U.S. Dept. of Commerce / NOAA / OAR / PMEL / Publications

Variability in the Eastern Equatorial Pacific Ocean During 1986-1988

Michael J. McPhaden and Stanley P. Hayes

NOAA/Pacific Marine Environmental Laboratory, Seattle, Washington

Journal of Geophysical Research, 95(C8), 13,195-13,208 (1990)
Not subject to U.S. copyright. Published in 1990 by the American Geophysical Union.

We examine variability in the eastern equatorial Pacific during 1986-1988 using conductivity-temperature-depth data, velocity and temperature data from equatorial moorings between 110W and 140W, and wind data from a basin scale zonal array of islands and moorings between 110W and 165E. The period studied coincides with the El Nio/Southern Oscillation (ENSO) event of 1986-1987 and a subsequent cold event in 1988. Weak warm sea surface temperature anomalies first appeared in the eastern equatorial Pacific in mid-1986 and increased to >1C in September-November 1986 in association with a 30 cm s weakening of the South Equatorial Current and a 20- to 40-m depression of the thermocline. These warm anomalies lasted until early 1988, after which a large-scale shoaling of the thermocline led to sea surface temperatures more than 3C colder than climatology. Year-to-year fluctuations in the eastern Pacific were related primarily to zonal wind variations in the central and western Pacific. Westerly wind stress anomalies of 0.02-0.05 N m were observed between 140W and 165E from the latter half of 1986 until the end of 1987; these were replaced by easterly wind anomalies of similar magnitude between 157W and 165E in 1988. Energetic intraseasonal fluctuations with periods of 2-3 months were also prominent in zonal current, temperature, and dynamic height time series. These fluctuations propagate eastward at approximately first baroclinic mode Kelvin wave phase speeds and are forced west of the date line by episodes of westerly winds. Extrema in several oceanic variables occurred in association with these waves, though their precise dynamical link to the ENSO cycle is unclear from our data. Sea surface temperature and thermocline depth anomalies at 0, 110W were less pronounced during the 1986-87 ENSO than during the 1982-1983 ENSO; the Equatorial Undercurrent, though weaker than normal in early 1987, did not disappear as it did in early 1983.

1. Introduction

Kousky and Leetmaa [1989] have described the planetary scale evolution of the 1986-1987 El Nio/Southern Oscillation (ENSO) event using satellite data and general circulation model analyses. Warm sea surface temperature (SST) anomalies began to develop in the equatorial Pacific Ocean in mid-1986, indicating the onset of the event. These anomalies persisted throughout 1987, during which the trade winds weakened in the central and western Pacific and anomalous deep atmospheric convection developed near the date line. The southern oscillation index (SOI), which is a measure of the driving force for the Pacific trade winds, reached its second lowest value in nearly 20 years in mid-1987. (The SOI was lower only in 1982-1983 during the most intense ENSO of the century.) The 1986-1987 ENSO was over by early 1988, after which the ocean-atmosphere system rebounded with the coldest equatorial SST anomalies and the highest SOI values since the mid-1970s [Climate Analysis Center, 1988].

Plate 1 illustrates the interannual changes in SST in the tropical Pacific from near-normal conditions in the central and eastern Pacific in June 1986 (patchy SST anomalies along the equator typically <1C in magnitude) to warm ENSO conditions in June 1987 (basin scale warm SST anomalies of >1C) to colder than normal conditions in June 1988 (SST anomalies of -3C in the eastern equatorial Pacific). The occurrence of warm ENSO SST anomalies in the equatorial Pacific is associated with a characteristic global pattern of climatic variability [Ropelewski and Halpert, 1986, 1987], and in many respects the global effects of the 1986-1987 ENSO fit this pattern [Kousky and Leetmaa, 1989]. In addition, the cold equatorial SST anomalies following the 1986-1987 ENSO were believed to contribute in part to the severe North American drought during the summer of 1988 [Trenberth et al., 1988; Palmer and Brankovic, 1989]. Understanding the mechanisms responsible for equatorial SST anomalies is therefore one of the central themes of the Equatorial Pacific Ocean Climate Studies (EPOCS) program [Hayes et al., 1986] and the Tropical Ocean-Global Atmosphere (TOGA) program [National Academy of Sciences, 1988].

Plate 1. Sea surface temperature anomalies for June 1986, June 1987, and June 1988 based on analyses from the National Meteorological Center Climate Analysis Center [Reynolds, 1988]. Superimposed are the locations of time series measurements from current meter moorings (squares), ATLAS moorings (diamonds), and island wind stations (circles). Also shown are the locations of CTD sections along the equator and along 110W.

The purpose of this study is to describe oceanic variability in the eastern equatorial Pacific during 1986-1988 using in situ measurements of wind, current, temperature, and salinity from moored buoys, islands, and research vessels. It is possible from this description to provide a preliminary, qualitative evaluation of processes responsible for some of the interannual changes in the current and hydrographic structure of the eastern equatorial Pacific. A complementary study [McPhaden et al., 1990a] discusses variability in the western equatorial Pacific during 1986-1988.

The paper is organized as follows. Data sources and processing are described in section 2. Some details dealing with data gaps, gridding, and the estimation of mean seasonal cycles are relegated to appendices for the sake of brevity. Section 3 discusses the regional scale hydrography of the eastern equatorial Pacific during 1986-1988. Time series variability at 110W and 140W is then described in section 4, and the extent to which this variability may be remotely versus locally forced is discussed in section 5. Section 6 compares variability in the eastern equatorial Pacific during the 1982-1983 ENSO and the 1986-1987 ENSO, using long time series from 0, 110W (Plate 2). The paper concludes with a summary and discussion of the most relevant results in section 7.

Plate 2. Contour plots of temperature (in degrees Celsius) and zonal velocity (in centimeters per second) for 1980-1989. Daily data have been gridded according to procedures described in Appendix A and smoothed to approximately monthly means using a 51-day filter [Bendat and Piersol, 1971].

2. Data

The moored time series used in this study consist of current, temperature, and wind data from the locations shown in Plate 1. These measurements were made from taut wire surface mooring at depths ranging from 3 to 5 km. The lengths of the moored time series and depths instrumented for the period 1986-1988 are shown in Figure 1. The number of depths instrumented with current meters was reduced from seven to five in boreal fall 1987 in response to changing scientific priorities. Typically, six additional depths were instrumented with SeaData temperature recorders (TRs). Data gaps during 1986-1988 were filled where appropriate according to procedures outlined in Appendix A in order to facilitate analysis. The data distribution for times prior to 1986 (used to generate Plate 2 and to compile mean seasonal cycles, as discussed in Appendix B) is given by Halpern [1987a], Halpern et al. [1988], and McPhaden and Taft [1988].

Figure 1. Record lengths of daily time series from moorings and island wind stations.

The current meter moorings were equipped with EG&G model 610 vector averaging current meters (VACMs) and vector measuring current meters (VMCMs) in the upper 250 m. The difference between VACM and VMCM measurements is generally <5 cm s [Halpern, 1987b], so for the purposes of this study they can be considered as interchangeable. Temperature and velocity data were recorded at 15-min intervals and then processed to daily averages. Instrumental accuracy of the VACM and TR temperature sensors is approximately 0.01C and 0.05C, respectively. SST is measured 1 m below the surface using either a Yellow Springs Instrument (YSI) model 44032 temperature sensor (calibrated accuracy of 0.01C) or a YSI model 44204 temperature sensor (calibrated accuracy of 0.05C). Additional information on the processing of data from equatorial current meter mooring is given by Freitag et al. [1987].

Winds from the equatorial current meter mooring were sampled 4 m above the mean water line on the surface toroid with either a vector averaging wind recorder (VAWR) or an Argos meteorological platform (AMP). The VAWR is an inverted VACM equipped with a Climet cup model 011-2B three-cup anemometer and pivoted vane [Freitag et al., 1989]. The AMP was designed at Pacific Marine Environmental Laboratory (PMEL) to transmit data in real time and is equipped with an R.M. Young model 05103 propeller and vane. Predeployment and postdeployment calibrations for the VAWR and AMP indicate expected instrumental errors in wind speed of about 0.1 m s. Comparison of the two wind systems in a field experiment near 0, 140W suggests that for our purposes the cup and vane and the propeller and vane systems can be considered interchangeable [Freitag et al., 1989].

