Low-oxygen and low-pH events are an increasing concern and threat in the Eastern Pacific coastal waters, and can be lethal for benthic and demersal organisms on the continental shelf. The normal seasonal cycle includes uplifting of isopycnals during upwelling in spring, which brings low-oxygen and low-pH water onto the shelf. Five years of continuous observations of subsurface dissolved oxygen off Southern California, reveal large additional oxygen deficiencies relative to the seasonal cycle during the latest La Niña event. While some changes in oxygen related to the isopycnal depression/uplifting during El Niño/La Niña are not unexpected, the observed oxygen changes are 2–3 times larger than what can be explained by cross-shore exchanges. In late summer 2010, oxygen levels at mid-depth of the water column reached values of 2.5 ml/L, which is much lower than normal oxygen levels at this time of the seasons, 4–5 ml/L. The extra uplifting of isopycnals related to the La Niña event can explain oxygen reductions only to roughly 3.5 ml/L. We find that the additional oxygen decrease beyond that is strongly correlated with decreased subsurface primary production and strengthened poleward flows by the California Undercurrent. The combined actions of these three processes created a La Niña-caused oxygen decrease as large and as long as the normal seasonal minimum during upwelling period in spring, but later in the year. With a different timing of a La Niña, the seasonal oxygen minimum and the La Niña anomaly could overlap to potentially create hypoxic events of previously not observed magnitudes.
 Marine ecosystems are affected by interannual and decadal climatic variations such as El Niño/La Niña events [Chavez et al., 1999], and by upper ocean warming and consequent stratification changes [Roemmich and McGowan, 1995]. The impacts of such climate processes on marine ecosystems are of particular interest in the productive upwelling ecosystem in the coastal Eastern Pacific (EP) where the world's largest oxygen minimum zone (OMZ) [Keeling et al., 2010; Levin et al., 2009; Feely et al., 2008; Stramma et al., 2008; Levin, 2003] is located. Coastal upwelling in the California Current lifts isopycnal (i.e., constant-density) surfaces and brings low oxygen and low pH water from intermediate depths onto the continental shelf [Feely et al., 2008]. In extreme cases this can lead to hypoxia causing mass die-offs of vulnerable species, and to corrosive conditions causing dissolution of shells and skeletons of calcifying organisms [Doney, 2010]. Since the dissolved oxygen (DO) and pH levels of this intermediate water have been decreasing over the past decades [Keeling et al., 2010; McClatchie et al., 2010; Stramma et al., 2008; Bograd et al., 2008], many benthic and demersal organisms in the upwelling ecosystem are already experiencing increased seasonal exposure to the low pH [Feely et al., 2008; Sabine et al., 2004] and low DO stresses due to the naturally occurring coastal upwelling. Interannual variability in the pH and DO levels, however, are not fully understood due to a lack of long and continuous pH and DO observations. In particular along the California Coast it is unclear whether, when, and for how long the pH and DO levels on the continental shelf reach the corrosive (e.g., pH < 7.75 [Feely et al., 2008]) and hypoxic (e.g., DO < 2.0 ml/L [Diaz and Rosenberg, 2008; Vaquer-Sunyer and Duarte, 2008]) thresholds in response to seasonal upwelling. The additional impact of interannual climate variability on these extreme conditions is largely unknown along the entire west coast of North America.
2. Data and Results
 We have analyzed a nearly 5-year long record of continuous DO measurements at mid-depth on a mooring at the 100 m isobath on the continental shelf off northern San Diego, California (Figure 1a). Other data collected on the mooring include currents, temperature/salinity, and chlorophyll fluorescence. This allows us to estimate the typical annual (i.e., seasonal) cycle of these quantities, and also to analyze departures from this. Of particular interest in this study is the impact of the 2009/2010 El Niño and 2010/2011 La Niña events.
Figure 1b shows seasonal anomalies, i.e., departures from the estimated mean seasonal cycle, of temperature (T), density (σθ), alongshore current (V), chlorophyll fluorescence (CHL), and DO concentration, all exhibiting large anomalies at the time of the 2009–2010 El Niño and the 2010–2011 La Niña. Based on a local DO-pH relationship (auxiliary material), approximate pH changes corresponding to the DO anomalies are also shown with an alternate y-axis. The Oceanic Niño index (ONI) is used as a measure for El Niño/La Niña occurrence and strength. During the 2009/2010 El Niño event, warmer T than normal conditions contributed to the negative density (σθ) anomalies, and vice versa during the 2010/2011 La Niña event (Figure 1b, curves B and C). The density and temperature anomalies peak in January–March 2010 with a time lag of 2–4 months after the corresponding ONI peak (November 2009) which yields a propagation speed of 0.8–2.3 m/s (Figure 1b, curves A, B, and C) from the equatorial region, which is consistent with the range of phase speeds of El Niño-related propagating coastally trapped waves [Todd et al., 2011a; Meyers et al., 1998; Ramp et al., 1997; Chelton and Davis, 1982; Enfield and Allen, 1980].
