes. It had previously been demonstrated that
relaxation phases advect warm water along shore (poleward) from regions of less intense
upwelling, thus adding to the heat balance in upwelling locations and providing a
rectification of oscillating heat fluxes. In the current paper it is shown that relaxations also
modify the dissolved oxygen (DO) budget of the lower layers. On a narrow shelf, this
provides enhanced DO values due to near-surface exposure, while on a wide shelf
decreased DO concentrations are created due to oxygen consumption on the inner shelf.
The resulting variations along the coast can lead to along-shore advection of high or low
DO during a relaxation event. Observations are presented from moorings off San Diego
and Del Mar, which show large departures from density-correlated DO values during
relaxations and which support the proposed mechanisms.
Citation: Send, U., and S. Nam (2012), Relaxation from upwelling: The effect on dissolved oxygen on the continental shelf,
J. Geophys. Res., 117, C04024, doi:10.1029/2011JC007517.

1. Introduction and the vertical shear decreases due to the isopycnals relaxing
toward a more horizontal slope [e.g., Send, 1989; Send et al.,
[2] The continental shelf of the west coast of North
1987]. These relaxation phases do not just reverse the
America lies next to a pronounced oxygen minimum zone
upwelling processes and changes, even though at first sight
(OMZ) in the Pacific Ocean [Helly and Levin, 2004;
the isopycnals return to their initial state. As shown by Send
Paulmier et al., 2006], and is also one of the most pro-
et al. [1987] the upwelling/relaxation cycles represent a rec-
nounced and well studied upwelling regimes in the world.
tification mechanism for heat flux, involving along-shore
Upwelling here takes place in the spring and summer advection. The upwelling thus is not reversible, and the
seasons, when the prevailing winds switch to equatorward
effect of a sequence of events with zero mean cross-shore
during the so-called “spring transition” [e.g., Huyer et al.,
flow and zero net isopycnal displacement is not zero. The
1979]. There is large spatial variability along the coast in
study presented here proposes somewhat similar processes
upwelling intensity, processes, and effects on the physical
also for the dissolved oxygen (DO) concentrations on the
and biogeochemical systems and the ecosystem. In addition,
deep middle/outer shelf.
there is pronounced temporal variability in the form of
[4] The DO conditions on the shelf of the west coast of
upwelling events which typically last for order of 1 week.
North America have received increasing attention over the
During these events, the isopycnals near and on the conti-
past years with the observation of hypoxic and anoxic events
nental shelf are lifted upward, bringing cold, low-oxygen,
which are lethal for some parts of the ecosystem and which can
high-CO2, low-pH, and high-nutrient water onto the shelf have dramatic impacts on commercial fisheries [McClatchie
and close to the surface. In addition, the sloped isopycnals
et al., 2010; Grantham et al., 2004]. For example, Grantham
imply a strongly sheared along-shore equatorward flow,
et al. [2004] reported severe shelf hypoxia and mass die-offs
strongest at the surface, sometimes called the “upwelling of fish and invertebrates in the northern California Current
jet” [e.g., Checkley and Barth, 2009].
System off Oregon, and Connolly et al. [2010] delineate the
[3] The upwelling conditions usually set up along-shore
role of physical and biological processes contributing to more
pressure gradients. When the local or remote upwelling
frequently occurring hypoxia off Washington shelf. In the
forcing ceases, these gradients drive poleward flow reversals,
southern California Bight, a recent trend toward hypoxia
since mid-1980s and its potential impacts on fisheries, e.g.,
1
loss of 18% of the habitat with 55% of the total habitat
Climate, Atmospheric Science, and Physical Oceanography Division,
Scripps Institution of Oceanography, University of California, San Diego,
exposed to hypoxia, were reported [McClatchie et al., 2010].
La Jolla, California, USA. [5] Trends are expected and starting to be observed in the
severity and frequency of hypoxic events, as a result of
Copyright 2012 by the American Geophysical Union. changes in the source water [Keeling et al., 2010; Stramma
0148-0227/12/2011JC007517

C04024 1 of 9
C04024 SEND AND NAM: RELAXATION FROM UPWELLING C04024

Figure 1. Bathymetry around (a) the southern California Bight (SCB) and (b) offshore of La Jolla. In
Figure 1a are shown major circulation features and stations in the SCB along CalCOFI lines 93, dates
of those measurements in 2005, and the inshore domain of Figure 1b. Locations of three inshore stations
of CalCOFI line 93 (x), multidisciplinary mooring at a depth of 100 m off Del Mar (DM) (pink circle), and
mini-bottom moorings off Imperial Beach (MBIB) (crosses) are shown in Figure 1b.