The 0, 125W and 2S, 165E moored data used in this study were collected with ATLAS (autonomous temperature line acquisition system) thermistor chains. ATLAS [Milburn and McClain, 1986] is a taut wire surface mooring which measures winds, air temperature, SST, and 10 subsurface temperatures to a maximum depth of 500 m. Winds are measured using an R.M. Young model 05103 propeller and vane assembly. Thermistors are calibrated prior to deployments to an accuracy of about 0.005C; in situ comparisons with nearby conductivity-temperature-depth (CTD) casts indicate a long-term accuracy of better than 0.1C. Data are telemetered to shore via Service Argos as 2-hour averages (or in some cases 1-hour averages). Normally, five unique data transmissions are received each day. The basic time series is taken to be daily averages of these data.

Winds are measured at Nauru (032 S, 16654E), Baker (012N, 17629W), and Christmas (159N, 15729W) islands from an R.M. Young model 05103 propeller and vane anemometer mounted on a tower 10 m above the ground. Anemometers are replaced every 6 months to 1 year and are calibrated prior to deployment to within 0.2 m s. Data are vector averaged for 40 min of each hour, and then three individual hourly samples are transmitted to shore via GOES geostationary satellite. Data are processed to daily means for this study.

Nauru winds tend to underestimate the amplitude of variations at the 0, 165E mooring site during periods of westerlies [McPhaden et al., 1988, 1990a] when the wind sensor, located on the northeast side of the island, is in the lee of a 40-m hill. Nonetheless, there is a high correlation between daily time series at the two locations for the period December 1986 to October 1987 (0.88 for zonal winds and 0.85 for meridional winds), so that Nauru winds can be used as an index for winds at the mooring site during the ENSO event. Nauru winds are a more exact indicator of open ocean conditions during periods of easterlies as occurred during early 1986 and during 1988.

CTD data were collected from National Oceanic and Atmospheric Administration (NOAA) research vessels using a Neil Brown Instrument Systems Mark III CTD along the transacts indicated in Plate 1. Stations were occupied approximately every 5 of longitude and 1 in latitude during spring and fall cruises each year. Casts were made to at least 1000 dbar and processed to 1-dbar resolution. Lynch et al. [1988] describe the acquisition and processing of these data in greater detail.

Moored temperature data were used to calculate surface dynamic height relative to 250 dbar. To estimate salinity in the dynamic height calculation, we used a mean temperature-salinity profile based on 17 CTDs at 125 and 140W and 44 CTDs at 110W. Figure 2 shows the dynamic height time series and individual dynamic height estimates based on eight CTD casts at 110W and six CTD casts at 140W. The rms difference between daily averaged dynamic heights estimated from moorings and from the hydrocast calculations is 1.0 dynamic centimeters (dyn. cm) (110W) and 2.0 dyn. cm (140W). These are comparable to the 0-/500-dbar rms differences found by Emery and Dewar [1982] from historical data. Also, 0-/250-dbar dynamic height variations at 110W (140W) were only 1% (2%) weaker than 0-/1000-dbar variations estimated from the eight (six) available CTD casts.

Figure 2. Zonal wind, meridional wind, sea surface temperature, dynamic height 0/250 dbar, and 20C isotherm depth at (a)0, 110W and (b) 0, 140W. Temperature at 10 m has seen substituted for the more gappy sea surface temperature record at 140W, since the two time series typically agree with one another to within about 0.1C. Daily data have been smoothed with an 11-day Hanning filter. Superimposed on the time series are estimates of the monthly mean seasonal cycle based on mooring data. Dynamic heights of 0/250 dbar from CTD casts are also indicated (triangles).

Mean seasonal cycles have been estimated from monthly averaged mooring data at 100W and 140W and plotted in Figures 2, 3, 4, 5, 8, and 10. Appendix B describes the derivation and comparison of these climatologies with Reynolds' [1988] SST climatology and Wyrtki and Meyers' [1975] wind climatology. Uncertainties in the estimates of climatological monthly means from the moorings are about 1C (SST), 10 m (20C isotherm depth), 4 dyn. cm (dynamic height), 20 cm s (zonal currents), and 1 m s (winds). Thus, when discussing anomalies from the mean seasonal cycles in the following sections, we will focus on those variations persisting longer than 1 month in excess of the values quoted above.

3. Hydrography

Plate 3 shows meridional temperature sections between 5N and 5S along 110W from six approximately semiannual EPOCS cruises. The May-June 1986 section shows conditions during boreal spring near the onset of the ENSO. Minimum SSTs at the surface near 24C are centered at 0.5S, and there is a pronounced weakening of the equatorial thermocline associated with the presence of a geostrophically balanced Equatorial Undercurrent in the upper 100 m [e.g., Picaut et al., 1989]. By October-November 1986, surface temperatures poleward of about 2N and 2S have cooled by about 1C, whereas SST near the equator is still near 24C. The lack of significant seasonal cooling near the equator indicates the presence of anomalously warm water associated with the developing ENSO. Weakening of the thermocline near the equator is still evident in October-November, though the upper thermocline (as measured, for example, by the depth of the 20C isotherm) has been depressed by about 50 m between about 2N and 2S.

Plate 3. Meridional CTD temperature sections along 110W during 1986-1988. Contour interval is 1C. Warmest temperatures are in magenta (>25C); coldest temperatures are in blue (10C). Intermediate temperatures are in orange (21-25C), green (16-20C), and aquamarine (11-15C).

Anomalous warming continues into 1987, with temperatures in May-June 1987 3-4C warmer in the upper 50 m than during May-June 1986. Moreover, no pronounced equatorial SST minimum appears in the May-June 1987 section, and the thermocline is 20-40 m deeper relative to the previous spring. From boreal spring to boreal fall 1987, significant seasonal cooling occurs south of 4N, and an equatorial minimum in SST <24C that is centered near 0.5S develops. As compared to the fall of 1986, a sharp SST front can be found to the north of the equatorial SST minimum, and the thermocline near the equator is slightly shallower.

Both 1988 sections show an equatorial SST minimum that is several degrees Celsius colder than either of the previous 2 years. In particular, the difference between the May-June 1987 and June-July 1988 minima is 7C in the center of the cold tongue near 0.5S. A well-developed SST front is evident north of the equator in both spring and fall 1988. As compared to 1986-1987, the thermocline is much shallower, and there is less evidence for a weakening of the equatorial thermocline, especially in fall 1988.

Variations in thermal structure below the upper 150-200 m can also be seen in these sections. For example, deeper isotherms such as the 10C isotherm undergo 50- to 100-m excursions, most notably near 5N. It is probable that some of this variability is related to the 1986-1987 ENSO and the 1988 cold event, given the importance of low baroclinic mode equatorial waves in the dynamics of interannual variability [e.g., Busalacchi and Cane, 1985].

Plate 4 shows zonal sections of temperature along the equator during 1986-1988. The tendency for equatorial SST warming from 1986 to 1987 and cooling from 1987 to 1988 at 110W is evident at all longitudes between 110W and 140W. Similarly, the interannual rise and fall of the thermocline at 110W ( Plate 3) is evident to the west of 110W. The thermocline in all cases is deeper near the western terminus of the sections than at the eastern terminus, though its slope is not monotonic from east to west. In 1988, for example, the thermocline undulates with an approximately 1000-km wavelength, which, though barely resolved with the 5 CTD zonal resolution, indicates a return of 20- to 30-day instability waves that are typically suppressed during ENSO [Philander et al., 1985]. These waves are characteristically associated with 100 cm s peak-to-peak meridional velocities [Halpern et al., 1988] whose zonal scale sets the pattern of temperature variations both in the thermocline and at the surface via meridional advection [e.g., Pullen et al., 1987].

Plate 4. Equatorial CTD temperature sections during 1986-1988. Contours are as in Plate 3.

