SciELO - Scientific Electronic Library Online

vol.70 suppl.1Oxygen deficiency and benthic communities in the peruvian upper continental marginGlobal warming projection of the change in dissolved oxygen concentrations in low oxygen regions of the oceans author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand




Related links


Gayana (Concepción)

Print version ISSN 0717-652XOn-line version ISSN 0717-6538

Gayana (Concepc.) vol.70  suppl.1 Concepción Oct. 2006 

Suplemento Gayana 70: 37-45, 2006

The Perú-Chile undercurrent and the oxigen minimum zone variability off central Chile

La corriente sub-superficial del Perú-Chile y la variabilidad de la zona de mínimo oxígeno frente a Chile central

Samuel Hormazábal1, Gary Shaffer1, Nelson Silva2 & Eduardo Navarro1

1Department of Geophysics, University of Concepción, P.O. Box 160-C, Concepción, Chile,

2Catholic University of Valparaíso, Valparaíso, Chile


Ten years of continuous current meter recordings near the core of the Peru-Chile Undercurrent (PCU) over the continental slope at 30°S off Chile, local and equatorial wind stress data, and monthly oceanographic sections off Valparaíso (33°S) are used to study the variability of the PCU and the Oxygen Minimum Zone (OMZ) off Chile. The physics that governs the PCU variability differs for intraseasonal periods and for longer periods: Coastal trapped waves dominate the intraseasonal band whereas Rossby waves dominate for periods greater than about 120 days. Semiannual and annual period fluctuations can largely be explained by Rossby waves forced by equatorial Kelvin waves and local winds. Seasonal variations in the PCU are strongly modulated over El Niño/La Niña cycles. During warm El Niño events, the PCU undergoes strong, semiannual variability whereas during La Niña events, annual variability dominates. These results together with oceanographic sections off Valparaíso suggest that Rossby waves provide a physical mechanism to explain a significant fraction of seasonal variability observed in the OMZ. During the downwelling (upwelling) phase of these waves the poleward (equatoward) flow is stronger (90° out of phase with respect to sea level), the isotherms are deeper (shallower), the OMZ thickness increases (decreases), the mean concentration of salinity and silicate increases (decreases) and oxygen concentration and nitrate deficit decrease (increase) in the OMZ.

Keywords: Rossby waves, oxygen minimum zone, El Niño/La Niña cycles, Peru-Chile Undercurrent.


Una década de registros continuos de corrientes medidas cerca del núcleo de la corriente subsuperficial de Perú-Chile (PCU), sobre el talud continental en los 30°S frente a Chile, datos del esfuerzo del viento a lo largo del ecuador y a lo largo de la costa de Chile, además de secciones oceanográficas realizadas mensualmente frente a Valparaíso (33°S), son usados para estudiar la variabilidad de la PCU y la zona de mínimo de oxígeno (OMZ) frente a Chile. La física que gobierna la variabilidad de la PCU es diferente entre periodos intraestacionales y periodos mayores. Ondas atrapadas a la costa dominan la banda intraestacional mientras que las Ondas de Rossby dominan en periodos mayores que 120 días. Fluctuaciones con periodos anuales y semi anuales pueden ser largamente explicadas por ondas de Rossby forzadas por ondas de Kelvin ecuatoriales y por vientos locales. Variaciones estacionales en la PCU son fuertemente moduladas por los ciclos de El Niño/La Niña. Durante los eventos calidos de El Niño, la variabilidad semi anual de la PCU se intensifica, mientras que durante los eventos fríos de La Niña la señal anual es dominante. Estos resultados junto a las secciones oceanográficas realizadas frente a Valparaíso, sugieren que las ondas de Rossby entregan el mecanismo físico para explicar una fracción significativa de la variabilidad estacional observada en la OMZ. Durante la fase de hundimiento (surgencia) de esas ondas el flujo se intensifica (90° desfasado respecto del nivel del mar) hacia el polo (ecuador), las isotermas son mas profundas (más someras), el espesor de la OMZ incrementa (disminuye), la concentración media de salinidad y silicato incrementa (disminuye), y el oxígeno y el déficit de nitrato disminuye (incrementa) en la OMZ.

Palabras Claves: Ondas de Rossby, Zona de mínimo de oxígeno, El Niño, Corriente subsuperficial Perú-Chile.


