SATELLITE MEASUREMENTS
183
interpolation (OI), or other analysis technique, to create the best daily
fields of SST fnd , without any cloud gaps.
The processing models required to generate these products are fully
defined in the GHRSST Data-processing System document, the current
version of which is GDS v1.5, available online from http://www.ghrsstpp.org. The model is an active document, still being refined in the light of
working experience.
4.3
Putting the GHRSST processing model into action
GHRSST-PP has evolved rapidly to meet the perceived need for a new
class of merged SST products. After four Workshops the Science Team
produced an implementation plan that was approved by GODAE, its parent
body. GHRSST-PP is now established through the creation at the UK Met.
Office of an International Project Office sponsored by the European Space
Agency (ESA). While this provides leadership and co-ordination, the
development comes from work at a regional level, through the GHRSST-PP
regional data assembly centres (RDAC)s. One of these already existed
through a Japanese SST programme for the Asian Pacific area. The function
of a European RDAC is now being fulfilled by a project within the ESA’s
Data Utilisation Envelope (DUE) Programme called “Medspiration”. The
Multi-sensor Improved SST Project (MISST), has been established by the
National Ocean Partnership Programme in USA to support the function of an
American RDAC. When fully implemented GHRSST-PP should make a
considerable impact on the quality of SST data available for operational
oceanography worldwide.
As the European RDAC for GHRSST-PP the Medspiration project,
which commenced in January 2004, contributes to the development of an
operational oceanography system within the European Union’s Global
Monitoring for Environment and Security (GMES) programme. The core
user need identified for Medspiration was for SST products at high
resolution (~10 km) over the Atlantic Ocean and adjoining Seas, which
combine the best of all the SST data products already available from
individual sensors. The desired temporal resolution is 6 hrs. The required
accuracy of SST data is to be better than 0.4 K, with a goal of 0.1 K. Since
Medspiration will use the SST sources in Table 4, which are already
supplied by various agencies, it will always depend on the quality of the data
being ingested.
However, it will for the first time provide European operational users not
only with a near-real time single point access to data from all the different
SST sources but also with robust error and confidence statistics attached to
the SST products. This capacity to assign quality values to all the SST data
sources being used, and to any future SST products that may be generated, is
a fundamental requirement of GHRSST-PP. It is needed in order to
Précédent

- 191/573

Suivant