et al., 2008; Bograd et al., 2008] and possibly of changes of
processes also on the continental shelves [Monteiro et al.,
2011; Connolly et al., 2010; Chan et al., 2008; Paulmier
et al., 2006; Grantham et al., 2004]. The uplifting and
onshore transport of low-DO water onto the shelf usually also
coincides with low-pH (corrosive) conditions [Feely et al.,
2008], equally harmful for marine life. The present study
analyzes the effects of relaxation events on the short-term
variability of DO, possibly contributing to hypoxic condi-
tions. The novel aspect is that it is shown that low DO not
only occurs when upwelling is intense (as a result of uplift-
ing of offshore dense water) but that this can also happen
during relaxation phases (when relaxing isopycnals should
return the higher-DO water to the outer shelf).
2. Methods
2.1. Area and Timescales
[6] The southern California Bight (SCB) (Figure 1a) is
known for complex systems of currents and bathymetry. It
is affected by water masses from both the subarctic and
tropical Pacific through the California Current, California
Undercurrent, and water recirculation [Dong et al., 2009;
Bray et al., 1999]. In the nearshore areas (Figure 1b),
coastal water is more affected by local and coastally prop-
agating remote forcing [Nam and Send, 2011; Dong et al.,
2009] than changes of the large-scale current systems on a
time scale of a few to tens of days while strengthening of
the California Undercurrent and seasonal upwelling from
spring to summer play a significant role on longer timescales
[McClatchie et al., 2010; Di Lorenzo, 2003]. Although the
well-known seasonal upwelling is one important physical
driver of hypoxia as is also shown here, the focus of the
current study is on shorter-term oxygen variability in the
coastal SCB, which has received less attention.
2.2. Measurements
[7] Temporal variability was investigated by analyzing
moored time series data of water temperature, salinity, and
DO at fixed depths and water column profiles of currents. We
collected the data using a multidisciplinary surface mooring
off Del Mar (DM: 32
56.10′N, 117
19.11′W) deployed on
the shelf break at a water depth of 100 m (Figure 1b). Current
data were obtained every 7.5 min at 5 m intervals from 5 to
100 m using a 300 kHz acoustic Doppler current profiler
(ADCP). Time series data of DO were collected every
20 min using SBE 43 conductivity-temperature-depth (CTD),
located at 35 and 90 m on the DM mooring. During May/June
2010, time series data of water temperature, salinity, and
DO were acquired at an additional location off Imperial
Beach (MBIB, sensor depths: 83 m, water depth: 85 m) by
attaching an Aandeeraa Optode-equipped CTD to mini-
bottom moorings (Figure 1b).
[8] Inside and outside of the SCB, hydrographic data are
available from the California Cooperative Oceanic Fisheries
Investigations (CalCOFI) program (http://www.calcofi.org/).
The density and DO observed at 30–40 m of CalCOFI station
in the vicinity of the DM mooring (station 093.26.7,
Figure 1b) are in close agreement with those at 35 m of the
mooring (Figure 2). Both data sets show a clear seasonal
cycle with peak upwelling in spring every year although the
sparse CalCOFI measurements do not always resolve the
peak upwelling, e.g., maximum density and minimum DO
(Figure 2). The continuous DM mooring time series provide
reliable data to understand the seasonal DO anomalies [Nam
et al., 2011] preventing also any aliasing in the CalCOFI

2 of 9
C04024 SEND AND NAM: RELAXATION FROM UPWELLING C04024

Figure 2. Time series of (a) density and (b) dissolved oxygen measured at 35 (red) and 88 m (blue)
on the DM mooring from 2006 to 2011 (thin solid line) and at 30–40 m at the ship-occupied CalCOFI
station 093.26.7 (shown in Figure 1b) from 2005 to 2010 (circle with dashed line). Here mean seasonal
cycles were superimposed with thick sky blue line.

