&51=-9:1;@6.$6<;03691,)&51=-9:1;@6.$6<;03691,)
$+063)964465:$+063)964465:
)915-$+1-5+-)+<3;@"<*31+);165: 633-/-6.)915-$+1-5+-
%0-%9671+)3!+-)536*)3;46:70-9-!*:-9=15/$@:;-4%0-%9671+)3!+-)536*)3;46:70-9-!*:-9=15/$@:;-4
-+),-6."96/9-::-+),-6."96/9-::
1+0)-3+"0),-5
")+1B+)915-5=19654-5;)3)*69);69@
5;6516<:)3)++01
$6,,)9,$7)+-31/0;-5;-9
#6*-9;0-5-@
);165)3!+-)5$-9=1+-
-)5#-5-65/<@
15:;1;<;9)5+)1:,-#-+0-9+0-$+1-5;1B8<-76<93--=-3677-4-5;-5667-9);165
-55-;0$)/-
-96564@)*69);69@
$--5-?;7)/-.69),,1;165)3)<;069:
6336>;01:)5,),,1;165)3>692:);0;;7::+063)9+64465:<:.-,<4:+(.)+7<*
")9;6.;0-)915-1636/@64465:
$+063)964465:1;);165$+063)964465:1;);165
+"0),-51+0)-3<:)3)++015;65160-5-@#6*-9;65/<@-)5#-5-)/--55-;0$
)37-95)=1,115/<31)5")<3-@-9:)9@1;+0<4)9@%"1+)<;6-3#-@563,:#1+0)9,'
$41;0 -=133-%)2-<+01-5:<2-)5,"133-9"-)95"%0-%9671+)3!+-)536*)3;46:70-9-!*:-9=15/
$@:;-4-+),-6."96/9-::
)915-$+1-5+-)+<3;@"<*31+);165:
0;;7::+063)9+64465:<:.-,<4:+(.)+7<*
%01:9;1+3-1:*96</0;;6@6<.69.9--)5,67-5)++-::*@;0-633-/-6.)915-$+1-5+-);$+063)964465:;0):
*--5)++-7;-,.6915+3<:16515)915-$+1-5+-)+<3;@"<*31+);165:*@)5)<;0691A-,),4151:;9);696.$+063)9
64465:69469-15.694);16573-):-+65;)+;:+063)9+64465:<:.-,<
<;069:<;069:
1+0)-3+"0),-55;6516<:)3)++01#6*-9;0-5-@-)5#-5-65/<@-55-;0$)/-)=1,
)37-9515/1")<3<31)5)9@-@-9:)9@%1;+0<46-3"1+)<;#1+0)9,'#-@563,: -=133-
$41;0-5:<2-%)2-<+01)5,"-)95""133-9
%01:)9;1+3-1:)=)13)*3-);$+063)964465:0;;7::+063)9+64465:<:.-,<4:+(.)+7<*
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 103, NO. C7, PAGES 14,169-14,240, JUNE 29, 1998
The Tropical Ocean-Global Atmosphere observing
system' A decade of progress
Michael J. McPhaden, 1 Antonio J. Busalacchi, 2 Robert Cheney, 3 Jean-Rend
Donguy, 4 Kenneth S. Gage, s David Halpern, 6 Ming Ji, 7 Paul Julian, 8 Gary
Meyers, 9 Gary T. Mitchurn, 1ø Pearn P. Niiler, TM Joel Picaut, TM Richard W.
Reynolds, 7 Neville Smith, TM and Kensuke Takeuchi ls
Abstract. A major accomplishment of the recently completed Tropical Ocean-
Global Atmosphere (TOGA) Program was the development of an ocean observing
system to support seasonal-to-interannual climate studies. This paper reviews the
scientific motivations for the development of that observing system, the technological
advances that made it possible, and the scientific advances that resulted from
the availability of a significantly expanded observational database. A primary
phenomenological focus of TOGA was interannual variability of the coupled ocean-
atmosphere system associated with E1 Nifio and the Southern Oscillation (ENSO).
Prior to the start of TOGA, our understanding of the physical processes responsible
for the ENSO cycle was limited, our ability to monitor variability in the troi•ical
oceans was primitive, and the capability to predict ENSO was nonexistent. TOGA
therefore initiated and/or supported efforts to provide real-time measurements of
the following key oceanographic variables: surface winds, sea surface temperature,
subsurface temperature, sea level and ocean velocity. Specific in situ observational
programs developed to provide these data sets included the Tropical Atmosphere-
Ocean (TAO) array of moored buoys in the Pacific, a surface drifting buoy program,
an island and coastal tide gauge network, and a volunteer observing ship network of
expendable bathythermograph measurements. Complementing these in situ efforts
were satellite missions which provided near-global coverage of surface winds, sea
surface temperature, and sea level. These new TOGA data sets led to fundamental
progress in our understanding of the physical processes responsible for ENSO and
to the development of coupled ocean-atmosphere models for ENSO prediction.