Plate 5 shows the salinity sections corresponding to Plate 4. A prominent feature of these sections is the subsurface salinity maximum (>35 practical salinity units (psu)) in 1986-1987 near depths of 100 m. This maximum is typically located in or just above the core of the Equatorial Undercurrent in the central and eastern Pacific [McPhaden, 1985; Mangum et al., 1986] and has been interpreted in terms of eastward advection of high-salinity southern hemisphere subtropical waters by the Undercurrent [Tsuchiya, 1968; Lukas, 1985]. High-salinity water penetrates further to the east in boreal spring 1986 and 1987 consistent with the seasonal cycle of zonal flow in the Undercurrent (Plate 2, lower panel; see also Halpern [1987a]). However, there is a clear freshening of the thermocline from 1987 to 1988 by 0.2-0.4 psu across 110-140W. The appearance of relatively fresh water in the eastern Pacific in 1988 coincides with a relatively shallow thermocline along the equator (Plate 4) and weak isothermal spreading near the equator at 110W. Direct velocity measurements discussed below (Figures 4 and 5) show 1988 to be a period of weaker than normal eastward flow in the Undercurrent at 110W.

Plate 5. Equatorial CTD salinity sections during 1986-1988. Contour interval is 0.1 psu. Saltiest water is in magenta (35.3 psu); freshest water is in blue ( 34.6 psu). Intermediate salinities are in orange (35.1-35.2 psu), green (34.9-35.0 psu), and aquamarine (34.7-34.8 psu).

Plate 5 also shows alternating pockets of high and low salinity near the surface with peak-to-peak amplitudes of about 0.3 psu at about 500-km intervals. They are most apparent in 1988, and like the thermocline and SST undulations on similar zonal scales (Plate 4), they are associated with the presence of energetic instability waves. The oscillating salinity pattern is set by meridional advection of a salinity front which separates relatively fresh water (>35 psu) north of the equator from relatively saline water south of the equator [McPhaden et al., 1990b]. Salinity variations below about 200 m, on the other hand, show little variability (<0.2 psu) over the 3-year sequence in Plate 5.

4. Temporal Variations at 110W and 140W

4.1 Winds

Figure 2 shows time series of winds, SST, dynamic height 0/250 dbar, and 20C isotherm depth at 110W and 140W. The zonal winds at both locations tend to follow the normal seasonal cycle of strong southeasterlies in the latter half of the year and weak southeasterlies in boreal spring. However, easterly winds were weaker by about 1 m s at 140W in 1987 and stronger by a similar amount at 110W from January to July 1987. Meridional winds also tend to be stronger from the south in 1986-1987 during the ENSO and weaker from the south during 1988. The variations in local meridional wind component may be due in part to enhanced (reduced) vertical mixing in the atmospheric boundary layer during periods of warm (cold) equatorial SST anomalies. As Wallace et al. [1989] have shown, northward wind speed just 100 m above sea level in the eastern equatorial Pacific are more than twice as strong as at the surface. Wallace et al. and Hayes et al. [1989] have argued that SST variations near the equator affect the stability of the atmospheric boundary layer and therefore the efficiency with which momentum aloft is mixed down to the surface.

4.2 Sea Surface Temperature

SST at both locations was near normal for the first 4 months of 1986, followed by a weak anomalous warming during May-July. Persistent anomalies of >1C then developed in September 1986 at both 110W and 140W and lasted until January-February 1988. The warmest SST anomalies occurred at 110W in September 1987 (+3.4C) and at 140W in June 1987 (+3.0C). Cold anomalies developed abruptly in early 1988 following the ENSO. From March to May 1988, SST dropped nearly 8C at 110W, and in April 1988, SST dropped 4C at 140W. Subsequently, 2-3C cold anomalies persisted throughout 1988, with the coldest anomalies at 110W in December 1988 (-3.7C) and at 140W in November 1988 (-4.0C).

The development of these anomalies is not obviously related to the local zonal winds. Coincident with the warm SST anomalies in 1986 and 1987, the zonal wind component was either close to or weaker than normal at 140W, whereas it was close to or stronger than normal at 110W. A consistent argument in terms of anomalous Ekman divergence and local upwelling is not possible for both locations. Also, the precipitous drop in SST in the spring of 1988 occurred at a time when the local zonal winds were near normal, again ruling out anomalous local Ekman divergence as the causative mechanism.

4.3 Dynamic Height and 20C Isotherm Depths

Surface dynamic height and the depth of the 20C isotherm (which is an index for the depth of the thermocline) tend to be mirror images of one another at 110W and 140W, similar to what Rebert et al. [1985] found throughout the equatorial Pacific. The mean seasonal cycle in both variables is dominated by the annual and semiannual harmonics with an annual range in dynamic height of 5 dyn. cm (110W) and 13 dyn. cm (140W) and an annual range in 20C isotherm depth of 19 m (110W) and 42 m (140W). In contrast, we note that during 1986-1988, variability is dominated by an interannual rise in the dynamic height from 1986 to 1987 and subsequent drop into 1988, and an intraseasonal variation with periods of 2-3 months. The amplitude range of both these time scales is larger than the annual range. The 20C isotherm depth, for example, shoaled by over 100 m at 110W from January 1987 to June 1988 when it broke the surface; similarly, the 20C depth ranged between 180 m and 0 m at 140W from 1987 to 1988. These changes and similar ones in dynamic height are large enough to obscure the mean seasonal cycle in these variables during 1986-1988.

On interannual time scales there is a tendency for warm SST anomalies to correspond to high surface height and depressed thermocline, and vice versa. This is most evident at 110W, where the mooring climatologies (and therefore anomalies) are better defined. For example, persistent warm anomalies, elevated dynamic heights, and depressed 20C isotherm depths develop in September 1986 and last until late 1987 to early 1988. Similarly, cold SSTs in March-May 1988 set in with a 40-m rise in the depth of the 20C isotherm and a 10 dyn. cm drop in dynamic height. As with SST, the interannual changes in dynamic height and 20C isotherm depths cannot be simply related to local zonal wind variations.

Much of the intraseasonal variability with periods of 2-3 months in Figure 2 is due to the excitation of first baroclinic mode Kelvin waves. McPhaden and Taft [1988] discussed these wave time series spanning 110-140W for the period November 1983 to June 1986, the last 5 months of which overlaps with the present study. They found a nearly in-phase relationship of zonal velocity, thermocline temperature, and dynamic height which propagated eastward at approximately 210 cm s. The eastward phase propagation can be seen visually in Figure 3a for dynamic height 0/250 dbar and in Figure 3b for dynamic height 80/250 dbar (in the thermocline near the mean position of the Equatorial Undercurrent core). Well-defined peaks and troughs at 140W lead those at 110W by about 2 weeks, implying phase speeds in approximate agreement with those reported by McPhaden and Taft [1988]. One can also see the expected Kelvin wave relationship between zonal currents and 20C isotherm depths by comparing Figures 2a and 4. The 2- to 3-month periodicity is most apparent in zonal currents in the thermocline at depths below the Undercurrent core where the seasonal cycle and the 20- to 30-day instability waves are of diminished amplitude. By overlaying the 20C isotherm depth time series in Figure 2a and the zonal currents at, say, 120 m in Figure 4, one would find that on intraseasonal time scales a depression of the thermocline corresponds to enhanced eastward velocity and vice versa. This is also evident in the decadal time series shown in Plate 2. There is no obvious evidence in the local winds for the energetic 2- to 3-month periods observed in the oceanic data sets (Figure 2), implying that the variability must be nonlocally forced as previously inferred by Enfield [1987] and McPhaden and Taft [1988].

Figure 3. Time series of dynamic height (a) at the surface relative to 250 dbar and (b) at 80 dbar relative to 250 dbar. The 0, 125W time series is indicated by a dashed line for clarity. Also shown are the dynamic height differences over 110-140W superimposed on the mean seasonal cycle of dynamic height difference. Wind stress at 0, 165E (solid line) and 2, 165E (dashed line) appear in Figure 3a superimposed on the Wyrtki and Meyers [1975] stress climatology. Daily data have been smoothed with an 11-day Hanning filter.

Figure 4. Time series of zonal velocity in the upper 250 m superimposed on estimates of the monthly mean seasonal cycle. Daily data have been smoothed with an 11-day Hanning filter. Shown at night are the depths of the velocity measurements relative to the 5-year mean profile of zonal velocity (solid line) and temperature (dashed line). Depths on the ordinates are for January 1986 to October 1987 (left) and October 1987 to October 1988 (right).