data. For a spatial view here we used water temperature,
salinity, and DO data collected at the CalCOFI stations of
lines 93 in the SCB (Figure 1a) during Cruise 0504 in
April 2005 [Scripps Institution of Oceanography, 2007]
to present typical seasonal upwelling since density and
DO at 30–40 m of the station near the DM mooring
(station 093.26.7, Figure 1b) are close to the values observed
by the mooring at peak upwelling in most years (Figure 2).
Using CalCOFI sections from other years does not qualita-
tively change the analyses and conclusions of this paper.
2.3. Data Processing
[9] From the CalCOFI Cruise 0504 measurements, cross-
sectional structures of density and DO were mapped using a
punctual kriging method [Middleton, 2000] with the weights
li taken from a linear variogram fulfilling the unbiasedness
condition (∑ li = 1) in grid sizes of 5 km and 5 m, respectively.
For each point, data Z at neighboring points xi within a search
radius of 100 km horizontally and 50 m vertically were used
to obtain the interpolated value Z′ using the equation
X covaries with DO at 35 m, but with higher density and lower
Z′ ¼ li Z ðxi Þ
[10] The maximum numbers of points to use from all four
sectors and from each sector are 64 and 16, respectively.
[11] We used the geographic Cartesian coordinates without
axis rotation for the horizontal current vectors (east–west;
u and north–south; v), because the coastline in the vicinity
of the DM mooring is close enough to the north–south
direction (hereafter along-shore direction) (Figure 1b). When
the along-shore current v at 88 m is poleward (v > 0) and the
magnitude of the depth-averaged vertical shear of the along-
shore current is below 0.003 s1 (∣dv/dz∣ < 0.003), we defined
the period as a “relaxation event.” This is our quantification
of the two qualitative features of a relaxation event: the
reversed poleward currents and the flattening of the iso-
pycnals which had been uplifted by the preceding upwelling
phase. With this small mean vertical shear, the isopycnals do
not have a significant slope toward the coast and the related
cross-shore density gradient from the thermal wind relation-
ship would be smaller than 0.2 kg m3 over a cross-shore
distance of 50 km. The upwelling events are defined as the
periods that do not correspond to the relaxation events and
when the magnitude of vertical shear of the along-shore
current increases in time with negative values (dv/dz < 0 and d
(dv/dz)/dt < 0) or when the shear has reached values below
0.003 s1 (∣dv/dz∣ > 0.003). This is our quantification of the
isopycnal uplifting process during upwelling (generating the
upward slope toward the coast characteristic of upwelling).
[12] Near-bottom density and DO time series collected for
1 year at the DM mooring at a depth of 88 m (blue line in
Figure 2) allow the analysis of the overall oxygen variability
during sequences of upwelling and relaxation events. Note
that the DO measured at 88 m in 2009–2010 generally
DO values (Figure 2). Figures 3a and 3b show the covaria-
bility of DO and density at 88 m depth, together with along-
shore currents, during 2010. As expected, the most common
behavior is that DO decreases while density increases during
upwelling phases, due to the uplifting of deeper water with
lower DO. A tight inverse correlation is found between
density and oxygen (R2 = 0.9 for linear regression) for most
phases of upwelling events. This relation reflects the rather
stable properties of the source water found from the historical
CalCOFI measurements within 100–200 km from the coast
[e.g., Nam et al., 2011, Figure S1A], also see Figure 4b here.
As a result one can calculate the oxygen that is “predicted”

3 of 9
C04024 SEND AND NAM: RELAXATION FROM UPWELLING C04024

Figure 3. Time series of (a) density (blue) and oxygen (red) at 88 m at DM, (b) along-shore currents
(positive poleward) measured at eight different depths at the DM, and (c) oxygen anomaly where the effect
of density excursion is removed. Here the shaded periods correspond to upwelling (blue) and relaxation
(green) phases labeled u1–u7 and r1–r8, respectively (see text for definition).

from the observed density, and then derive DO anomalies
relative to these values that are expected from just the vertical
excursion of isopycnals. These anomalies are shown in
Figure 3c.
3. Results
[13] Based on the above definitions, we have identified
and labeled seven upwelling (u1–u7) and eight relaxation
(r1–r8) events from February to August 2010, all marked in
Figure 3. The results of the analysis of variability in DO at
88 m in 2009–2010 reveal large departures of both signs
from the density-correlated DO values, and that the largest
departures all occur during or at the end of relaxation periods
(Figure 3c). For most cases (particularly r5, r6, and r7), the
positive anomalies (DO anomaly higher than 0.2 ml/L)
occur in the middle of relaxations, while negative anomalies

Figure 4. Cross sections of density and oxygen: (a) density from line 93, (b) oxygen from line 93 (color)
with density contour superimposed, and (c) expanded scale of the coastal section of Figure 4b. Here the
density contour interval is 0.05 kg/m3 with a thick line every 0.5 kg/m3. The oxygen concentration below
1.5 ml/L and between 1.5 and 2.0 ml/L are highlighted with magenta and pink colors, respectively. The
corresponding location of the oxygen sensor at the DM mooring is marked by a white square at 88 m depth
in each section. The arrows in the bottom axes denote the scale of shelf width in the south off Imperial Beach.

4 of 9
C04024 SEND AND NAM: RELAXATION FROM UPWELLING C04024

Figure 5. Vertical depth profiles of cross-shore currents averaged over (a) upwelling and (b) relaxation
events shown in Figure 3. Here the mean profiles over all the upwelling and relaxation periods are plotted
with thick black lines. Major and minor tick intervals in x axis are 2.0 and 1.0 cm/s, respectively.

