DATA ASSEMBLY AND PROCESSING FOR OPERATIONAL OCEANOGRAPHY: 10 YEARS OF ACHIEVEMENTS

1
DATA ASSEMBLY AND PROCESSING FOR OPERATIONAL
OCEANOGRAPHY 10 YEARS OF ACHIEVEMENTS

• P.Y. Le Traon, G. Larnicol, S. Guinehut, S.
  Pouliquen, A. Bentamy, D. Roemmich, C. Donlon, H.
  Roquet, G. Jacobs, D. Griffin, F. Bonjean, N.
  Hoeppfner, L.A. Breivick
• Outline
• The role for data processing centers for
  operational oceanography
• Review of main achievements over the past 10
  years
• In-situ and satellite data centers
• Joint use of in-situ and satellite data
• Conclusions and prospects
•  

2
Data centers an essential component of
operational oceanography infrastructure

• The quantity, quality and availability of data
  sets and data products directly impact the
  quality of ocean analyses and forecasts.
  Products derived from the data themselves are
  also used directly for applications.
• Key data processing needs for operational
  oceanography
• more effective data assembly
• more timely data delivery,
• improvements in data quality,
• better characterization of data errors,
• development of new or high level data products
• Over the past 10 years, major progresses have
  been achieved. These improvements are mandatory
  for an effective use of data in assimilation
  systems and for OO services.

3
Role of data processing centers

• provide data sets and products required for data
  assimilation both in real-time and delayed mode
  (reanalyses). QC, validation, error
  characterization
• provide data syntheses and high level data
  products (e.g. merging of multiple satellite data
  sets, climatologies)
• monitor performances of the observing system
  (e.g. data availability, possible degradation of
  performances and/or sampling)
• organize interfaces with data assimilation
  centers (e.g. feedback on QC, on the impact of
  data sets and on new or future requirements)

4
Main achievements

• Main GODAE efforts were focused on altimeter, SST
  and in-situ data needed and used by most GODAE
  global data assimilation systems, e.g.
• Argo data system
• Coriolis in-situ data assembly and processing
• SSALTO/DUACS, Navocean for altimetry
• GHRSST-PP for sea surface temperature
• Advances were also made for other data sets that
  are needed for specific applications (e.g. sea
  ice) or will become more and more important for
  operational systems (e.g. ocean color, high
  resolution winds) and/or for reanalysis
  activities (satellite windsfluxes).

5
Argo a breakthrough in data management and data
processing
Argo data flows and actors for real-time (left)
and delayed mode (right) processing Data freely
available in real time (12-24h) for all users.
6
In-situ data assembly Coriolis
Surface drifters Argo R.
Vessels and VOS
March 2004 April 2008
Moorings2004 2008
Gliders
Data delivered to models x 3 between 2004 and 2008
7
Altimetry SSALTO/DUACS

• Homogeneous, inter-calibrated and directly
  usable high quality altimeter data from all
  missions
• Along-track gridded products in near real time
  and delayed mode.
• New and improved products (e.g. MSLAs, MDTs).
• Timeliness improved (e.g. use of OSDR).

8
Merging multiple altimeter missions

• Intercalibration orbit error reduction by using
  T/P-Jason as a reference for the other missions
• Improvements in mappingmerging methodologies
  (e.g. preprocessing, noise and signal
  covariances)
• Use of Mean Dynamic Topographies (geoids from
  GRACE, synthetic methods) to get absolute dynamic
  topography (major impact)

ERS ERS adjusted onto T/P T/P
9
Surface current mapping using 4 altimeters
(Pascual et al., GRL, 2006)
2 alt
4 alt
Geostrophic Velocity Anomalies Absolute Geostrophic Velocity Absolute Velocity (Ekman component) Improvement using 4 sat missions
U 59.6 34.2 24.3 9
V 39.2 32.1 28.4 15
Need gt 2 altimeters, precise mean currents
(GRACE, GOCE) and scatterometers Requirement is
much higher for real time applications (Pascual
et al., 2008)
10
GODAE SST need Global high resolution in time
(lt1 day) and space (lt10 km) A GODAE pilot
project (see Donlon et al)

