Ga. Maul et al., OBSERVED SEA-SURFACE HEIGHT AND MODELED DYNAMIC HEIGHT ANOMALY DEPARTURES IN THE TROPICAL PACIFIC-OCEAN - 1986-1989, Oceanologica acta, 20(4), 1997, pp. 569-584
Observations of sea surface height departure (SSH') from November 1986
through September 1989 in the tropical Pacific from the improved (T2)
GEOSAT data set (Cheney et al., 1991) are compared with monthly mean
dynamic height anomaly departure (Delta D') from the ''second reanalys
is'' using the NOAA ocean analysis system (Ji et al., 1994, 1995). For
comparisons, Delta D' is calculated by removing north-south tilt and
bias as in SSH orbit error removal, giving standard deviation fields s
igma(SSH') and sigma(Delta D') that quantitatively reproduce variabili
ty of the North Equatorial Current (NEC)/North Equatorial Countercurre
nt (NECC)/South Equatorial Current (SEC) system between El Nino and no
n-El Nino years. Hovmoller diagrams of SSH' and Delta D' variability a
t 110 degrees W, 140 degrees W, 170 degrees W, and 165 degrees E betwe
en 20 degrees S and 20 degrees N, and along the equator from 120 degre
es E to 80 degrees W, display the amplitude and phase of the 1986-1987
El Nino-Southern Oscillation (ENSO) event as distinguished from the y
ears following, with the NECC significantly weakened in 1987 as compar
ed to 1988. Cross-correlations (r) between Delta D' and SSH' are highe
st near the Equator in the vicinity of the TOGA-TAO (Hayes et al., 199
1; McPhaden, 1993) in situ mooring arrays, with values above r = 0.7 i
n much of the region +/- 7 degrees of the Equator across about half of
the Pacific basin. Differences between Delta D' and SSH' are typicall
y less than +/- 5 cm RMS in this same equatorial band, but there are t
wo regions of differences in excess of +/- 15 cm RMS: off Central Amer
ica and east of New Guinea. The reason for these large RMS differences
is uncertain, but it is inferred from Lagrangian buoy data that inten
se eddy activity is unresolved in the model as compared to GEOSAT. For
the 35 monthly realizations, the ensemble cross correlation r = 0.5 h
as +/- 7 cm RMS within +/- 15 degrees of the Equator, and peaks in 198
8 with decay towards the end of the GEOSAT Exact Repeat Mission (ERM)
in September 1989. Seasonally, both SSH' and Delta D' show that the bo
real winter (DJF) of 1987 is dominated by the ENSO-related sea level m
aximum in mid-basin, which is absent in the following two winters; spr
ing (MAM) and summer (JJA) of all three years have little resemblance
to each other, with 1987 and 1989 being more similar than 1988; autumn
(SON) of 1988 shows a well developed NECC across much of the eastern
Pacific centered at similar to 7 degrees N. The ERM altimeter data and
the model calculations of the annual cycle both show east-west height
amplitude maxima along 5 degrees N and 10 degrees N that explain more
than 50% of the variance for a substantial portion of the central Pac
ific Ocean. Concurrent drogued buoy tracks show more eddies centered a
long similar to 5 degrees N in the eastern Pacific in 1988 as compared
to 1987 or 1989, but Delta D' is unable to resolve these features.