And thorough this distemperature we see the
seasons alter...
Shakespeare's "A Midsummer Night's Dream"
Act 2, Scene 1
1. Introduction
E1 Nifio (EN) is characterized by a large-scale weak-
ening of the trade winds and warming of the surface l•v-
ers in the eastern and central equatorial Pacific Ocean.
• Pacific Marine Environmental Laboratory, NOAA, Seat-
tle, Washington.
2NASA Goddard Space Flight Center, Greenbelt, Mary-
land.
3National Ocean Service, NOAA, Silver Spring, Mary-
land.
4institut Fran•ais de Recherche Scientifique pour le D•vel-
oppement en Cooperation, Plouzane, France.
5Aeronomy Laboratory, NOAA, Boulder, Colorado.
6Jet Propulsion Laboratory, California lnstitute of Tech-
nology, Pasadena.
Copyright 1998 by the American Geophysical Union.
Paper number 97JC02906.
0148-0227 / 98 / 97J C-02906 $09.00
E1 Nifio events occur irregularly at intervals of roughly
2-7 years, although the average is about once every 3-
4 years [Quinn et al., 1987]. They typically last 12-18
months, and are accompanied by swings in the Southern
Oscillation (SO), an interannual seesaw in tropical sea
level pressure between the eastern and western hemi-
spheres [Walker, 1924]. During E1 Nifio, unusually high
atmospheric sea level pressures develop in the western
?National Centers for Environmental Prediction, NOAA,
Camp Springs, Maryland.
8Suitland, Maryland.
9Commonwealth Scientific and Industrial Research Orga-
nization, Tasmania, Australia.
•øDepartment of Marine Science, University of South
Florida, Saint Petersburg.
• Scripps Institution of Oceanography, La Jolla, Califor-
nia.
•2Institut Fran•ais de Recherche Scientifique pour le D•-
veloppement en Cooperation.
•3Now at NASA Goddard Space Flight Center, Greenbelt,
Maryland
•4Bureau of Meteorology Research Centre, Melbourne,
Victoria, Australia.
•5Institute of Low Temperature Science, Hokkaido Uni-
versity, Sapporo, Japan.
14,169
14,170 MCPHADEN ET AL.: TOGA OBSERVING SYSTEM
tropical Pacific and Indian Ocean regions, and unusu-
ally low sea level pressures develop in the southeast-
ern tropical Pacific. Bjerknes [1966, 1969] was the first
to link swings in the Southern Oscillation to E1 Nifio
events, proposing that the two phenomena were gen-
erated by coupled ocean-atmosphere interactions. SO
tendencies for unusually low pressures west of the date
line and high pressures east of the date line have also
been linked to periods of anomalously cold equatorial
Pacific sea surface temperatures (SSTs) sometimes re-
ferred to as La Nifia [Philander, 1990]. The full range
of SO variability, including both anomalously warm and
cold equatorial SSTs, is often referred to as ENSO.
ENSO is associated with shifts in the location and
intensity of deep convection and rainfall in the tropi-
cal Pacific. During E1 Nifio events, drought conditions
prevail in northern Australia, Indonesia, and the Philip-
pines, and excessive rains occur in the island states of
the central tropical Pacific and along the west coast of
South America. Shifts in the pattern of deep convection
in the tropical Pacific also affect the general circulation
of the atmosphere and extend the impacts of ENSO to
other tropical ocean basins and to midlatitudes [Ras-
musson and Wallace, 1983; Ropelewski and Halpert,
1986, 1987; Halpert and Ropelewski, 1992; Trenberth et
al., this issue]. During E1 Nifio most of Canada and
the northwestern United States tend to experience mild
winters, and the states bordering the Gulf of Mexico
tend to be cooler and wetter than normal. California
has experienced a disproportionate share of episodes of
heavy rainfall during E1 Nifio winters such as 1982-
1983, 1991-1992, and 1994-1995. Atlantic hurricanes
tend to be less frequent during warm events and more
frequent during cold events [Gray et al., 1993]. E1 Nifio
events also disrupt the marine ecology of the tropical
Pacific and the Pacific coast regions of the Americas,
affecting the mortality and distribution of commercially
valuable fish stocks and other marine organisms [Barber
and Chavez, 1983; Dessier and Donguy, 1987; Pearcy
and $choener, 1987; Lehodey et al., 1997]. Thus, though
originating in the tropical Pacific, ENSO has socioeco-
nomic consequences that are felt worldwide.
