coverage of the surface to ongoing large-scale coverage of the upper ocean, to the global description
of properties, tracers and circulation at depth.
Because of the massive scale of WOCE, it has generally been much more difficult to isolate specific
urgent questions, leading to concerns about its
being an unfocused experiment, unlike for instance
the tropical Pacific experiments with a unified goal –
to describe and predict El Niño. The WOCE data
sets are being used to describe a plethora of questions of the same scope as El Niño; some of these
problems that are first being addressed with
WOCE are blossoming into larger ongoing projects. We have made an attempt in this chapter,
and the attempt is made in more depth throughout
this book, to cover the many different thrusts
taken with WOCE data. The results are clearly
accretionary. Taken as whole, even at this time,
and clearly further along in the analysis, WOCE
will provide many significant advances for largescale oceanography.
A major strength of WOCE has been the collection of many different types of complementary
data and development of models and investigator
expertise for every basin. WOCE ‘synthesis’ means
approaching individual ocean descriptions or
processes using as many of the data types and
modelling tools as are available and useful to the
specific problem. The existence of concurrent
upper ocean and satellite time series along with all
of the background information at the largest scale
for the WOCE hydrographic programme, as well
as basin-scale and global models, has allowed a
description of the time-varying response of the
ocean that was not possible before. The coexistence of a complete mid-depth mean circulation
from floats, along with hydrographic and processoriented measurements and model analysis of the
same regions, has meant that new and potentially
important circulation features have been exposed.
The coexistence of full tracer data sets, temperature/salinity and direct velocity measurements at
depth is allowing significant advances in describing the mean circulation and ventilation time
scales. Well-designed local current arrays with
hydrographic and basin-scale float measurements
have enabled major circulation elements to be
quantified. One of the major advances of WOCE,
quantification of the diapycnal diffusion rate
(Ledwell et al., 1993) and its geographic variations
(e.g. Ledwell et al., 2000), has been considerably
strengthened by the existence of concurrent largerscale data sets in addition to those observations
specifically designed for measuring diffusivity.
Was it useful to have undertaken the whole
exercise of observing the global ocean? Models are
becoming global, climate-relevant processes are
global, and the existence of global data sets to test
hypotheses and provide context for regional
process descriptions is critical. The latter has not
perhaps received proper credit. Every analysis of
every regional process (using WOCE data) requires
a larger-scale context (provided also by the larger
WOCE data sets). Many basin-scale analyses are
underway, all drawing to some extent on the context/boundary conditions provided by the larger
data sets. The importance, persistence, and basic
physics of various ocean processes are much more
clearly revealed if they can be found and studied in
more than one setting, meaning more than one
basin. The particular scientific questions that will
be answered by the global data sets per se are of
the same ilk as those for each region – describing
the global thermohaline circulation, global heat
and property transports, global variability. All
require global coverage. But how important was
synopticity, even in the loose WOCE sense of coverage over about 10 years? It was important for
two reasons – the most efficient way to obtain
enough information to make the global estimates
and the regional descriptions with global context
was to do it as quickly as possible, using the same
technology throughout with the same sampling
philosophy. Some data sets, such as transient tracers used for ventilation studies, require a reasonable amount of synopticity.
What would have been better done differently, if
anything? Better coordination could have reduced
the amount of time required to cover regions. Satellite programmes that were an integral part of the
description did not begin until after WOCE had
started, resulting in unfortunately incomplete
coverage overlap. Throughout the design and
execution of WOCE there was pressure to have
modellers and observers coordinate their efforts
more seamlessly. However, at a practical level, this
can only be accomplished if investigators have an
immediate common goal that requires the integration of different approaches that each provide
enough valid information to allow synergy. That
this did not happen was perhaps a consequence
of the state of ocean modelling, in particular
7.1 Towards a WOCE Synthesis
543
Talley, Stammer and Fukumori
of properties, tracers and circulation at depth.
Because of the massive scale of WOCE, it has generally been much more difficult to isolate specific
urgent questions, leading to concerns about its
being an unfocused experiment, unlike for instance
the tropical Pacific experiments with a unified goal –
to describe and predict El Niño. The WOCE data
sets are being used to describe a plethora of questions of the same scope as El Niño; some of these
problems that are first being addressed with
WOCE are blossoming into larger ongoing projects. We have made an attempt in this chapter,
and the attempt is made in more depth throughout
this book, to cover the many different thrusts
taken with WOCE data. The results are clearly
accretionary. Taken as whole, even at this time,
and clearly further along in the analysis, WOCE
will provide many significant advances for largescale oceanography.
A major strength of WOCE has been the collection of many different types of complementary
data and development of models and investigator
expertise for every basin. WOCE ‘synthesis’ means
approaching individual ocean descriptions or
processes using as many of the data types and
modelling tools as are available and useful to the
specific problem. The existence of concurrent
upper ocean and satellite time series along with all
of the background information at the largest scale
for the WOCE hydrographic programme, as well
as basin-scale and global models, has allowed a
description of the time-varying response of the
ocean that was not possible before. The coexistence of a complete mid-depth mean circulation
from floats, along with hydrographic and processoriented measurements and model analysis of the
same regions, has meant that new and potentially
important circulation features have been exposed.
The coexistence of full tracer data sets, temperature/salinity and direct velocity measurements at
depth is allowing significant advances in describing the mean circulation and ventilation time
scales. Well-designed local current arrays with
hydrographic and basin-scale float measurements
have enabled major circulation elements to be
quantified. One of the major advances of WOCE,
quantification of the diapycnal diffusion rate
(Ledwell et al., 1993) and its geographic variations
(e.g. Ledwell et al., 2000), has been considerably
strengthened by the existence of concurrent largerscale data sets in addition to those observations
specifically designed for measuring diffusivity.
Was it useful to have undertaken the whole
exercise of observing the global ocean? Models are
becoming global, climate-relevant processes are
global, and the existence of global data sets to test
hypotheses and provide context for regional
process descriptions is critical. The latter has not
perhaps received proper credit. Every analysis of
every regional process (using WOCE data) requires
a larger-scale context (provided also by the larger
WOCE data sets). Many basin-scale analyses are
underway, all drawing to some extent on the context/boundary conditions provided by the larger
data sets. The importance, persistence, and basic
physics of various ocean processes are much more
clearly revealed if they can be found and studied in
more than one setting, meaning more than one
basin. The particular scientific questions that will
be answered by the global data sets per se are of
the same ilk as those for each region – describing
the global thermohaline circulation, global heat
and property transports, global variability. All
require global coverage. But how important was
synopticity, even in the loose WOCE sense of coverage over about 10 years? It was important for
two reasons – the most efficient way to obtain
enough information to make the global estimates
and the regional descriptions with global context
was to do it as quickly as possible, using the same
technology throughout with the same sampling
philosophy. Some data sets, such as transient tracers used for ventilation studies, require a reasonable amount of synopticity.
What would have been better done differently, if
anything? Better coordination could have reduced
the amount of time required to cover regions. Satellite programmes that were an integral part of the
description did not begin until after WOCE had
started, resulting in unfortunately incomplete
coverage overlap. Throughout the design and
execution of WOCE there was pressure to have
modellers and observers coordinate their efforts
more seamlessly. However, at a practical level, this
can only be accomplished if investigators have an
immediate common goal that requires the integration of different approaches that each provide
enough valid information to allow synergy. That
this did not happen was perhaps a consequence
of the state of ocean modelling, in particular
7.1 Towards a WOCE Synthesis
543
Talley, Stammer and Fukumori