Figure 3 also shows the difference between dynamic height at 110W and 140W, which is proportional to the zonally averaged zonal pressure gradient across these longitudes. The mean seasonal cycle indicates dynamic height differences at the surface (0/250 dbar) and in the thermocline (80/250 dbar) that are weakest in the spring and strongest in the fall, in approximate balance with the mean seasonal cycle of zonal wind stress [McPhaden and Taft, 1988]. During 1986-1988, dynamic height differences are modulated by intraseasonal wave activity with peak-to-peak amplitudes larger than the annual range. There are in fact a few occasions of 1 month duration when the height difference across 110W-140W nearly vanishes because of these waves. Height differences vary interannually as well across 110W-140W: magnitudes are larger than normal in 1986 and smaller than normal in 1988, during which near-zero values are observed for several consecutive months in the thermocline in the early part of the year.

Dynamic height differences on a smaller zonal scale can be inferred from additional data at 0, 125W during 1987-1988. Differences are generally smaller between 110W and 125W than between 125W and 140W, and there are periods of 1 month when the 0-/250-dbar and 80-/250-dbar surfaces are 1-2 dyn. cm higher at 110W than at 125W. This is consistent with the zonal variations in thermocline slope along the equator in 1988 as seen in Plate 4 and implies a weaker driving force for the Equatorial Undercurrent at 110W as compared to 140W.

4.4 Undercurrent Variations

Figure 4 shows time series of zonal velocity in the upper 200-250 m at 110W. The location of the current meters relative to the mean current and temperature structure at 110W is shown at night. There is a significant amount of 20- to 30-day instability wave energy in these time series which tends to obscure lower-frequency fluctuations. However, two periods of pronounced Equatorial Undercurrent weakening are evident. The first is February-March 1987 in the upper 250 m when flow is on average 25-50 cm s weaker than normal; the second is December 1987 to December 1988 at 80-120 m (in and below the mean depth of the Undercurrent core) when flow is 20-25 cm s weaker than normal. In part these reductions in eastward flow at fixed levels are due to vertical excursions of the thermocline in which the Undercurrent is embedded (compare the upper and lower panels of Plate 2). Depression of the thermocline in early 1987 leads to anomalously westward flow above the Undercurrent core, whereas elevation of the thermocline in 1988 leads to anomalously westward flow below the Undercurrent core. However, these westward flow anomalies are not entirely compensated by eastward flow anomalies below the Undercurrent core in early 1987 or above the Undercurrent core in 1988, as would be expected if vertical displacements were the only influence on the velocity records. Thus there is a large net reduction in transport per unit width during February-March 1987 and during 1988 relative to the normal seasonal cycle at 110W (Figure 5a). The prolonged reduction in 1988 moreover is consistent with diminished isothermal spreading in the thermocline near the equator (Plate 3), and diminished zonal pressure gradient at the surface and in the thermocline (Figure 3). At 140W in contrast, the transport per unit width is near normal during most of 1988 (Figure 5b), possibly due to the presence of stronger zonal pressure gradients in the thermocline between 140W and 125W vis--vis between 125W and 110W, as suggested by Plate 4 and Figure 3b.

Figure 5. Zonal transport per unit width (10-200 m) at (a) 0, 110W and (b) 0, 140W superimposed on estimates of the monthlv mean seasonal cycle. Daily data have been smoothed with an 11-day Hanning filter.

The reduction in Equatorial Undercurrent speed and eastward transport at 110W in February-March 1987 is mirrored in an even more dramatic collapse of eastward transport in the upper 200 m at 140W about 1 month earlier (Figure 5b). This collapse roughly coincides with the dynamic height trough associated with the strongest cycle of intraseasonal wave variability during 1986-1988. Figures 6 and 7 show the vertical profiles of velocity and temperature for February 1987, which is the period bracketed by the minimum transports at 110W and 140W. Also plotted are February profiles for 1983-1988 at 110W and 1984-1988 at 140W. It is clear that February 1987 zonal currents are the weakest of the 5 to 6-year period shown in Figures 6 and 7, with the exception of February 1983 at 110W during the previous ENSO when the Undercurrent disappeared [Halpern, 1987a]. The Undercurrent does not disappear during the 1986-1987 ENSO, but in February 1987 it is significantly reduced in speed especially at 140W, where its core velocity is only 45 cm s . It is also evident that this reduction is not due to a meridional advection of the Undercurrent off the equator. Average meridional velocities in February 1987 are near zero in the thermocline and are not significantly different from the meridional velocities during other years.

Figure 6. Profiles of monthly mean zonal velocity, meridional velocity, and temperature in the upper 250 m for each February from 1983 to 1988 at 0, 110W.

Figure 7. Profiles of monthly mean zonal velocity, meridional velocity, and temperature in the upper 250 m for each February from 1984 to 1988 at 0, 140W.

4.5 Relationship of Zonal Currents to Surface Temperature

Typically at 110W, flow at 10 m is eastward from March to June (the springtime reversal of the South Equatorial Current) and westward the remainder of the year ( Figure 4). However, in 1986 the springtime reversal in March-June is 31 cm s stronger to the east, and the South Equatorial Current in September-November is 32 cm s weaker to the west. (See Figure 10 for low-pass versions of 10-m zonal currents and SST at 110W.) Both these periods correspond to periods of enhanced eastward transport as well ( Figure 5a).

The stronger than normal springtime reversal occurs when the magnitude of the zonal SST gradient is typically weakest along the equator (e.g., -T/x ~ 0.2-0.3 10 C m during boreal spring between 110W and 140W based on Reynolds' [1988] climatology), so that one might expect anomalous advective warming during this time of <1C. Consistent with this, the observed SST anomaly was 0.5C during March-June 1986 at 110W; allowing for a 1-month lag in response to changing zonal currents, the anomaly was 0.7C during April-July 1986. These are of questionable significance given the uncertainties in estimating anomalies as discussed in section 2 and Appendix B. Conversely, during the boreal fall, SST gradients along the equator between 110W and 140W are typically strongest (i.e., -T/x ~ 0.8-1.0 10 C m). At 32 cm s zonal current anomaly acting on a gradient of this magnitude for 3 months would produce a warm SST anomaly of approximately 2C, comparable to the average 1.4C (1.3C) anomaly actually observed in September-November (October-December) 1986. These observations are consistent with the model results of Harrison and Schopf [1984], who argued that advective warming in the eastem Pacific by anomalous currents should be more pronounced in boreal fall than in boreal spring because of the relative strength of the zonal SST gradients during these seasons. Note also that in boreal fall 1986, anomalously weak westward flow was accompanied by a thermocline depression of 20-40 m at 110W (Figure 2a). Hence local upwelling, which would continue unabated because of near-normal local easterlies, would bring warmer water to the surface and contribute to the development of SST anomalies. A similar sequence of events occurred at 140W in 1986 (not shown), with weaker than normal westward flow in the South Equatorial Current and a deeper than normal thermocline associated with the development of SST anomalies of >1C.

Zonal current anomalies at 10 m are less obviously related to SST anomalies at other times during 1986-1988. In particular, rapid cooling that occurs during the termination of the warm event in early 1988 coincides with a near-normal eastward springtime reversal at 110W and 140W (Figures 4 and 10). At 110W the reversal starts late but finishes strong such that the eastward flow from March to June 1988 is within 1 cm s of the mean seasonal average. Thus SST cooling of nearly 8C from March to May at 110W cannot be due to surface zonal advection, which would tend to produce weak warming. It is more likely that uplift and entrainment of thermocline waters into the surface layer as suggested by the rapid coincidental shoaling 20C isotherm in Figure 2 is the mechanism generating anomalously cold SSTs in early 1988. This upwelling may in part be related to the anomalous zonal convergence in eastward mass transport that develops in early 1988 between 110W and 140W (Figure 5). In the mean, the equatorial SST minimum is maintained by eastward transport of cold thermocline water which is continuously upwelled and transported poleward in the surface layer [Bryden and Brady, 1985]. Mass continuity requires that anomalous zonal convergence be balanced by enhanced divergence on the meridional plane, which could intensify this upwelling circulation and lead to additional surface cooling. Consistent with this hypothesis, the coldest SST anomalies in 1988 were found between 110W and 140W as, for example, can be seen in the June 1988 chart shown in Plate 1.

