GCOS GOOS WCRP/OOPC IX/3
Action: NWP centres and VOS operators should be reminded of the importance of collecting
data in line with the GCOS Climate monitoring principles. The importance of VOS
observations for the calculation of surface fluxes in regions away from flux reference sites
should be stressed in revisions to the GCOS Draft Implementation Plan.
Convergence of VOS and VOSClim
Ideally all VOS observations would be of the quality aspired to by the VOSClim
project. However, a significant minority of VOS observations are of poor quality and to some
extent this undermines the usefulness of the remainder of the better quality observations. For
example, SST observations from VOS are assimilated into SST analyses with a relatively low
weight. This is partly because some VOS SST reports contain gross errors that could have a
serious adverse effect on the analysis. Numerical Weather Prediction (NWP) centres therefore
prefer to rely on SST reports from drifting buoys, which contain fewer gross errors but are
prone to drift with time (potentially causing problems with large quantities of biased data in
data sparse regions). VOS reports are therefore given a low priority by NWP centres and
there is little incentive to improve the quality as little use is made of the reports. Demand from
the NWP community is therefore shifting from VOS to alternative data sources. In the past
the requirements of the NWP and climate researchers have to a large extent coincided, but
more recently there have been moves by NWP centres to develop methods of targeting
observations in sensitive areas for forecasting in order to reduce costs and improve forecasts.
In addition new initiatives such as the Network of European Meteorological Services
(EUMETNET) Composite Observing System (EUCOS) (http://www.eucos.net/) and its
surface marine programme (E SURFMAR) are addressing their efforts to producing better
quality NWP forecasts over Europe. As a consequence National Met Services are likely to re
focus their VOS and buoy resources to meet the demands of NWP, possibly at the expense of
climatological requirements7. The importance of VOS for climate is that ships provide a wide
range of parameters from which the four components of the heat budget can be calculated
(long and shortwave radiation, sensible and latent heat flux). Whilst some moored buoys
provide similar information, typically alternative systems provide a subset of the information
required for climate studies, concentrating on the variables most important for NWP, such as
pressure and SST and possibly winds and air temperature. Limited automatic systems have
been installed on VOS providing similar information to that obtained from a drifting buoy.
The needs of the NWP and climate researchers for VOS data appear to be diverging.
However, both communities would be well served by a smaller number of reliably high
quality VOS providing a full range of meteorological variables. This would be supplemented
for NWP by information from other systems such as the buoy network, satellites and
radiosondes. The VOSClim project, although much smaller in size, provides a possible model
for a future transition to make the VOS a high quality data system. Dialogue between the
NWP and climate communities is limited, especially when one considers that many National
Meteorological Agencies have both NWP and climate responsibility. Improving this dialogue
The UK Met Office held a "User forum for observations" in May at which participants were invited to
present their requirements and priorities for observations collected by the Met Office. We are now
invited to submit our requirements for VOS sampling, and we urgently need to define requirements in
terms of numbers of observations and target accuracies for basic meteorological variables rather than
mean flux values. Until this is done our requirements cannot easily be incorporated into observing
system design either by the Met Office or other Meteorological Agencies.
Action: OOPC should urgently convene a small taskgroup to define VOS sampling and accuracy
requirements for basic meteorological variables to meet surface flux uncertainty targets. ????? to
report by end July ??????