Much of the intraseasonal and interannual variability of slope currents off Chile at 30°S is ultimately forced in the equatorial Pacific (Shaffer et al. 1997; Hormazabal et al. 2002; Pizarro et al. 2002). Wind-forced Kelvin waves propagate eastward along the equator to the South American coast. At intraseasonal periods, such waves force poleward-propagating, coastal trapped waves which explain much of the observed intraseasonal variability of sea level and coastal currents in the Peru-Chile current system (Enfield 1987; Shaffer et al. 1997, Hormazábal et al. 2002). However, south of 20°S, some oceanic intraseasonal variability is forced by local winds due to equatorial-midlatitude teleconnections in the atmosphere (Hormazábal et al. 2002). At seasonal and interannual periods, the equatorial Kelvin waves force Rossby waves which explain much of the observed seasonal and interannual variability of sea level along the South American coast and of the Peru-Chile Undercurrent (PCU) near 30°S (Pizarro et al. 2002, Ramos et al. 2006). On the other hand, model results indicate that poleward undercurrents and Rossby waves can be forced by local winds at eastern boundaries (McCreary 1981, Philander & Yoon 1982). Nutrient-rich but oxygen-poor Equatorial Subsurface Water (ESSW) is transported poleward in the PCU, forming the core of the Oxygen Minimum Zone (OMZ). Rossby and coastal trapped waves can produce strong current variability and vertical oscillations of several tens of meters in the isotherm depth and the upper limit of the OMZ. These vertical oscillations can extend more than one hundred meters during El Niño periods (Morales et al. 1999; Thomas et al. 2001). Here we use current, sea level and winds observations to identify the dominant physical processes in the region and then use oceanographic sections to study how these physical processes can modulate the spatial and temporal variability of the OMZ.


Ocean current data were obtained at 220 m depth (near the PCU core) between November 1991 and June 2001 from a slope mooring off Chile (30° 19'S, 71° 47'W; ~875 m water depth and 10 km from the coast). Sea level data at Caldera (27°S) for this period were supplied by the Sea Level Center, University of Hawaii. Mean deviation of surface atmospheric pressure were taken from NCEP/NCAR reanalysis for a 2.5° x 2.5° grid off Caldera and added to the sea level data (scale factor 1 cm/1 mb) to form adjusted sea level. Sea surface wind stress was calculated for this period on 1° x 1° grids in the equatorial Pacific and along the South American west coast using ERS-1, ERS-2 and QuikSCAT (after August 1999) satellite scatterometer measurements (Département d'Océanographie Spatiale, IFREMER, and SeaWinds Project, Jet Propulsion Laboratory). The resulting daily time series were analyzed using the wavelet method (e. g., Torrence & Compo, 1998).

Between July 1994 and September 1995, 14 monthly oceanographic cruises were carried out off Valparaiso (32°55'S). On each cruise, an offshore section with six CTD stations was occupied, using a Seabird SB-25 CTD. At an oceanographic station located ~26 km offshore, water samples at oceanographic standard depth were taken with Niskin bottles between sea surface and 800 m depth. Water samples were used to determine salinity (with an induction salinometer), dissolved oxygen (by the Winkler method modified by Carpenter (1965)), and nutrient concentrations (phosphate, nitrate+nitrite and silicate; using visible spectrophotometry). Nitrate deficit (N*) was estimated according to Deutsch et al. (2001).

Peru-Chile Undercurrent Variability

Figure 1 shows time series of along equator wind stress (170°E), local alongshore wind stress, local adjusted sea level (ASL) and local slope currents at 220 m depth, near 30°S off Chile . Along equator winds reflect interannual modulation of the westward trade winds, interspersed with intraseasonal wind events associated with the Madden-Julian Oscillation. Alongshore winds are equatorward (upwelling-favorable), interspersed with intraseasonal and synoptic-scale wind events. There is mean poleward flow at 220 m of 12.4 ± 1.9 cm s-1, within the PCU core. This 10-year mean flow is very similar to 6-year mean PCU flow reported earlier for this site (Shaffer et al, 1999).