5. Basin Scale Forcing

In the preceding section it was suggested that much of the variability observed in the eastern equatorial Pacific during 1986-1988 could not be explained solely in terms of local wind forcing. We explore this issue in more detail by examining time series of wind stress from six locations along the equator spanning 110W to 165E. Stress has been calculated assuming a drag coefficient of 1.2 10 [Large and Pond, 1981] and air density of 1.2 kg m. The mooring wind time series have been adjusted to 10 m anemometer height assuming a neutral stability, constant stress layer.

Figure 8 shows the zonal wind stress component overplotted on the mooring climatologies (110W and 140W) and the Wyrtki and Meyers [1975] stress climatology (165E, Nauru, Baker, and Christmas islands). The 165E time series (which is also plotted in Figure 3) shows westerly or anomalously weak easterly winds for most of mid-1986 to early 1988. This is followed by anomalously strong easterly wind stresses of up to 0.05 N m during March-December 1988. The winds in 1986-1987 are also frequently punctuated by episodes of strong westerlies lasting 10 days to 2 months, most notably in May 1986, November-December 1986, and June-July 1987. Variations on both these intraseasonal time scales and on interannual time scales exceed those for the mean seasonal cycle, which has a range of only 0.03 N m at this location.

Figure 8. Time series of zonal wind stress at various locations in the equatorial Pacific. Data at 165E are from 0 (solid line) and 2S (dashed line). Superimposed are estimates of the monthly mean seasonal cycle from Wyrtki atid Meyers [1975] for Nauru, the moorings at 165E, Baker, and Christmas. Mooring mean seasonal cycles are overplotted on the 140W and 110W time series. Daily data have been smoothed with an 11-day Hanning filter.

The remote response to the May 1986 wind burst has been discussed by Miller et al. [1988], McPhaden et al. [1988, 1989], and Harrison and Giese [1989]. The May westerly wind burst excited a Kelvin wave which could be traced all the way to the South American coast. Its signature may be seen in the dynamic height time series in Figure 3, in the current time series in Figure 4, and in the transport time series in Figure 5. Eastward propagating pulses in dynamic height appear at other times as well, as, for example, following the November-December 1986 and June-July 1987 westerly wind events. In each case the time lag between the maximum wind stress and maximum dynamic height at 140W implies a zonal phase speed of approximately 2-3 m s, comparable to first baroclinic mode Kelvin wave phase speeds along the equator.

Figure 8 shows that pronounced westerly wind events are confined to west of the date line. However, lower-frequency interannual variations are not. In particular, a weakening of the easterlies can be seen as far west as 140W in 1987, and stronger than normal easterlies prevail between 165E and 157W in 1988. To highlight the large-scale interannual variations, we computed monthly averaged wind anomalies in both the meridional and zonal directions and then smoothed them with a 1-2-1 filter in time. The results are shown in Figure 9. As discussed in Appendix B, the Wyrtki and Meyers [1975] climatology is weaker at 110W and 140W than the mooring climatologies. Hence, if Figure 9 were plotted using the Wyrtki and Meyers stress climatology at all locations, 110W and 140W would show stronger easterly anomalies and weaker westerly anomalies. Also recall that westerly winds are underestimated at Nauru because of topographic shielding.

Figure 9. Monthly mean vector wind stress anomalies corresponding to the time series in Figure 8.

Figure 9 shows that easterlies were stronger than normal by about 0.01-0.02 N m between 167E and 157W in early 1986. Then in mid-1986, westerly anomalies of similar magnitude developed at 165E, Nauru, and Baker islands; at 165E these anomalies intensified to nearly 0.05 N m during the November-December 1986 westerly wind event. During 1987, westerly anomalies of 0.02 N m could be found between 165E and 140W, the largest of which (0.05 N m) occurred between 165E and 176W. Westerly anomalies diminished in late 1987 and were replaced by easterly anomalies which persisted until the end of 1988 between 165E and 157W.

It is interesting that the largest zonal wind stress anomalies at 110W tend to be directed opposite to those in the central Pacific during 1987-1988. Specifically, when the winds are anomalously westerly (easterly) in 1987 (1988) at Nauru, Christmas, and Baker islands, the winds at 110W are anomalously easterly (westerly). The distribution of zonal winds along the equator in 1987 is similar to that documented by Rasmusson and Carpenter [1982] during the mature phase of their composite El Nio. These zonal variations in this wind field may be related to east-west displacements of the "Walker cell," a thermally direct atmospheric circulation on the equatorial place driven by SST-induced heating [Gill, 1982]. A detailed analysis of the relationship between surface winds, SST, and atmospheric heating is beyond the scope of this study, however.

6. Comparison of ENSO Variations During 1982-1983 and 1986-1987

A general discussion of the 1982-1983 ENSO is given by Cane [1983], and a detailed description of moored wind, current, and temperature measurements in the eastern equatorial Pacific during 1982-1983 is given by Halpern [1987a]. The purpose of this section is to point out some of the most significant similarities and differences between the 1982-1983 and 1986-1987 ENSO events in the eastern equatorial Pacific, relying primarily on the long time series at 0, 110W (Plate 2).

Figure 10 shows 3-year segments of data (1982-1984, 1986-1988) overplotted on the mooring climatologies for 110W. These data show that in terms of thermal and surface height anomalies the 1982-1983 ENSO was more intense than the 1986-1987 ENSO. Table 1, for example, indicates that the maximum monthly mean SST anomaly was 5.6C (January 1983) compared to 3.4C (September 1987) and the maximum depression of the thermocline was 96 m (December 1982) compared to 67 m (January 1987). Zonal winds were close to or slightly stronger than normal during the 1986-1987 event, whereas they were weaker than normal during late 1982 (Figure 10). Moreover, according to Halpern [1987a], zonal winds became westerly in early 1983 as far east as 95W. This local relaxation and reversal of the trade winds may have contributed to the intensity of thermal and flow anomalies that developed during the 1982-1983 ENSO at 110W.

Figure 10. Time series of zonal winds, 10-m zonal velocity, SST, dynamic height 0/250 dbar, and 20C isotherm depth for 1986-1988 (solid curves) and 1982-1984 (dashed curves). Daily data have been smoothed with a 51-day Hanning filter to remove variability at periods shorter than about 1 month. SST, 20C isotherm depths, and 10-m zonal velocities are identical to those plotted in Plate 2. Monthly mean mooring climatologies are indicated by dotted lines.

If your browser cannot view the following table correctly, click this link for a GIF image of Table 1.
TABLE 1.  Selected Monthly or Seasonally Averaged Anomaly Extrema During the 1982–1983
ENSO and the 1986–1987 ENSO

1982–1983 1986–1987

SST, C     5.6 (Jan. 1983)        3.4 (Sept. 1987)
20C isotherm depth, m 96 (Dec. 1982) 67 (Jan. 1987)
Dynamic height 0/250 dbar, dyn. m 36 (Dec. 1982) 21 (Jan. 1987)
SEC, cm s          32 (Aug.–Dec. 1982)            32 (Sept.–Nov. 1986)

   All values are positive, implying elevated SST, depressed 20C isotherm depth, elevated dynamic
height, and weaker South Equatorial Current (SEC). Time of the anomalies is indicated in parentheses.

During both events, warm thermal anomalies began to develop in midyear, prior to the South American coastal warming [Kousky and Leetmaa, 1989]. The 1982-1983 ENSO was of shorter duration, though, terminating after about 12 months in mid-1983, whereas the 1986-1987 ENSO lasted for about 18 months from mid-1986 to early 1988. Following both ENSOs, the ocean rebounded to significantly colder than normal conditions, with SSTs plummeting to <20C and dynamic heights dropping to 10 dyn. cm below normal.

Near-surface flow variations show a weakening of the South Equatorial Current during August-December 1982 and September-November 1986. In both cases the average anomaly relative to climatology is 32 cm s though during 1982 it lasts for 5 months whereas during 1986 it lasts for only 3 months. Compared to the 1986-1987 ENSO, prolonged weakening of the South Equatorial Current, combined with a much deeper thermocline, implies more pronounced anomalous zonal and vertical heat advection. This in turn may account for the larger SST anomalies observed during the 1982-1983 ENSO vis--vis the 1986-1987 ENSO at 110W.