are found later. As a consequence a new upwelling event
starts with lower-than-normal DO, therefore the density-DO
correlation decreases when the first few days of upwelling
events are included in the calculation, and we have omitted
those initial days in the density-DO relation used below.
3.1. Narrow Shelf
[14] Much of the above DO anomalies can be understood
in terms of cross-shore movement of water near the bottom
which can be estimated using the ADCP velocity data.
Figure 5 shows the mean depth profiles of cross-shore flow
averaged over all upwelling and relaxation phases. Upwell-
ing periods exhibit a mean onshore flow of approximately
1.0 cm/s over much of the water column, while relaxation
periods have a clear sheared flow with a mean offshore
movement in the lower layer and onshore movement above,
again with order 1.0 cm/s amplitudes. The shelf is narrow at
the mooring location, with a distance of just 3.5 km from
95 m to 35 m water depth, giving a bottom slope of close to
1
. A typical upwelling event of 1 week length would push a
deep water parcel onshore along the bottom by several km,
and thus into shallow water (i.e., the mixed layer and
euphotic zone) if it starts at the mooring location (Figure 6a).
Even without the assumption of the velocity measured at the
fixed mooring location applying to a water parcel while it
moves onshore, the same result can be obtained by calculating
the rate dh/dt at which the starting isopycnal (isopycnal ini-
tially at depth 90 m at the mooring location) is lifted if the
volume below it gets filled by the onshore flow at the
mooring location, assuming two-dimensional flow: dh/dt =
u tan q, where q is the bottom slope and u the deep onshore
flow measured at the mooring. This assumes a long and
uniform shelf without divergences in the along-shore flow.
For u = 1 cm/s one obtains a dh/dt of 15 m/d or 105 m over
7 days.
[15] The same calculations apply for the reverse case, the
offshore relaxation currents in the lower layer. Water
reaching the mooring after 7 days of those flows, must have
come from far inshore, which on the steeply sloping bottom
means near the surface. This can be quantified by calculating
cumulative lower layer offshore displacements for each
relaxation event, and the numbers found are typically 5–7 km
(Figure 6b). This means that under uniform along-shore
conditions, the water reaching 90 m depth during a relaxation
event must have come from within the mixed layer and the
euphotic zone, where DO enhancement by surface exchange
and primary productivity take place. This is the only pos-
sible source of the positive DO anomalies seen in Figure 3c
during relaxation phases, since along-shore advection will
act in the opposite sense as shown in section 3.3. This means
that at least to some extent the two-dimensional picture is
valid here and not all the cross-shore flow goes into/comes
from along-shore divergences. The DM mooring includes
DO sensors at 35 m depth, and these show DO values large
enough to explain the enhanced concentrations seen at 88 m
as a result of the offshore/downward advection along the
bottom slope during relaxations, assuming the relaxation
water comes from depths of 35 m or shallower.
3.2. Wide Shelf
[16] Now consider a wide shelf, as found 20–30 km farther
south, off San Diego and Imperial Beach (Figure 1b). The
slope there is approximately three times less than at the DM
mooring. It is assumed that the deep cross-shore flow at the
edge of the shelf in 90–100 m depth is of same magnitude
as at DM (order 1.0 cm/s), since this circulation compensates
a similar Ekman divergence at the surface (which is expected
to vary at along-shore scales larger than the separation of
20–30 km). By the same arguments as above, a water parcel
moving up the slope at 1.0 cm/s will now end up 20–30 m
shallower after 7 days, or an initially deep isopycnal will be
lifted from 90 m to 60–70 m depth. During relaxation there
is downward/offshore motion again, but the water found at
90 m this time comes from typically 60–70 m depth and
has remained below productive and mixed layer depths for
1–2 weeks. During this time, oxygen consumption is expected
to lower the DO values by 0.4–0.8 ml/L, or more in shal-
lower water (J. Barth and F. Chan, personal communication,
2011) (see below).
[17] In order to test such processes on a wide shelf, and the
connection with the DM observations, the short-term near-
bottom mini-mooring MBIB was deployed for one month in
May/June 2010 (see section 2.2). Data from MBIB, covering
one strong upwelling and relaxation cycle, are presented in
Figure 7. The density excursion from s = 26.0 to s = 26.3
corresponds to approximately 30–35 m uplifting in the
vertical, as judged from Figure 4 and from density profiles
from the DM mooring. This is consistent with the above