The wavelet transforms of currents observed at 220 m depth, near 30°S off Chile (Fig. 2) shows significant variance in the intraseasonal band (30-90 days) with a strong energy enhancement during the 1991-92 and 1997-98 El Niño events. Strong intraseasonal variability of about 50 day period in currents and sea level was found during the 1991-92 event while such variability exhibited about 70 day periods followed by about 40 day periods during the 1997-98 event (Fig. 2). A similar but weaker sequence from longer to shorter period, intraseasonal oscillations was also observed in 1994/1995 and such dual period structure is also evident in the equatorial winds (not shown). In fact, there are counterparts in currents and sea level of central Chile to much of the time-frequency structure of the equatorial winds. This is consistent with earlier conclusions (i.e. Hormazábal et al. 2002) that most of the intraseasonal variability off central Chile has its origin in equatorial wind forcing by way of remotely-forced coastal trapped waves. Cross-wavelet spectra show phase angles of about 180° at high coherency between local ASL and slope currents in the intraseasonal band (Fig. 2). This result is consistent with cross slope geostrophic balance in coastal trapped waves.

Figure 1. Low passed, zonal wind stress on the equator at 170°W (from the TAO project), and alongshore wind stress at a nearshore 1° x 1° grid near the mooring site (see below), from ERS 1/2 (1991-1999) and QuikSCAT (1999-2001) observations, low passed adjusted sea level at Caldera (27°S), and low passed, time series of vector currents observed at 220 m depth at the slope mooring site (30° 19'S, 71° 47'W, water depth 875 m), over the 1991-2001 study period. Sea level and wind data were low passed filtered with a 51 day, Cosine-Lanczos filter with a half amplitude point at 15 days. Original hourly current data were low passed filtered with a 181 hour, Cosine-Lanczos filter with half amplitude point at 60 h and then daily averaged. Gray color in the vector current marks interpolated data except for the period February 12, 1999 to March 2, 2000 at 220 m when only current speeds were observed. Current directions for this period were estimated from observations at 485 m using a linear estimate based on observations from the entire study period (see Hormazábal et al. 2004)

Figure 2. (a) Low-low pass (twice filtered with 3 month running means) anomalies of sea surface temperature for the Niño 3.4 region (SST Niño 3,4), and low low-pass Southern Oscillation Index (SOI) multiplied by _1, (b) wavelet power spectrum (Morlet wavelet) for current observations at 220 m depth (core of Peru-Chile Undercurrent off Coquimbo), and Wavelet phase angle between Caldera ASL and alongshore current at 220 m of depth. Phase results are only plotted for coherence squared over 0.6. Vertical dashed lines mark warm and cold periods, and horizontal dashed-dotted lines mark periods of 50, 64, 90, 128, 182, and 365 days. The black contour in panel (b) indicates the 95% significance level.

Figure 3. Comparison for the seasonal band of model results (dashed lines) with observations of the Peru-Chile Undercurrent at 30°S off Chile (full lines) for the period January 1, 1993 to June 1, 2001. These signals have been band-pass-filtered using the Morlet wavelet (140-520 day periods retained).

Wavelet spectra clearly demonstrate that seasonal variations in slope currents, and particularly within the PCU (220 m depth) are strongly modulated over El Niño/La Niña cycles (Fig. 2). Semiannual variability is most prominent during warm events while strong annual variability also appears during cold events. This structure is also evident in the ASL data although it is somewhat masked by an annual cycle associated with steric effects. Likewise, cross wavelet spectra between local ASL and flow in the PCU show significant coherence in the semiannual band for warm events and in the annual band for cold events. For periods increasing from 120 days into the semiannual and annual bands, phase angles between local sea level and flow in the PCU remain quite uniform, close to 90°, as found in previous results based on shorter records and traditional spectral analysis (e.g. Pizarro et al. 2002). These phase angles strongly suggest seaward propagation of baroclinic Rossby waves for both semiannual and annual bands.

To further evaluate the remote forcing contribution to seasonal-scale, coastal current variability off Chile, a simple, linear, ocean model of equatorial Kelvin waves, forced by satellite wind stress data from satellite observations, is used together with simple Rossby and coastal trapped wave forcing (McPhaden & Yu 1999, Hormazábal et al. 2002; Pizarro et al. 2002). Integration of the equatorial Pacific model was started with zero Kelvin wave amplitude at 150°E and considered a band between 13°N and 13°S. The model retained the first two baroclinic modes and model parameters were taken from Kessler & McPhaden (1995) with damping coefficients from McPhaden & Yu (1999). Modelled seasonal-scale variability of alongshore flow at 220 m depth at 30°S agree best with observations during warm El Niño periods (Fig. 4). During warm periods (January 1, 1993 -July 1, 1995 and March 1, 1997 -June 1, 1998) model results lead observations by about one month, the maximum correlation between model results and PCU observations were 0.90 with a lag time of 36 days and RMS ratio (results to observations) of 1.23. During cold, La Niña periods (July 1, 1995 -March 1, 1997 and June 1, 1998 -June 1, 2001), our simple, remote-forcing model does not reproduce alongshore flow well (maximum correlation of 0.24 with a lag time of 49 days and RMS ratio of 0.4). Wavelet analysis suggests that winds along the South American coast play a larger role in forcing seasonal scale variability during La Niña periods. This should be tested in the future with a more complete model.