It is interesting to note the opposite behavior of the springtime reversal in the South Equatorial Current near the onset of these two events. In March-June 1982, flow is weaker than normal by 19 cm s, whereas in March-June 1986 it is stronger than normal by 31 cm s. Warm SST anomalies appear during both periods, so that there is no consistent argument relating advective warming and zonal flow anomalies during the spring season prior to both events. Similarly, the strongest eastward flow (in both an absolute sense and an anomaly sense) for the 6 years of data shown in Figure 10 occurs in March-May 1983 (average anomaly of 54 cm s). During this period, SST changes little because the SST gradient across the Pacific is nearly zero [Kousky and Leetmaa, 1989].

We noted in section 4.5 that the 1986-1987 ENSO terminated in the eastern equatorial Pacific with surface cooling related to remotely forced uplift of the thermocline. Figure 10 shows that at the end of the 1982-1983 ENSO, failing surface temperatures were also associated with a shoaling thermocline (and decreasing surface heights). The precipitous drop in SST in June-July 1983, however, coincided with the strongest westward flow in the 6 years of data shown in Figure 10. Average anomalies of 49 cm s to the west were observed during this period, suggesting that westward advection of cold upwelled water may have contributed to the termination of the 1982-1983 ENSO. Local winds were weaker than normal throughout 1983 at 110W [Inoue and O'Brien, 1984], indicating that like the 1986-1987 ENSO event, nonlocal forcing was involved in the termination of the 1982-1983 ENSO in the eastern Pacific.

The Equatorial Undercurrent disappeared and was replaced by westward flow in the thermocline at 159W in September 1982 [Firing et al., 1983] and in January-February 1983 at 110W (Plate 2 and Figure 7; see also Halpern, [1987a]). This followed several months of nearly flat or reversed sea surface slope along the equator in the central and eastern Pacific [Wyrtki, 1984]. There were also periods of nearly flat or reversed pressure gradient at the surface and in the thermocline in the eastern Pacific during 1986-1988. During some of these periods (e.g., 1988 at 110W) the Undercurrent was significantly weaker than normal, though it did not disappear at either 110W or 140W during 1986-1988. The difference may be due to the duration, intensity, and/or zonal scale of the pressure gradient anomalies, which are likely to have been smaller in 1986-1988 than during 1982-1983. As noted above, westerly winds, which favor the establishment of a westward zonal pressure gradient force along the equator, penetrated to at least 95W in 1983 [Halpern, 1987a], whereas they were confined to the west of Christmas Island (157W) during the 1986-1988 (Figure 8). The Undercurrent disappeared in the western Pacific during October-November 1987 in response to these westerlies when the zonal pressure gradient at the surface and in the thermocline reversed [McPhaden et al., 1990a].

7. Summary and Discussion

The eastern equatorial Pacific was characterized by warmer than usual SSTs from mid-1986 until early 1988. Weak initial anomalies in mid-1986 increased to more than 1C in September-November 1986 in association with a 30 cm s reduction in speed of the South Equatorial Current and a 20- to 40-m depression of the thermocline. These warm anomalies persisted until early 1988, after which cold anomalies of >3C developed in response to large-scale, remotely forced upwelling. Interannual fluctuations observed in the eastern equatorial Pacific appear for the most part to be determined by year-to-year variations in the trade winds over 40-50 of longitude in the central and western Pacific.

The importance of remotely forced variability in the eastern Pacific during ENSO events is well known [e.g., Cane, 1983]. However, it is interesting to note that both interannual and intraseasonal time scales contribute to the observed variability, and that for several variables, extrema during the 1986-1987 ENSO coincided with the occurrence of remotely forced intraseasonal Kelvin waves. Specifically, the highest dynamic heights, deepest thermocline depths, and largest zonal current transport anomalies were observed in early 1987 as a result of the strong 2-month westerly wind event west of the date line at the end of 1986.

The dynamical relevance of these intraseasonal fluctuations to the evolution of ENSO is not well understood. Statistically, for example, the frequent occurrence of western wind events may result in a seasonally averaged westerly wind anomaly [Luther et al., 1983]. However, the variations themselves do not appear to be essential to the generation of ENSO events based on simple coupled ocean-atmosphere model analyses [e.g., Zebiak and Cane, 1987; Schopf and Suarez, 1988; Battisti, 1988; Zebiak, 1989]. Also, intraseasonal time scale variations are not as prominent in SST as in subsurface temperature, dynamic heights, and thermocline currents (for which, see Figures 2 and 4). Nonetheless, in general, one might expect a degree of nonlinear coupling between high- and low-frequency phenomena because intraseasonal waves like those excited by the November-December 1986 wind event can have very large amplitude. It is also possible that the timing of these waves relative to the seasonal cycle could be important in the development of ENSO, if, for example, the coupled system is particularly sensitive to perturbation at certain times of the year.

In section 6 we showed that the 1986-1987 ENSO was less intense and evolved differently than the 1982-1983 ENSO. Particularly noteworthy is the fact that the Equatorial Undercurrent did not disappear in the eastern Pacific during 1986-1987 as it did during 1982-1983, which probably reflects the smaller amplitude and scale of the zonal wind and pressure gradient anomalies along the equator during 1986-1987. In contrast, a feature common to both events was a weakening of the South Equatorial Current by about 30 cm s in the latter part of 1982 and 1986. Sea level differences across the South Equatorial Current in the central and western Pacific have previously shown this weakening, albeit inferentially by geostrophy, for ENSO events from the 1950s through 1982-1983 [Wyrtki, 1974; Philander, 1990]. Direct measurements on the equator suggest that the duration and/or intensity of this weakening may be related to the magnitude of SST anomalies via anomalous zonal advection during the latter part of the year when zonal SST gradients are strongest.

The 1986-1987 ENSO, like the 1982-1983 ENSO, evolved differently than the Rasmusson and Carpenter [1982] composite. Warm SSTs of >1C developed in the eastern Pacific during September-November 1986 prior to South American coastal warming and persisted for 18 months (through December 1987 up to February 1988) instead of for only 12 months as in the composite. Moreover, Rasmusson and Carpenter found that westerly wind anomalies develop as early as January of the ENSO year near the date line, intensify in the summer, and last until the following February. In contrast, the observations for 1986-1988 show weak westerly anomalies developing near the date line in mid-1986 and persisting until the end of 1987. Lander [1989] discusses the differences between the 1986-1987 ENSO and the Rasmusson and Carpenter [1982] composite in more depth. It is clear, however, that although there are similarities between ENSO events, there is no single scenario that describes them all. One must be cautious, therefore, when generalizing about ENSO on the basis of observations from a single event.

Appendix A: Data Gaps and Gridding

The depths at which moored velocity and temperature measurements have been made at 0, 110W and 0, 140W have changed periodically due to shifting scientific priorities in EPOCS. Also, although VACM/VMCM temperature and velocity data return has been high (>90%), occasional instrument failures and the loss of a mooring at 0, 110W between April and October 1982 have led to data gaps. We have filled these data gaps where appropriate and gridded the time series to a set of standard levels (Plate 2) to facilitate the computation of mean seasonal cycles (Appendix B), and the comparison of the 1982-1983 and 1986-1987 ENSO events (section 6). For example, at 0, 110W we substituted SST as well as temperature and velocity data at 15 m, 50 m, and 100 m from from a backup mooring at 0, 108W during April-October 1982; and we substituted velocity at 80 m from the backup mooring during July-November 1986. We also used linear interpolation between vertically adjacent instruments and, where this was not possible or advisable, linear extrapolation and/or interpolation by least squares regression. Some specific examples of these procedures and how they affect subsequent analyses are presented in this appendix. In each case the examples are in terms of daily or weekly averaged data; results (quoted in terms of cross correlations and rms differences) improve with further temporal averaging. Additional examples are given by McPhaden et al. [1988]; similarly, Halpern [1987a] discusses the representativeness of 108W data as a substitute for 110W data for variations with periods longer than 1 week.