The widespread and systematic influence of ENSO on
the ocean-atmosphere system, and the potential that
it might be predictable seasons to years in advance,
led to initiation of the international Tropical Ocean-
Global Atmosphere (TOGA) Program, a 10-year study
(1985-1994) of seasonal-to-interannual (also referred to
as short-term) climate variability. The goals of the
TOGA program were [World Climate Research Pro-
gram, 1985, p. vii]
[1.] to gain a description of the tropical
oceans and the global atmosphere as a time de-
pendent system, in order to determine the ex-
tent to which this system is predictable on time
scales of months to years, and to understand
the mechanisms and processes underlying that
predictability;
[2.] to study the feasibility of modeling the
coupled ocean-atmosphere system for the pur-
pose of predicting its variability on timescales
of months to years; and
[3.] to provide the scientific background for
designing an observing and data transmission
system for operational prediction if this capa-
bility is demonstrated by the coupled ocean-
atmosphere system.
The scientific background and rationale for TOGA
was spelled out in several planning documents [e.g.,
World Climate Research Program, 1985; National Re-
search Council, 1983, 1986]. Prior to TOGA, a ba-
sic description of oceanic and atmospheric variability
associated with E1 Nifio existed [e.g., Rasmusson and
Carpenter, 1982], as did a basic description of trop-
ical/extratropical atmospheric teleconnections in the
northern hemisphere [e.g., Hotel and Wallace, 1981].
Atmospheric general circulation models had shown a
sensitivity both in the tropics and at higher latitudes
to underlying equatorial Pacific SST anomalies, and
theories were emerging on how tropical forcing gave
rise to observed teleconnection patterns [e.g., Hoskins
and Karoly, 1981]. Relatively simple wind-forced ocean
models prior to TOGA were capable of simulating some
aspects of seasonal-to-interannual variability associated
with sea level variations in the Pacific [e.g., Busalac-
chi and O'Brien, 1980; Busalacchi and O'Brien, 1981;
Busalacchi et al., 1983]. Initial attempts to quantita-
tively assess the role of ocean dynamics in controlling in-
terannual variations in SST were underway [Gill, 1983].
Also, ocean general circulation models with explicit
mixed layer thermodynamics were being developed for
improved simulations of SST variability [e.g., $chopf
and Cane, 1983]. Coupled tropical ocean-atmosphere
models were in their infancy prior to TOGA. They
showed promise though in their ability to elucidate
possible mechanisms responsible for ocean-atmosphere
feedbacks and in their ability to crudely simulate as-
pects of the ENSO cycle [McCreary, 1983; Philander et
al., 1984].
Theories regarding the mechanisms responsible for E1
Nifio variations in the ocean were likewise developing
[e.g., Wyrtki, 1975; McCreary, 1976; Hurlbutt et al.,
1976]. The roles of ocean dynamics and, in particular,
wind-forced equatorial Kelvin and Rossby waves in af-
fecting large-scale redistribution of mass and heat in the
equatorial band were widely regarded as crucial aspects
of the ocean's role in the ENSO cycle. The rapid re-
sponse of the equatorial ocean to wind forcing and the
ability of equatorial waves to affect remote parts of the
basin on relatively short timescales distinguish the trop-
ics from higher latitudes where planetary scale waves
propagate much more slowly. Substantial responses in
equatorial currents and sea surface heights to relatively
short-duration wind events were evident in observations
before the start of TOGA [Knox and Halpern, 1982;
Eriksen et al., 1983]. These observations suggested the
potential for remotely forced changes in SST due to
wave-induced changes in horizontal and vertical advec-
tion and upper ocean mixing. Thus understanding the
MCPHADEN ET AL.: TOGA OBSERVING SYSTEM 14,171
oceanic processes giving rise to SST variability in the
tropical Pacific was a more challenging problem than
at midlatitudes, where SST variations on seasonal and
interannual timescales are generated primarily by local
air-sea heat exchange [Gill and Niiler, 1973].