Figure 4. Seasonal- band time series of alongshore Peru-Chile Undercurrent observed at 30°S off Chile (upper panel) and time-depth diagram of temperature, salinity, oxygen and nitrate deficit at an oceanographic station ~ 26 km off Valparaíso (33ºS).


Figure 4 shows time-depth variability of physical and chemical water column properties off Valparaíso. Higher (lower) salinities and silicates, lower (higher) dissolved oxygen and nitrate deficit concentrations are observed when a thicker (thinner) ESSW layer is present. During these periods, physical and chemical ESSW characteristics are stronger (weaker), and the offshore extension of ESSW increases (decreases). Taken together with the PCU records, these results suggest that, for seasonal periods, extremes of water column properties are associated with maximum (minimum) poleward displacement perturbations in the PCU, following maximum (minimum) poleward flow by about 90° in phase.

The OMZ thickness exhibits a significant linear relationship with mean dissolved oxygen, salinity and silicate concentrations, averaged vertically over the OMZ layer (Fig. 5). An OMZ thickness increase (decrease) is associated with an increase (decrease) in salinity and silicate and is associated with a decrease (increase) in oxygen and nitrate deficit. A similar significant linear relationship is observed between the 11ºC isotherm depth and mean dissolved oxygen, salinity and silicate concentrations within the OMZ. These linear relationships imply a significant linear relationship between the 11ºC isotherm depth and the OMZ thickness, as shown in Fig. 6. In our study region, isotherm depth variability is usually associated with passing waves and, for this particular study period, oscillations of the 11°C isotherm depth are associated with semiannual Rossby waves, as showed by the PCU variability analysis.

Figure 5. Relationships between Oxygen Minimum Zone thickness (left panels), 11ºC isotherm depth (right panels) and mean oxygen, salinity and silicate concentration in the OMZ. The OMZ boundaries were taken to be the upper and lower depths of the 1 ml/L dissolved oxygen concentration.

Figure 6. Relationship between the Oxygen Minimum Zone thickness and the 11ºC isotherm depth.

Figure 7. Conceptual model for Rossby waves effects on the physical and chemical water column properties variability. During the downwelling (upwelling) phase of Rossby waves the poleward (equatoward) flow is stronger (90° out of phase respect to sea level), the isotherms are deeper (shallower), the OMZ thickness increases (decreases), the mean concentrations of salinity and silicate increase (decrease) and oxygen and nitrate deficit decrease (increase) in the OMZ. Arrows show flow direction (h equator, i pole). SAAW is Subantarctic Water, ESSW is Equatorial Sub-surface water and AAIW is Antarctic Intermediate Water.

A conceptual model for the effects of passing Rossby waves on OMZ variability is shown in Fig. 7: Rossby waves raise (depress) the sea surface by a few centimeters, depress (raise) the thermocline tens of meters, strengthen (weaken) ESSW characteristics and increase (decrease) OMZ thickness. During the downwelling (upwelling) phase of these waves the poleward (equatoward) flow is stronger (90° out of phase with respect to sea level).


Here we have shown how seasonal scale winds along the equator in the Pacific strongly influence sea level and the Peru-Chile Undercurrent off central Chile by way of equatorial Kelvin waves and subtropical Rossby waves. During warm El Niño periods, seasonal scale variability in sea level and in the Peru-Chile Undercurrent off Chile is dominated by semi-annual variability forced by winds along the equator. During cold La Niña periods, both annual and semi-annual periods and both remote and local forcing are important.

These results, together with oceanographic data off Valparaiso, suggest that at seasonal periods the variability of water column properties in the OMZ is strongly modulated by Rossby waves. During the downwelling (upwelling) phase of Rossby waves the isotherms are deeper (shallower), the OMZ thickness increases (decreases), the OMZ upper limit is shallower (deeper), the mean concentration of salinity and silicate increases (decreases) and oxygen and nitrate deficit decreases (increases) in the OMZ.