At 45 m at 110W we used a regression fill for zonal velocity during October 1987 to December 1988, because mean zonal velocity curvature precludes linear interpolation (Figure 4). The regression equation was based on 3 years (November 1983 to November 1986) of simultaneous measurements at 25 m and 45 m. The correlation between weekly averaged data and simulated data derived from this regression fill was 0.95 with a mean (rms) difference of 0 cm s (15 cm s). We likewise used regression fills for 10-m data based on 25-m or 45-m data at 140W for March-June 1986, March-May 1987, and May-November 1988. The correlation of simulated daily 10-m data based on the regression formula with actual daily data at 10 m for a 4 year period (April 1983 to October 1987) was 0.94 with a mean (rms) difference of 0 cm s (14 cm s). Zonal velocity data prior to November 1983 at 110W were linearly extrapolated based on the underlying vertical shear from 15 m or 20 m depth to 10 m depth for consistency with the more recent 10-m time series.

The SST time series at 140W was very gappy, so we have substituted 10-m temperatures in Figures 2b and B1. The correlation between 576 contemporaneous daily averaged SST and 10-m temperatures spanning April 1984 to June 1986 was 0.99 with a mean difference of 0.02C (10 m colder) and an rms difference of 0.14C. Thus, for our purposes, 10-m temperatures are equivalent to SST at 140W. At 110W the correspondence between SST and 10- to 20-m temperatures is less perfect because of greater near-surface stratification. Thus SST data gaps were filled with regression formulae at 110W using the closest near-surface temperature record. For instance, several months of missing SST data between July 1981 and April 1983 were filled with a regression formula based on 15-m temperatures. The correlation between 233 days of regression-derived SST data and actual SST data during this period was 0.99 with a mean (rms) difference of 0.00C (0.30C).

Figure B1. Monthly mean climatologies of sea surface temperature and zonal winds at 0, 110W and 0, 140W (solid lines). Superimposed are monthly means from the Reynolds [1988] sea surface temperature climatology and Wyrtki and Meyers [1975] wind climatology (dotted lines).

Temperature time series data were sometimes missing below 200 m (e.g., June-November 1988 at 140W and February-October 1988 at 110W). For dynamic height calculations relative to 250 dbar at these times, we appended the mean temperature gradient below 200 m to the 200-m temperature record to create an artificial time series at 250 m. This method, discussed by Kessler et al. [1985] and McPhaden et al. [1990a], assumes that temperature changes at depth are due to vertical displacements of the mean thermal structure. Sensitivity experiments in which we substituted a mean gradient estimate for existing moored temperature data at 250 m indicate that the method leads to daily averaged 0-/250-dbar dynamic height errors of O(0.1 dyn. cm).

Appendix B: Mean Seasonal Cycles

Mean seasonal cycles have been estimated from monthly averaged equatorial mooring time series measurements at 110W and 140W. Monthly averages were determined by smoothing the time series with a 31-day running mean filter and subsampling on the fifteenth of each month. Three years of data were used at 140W (January 1984 to December 1986), and 5 years of data at 110W (April 1980 to March 1982; January 1984 to December 1986). The first several months of 1986-1987 ENSO variability are included in these estimates, but the large interannual variations in 1987 and 1988 are not. Similarly, variability from the 1982-1983 ENSO episode is not included in the 110W mean seasonal calculation.

One can compute a standard deviation for each month in the mean seasonal cycle at 110W where 5 years of data were used. Characteristically, these deviations are about 1C (SST), 10 m (20C isotherm depth), 4 dyn. cm (dynamic height), 20 cm s (zonal currents), and 1 m s (winds). These numbers typify non-ENSO interannual variations on a monthly basis. Thus, when discussing anomalies from the mean seasonal cycle, emphasis should be given to those variations persisting longer than 1 month in excess of the values quoted above.

Note that we expect differences between these climatologies and others that may exist in the eastern equatorial Pacific. Figure B1, for example, shows a comparison of estimates based on moored measurements with the Reynolds [1988] SST climatology and the Wyrtki and Meyers [1975] wind climatology. The moored data show SSTs that are consistently colder by l-2C and zonal winds that are consistently stronger by 1-3 m s. There are several possible explanations for these differences. First, they could represent real interdecadal variations since the Reynolds and the Wyrtki and Meyers climatologies are based on pre-1980 data. Second, Reynolds and Wyrtki and Meyers include ENSO years in their climatologies, whereas we have excluded data from the 1982-1983 ENSO and from most of the 1986-1987 ENSO. This could lead to slightly warmer temperatures and, at 140W, slightly weaker easterlies compared to the mooring climatologies. (At 110W easterlies tend to be stronger than usual during ENSO, with the notable exception of the 1982-1983 event, so that ENSO biases cannot explain the wind differences at 110W.) Third, the previously published climatologies are spatially smoothed over several degrees of latitude and longitude. In the case of the Reynolds climatology this would lead to warmer equatorial SSTs because of the strong meridional SST gradients flanking the equatorial minimum. Fourth, the Reynolds and the Wyrtki and Meyers climatologies are based on different measurement techniques than the mooring climatologies; i.e., merchant ship wind observations were often reported on the Beaufort scale, and SST observations were based on bucket or engine room intake temperatures. These climatologies may be biased because of uneven spatial and temporal sampling and, in the case of SST, because of engine room heating from the intake method [Saur, 1963]. We have therefore favored the mooring-based mean seasonal cycles in this study in spite of the relative shortness of the records, because they allow for a consistent discussion of interannual anomalies in all the moored measurements (including subsurface velocity, for which no other climatology exists).

Acknowledgments. Special thanks to Paul Freitag, Margie McCarty, and Nancy Soreide for processing the mooring time series. We also appreciate the technical assistance of Andy Shepherd, Carol Coho, Doug Fenton, Ben Moore, and Rick Miller for instrument preparation and mooring operations at sea. Linda Mangum and Linda Stratton were responsible for CTD data processing. Thanks also to the anonymous reviewer who was very detailed and thorough in his critique. This work has been supported by NOAA's Equatorial Pacific Ocean Climate Study (EPOCS) program and the U.S. TOGA Project Office. NOAA PMEL contribution 1120.


Battisti, D., Dynamics and thermodynamics of a warming event in a coupled ocean/atmosphere model, J. Atmos. Sci., 45, 2889-2919, 1988.

Bendat, J. S., and A. G. Piersol, Random Data: Analysis and Measurement Procedures, 407 pp., Wiley Interscience, New York, 1971.

Bryden, H. L., and E. L. Brady, Diagnositic model of the three dimensional circulation in the upper equatorial Pacific Ocean, J. Phys. Oceanogr., 15, 1255-1273, 1985.

Busalacchi, A. J., and M. A. Cane, Hindcasts of sea level variations during the 1982-83 El Nio, J. Phys. Oceanogr., 15, 213-221, 1985.

Cane, M. A., Oceanographic events during El Nio, Science, 222, 1189-1202, 1983.

Climate Analysis Center, Global analyses and indices, climate diagnostic bulletin, Natl. Oceanic and Atmos. Admin., Washington, D. C., 1988.

Emery, W. J., and J. S. Dewar, Mean temperature-salinity, salinity-depth and temperature-depth curves for the North Atlantic and North Pacific, Prog. Oceanogr., 11, 219-305, 1982.

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

Firing, E., R. Lukas, J. Sadler, and K. Wyrtki, Equatorial Undercurrent disappears during the 1982-83 El Nio, Science, 222, 1121-1123, 1983.

Freitag, H. P., M. J. McPhaden, and A. J. Shepherd, Equatorial current and temperature data: 108W to 110W; October 1979 to November 1983, NOAA Data Rep. ERL PMEL-17, Pac. Mar. Environ. Lab., Seattle, Wash., 1987.

Freitag, H. P., M. J. McPhaden, and A. J. Shepherd, Comparison of equatorial winds as measured by cup vs. propeller anemometers, J. Atmos. Oceanic Technol., 6, 327-332, 1989.

Gill, A. E., Atmosphere-Ocean Dynamics, 662 pp., Academic, San Diego, Calif., 1982.

Halpern, D., Observations of annual and El Nio flow variations at 0, 110W and 0, 95W during 1980-1985, J. Geophys. Res., 92, 8197-8212, 1987a.

Halpern, D., Comparison of upper ocean VACM and VMCM observations in the equatorial Pacific, J. Atmos. Oceanic Technol., 4, 84-93, 1987b.

Halpern, D., R. A. Knox, and D. S. Luther, Observations of 20-day period meridional current oscillations in the upper ocean along the Pacific equator, J. Phys. Oceanogr., 18, 1514-1534, 1988.