 The anomalies in DO from 35 m depth are very pronounced in 2009/2010, ranging from 0.7 ml/L above the seasonal average in the second half of 2009, to 1.5 ml/L below normal in the fall of 2010 (Figure 1b, curve F). Some of these large changes result from the unusual slope of density levels across the continental margin in 2009 and 2010, as can be judged from CalCOFI ship surveys1 which were conducted in July/August of both years (Figure 2). As expected in El Niño years, the isopycnals are depressed towards the coast in 2009 bringing with it higher DO from nearer the surface (Figures 2b and 2d), and in 2010 isopycnal surfaces are uplifted due to La Niña bringing low DO and low pH water from deeper and more offshore onto the shelf (Figures 2c and 2e). The difference plots Figures 2d and 2eshow the spatial distribution of the El Niño/La Niña related anomalies during summers, of particular interest here are the low DO and low pH values near the coast in July/August 2011. Note the deeper-reaching DO anomalies along the coast, at 50–150 m depth inFigures 2d and 2e. This is the depth of the poleward flowing California Undercurrent (CUC) or the Inshore Countercurrent (ICC) which is known to advect low-DO water from the south [Gay and Chereskin, 2009]. The alongshore current anomalies from Figure 1b, curve D appear to modulate the DO signal in this layer (negative, i.e., equatorward, flow anomalies weaken the advection and thus show higher DO; positive flow anomalies strengthen the advection leading to lower DO). Another tracer for the CUC/ICC is high spiciness (the quantity orthogonal to density in a temperature-salinity diagram). The high spiciness values close to the continental slope inFigures 2d and 2e imply strengthened advection from the south by the CUC/ICC system in July/August 2010.
 While the depression and uplifting of isopycnals in the cross-shore direction is an expected process during El Niño/La Niña, the following analysis shows that the DO and pH anomalies are 2–3 times larger than what can be explained with this cross-shore uplifting/depression of water masses.Figure 3ashows that the absolute DO concentrations observed at 35 m depth covary with the average seasonal cycle (“seasonal climatology”) of DO until July 2009. The seasonal cycle mainly results from the upwelling activity with a maximum in spring and with weakest amplitudes in fall/winter of each year. When the El Niño effect sets in, however, the observed DO departs from the seasonal cycle leading initially to higher-than-normal DO levels (Figure 3a). By spring 2010 the actual DO levels are back to the normal concentrations again, but in summer they begin to diverge in the opposite sense, leading to much lower values than the seasonal average in fall 2010. In August/September 2010 values close to 2.5 ml/L are reached in the low-pass filtered data set and 2.0 ml/L in daily values (hourly data reach 1.5 ml/L). How much of these anomalies relative to the seasonal cycle can be explained by cross-shore depression/uplifting of isopycnals due to El Niño/La Niña?
 During non-El Niño/La Niña months, and away from the surface but close to the coast, there is a stable relation between density and DO along CalCOFI line 93, which allows to predict DO from density with an rms error of 0.35 ml/L (Figure S1a in theauxiliary material). Thus the measured density at the mooring can be used to calculate what DO levels would be observed if all the changes resulted from only isopycnal depression/uplifting. This “density-predicted DO” (DODen) is shown as the orange line of Figure 3a. In summer-fall 2009, DO levels are 0.75 ml/L higher than during the normal seasonal cycle, and approximately half of this can be explained by the depression of isopycnals, bringing water with higher DO from nearer the surface to these depths (Figure 3a). A large negative anomaly in DO is observed in fall 2010 (difference between heavy black and dashed black line in Figure 3a), over 1.5 ml/L in amplitude. Here the uplifting of density related to La Niña can only explain 30–50% of this DO drop relative to the normal seasonal cycle. The absolute DO levels reached are 2.5 ml/L on a 1-month timescale, virtually identical to minimal values during the typical upwelling maximum in spring (Figure 3a). Note that the near-bottom DO sensor which had been deployed since fall 2009 showed even lower values, both at the upwelling maximum and during the fall La Niña event (thin black line inFigure 3a).
