The 172 synoptic and nominally meridional CTD/ADCP sections used in this study were observed from June 1985 through December 2000 between 138°E and 86°W (Fig. 1). Of these, 103 sections were taken during TAO mooring array maintenance cruises from June 1991 through December 2000 between 165°E and 95°W. The cruises were on the NOAA Ship Discoverer in the first half of the decade (e.g. Johnson & Plimpton, 1999) and the NOAA Ships Ka’imimoana and Ronald H. Brown for the second half of the decade (e.g. McTaggart & Johnson, 1999). Another 29 sections were taken on the R/V Kaiyo as part of the Tropical Ocean Climate Study (TOCS; Kashino et al., 2001). The bulk of the remaining sections came from the World Ocean Circulation Experiment (WOCE; Firing, Wijffels, & Hacker, 1998) one-time hydrographic survey, the United States/People’s Republic of China cooperative program for Air–Sea Interaction (US/PRC; Gouriou & Toole, 1993), the US Joint Global Ocean Flux Study (JGOFS) Equatorial Pacific Process Study (Murray, Johnson, & Garside, 1995), the Tropical Ocean Global Atmosphere Coupled Ocean–Atmosphere Response Experiment (TOGA–COARE; Eldin et al., 1994), and the Western Equatorial Pacific Ocean Circulation Study (WEPOCS; Tsuchiya, Lukas, Fine, Firing, & Lindstrom, 1989).
Fig. 1. CTD/ADCP section data distribution. Top panel shows latitudes and longitudes of the 172 meridional sections used in this study. Middle panel shows times when these sections crossed the equator. Sections occupied between 1991 and 2001 are shown by (+). Sections occupied between 1985 and 1991 are shown by (O), with 10 years added for compactness. Bottom left panel shows (contoured at intervals of 10 with increasingly dark contours indicating more data) number of sections where ADCP and CTD data are available in the meridional-vertical plane. Bottom right panel follows bottom left but for number of sections with CTD data alone.
The sections were occupied along ten nominal longitudes (Fig. 1; 143°E, 156°E, 165°E, 180°E, 170°W, 155°W, 140°W, 125°W, 110°W, and 95°W). At each longitude occupations tended to be separated by roughly 6 months or more. Except for 18 sections occupied from 1985 though 1990 along 165°E as part of US/PRC and WEPOCS, the sections were fairly evenly distributed by year through the 1990s, with as few as six in 1991, and as many as 23 in 1997. In longitude the sections were somewhat more concentrated in the western and central Pacific than in the eastern Pacific, with as many as 34 at three closely-spaced longitudes around 143°E and 21 at 140°W, and as few as 12 at both 110°W and 95°W. The distribution through the seasonal cycle was biased toward boreal fall. The most uneven 3-month groupings had 61 sections in September–November, 37 sections in December–February, 35 in March–May, and 39 in June–August. The section times also tended slightly to favor El Niño conditions, as did the 1990s. The mean and standard deviation of the SOI at the section times was –0.3 ± 1.0, the same statistics as for the period of mid-1991 through 2000.
First, the ADCP and CTD data for each individual, quasi-synoptic section have been gridded between 8°S and 13°N following Johnson, McPhaden, Rowe, and McTaggart (2000). The ADCP generally measures velocity at depths between about 20 and 300–450 m in 8-m bins, although some earlier measurements extend down only to about 200 m. These velocities are objectively mapped assuming a Gaussian covariance, a meridional correlation length scale of 1°, a vertical correlation length scale of 25 m, and a noise-to-signal energy ratio of 0.01. The objective mapping uses shear in the upper water column to extrapolate velocities to the surface. The CTD temperature, salinity, and pressure data are usually sampled from the surface to at least 1000 dbar at 1-dbar spacing in the vertical and 1° latitude in the horizontal. The station spacing is often 0.5° within ±3° of the equator, but occasionally exceeds 1°, mostly in the earlier part of the decade. These fields are gridded somewhat differently from the ADCP data. They are splined in latitude on isopycnals to a regular, closely spaced grid. Both the CTD and the ADCP data are concentrated about the equator (Fig. 1).
Once the individual, quasi-synoptic sections have been gridded, their potential temperature, salinity, and zonal velocity fields are then analyzed around each of the ten nominal longitudes listed above. For this analysis an isopycnal averaging procedure is applied to potential temperature, salinity, depth, and zonal velocity. This procedure better preserves a sharp pycnocline, water properties, and velocity extrema by reducing the smearing effects of isopycnal heave (Lozier, McCartney, & Owens, 1994). By following isopycnals, the averaging is quasi-Lagrangian in the vertical. In contrast to isobaric averaging, this process results in a very sharp thermocline, typical of the strength of a synoptic section. The mixed layer temperature, salinity, and depth as well as zonal velocity are analyzed separately, allowing for vertical shear in the mixed layer zonal velocity. The mixed layer analysis is connected to the permanent pycnocline with a shape-preserving spline (Akima, 1970).
For the averaging, potential temperature, salinity, zonal velocity and depth
are independently interpolated onto a closely-spaced latitudinal and isopycnal
grid (just latitudinal for the mixed-layer). Data at each latitude and isopycnal
from all sections within about 20° of the target longitude are least-squares
fitted to a polynomial function of longitude, an annual harmonic, and the SOI.
Fitting sections from two or three nominal longitudes means data from more
months of the year and in more phases of the SOI are included. At most locations
there are data from three nominal longitudes, and the polynomial is second
order. However, at the westernmost and easternmost lines at 143°E and
95°W, and anywhere else where only data from two nominal longitudes are
available, the polynomial reverts to first order. For the second order cases
the fit for the data D looks like D = a + bx + cx2 + d sin(2
t/T)
+ f cos(2t/T)
+ gS, where a–g are the fit coefficients, x is
the longitude, t is the time (in years) when the data were taken, T is
the period of a year, and S is the value of the SOI interpolated to
the time t. Averages are displayed only where more than 16 sections
are locally available for fitting.
Return to previous section or go to next section