5 of 9
C04024 SEND AND NAM: RELAXATION FROM UPWELLING
Figure 6. Time integrals of cross-shore currents averaged
over (a) 20–90 and (b) 70–90 m of the DM mooring during
the upwelling (Figure 6a) and relaxation (Figure 6b) periods,
shown as both cross-shore displacement in kilometers (left
axis) and corresponding vertical isopycnal displacement in
meters for both cases of narrow and wide shelves (right
axis). Here u1–u7 and r1–r8 denote the upwelling and relax-
ation events shown in Figure 3.
estimate for the horizontal/vertical displacement on this wide
shelf, resulting from 1 week of onshore flow at 1 cm/s.
Figure 7c shows the DO anomaly, as in Figure 3c, i.e., the
departure from a conservative DO concentration following
just the isopycnals. During the relaxation event, after 5–
6 days, lower-than-expected DO values start to appear, and
by day 9 of the relaxation an oxygen deficiency of 0.5 ml/L is
observed. This is consistent with oxygen consumption values
from other shelf regions. On the Oregon shelf, which is
more productive than the southern California Bight and
6 of 9
C04024
thus produces more organic matter for respiration, typical
values are 0.05 ml/l/d (J. Barth and F. Chan, personal
communication, YEAR), which would give 0.9 ml/L decay
over the 18 days at MBIB from start of the upwelling until
the large observed DO decrease. Based on the observed
drop in DO from 2.0 ml/L to 1.6 ml/L at MBIB, a value of
0.02 ml/l/d for the San Diego region is used here.
3.3. Along-Shore Advection
[18] Also included in Figure 7c are the DO anomalies
from the DM mooring during the same period. Three days
into the relaxation, enhanced DO values are found at DM,
resulting from the mechanism explained above for a narrow
shelf. At day 5, however, the anomaly goes back to 0 and
negative, at the end of a 3–4 day phase of northward near-
bottom flow of about 10 cm/s (Figure 7b). This suggests
along-shore advection of water from 30 km to the south,
which corresponds to the location of the Imperial Beach
mooring (MBIB). Figures 7a and 7c show that MBIB had
low-enough DO values 3 days before the drop in DO at DM
to explain this decrease at DM. Thus it appears that the low
DO concentrations from MBIB arrived at DM 5 days into
the relaxation. After day 5, the along-shore flow goes to
zero, and within 3 days the high DO anomalies return at
DM, again implying water brought down from shallower
depths during the relaxation on the narrow shelf. There is
however another episode of northward near-bottom flow,
and after 3 days of this, the negative DO anomaly from
Imperial Beach arrives at DM with a time lag of 3 days
(again 30 km in 3 days at the observed 10 cm/s). This
demonstrates that northward relaxation currents can bring
anomalously low-DO water into a region from a wider shelf
to the south. In the case of the DM mooring observations,
this is the only possible source of the anomalously low-DO
water, given the observed offshore and poleward flow during
the relaxations. Similar negative DO anomalies are also
observed during other relaxation events at DM (see Figure 3).
4. Conclusion and Discussions
[19] We have argued that at the locations studied, the cross-
shelf upwelling/relaxation oscillations can generate increased
oxygen concentrations on a narrow shelf and diminished
oxygen concentrations on a wide shelf. This provides a
mechanism and explanation which is consistent with our
density, DO, and along-shelf/cross-shelf flow data. Note
that stationary water at the outer shelf edge on a wide shelf
would experience much lesser DO reductions than what
results from the advection from shallower water during
relaxations, since the DO consumption increases toward the
shore and in shallower water. The observational data and
proposed mechanisms can be summarized in phase diagrams,
by following the trajectory of an isopycnal for a given density
in the depth-oxygen plane through an upwelling/relaxation
cycle at the fixed location of the Del Mar mooring (The
phase trajectory is not for a water parcel, since along-shore
advection will be invoked at this fixed location. Instead it
shows how DO changes on an isopycnals at a fixed location).
Figure 8a shows, for a broad shelf, a conceptual trajectory
of water, here for approximately the s = 26.0 isopycnal as
observed at MBIB. It starts out at 85 m depth with a DO
value near 2.0 ml/L (red triangle), as observed, e.g., on
C04024 SEND AND NAM: RELAXATION FROM UPWELLING C04024

Figure 7. Time series of (a) near-bottom oxygen (DM, pink; MBIB, red) and density (DM, green; MBIB,
blue), (b) along-shore currents measured at seven different depths at the DM mooring, and (c) oxygen
anomaly (DM, pink; MBIB, red) where the effect of density excursion is removed as in Figure 3c. The
shaded periods again correspond to upwelling (blue) and relaxation (green) events. Arrows in Figures 7a
and 7c denote the times of low-DO anomalies at DM (pink) and similar values at MBIB (red) indicating
time lag of 3 days.

Figure 8. Conceptual phase diagram following the trajectory of an isopycnal for a given density in the
depth-DO plane through an upwelling/relaxation cycle at a fixed location (a) on a broad shelf (here MBIB)
and (b) on a narrow shelf (here DM). See text for more details.