 In order to investigate the origin of the additional DO anomalies which cannot be explained by uplifting (depression) and onshore (offshore) advection of low (high) DO water during La Niña (El Niño), it is revealing to form the DO residual, by subtracting the density-predicted DO values from the observed ones (DOObs). This residual (DOObs- DODen) is compared with the time series of both CHL and V in Figures 3b and 3crespectively. The CHL follows the low-frequency (seasonal period) behavior of the DO residual (Figure 3b), with CHL being anomalously high during the El Niño phase, and anomalously low during La Niña. In particular the usual seasonal fall maximum in chlorophyll is completely absent in the La Niña year 2010, which would lead to lower-than-normal DO values. It thus seems that the overall evolution of the DO residual, i.e., the seasonal anomalies during El Niño/La Niña which cannot be explained by cross-shelf advection or isopycnal shoaling, is governed by the increase/decrease of primary production during El Niño/La Niña. In addition there is a strong response of the alongshore currents in La Niña year 2010 (Figure 3c). The usually weak poleward flow at this location and depth becomes abnormally strong during fall 2010 (blue solid vs blue dashed in Figure 3c), with a time evolution that closely matches the occurrence of the large negative DO residual in summer/fall/winter 2010 (blue vs black in Figure 3c). The spiciness section of Figure 2e supports the notion that there was a poleward advection of low DO water which is suggested to occur near the coast [Todd et al., 2011b; Gay and Chereskin, 2009], by the CUC/ICC.2. The nearshore spiciness in Jul/Aug 2010 (Figure 2e) provides further evidence of enhanced nearshore poleward advection during the La Niña.
 A simple linear combination of DODen, chlorophyll fluorescence CHL (DOChl), and alongshore advection V (DOV), has remarkable skill in reproducing the complete evolution of the observed oxygen concentrations during the El Niño/La Niña (Figure 3d). This strongly suggests that the large anomalous DO decline at 35 m depth during August/September 2010, leading to hypoxic conditions can causally be explained by La Niña related uplifting/cross-shore advection of deeper water masses, together with reduced primary production, and a short-term intensification of the poleward advection by CUC or ICC. While the latter effect of enhanced poleward advection is expected based on the known properties of the CUC/ICC and on the high spiciness and low DO (and low pH) inFigure 2e, the decreased chlorophyll concentration during the La Niña is surprising.
 The normal state of La Niña is similar to enhanced upwelling conditions (more uplifted isopycnals) and thus also leads to enhanced CHL concentrations near the surface (which we verified with MODIS satellite images, not shown here). Strong upwelling conditions however create a bloom at the surface (only in spring during normal years) which is not seen at 35 m since the nitracline is near the surface. In summer/fall however, isopycnals relax and the nitracline is deeper, creating high CHL values at the top of the nitracline [Mantyla et al., 2008]. This is the Aug/Sep maximum seen in the seasonal CHL concentration at 35 m in Figure 3b. A La Niña, however, keeps isopycnals uplifted even in summer (Figure 2c), thus the nitracline is near the surface similar to the spring bloom, with an absence of primary production at 35 m depth.
 The large amplitudes of interannual DO (and pH) changes associated with the El Niño/La Niña dynamics imply that the carbonate saturation state [Juranek et al., 2009] would also be significantly affected by these events. The pH threshold of 7.75 (magenta lines in Figures 2a–2c and 3a) corresponds to an aragonite saturation level of Ω = 1, which means below this the water is corrosive for calcium forming organisms [Feely et al., 2008]. At 35 m depth, pH falls far below this threshold both during seasonal upwelling and for another 2 months during the La Niña phase as shown by the additional pH axes drawn in Figure 3.
 With the phasing of the La Niña impacts during the seasonal cycle as observed in 2010, the net effect is to approximately double the normal seasonal exposure of organisms to hypoxic and corrosive conditions, by introducing a second period of such conditions a few months after the typical upwelling, below the layers of the surface upwelling bloom. The phase of La Niña during the calendar year is highly variable. In particular, the ONI timeseries from the equator show that some of the past La Niña events (e.g., 1964, 1973, 1988) initiated 3–4 months earlier than the 2010 La Niña did (auxiliary material). If this happens, the potential exists for the seasonal upwelling and the La Niña impacts on hypoxia and low-pH to overlap and add, which could lead to more extreme hypoxic and corrosive conditions than observed in 2010.
 However, our observations are limited to this particular El Niño/La Niña event, and some caution is necessary when extending the results to other events. Anomalous poleward flows and southern water mass characteristics were observed during the 1997–1998 El Niño and there was anomalous equatorward transport during the 1998–1999 La Niña [e.g., Lynn and Bograd, 2002]. This different behavior compared to our observations may be related to the abrupt switch of the Central Pacific-type El Niño [e.g.,Yeh et al., 2009] to the La Niña in 2010 in contrast to those of the Eastern Pacific-types like 1997–1998 or 1982–1983 events. These differences will require further investigation.
 We are grateful to all members of the SIO Ocean Time-series Group for their dedicated field work to maintain the mooring for more than five years. The CalCOFI data are also greatly appreciated. We acknowledge funding provided by NSF (OCE0551363) and NOAA (NA17RJ1231). The first author was partially supported by a JIMO postdoctoral fellowship at SIO.
 The Editor thanks two anonymous reviewers for their assistance in evaluating this paper.