Harrison, D. E., and B. S. Giese, Comment on "The response of the equatorial Pacific Ocean to a westerly wind burst in May 1986" by M. J. McPhaden et al., J. Geophys. Res., 94, 5024-5026, 1989.

Harrison, D. E., and P. Schopf, Kelvin wave induced anomalous advection and the onset of surface warming in El Nio events, Mon. Weather Rev., 112, 923-933, 1984.

Hayes, S. P., et al., The Equatorial Pacific Ocean Climate Studies (EPOCS) plans: 1986-1988, Eos Trans. AGU, 67, 442-444, 1986.

Hayes, S. P., M. J. McPhaden, and J. M. Wallace, The influence of sea surface temperature upon surface wind in the eastern equatorial Pacific: Weekly to monthly variability, J. Clim., 2, 1500-1506, 1989.

Inoue, M., and J. J. O'Brien, Predictability of the decay of the 1982/83 El Nio, Mon. Weather Rev., 114, 967-972, 1984.

Kessler, W. S., B. A. Taft, and M. J. McPhaden, An assessment of the XBT sampling network in the central Pacific, Tech. Rep. USTOGA-4, 62 pp., Univ. Corp. for Atmos. Res., Boulder, Colo., 1985.

Kousky, V. E., and A. Leetmaa, The 1986-87 Pacific warm episode: Evolution of oceanic and atmospheric anomaly fields, J. Clim., in press, 1989.

Lander, M. A., A comparative analysis of the 1987 ENSO event, Trop. Ocean Atmos. Newsl., 49, 3-6, 1989.

Large, W. G., and S. Pond, Open ocean momentum flux measurements in moderate to strong winds, J. Phys. Oceanogr., 11, 324-336, 1981.

Lukas, R., The termination of the Equatorial Undercurrent in the eastern Pacific, Prog. Oceanogr., 16, 63-90, 1985.

Luther, D. S., D. E. Harrison, and R. A. Knox, Zonal winds in the central equatorial Pacific and El Nio, Science, 222, 237-230, 1983.

Lynch, J. M., L. J. Mangum, and S. P. Hayes, CTD/0 measurements during 1986 as part of the Equatorial Pacific Ocean Climate Studies (EPOCS), NOAA Data Rep. ERL PMEL-24, 261 pp., Pac. Mar. Environ. Lab., Seattle, Wash., 1988.

Mangum, L. J., S. P. Hayes, and J. M. Toole, Eastern Pacific circulation near the onset of the 1982-1983 El Nio, J. Geophys. Res., 91, 8428-8436, 1986.

McPhaden, M. J., Fine-structure variability observed in CTD measurements from the central equatorial Pacific, J. Geophys. Res., 90, 11,726-11,740, 1985.

McPhaden, M. J., and B. A. Taft, On the dynamics of seasonal and intraseasonal variability in the eastern equatorial Pacific, J. Phys. Oceanogr., 18, 1713-1732, 1988.

McPhaden, M. J., H. P. Freitag, S. P. Hayes, B. A. Taft, Z. Chen, and K. Wyrtki, The response of the equatorial Pacific Ocean to a westerly wind burst in May 1986, J. Geophys. Res., 93, 10,598-10,603, 1988.

McPhaden, M. J., H. P. Freitag, S. P. Hayes, B. A. Taft, Z. Chen, and K. Wyrtki, Reply, J. Geophys. Res., 94, 5027-5028, 1989.

McPhaden, M. J., S. P. Hayes, L. J. Mangum, and J. M. Toole, Variability in the western equatorial Pacific Ocean during the 1986-87 El Nio/Southern Oscillation event, J. Phys. Oceanogr., 20, 190-208, 1990a.

McPhaden, M. J., H. P. Freitag, and A. J. Shepherd, Moored salinity time series measurements from 0, 140W, J. Atmos. Oceanic Technol., in press, 1990b.

Milburn, H. B., and P. D. McClain, ATLAS--A low cost satellite data telemetry mooring developed for NOAA's climate research mission, in Proceedings MDS 1986; Marine Data Systems International Symposium, pp. 393-396, Marine Technology Society, Washington, D. C., 1986.

Miller, L., R. Cheney, and B. Douglas, GEOSAT altimeter observations of Kelvin waves and the 1986-87 El Nio, Science, 239, 52-54, 1988.

National Academy of Sciences, Report of a drafting workshop, AOML, Miami, on elements of the ten-year monitoring component for El Nio and the Southern Oscillation (ENSO), Tech. Rep. USTOGA-1, 42 pp., Univ. Corp. for Atmos. Res., Boulder, Colo., 1988.

Palmer, T. N., and C. Brankovic, The 1988 US drought linked to anomalous sea surface temperature, Nature, 338, 54-57, 1989.

Philander, S. G. H., El Nio, La Nia, and the Southern Oscillation, 289 pp., Academic, San Diego, Calif., 1990.

Philander, S. G. H., D. Halpern, D. Hansen, R. Legeckis, L. Miller, C. Paul, R. Watts, R. Weisberg, and M. Wimbush, Long waves in the equatorial Pacific Ocean, Eos Trans. AGU, 66, 154, 1985.

Picaut, J., S. P. Hayes, and M. J. McPhaden, Use of the geostrophic approximation to estimate time-varying zonal currents at the equator, J. Geophys. Res., 94, 3228-3236, 1989.

Pullen, P. E., R. L. Bernstein, and D. Halpern, Equatorial long-wave characteristics determined from satellite sea surface temperature and in situ data, J. Geophys. Res., 92, 742-748, 1987.

Rasmusson, E. M., and T. H. Carpenter, Variations in tropical Pacific sea surface temperature and surface wind fields associated with the Southern Oscillation/El Nio, Mon. Weather Rev., 110, 354-384, 1982.

Rebert, J. P., J. R. Donguy, G. Eldin, and K. Wyrtki, Relations between sea level, thermocline depth, heat content, and dynamic height in the tropical Pacific Ocean, J. Geophys. Res., 90, 11,719-11,725, 1985.

Reynolds, R. W., A real-time global sea surface temperature analysis, J. Clim., 1, 75-86, 1988.

Ropelewski, C. F., and M. S. Halpert, North American precipitation and temperature patterns associated with the El Nio-Southern Oscillation (ENSO), Mon. Weather Rev., 114, 2352-2362, 1986.

Ropelewski, C. F., and M. S. Halpert, Global and regional scale precipitation patterns associated with the El Nio-Southern Oscillation, Mon. Weather Rev., 115, 1606-1626, 1987.

Saur, J. F. T., A study of the quality of sea surface temperatures reported in logs of ships' weather observations, J. Appl. Meteorol., 2, 417-425, 1963.

Schopf, P. S., and M. J. Suarez, Vacillations in a coupled ocean-atmosphere model, J. Atmos. Sci., 45, 549-566, 1988.

Trenberth, K. E., G. W. Branstator, and P. A. Arkin, Origins of the 1988 North American drought, Science, 242, 1640-1645, 1988.

Tsuchiya, M., Upper waters of the intertropical Pacific Ocean, Johns Hopkins Oceanogr. Stud., 4, 50 pp., 1968.

Wallace, J. M., T. P. Mitchell, and C. Deser, The influence of sea-surface temperature upon surface wind in the eastern equatorial Pacific: Seasonal and interannual variability, J. Clim., 2, 1492-1499, 1989.

Wyrtki, K., Sea level and the seasonal fluctuation of equatorial currents in the western Pacific Ocean, J. Phys. Oceanogr., 4, 91-103, 1974.

Wyrtki, K., The slope of sea level along the equator during the 1982-1983 El Nio, J. Geophys. Res., 89, 10,419-10,424, 1984.

Wyrtki, K., and G. Meyers, The trade wind field over the Pacific Ocean, I, The mean field and the annual mean variation, Tech. Rep. HIG-75-1, 26 pp., Hawaii Inst. of Geophys., Univ. of Hawaii, Honolulu, 1975.

Zebiak, S. E., On the 30-60 oscillation and the prediction of El Nio, J. Clim., 2, 1381-1387, 1989.

Zebiak, S. E., and M. A. Cane, A model El Nio-Southern Oscillation, Mon. Weather Rev., 115, 2262-2278, 1987.

PMEL Outstanding Papers

PMEL Publications Search

PMEL Homepage