According to our results, surface layer properties in the coastal upwelling system off central Chile could be significantly influenced by remotely-forced Rossby waves. The upwelling source water could be modulated by the passage of these low frequency waves such that the surface layer would tend to be cooler (warmer) and more nutrient rich (poor) in association with the upwelling (downwelling) phase of these waves. These waves could be one of the factors contributing to the supply of nutrients to the upper layer and to observed variations in primary production.


This work was supported by grants from the Danish National Research Foundation, the Chilean National Research council (FONDAP, FONDAP-COPAS, FONDECYT 1040618, PBCT ACT-19), DGIP 223.759-94, Fundación Andes and the Swedish Agency for Research Cooperation with Developing Countries.


Carpenter, J. H. 1965. The Chesapeake Bay Institute technique for the Winkler dissolved oxygen method. Limnology and Oceanography 10, 141-143.         [ Links ]

Deutsch, C., N. Gruber, R. M. Key, J. L. Sarmiento. 2001. Denitrification and N2 fixation in the Pacific Ocean. Global Biogeochemical Cycles, 15(2), 483-506.         [ Links ]

Enfield, D. B. 1987. The intraseasonal oscillation in eastern Pacific sea levels: How is it forced?, J. Phys. Oceanogr., 17, 1860-1876.         [ Links ]

Hormazabal, S., G. Shaffer & O. Pizarro. 2002. Tropical Pacific control of intraseasonal oscillations off Chile by way of oceanic and atmospheric pathways. Geophys. Res. Lett., 29(6), 10.1029/2001GL013481.         [ Links ]

Hormazabal, S., G. Shaffer & O. Leth. 2004. The Coastal Transition Zone off Chile. J. Geophys. Res., 109(C01021), doi:10.1029/2003JC001956         [ Links ]

Kessler, W. S. & M. J. McPhaden. 1995. Oceanic Equatorial waves and the 1991-1993 El Niño, J. Climate, 8, 1757-1774.         [ Links ]

McCreary. J. A linear stratified ocean model of the coastal Undercurrent, Phil. Trans. Roy. Soc. London, A302, 385-413, 1981.         [ Links ]

McPhaden, M. J., & X. Yu. 1999. Equatorial waves and the 1997-1998 El Niño, Geophys. Res. Lett., 26, 2961-2964.         [ Links ]

Morales, C., S. Hormazabal & J. Blanco. 1999. Interannual variability in the mesoscale distribution of the depth of the upper boundary of the oxygen minimum layer off northern Chile (18º-24ºS): implications for the pelagic system and biogeochemical cycling. J. Mar. Res., 57, 909-932.         [ Links ]

Philander, S. G. & J.-H. Yoon. 1982. Eastern Boundary Currents and Coastal Upwelling, J. Phys. Oceanogr., 12, 862—879.         [ Links ]

Pizarro, O., A. J. Clarke, & S. Van Gorder. 2001. El Niño sea level and currents along the South American coast: Comparison of observations with theory. J. Phys. Oceanogr. 31, 1891-1903.         [ Links ]

Pizarro, O., G. Shaffer, B. Dewitte & M. Ramos. 2002. Dynamics of seasonal and interannual variability of the Peru-Chile Undercurrent. Geophys. Res. Letters, 29(12), 10.1029/2002GL014790.         [ Links ]

Shaffer, G., O. Pizarro, L. Djurfeldt, S. Salinas & J. Rutllant. 1997. Circulation and low frequency variability near the Chile coast: Remotely-forced fluctuations during the 1991-1992 El Niño. J. Phys. Oceanogr., 27, 217-235 .         [ Links ]

Shaffer, G., S. Hormazabal, O. Pizarro & S. Salinas. 1999. Seasonal and interannual variability of currents and temperature over the slope of central Chile. J. Geophys. Res., 104, C12, 29,951-29,961, 1999.         [ Links ]

Thomas, A. C., J. L. Blanco, M. E. Carr, P. T. Strub, & J. Osess. 2001. Satellite-measured chlorophyll and temperature variability off northern Chile during the 1996-1998 La Niña and El Niño. J. Geophys. Res., 106, 899-915.         [ Links ]

Torrence, C., & G. P. Compo. 1998. A practical guide to wavelet analysis. Bull. Amer. Meteor. Soc., 79, 61-78.         [ Links ]

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License