7 of 9
C04024 SEND AND NAM: RELAXATION FROM UPWELLING
22 May (Figure 7a). The cross-shore flow data suggest that
this isopycnal gets uplifted by approximately 20–30 m over
a 6 day period (see section 3.2; phase A in Figure 8a).
Staying below 40 m depth, it is thus away from oxygen
sources and will experience respiration of order 0.20 ml/L,
using the value derived above. This puts it at the estimated
point marked with the purple square and orange cross, e.g.,
1.80 ml/L. The relaxation then carries the water back to the
mini-mooring over 6 days, with respiration of another
0.20 ml/L (phase B). It then arrives at the mini-mooring
with the observed properties marked by the blue diamond,
0.4 ml/L lower in DO than when it started (Figure 8a), as
observed, e.g., on 8 June (Figure 7a).
[20] Figure 8b shows an equivalent cycle for the narrow
shelf location at the DM mooring. The starting point is the
observed depth and DO concentration of 90 m and 2.0 ml/L,
as observed, e.g., on 1 or 17 May for the s = 26.0 (Figures 3a
and 7a). During phase A the water parcel is uplifted by up to
100 m over a week. We do not know what depth was actually
reached, but can assume it was at least 35 m, where we have
a DO sensor, and where primary productivity and surface
mixing will generate enhanced DO values for this isopycnal
(phase B in Figure 8b) (At the DM mooring the s = 26.0
isopycnal does not reach 35 m depth, but the water parcel
considered here moves southward as it is pushed up the shelf
and is believed to approach the surface farther south. The
fact that water of this density returns with higher DO must
mean it has been in the euphotic zone or the surface mixed
layer during the peak of the upwelling). After this, the
relaxation phase sets in with offshore and downslope
movement of water (phase C in Figure 8b). When the iso-
pycnal considered here reaches 90 m depth, the observed
DO value is, e.g., 2.2 ml/L (on 5 June, Figures 3, 7a and 7c),
higher than it started and similar to the 35 m DO concentra-
tions during the upwelling peak (Figure 8b). Afterward,
along-shore advection from the south lowers the DO values
by up to 0.5 ml/L (as observed, e.g., on 12 June, Figures 3,
7a, and 7c), generating hypoxic levels (Figure 8b).
[21] Much of the present scientific interest focuses on
hypoxic conditions, and the data and mechanisms presented
here contain two ways in which relaxation events can lower
the oxygen concentrations on the continental shelf. On a
broad shelf, this can happen entirely via cross-shore motions,
even under uniform conditions along the coast. On a narrow
shelf, advection during near-bottom poleward flow in a
relaxation phase, can bring low-oxygen water (resulting from
respiration) from more equatorward regions where these
processes dominate (here a broader shelf). Both can lead to
hypoxic conditions during relaxation phases, even though
this is the time where normally higher DO values would be
expected (from isopycnal movements).
[22] Somewhat similar to the temperature variability during
relaxations discussed by Send et al. [1987], the oxygen pro-
cesses presented here imply an important rectification pro-
cess resulting from upwelling/relaxation cycles. With large
but zero mean cross-shore flow oscillations, there is a net
DO enhancement on uniform narrow shelves which would
not exist with small or absent cross-shore flow variability.
This represents a pumping of DO downward into isopycnals
whose mean and offshore depth is far below the surface
layers. On broad shelves, DO levels are lowered at the outer
edge of the shelf (90–100 m depth) compared to stationary
8 of 9
C04024
water masses, since the relaxations bring water back that
had spent time in shallower regions where oxygen con-
sumption is several times higher. For the same reason,
upwelling/relaxation oscillations generate lower values on a
narrow shelf, when it is located poleward of a region with a
wider shelf, due to the poleward flow reversals occurring
during relaxations. Overall, the oxygen conditions thus depend
on the amplitude of the upwelling/relaxation oscillations
(even if they have zero mean displacement).
[23] The present study puts forward and documents the
impacts which relaxation events and along-shore changes in
shelf width can have on the dissolved oxygen concentrations
in the bottom layers of the middle and outer shelves. This adds
additional information and potential mechanisms to previous
observations of the DO impact of such temporal and spatial
variability. Monteiro et al. [2011] explained shelf hypoxia
variability on the basis of the role of shelf geometry in the
retention of shelf-based particulate organic carbon fluxes.
Cross-shelf carbon exchanges, e.g., export of new carbon
