equal or greater importance. CLIVAR is in essence a
research programme for coupled modes of the physical climate system, and the oceans are a key element.
The question then arises as to the relevance of
the wide spectrum of ocean circulation and variability to these coupled modes. For ENSO at least,
it is accepted that the tropical oceans play an
important role, though it is clearly the low-frequency, large-scale signals that are of greatest
interest. For decadal and longer time scale variations, it is reasoned that the sequestering and subsequent release of heat (and moisture/salinity),
involving the global ocean, is a key process (Bryden
and Imawaki, Chapter 6.1; Wijffels, Chapter 6.2).
In general, as the time scales of interest grow
longer, we expect the ocean to play a more critical
role. However, one must develop these ideas more
deeply if the true legacy of WOCE is to be understood. In Section 7.4.5 I will argue that we not
only need to build upon the knowledge (and
observing systems) that impact directly on the
scales of interest, but also to exploit the improved
understanding of the oceanic variability itself in
order to estimate properly these large-scale, lowfrequency signals. We might state this most simply
by renovating and restating WOCE Goal 2:
In order to find methods for determining longterm changes in climate, we must first determine
how well the ocean circulation can be determined on the basis of irregular, unevenly spaced
observations that do not resolve many of the
energetic modes of ocean variability.
It is conceivable that many oceanic modes of
variability do not directly impact useful predictability of the climate system, just as we accept (for the
time being at least) that the details of day-to-day
weather changes do not affect ENSO forecasts to
first order. Yet, just as we are finding the regular
analyses from Numerical Weather Prediction are
useful for climate analysis and prediction, we may
also find that analysing and understanding ocean
‘weather’ is important for determining ocean climate. It is this aspect that is being pursued within
the Global Ocean Data Assimilation Experiment
(see, for example, Smith, 2000, and Section 7.4.5).
7.4.3 WOCE observations
Other chapters have discussed in detail the various
efforts mounted in WOCE to observe the ocean,
using both direct and remote techniques, and discussed the way these observations have been
applied to improving our knowledge of the ocean
circulation. My purpose here is to discuss briefly
some elements and their importance for the subject
of this chapter, elements that were designed and
implemented for the purposes of WOCE, but
which through good practice and judgement have
proven useful for many other purposes.
7.4.3.1 General strategy
The word ‘World’ in WOCE immediately shifts
the focus to the global domain. The WOCE objectives also make clear that the full depth of the
ocean must also be considered. At the risk of oversimplifying the aims and implementation of the
WOCE observational programme, the strategy can
be reduced to a few tasks:
¥ Observe the fluxes of momentum, heat and
moisture through the surface to determine the
forcing of the ocean circulation;
¥ Observe the full-depth, global physical and
dynamical properties and relevant chemical
properties, including tracers, in order to make a
first estimate of the total circulation, including
the oceanic fluxes of heat and fresh water (historical observations could in places aid this
task);
¥ Gather selected other observations and conduct
process studies in order to understand likely
errors in these estimates. These included:
¥ high-density cross-sections to quantify the
impact of the mesoscale field on fluxes;
¥ dedicated boundary measurements to understand the influence of narrow currents at the
land–ocean interface and in narrow constrictions of the bathymetry;
¥ full-depth and selected depth fixed-point measurements to quantify high-frequency effects;
¥ various broad-scale strategies that sampled
the seasonally varying upper ocean;
¥ repeated deep sections in order to provide
some estimate of the first-order variability
about the long-term mean;
¥ selected process experiments; and
¥ dedicated measurements in regions known to
have distinctive modes of variability (e.g. at
high latitudes).
At the heart was the aim to provide an estimate
of the mean state (a ‘snapshot’) of the ocean
7.4 Ocean and Climate Prediction – the WOCE Legacy
589
Smith
research programme for coupled modes of the physical climate system, and the oceans are a key element.
The question then arises as to the relevance of
the wide spectrum of ocean circulation and variability to these coupled modes. For ENSO at least,
it is accepted that the tropical oceans play an
important role, though it is clearly the low-frequency, large-scale signals that are of greatest
interest. For decadal and longer time scale variations, it is reasoned that the sequestering and subsequent release of heat (and moisture/salinity),
involving the global ocean, is a key process (Bryden
and Imawaki, Chapter 6.1; Wijffels, Chapter 6.2).
In general, as the time scales of interest grow
longer, we expect the ocean to play a more critical
role. However, one must develop these ideas more
deeply if the true legacy of WOCE is to be understood. In Section 7.4.5 I will argue that we not
only need to build upon the knowledge (and
observing systems) that impact directly on the
scales of interest, but also to exploit the improved
understanding of the oceanic variability itself in
order to estimate properly these large-scale, lowfrequency signals. We might state this most simply
by renovating and restating WOCE Goal 2:
In order to find methods for determining longterm changes in climate, we must first determine
how well the ocean circulation can be determined on the basis of irregular, unevenly spaced
observations that do not resolve many of the
energetic modes of ocean variability.
It is conceivable that many oceanic modes of
variability do not directly impact useful predictability of the climate system, just as we accept (for the
time being at least) that the details of day-to-day
weather changes do not affect ENSO forecasts to
first order. Yet, just as we are finding the regular
analyses from Numerical Weather Prediction are
useful for climate analysis and prediction, we may
also find that analysing and understanding ocean
‘weather’ is important for determining ocean climate. It is this aspect that is being pursued within
the Global Ocean Data Assimilation Experiment
(see, for example, Smith, 2000, and Section 7.4.5).
7.4.3 WOCE observations
Other chapters have discussed in detail the various
efforts mounted in WOCE to observe the ocean,
using both direct and remote techniques, and discussed the way these observations have been
applied to improving our knowledge of the ocean
circulation. My purpose here is to discuss briefly
some elements and their importance for the subject
of this chapter, elements that were designed and
implemented for the purposes of WOCE, but
which through good practice and judgement have
proven useful for many other purposes.
7.4.3.1 General strategy
The word ‘World’ in WOCE immediately shifts
the focus to the global domain. The WOCE objectives also make clear that the full depth of the
ocean must also be considered. At the risk of oversimplifying the aims and implementation of the
WOCE observational programme, the strategy can
be reduced to a few tasks:
¥ Observe the fluxes of momentum, heat and
moisture through the surface to determine the
forcing of the ocean circulation;
¥ Observe the full-depth, global physical and
dynamical properties and relevant chemical
properties, including tracers, in order to make a
first estimate of the total circulation, including
the oceanic fluxes of heat and fresh water (historical observations could in places aid this
task);
¥ Gather selected other observations and conduct
process studies in order to understand likely
errors in these estimates. These included:
¥ high-density cross-sections to quantify the
impact of the mesoscale field on fluxes;
¥ dedicated boundary measurements to understand the influence of narrow currents at the
land–ocean interface and in narrow constrictions of the bathymetry;
¥ full-depth and selected depth fixed-point measurements to quantify high-frequency effects;
¥ various broad-scale strategies that sampled
the seasonally varying upper ocean;
¥ repeated deep sections in order to provide
some estimate of the first-order variability
about the long-term mean;
¥ selected process experiments; and
¥ dedicated measurements in regions known to
have distinctive modes of variability (e.g. at
high latitudes).
At the heart was the aim to provide an estimate
of the mean state (a ‘snapshot’) of the ocean
7.4 Ocean and Climate Prediction – the WOCE Legacy
589
Smith