El Ni ˜
no and Ocean Observations
89
with Joel Picaut in Noumea, New Caledonia, for deployments on French cruises
along 165
◦ E; and with Akimasa Sumi (University of Tokyo) and Kensuke Takeuchi
(Hokkaido University) for deployments on Japanese cruises along 147
◦ E. By the
TOGA midlife review in Honolulu in July 1990, there were 17 ATLAS moorings
deployed across the Pacific as part of the TOGA Observing System (Figure 6.2;
McPhaden and Hayes, 1990; Hayes et al., 1991). Stan named this array the TOGA
Thermal Array in the Ocean (TAO) (Nova University, 1989).
The first half of TOGA witnessed significant progress on a number of other
fronts as well. Between 1985 and 1990, the number of XBT lines regularly sampled
in the tropical Pacific nearly doubled from 10 to 18, the number of TOGA Pacific tide
gauges nearly doubled from 42 to 80, and surface drifters drogued at 15 m depth to
measure mixed layer velocity and SST increased from only 8 to 164 (McPhaden et al.,
1998). Declassification of the U.S. Navy’s GEOSAT satellite altimetry data made it
possible to observe Kelvin wave variations associated with El Ni˜ no from space for
the first time in 1986–88 (Miller et al., 1988).
Early in TOGA, Cane et al. (1986) made the first successful prediction of an
El Ni˜ no (the 1986–87 event) using a simple dynamical coupled ocean–atmosphere
model. Two other research groups likewise issued successful El Ni˜ no forecasts during
this time using statistical and statistical–dynamical hybrid models (Barnett et al.,
1988). Following the 1986–87 El Ni˜ no, unusually cold tropical Pacific SSTs in 1988–
89 and their link to drought in the U.S. Midwest (Trenberth and Branstator, 1992)
focused attention on the cold phase of ENSO, dubbed La Ni˜ na (Philander, 1990). New
theories of the ENSO cycle between warm and cold events also emerged in which
wind forced changes in thermocline depth, mediated by equatorial Kelvin and Rossby
waves, were identified as the mechanism governing delayed negative feedbacks on
the growth of tropical Pacific SST anomalies (Schopf and Suarez, 1988; Battisti
and Hirst, 1989). The accumulation of excess upper ocean heat content at equatorial
latitudes associated with these wave processes was also identified as a precondition
for the occurrence of El Ni˜ no and as the source of predictability for ENSO time scale
variations (Wyrtki, 1985; Cane et al., 1986; Zebiak, 1989).
Building on studies prior to TOGA using simple “two-layer” ocean models
(Busalacchi et al., 1983), realistic wind-forced simulations of the El Ni˜ no were
made for the first time from ocean general circulation models (e.g., Philander and
Seigel, 1985). These modeling advances emphasized the need for accurate basin scale
wind measurements to simulate ENSO-related changes in ocean circulation and SST
(Harrison et al., 1989). Also, NOAA’s National Meteorological Center established
the rudiments of an operational ocean data assimilation system for routine analyses
of upper ocean thermal and current structures (Leetmaa and Ji, 1989). These analyses could be used to produce ocean initial conditions for coupled ocean–atmosphere
model forecasts of El Ni˜ no and La Ni˜ na, but likewise depended on the availability
of accurate surface wind forcing. Unfortunately, by 1990 it had become clear that
the launch of the NASA scatterometer for high-precision surface wind measurements
from space, anticipated for the second half of TOGA, would be delayed until after
no and Ocean Observations
89
with Joel Picaut in Noumea, New Caledonia, for deployments on French cruises
along 165
◦ E; and with Akimasa Sumi (University of Tokyo) and Kensuke Takeuchi
(Hokkaido University) for deployments on Japanese cruises along 147
◦ E. By the
TOGA midlife review in Honolulu in July 1990, there were 17 ATLAS moorings
deployed across the Pacific as part of the TOGA Observing System (Figure 6.2;
McPhaden and Hayes, 1990; Hayes et al., 1991). Stan named this array the TOGA
Thermal Array in the Ocean (TAO) (Nova University, 1989).
The first half of TOGA witnessed significant progress on a number of other
fronts as well. Between 1985 and 1990, the number of XBT lines regularly sampled
in the tropical Pacific nearly doubled from 10 to 18, the number of TOGA Pacific tide
gauges nearly doubled from 42 to 80, and surface drifters drogued at 15 m depth to
measure mixed layer velocity and SST increased from only 8 to 164 (McPhaden et al.,
1998). Declassification of the U.S. Navy’s GEOSAT satellite altimetry data made it
possible to observe Kelvin wave variations associated with El Ni˜ no from space for
the first time in 1986–88 (Miller et al., 1988).
Early in TOGA, Cane et al. (1986) made the first successful prediction of an
El Ni˜ no (the 1986–87 event) using a simple dynamical coupled ocean–atmosphere
model. Two other research groups likewise issued successful El Ni˜ no forecasts during
this time using statistical and statistical–dynamical hybrid models (Barnett et al.,
1988). Following the 1986–87 El Ni˜ no, unusually cold tropical Pacific SSTs in 1988–
89 and their link to drought in the U.S. Midwest (Trenberth and Branstator, 1992)
focused attention on the cold phase of ENSO, dubbed La Ni˜ na (Philander, 1990). New
theories of the ENSO cycle between warm and cold events also emerged in which
wind forced changes in thermocline depth, mediated by equatorial Kelvin and Rossby
waves, were identified as the mechanism governing delayed negative feedbacks on
the growth of tropical Pacific SST anomalies (Schopf and Suarez, 1988; Battisti
and Hirst, 1989). The accumulation of excess upper ocean heat content at equatorial
latitudes associated with these wave processes was also identified as a precondition
for the occurrence of El Ni˜ no and as the source of predictability for ENSO time scale
variations (Wyrtki, 1985; Cane et al., 1986; Zebiak, 1989).
Building on studies prior to TOGA using simple “two-layer” ocean models
(Busalacchi et al., 1983), realistic wind-forced simulations of the El Ni˜ no were
made for the first time from ocean general circulation models (e.g., Philander and
Seigel, 1985). These modeling advances emphasized the need for accurate basin scale
wind measurements to simulate ENSO-related changes in ocean circulation and SST
(Harrison et al., 1989). Also, NOAA’s National Meteorological Center established
the rudiments of an operational ocean data assimilation system for routine analyses
of upper ocean thermal and current structures (Leetmaa and Ji, 1989). These analyses could be used to produce ocean initial conditions for coupled ocean–atmosphere
model forecasts of El Ni˜ no and La Ni˜ na, but likewise depended on the availability
of accurate surface wind forcing. Unfortunately, by 1990 it had become clear that
the launch of the NASA scatterometer for high-precision surface wind measurements
from space, anticipated for the second half of TOGA, would be delayed until after