produced by photosynthesis to depth, can be also affected by
the duration and frequency of upwelling/relaxation as sug-
gested by Hales et al. [2006]. Connolly et al. [2010] showed
the importance of along-shore as well as cross-shore advection
on short-term changes in DO concentrations, e.g., 1.5 ml/L
or more over 5 days on the Washington shelf. We suggest
the relaxation from upwelling and the important rectifica-
tion process resulting from upwelling/relaxation cycles as
mechanisms relevant to (1) contrasting the hypoxia regimes
associated with low ventilation-wide shelf systems and high
ventilation-narrow shelf systems described by Monteiro et al.
[2011], (2) controlling cross-shelf exchanges/fluxes of par-
ticulate organic carbon as suggested by Hales et al. [2006],
and (3) short-term DO changes observed on the Washington
shelf [Connolly et al., 2010].
[24] We have demonstrated that the interaction of
upwelling/relaxation circulation with shelf geometry can be
one important mechanism for driving the system toward
hypoxia on upwelling shelves although we did not provide
any quantitative estimate on the fluxes and rates of ventilation
by ocean boundary forcing through ocean-shelf exchange.
The observational database is still sparse, but more focused
follow-up experiments can now be designed in various
locations to explore this process further. Their goal would be
to test the outlined mechanisms more thoroughly, to better
quantify the oxygen consumption rates and particulate
organic carbon fluxes from time series observations, and to
establish a more quantitative understanding and predictive
capability based on more complete observations of the
along-shore/cross-shore flows, and of the density and oxygen
distributions at more locations along and across the shelf.
[25] Our suggested mechanism relies on along-shore uni-
formity within 20–30 km which is a widely accepted
assumption for subtidal timescales (days and weeks). Along-
shore correlation scales of subtidal currents on the southern
California shelf are much longer than cross-shore scale
(at least in the order of 30 km during summer [Winant,
1983] and excess of 50 km during winter versus cross-
shore scales of only 5 km or shorter [Lentz and Winant,
1986]) as found also on many other shelves. Dever [1997]
estimated correlation scales of subtidal currents using
moored time series data over the northern California shelf.
The along-shore correlation scale of the along-shore current
C04024 SEND AND NAM: RELAXATION FROM UPWELLING
was estimated to 60 km which is longer than the separation of
the DM and MBIB moorings. Though the along-shore scales
of cross-shore current can be shorter and vary in time, it is
still generally longer than 30 km when the correlation with
wind stress is high [Dever, 1997]. Since the profiles of cross-
shore currents we observed at DM mooring (Figure 5) are
consistent with two-dimensional (cross-sectional) circula-
tion, i.e., onshore subsurface flow during upwelling and
offshore flow at lower water column during relaxation with
compensating onshore flow at the upper levels, it is reason-
able to assume that the along-shore correlation scale of cross-
shore as well as along-shore currents is longer than 30 km.
Thus, the current profiles observed at the DM mooring
during the events can be applicable to the shelf within
30 km, e.g., the southern broad shelf, though not that
straightforward to far northern narrow shelf, e.g., Palos
Verdes shelf, 135 km distant from the DM [Noble et al.,
2002]. Further tests for this assumption would be enabled
with simultaneous time series at more locations.
[26] Acknowledgments. The dedicated effort of all the members of
SIO Ocean Time Series Group for technical developments, field work,
and data handling is greatly appreciated. The project benefited from partial
support by NOAA and the NSF. The authors thank two anonymous
reviewers for their careful and detailed reviews of this manuscript.
References
Bograd, S. J., C. G. Castro, E. Di Lorenzo, D. M. Palacios, H. Bailey,
W. Gilly, and F. P. Chavez (2008), Oxygen declines and the shoaling
of the hypoxic boundary in the California Current, Geophys. Res. Lett.,
35, L12607, doi:10.1029/2008GL034185.
Bray, N. A., A. Keyes, and W. M. L. Morawitz (1999), The California
Current system in the southern California Bight and the Santa Barbara
Channel, J. Geophys. Res., 104, 7695–7714, doi:10.1029/1998JC900038.
Chan, F., J. A. Barth, J. Lubchenco, A. Kirincich, H. Weeks, W. T. Peterson,
and B. A. Menge (2008), Emergence of anoxia in the California Current
large marine ecosystem, Science, 319, 920, doi:10.1126/science.1149016.
Checkley, D. M., and J. A. Barth (2009), Patterns and processes in the
California Current System, Prog. Oceanogr., 83, 49–64, doi:10.1016/