Much of the progress in oceanographic studies re-
lated to E1 Nifio in the 1970s and early 1980s was
stimulated by fieldwork and modeling efforts as part of
the Equatorial Pacific Ocean Climate Studies (EPOCS)
program [Hayes et al., 1986], the North Pacific Exper-
iment (NORPAX) [Wyrtki et al., 1981], and the Pa-
cific Equatorial Ocean Dynamics (PEQUOD) experi-
ment [Eriksen, 1987]. These programs provided new
data for basic description of phenomenology, for devel-
oping and testing dynamical hypotheses, and for model
development and validation [Halpern, 1996]. Impressive
though the scientific advances were during this period,
they were still inadequate in many respects. To quote
from the document U.S. Participation in the TOGA
Program [National Research Council, 1986, p. 6-7]:
[1.] The subsurface signature of E1 Nifio
events and the time-dependent fluxes of mo-
mentum and energy at the air-sea interface are
known only qualitatively, and existing observa-
tions are inadequate to define them with the
accuracy needed for initializing and verifying
models.
[2.] Major uncertainties still exist concern-
ing the tropical and southern hemisphere atmo-
spheric circulations and their interannual vari-
ability.
[3.] The processes that determine the sea sur-
face temperature distribution and the surface
wind field over the tropics are not yet well un-
derstood.
[4.] The fundamental behavior and predict-
ability of the coupled climate system are just
beginning to be understood.
TOGA, initiated by the World Climate Research Pro-
gram [1985], provided a flamework for coordinated, sus-
tained international efforts aimed at addressing these
shortcomings. Implementation of TOGA was to be
carried out with major new initiatives in modeling,
process-oriented field studies, and long-term observa-
tions. Efforts in these areas were to be highly interac-
tive and mutually reinforcing. Models and the results
of process studies would be used to help guide the de-
velopment of long-term observational systems. Long-
term observations in turn would provide a large-scale,
long-term flamework in which to interpret the results of
shorter-duration, geographically focused, intensive pro-
cess studies. Long-term observations would also be used
to validate models, to aid in the development of param-
eterization schemes for subgrid scale model physics, and
to initialize dynamical model-based climate forecasting
schemes.
The need for an improved observing system was un-
derscored during the planning stages of TOGA in the
early 1980s, when the scientific community was caught
completely off guard by the 1982-1983 E1 Nifio, the
strongest in over a hundred years (see Appendix A for
details). This E1 Nifio was neither predicted nor even
detected until several months after it had started. The
lesson from this experience was obvious: an in situ ob-
serving system capable of delivering data in real time
was urgently needed for improved monitoring, under-
standing, and prediction of E1 Nifio and related phe-
nomena. To meet these requirements, the TOGA Im-
plementation Plan called for the development of a "thin
monitoring" array of in situ measurements based on
the enhancement of existing capabilities [International
TOGA Project OJfice, 1992]. This observing system was
to provide data on a basin scale for at least 10 years
without significant temporal gaps, so that a continu-
ous record of climate variability could be assembled.
Ten years was considered the minimum length of time
needed for a comprehensive study of interannual vari-
ability, the dominant mode of which was ENSO cycle.
The purpose of this paper is to describe the devel-
opment of the TOGA observing system, to highlight
scientific advances that have resulted from implemen-
tation of this system, and to summarize how data from
this system have contributed to progress in developing
models for improved climate analysis and prediction.
We will emphasize oceanic, rather than atmospheric,
components of the observing system, reflecting relative
levels of effort expended on implementation during the
TOGA decade. However, we will discuss TOGA efforts
to augment the World Weather Watch for atmospheric
measurements and to establish a specialized network of
island-based wind profilers.
We will also emphasize in situ rather than satellite
data. Satellite missions were generally initiated for pur-
poses other than, or only partially motivated by, short-
term climate research (e.g., operational weather pre-
diction, national defense, general oceanographic and/or
meteorological applications). Also, delays in satellite
missions and/or temporal discontinuities in satellite
data coverage heightened reliance on in situ measure-
ments during the TOGA decade. For example, launch
of the National Aeronautics and Space Administration's
scatterometer (NSCAT) for surface wind velocity esti-
mates, originally scheduled for 1989, was repeatedly de-
layed until August 1996, almost 2 years after the end of
TOGA. The satellite carrying NSCAT then failed pre-
maturely, in June 1997, after being operational for only
8 months. Similarly, there was a 2-year hiatus in satel-
lite sea level altimetry measurements between the end
of the U.S. Navy's Geodetic Satellite (Geosat) mission
in 1989 and the launch of European Space Agency's
European Remote Sensing Satellite (ERS-1) in 1991.
Nonetheless, we will discuss those satellite missions that
contributed directly to TOGA objectives, particularly
with regard to oceanic variability. Satellite measure-
ments targeted more toward documenting and under-
standing atmospheric variability during TOGA, namely
those for precipitation, water vapor, clouds, radiation,
and evaporation [Lau and Busalacchi, 1993], are dis-
cussed in work by Wallace et al. [this issue].
Originally, it was anticipated that TOGA would de-