j.pocean.2009.07.028.
Connolly, T. P., B. M. Hickey, S. L. Geier, and W. P. Cochlan (2010), Pro-
cesses influencing seasonal hypoxia in the northern California Current
system, J. Geophys. Res., 115, C03021, doi:10.1029/2009JC005283.
Dever, E. P. (1997), Subtidal velocity correlation scales on the northern
California shelf, J. Geophys. Res., 102, 8555–8571, doi:10.1029/
96JC03451.
Di Lorenzo, E. (2003), Seasonal dynamics of the surface circulation in
the southern California Current system, Deep Sea Res., Part II, 50,
2371–2388, doi:10.1016/S0967-0645(03)00125-5.
Dong, C., E. Y. Idica, and J. C. McWilliiams (2009), Circulation and
multiple-scale variability in the southern California Bight, Prog. Oceanogr.,
82, 168–190, doi:10.1016/j.pocean.2009.07.005.
Feely, R. A., C. L. Sabine, J. M. Hernandez-Ayon, D. Ianson, and B. Hales
(2008), Evidence for upwelling of corrosive “acidified” water onto the
continental shelf, Science, 320, 1490–1492, doi:10.1126/science.1155676.
9 of 9
C04024
Grantham, B. A., F. Chan, K. J. Nielsen, D. S. Fox, J. A. Barth, A.
Huyer, J. Lubchenco, and B. A. Menge (2004), Upwelling-driven near-
shore hypoxia signals ecosystem and oceanographic changes in the north-
east Pacific, Nature, 429, 749–754, doi:10.1038/nature02605.
Hales, B., L. Karp-Boss, A. Perlin, and P. A. Wheeler (2006), Oxygen
production and carbon sequestration in an upwelling coastal margin,
Global Biogeochem. Cycles, 20, GB3001, doi:10.1029/2005GB002517.
Helly, J. J., and L. A. Levin (2004), Global distribution of naturally occur-
ring marine hypoxia on continental margins, Deep Sea Res., Part I, 51,
1159–1168, doi:10.1016/j.dsr.2004.03.009.
Huyer, A., E. J. C. Sobey, and R. L. Smith (1979), The spring transition
in currents over the Oregon continental shelf, J. Geophys. Res., 84,
6995–7011, doi:10.1029/JC084iC11p06995.
Keeling, R. F., A. Kortzinger, and N. Gruber (2010), Ocean deoxygenation
in a warming world, Annu. Rev. Mar. Sci., 2, 199–229, doi:10.1146/
annurev.marine.010908.163855.
Lentz, S. J., and C. D. Winant (1986), Subinertial currents on the southern
California shelf, J. Phys. Oceanogr., 16, 1737–1750, doi:10.1175/1520-
0485(1986)016<1737:SCOTSC>2.0.CO;2.
McClatchie, S., R. Goericke, R. Cosgrove, G. Auad, and R. Vetter (2010),
Oxygen in the southern California Bight: Multidecadal trends and impli-
cations for demersal fisheries, Geophys. Res. Lett., 37, L19602,
doi:10.1029/2010GL044497.
Middleton, G. V. (2000), Data Analysis in the Earth Science Using
MATLAB, 260 pp., Prentice Hall, Upper Saddle River, N. J.
Monteiro, P. M. S., B. Dewitte, M. I. Scranton, A. Paulmier, and
A. K. van der Plas (2011), The role of open ocean boundary forcing on
seasonal to decadal-scale variability and long-term change of natural shelf
hypoxia, Environ. Res. Lett., 6, 025002, doi:10.1088/1748-9326/6/2/
025002.
Nam, S. H., and U. Send (2011), Direct evidence of deep water intrusions
onto the continental shelf via surging internal tides, J. Geophys. Res.,
116, C05004, doi:10.1029/2010JC006692.
Nam, S. H., H.-J. Kim, and U. Send (2011), Amplification of hypoxic and
acidic events by La Niña conditions on the continental shelf off California,
Geophys. Res. Lett., 38, L22602, doi:10.1029/2011GL049549.
Noble, M. A., H. F. Ryan, and P. L. Wiberg (2002), The dynamics of
subtidal poleward flow over a narrow continental shelf, Palos Verdes,
CA, Cont. Shelf Res., 22, 923–944, doi:10.1016/S0278-4343(01)00112-1.
Paulmier, A., D. Ruiz-Pino, V. Garcon, and L. Farıas (2006), Maintaining
of the eastern South Pacific Oxygen Minimum Zone (OMZ) off Chile,
Geophys. Res. Lett., 33, L20601, doi:10.1029/2006GL026801.
Scripps Institution of Oceanography (2007), Physical, chemical and bio-
logical data: CalCOFI cruise 0704, 28 March–29 April 2007, CC Ref.
08-06, Univ. of Calif., San Diego, La Jolla.
Send, U. (1989), Vorticity and instability during flow reversals on the con-
tinental shelf, J. Phys. Oceanogr., 19, 1620–1633, doi:10.1175/1520-
0485(1989)019<1620:VAIDFR>2.0.CO;2.
Send, U., R. C. Beardsley, and C. D. Winant (1987), Relaxation from
upwelling in the Coastal Ocean Dynamics Experiment, J. Geophys.
Res., 92, 1683–1698, doi:10.1029/JC092iC02p01683.
Stramma, L., G. C. Johnson, J. Sprintall, and V. Mohrholz (2008),
Expanding oxygen-minimum zones in the tropical oceans, Science, 320,
655–658, doi:10.1126/science.1153847.
Winant, C. D. (1983), Longshore coherence of currents on the southern
California shelf during the summer, J. Phys. Oceanogr., 13, 54–64,
doi:10.1175/1520-0485(1983)013<0054:LCOCOT>2.0.CO;2.
S. Nam and U. Send, Climate, Atmospheric Science, and Physical
Oceanography Division, Scripps Institution of Oceanography, University
of California, San Diego, 9500 Gilman Dr., La Jolla, CA 92093-0230,
USA. (usend@ucsd.edu)