• Global High Resolution SST pilot project (GHRSST)
• combination of data from various sources
• modern data serving
• entrain scientific expertise for quality products

Outstanding progress efficient activation
through regional data assembly centres (R-DAC)
international cooperation, new high resolution
global and regional products (L2P, L4)
http//www.ghrsst-pp.org/
11
Bias correction based on AATSR SSTs
SST MSG SST AATSR Observed 17-19/02/2008
SST MSG SST AATSR Analysed 17-19/02/2008
SST MSG SST AATSR 19/02/2008 Before cor. std
0.46 After cor. std 0.38
12
Ocean colour

• Objectives to provide an accurate and
  consistent stream of ocean colour data required
  for monitoring forecasting (MERSEA)

Assemble a complete data base of phytoplankton
biomass, diffuse attenuation coefficient and
primary production Evaluate the quality of the
data through sensors comparison and validation
exercises Implement regional algorithms when
appropriate Provide a critical assessment on
use of OC into biogeochemical models R D on OC
multi-sensor merging techniques
13
Ice concentration Ice drift

• Large-scale ice drift fields derived from
  scatterometer data (QuikSCAT and ASCAT) and
  passive microwave data (SSMI and AMSR-E) for the
  winter months. Ice drift in the Fram Strait from
  SAR data (higher resolution).
• Assimilation of ice drift data together with ice
  concentration has started in some of the GODAE
  models.

14
High resolution winds

• To enhance the spatial and temporal resolutions
  of surface wind, several attempts have been made
  to merge the remotely sensed data to the
  operational NWP wind analyses over the global
  oceans
• Blended wind speed (QuickScat, SSM/I, ECMWF) and
  direction for 29 September 2008 (Bentamy et al.,
  2006).

15
Joint use of in-situ and remote sensing data

• Validation
• Comparison needed to validate satellite data but
  also in-situ data. Needed to check the
  consistency between the different data sets
  before assimilation in an ocean model.
• Examples tide gauges and altimeter data for
  validation (MSL), Argo and altimetry (QC),
  satellite SST validation.
• Merged products
• Products merging in-situ and satellite data
  through statistical methods have been developed
  by several GODAE groups.
• They have been used both for comparison with
  data assimilation systems (statistical vs
  dynamical) and to serve applications. Very
  useful.
• Examples Mercator-OceanCLS (Armor, Surcouf),
  Bluelink and NOPP (Oscar), Navocean (Modas).

16
Altimeter Sea Level Anomaly and Argo Dynamic
Height Anomaly time series for two Argo
floats (from Guinehut et al., 2008)

a.
Float 5900026


R
0.88

rms
-
diff 26.8



b. Float 1900249

R 0.00

rms
-
diff 1538.0




17
Ocean Surface Current Analyses Real-time
(OSCAR) (F. Bonjean)
18
PSY3V1R1
GLB/NCODA
19
Joint use of SST and altimetry (BlueLink D.
Griffin)
20
Merging altimeter, SST and in-situ data through
statistical techniques (Armor, Modas, Bluelink)
Half of the variance at depth can be explained
from altimetry and SST data.
Rms and mean error in predicting sub-surface T
using Levitus (red), synthetic fields (blue),
combined fields (green).
Instantaneous T field at 50 m from (a) in-situ
data, (b) synthetic T and (c) combined T in the
Kuroshio region (in C).
21
New theoretical frameworks to better exploit high
resolution information from satellite data
Relative vorticity and currents at 500 m depth
from the POP 1/10 model (left) and reconstructed
from SST only from surface quasigeostrophy (Klein
et al., 2006) equations. Isern-fontanet et al,
2006, 2008.
22
Conclusions and prospects

• Over the past 10 years, capabilities of data
  assembly and processing centers have been
  dramatically improved
• New or improved data sets and products needed by
  the modeling and data assimilation systems and
  for applications have been developed.
• Accuracy and timeliness of products have been
  improved.
• In-situ and remote sensing data are now jointly
  used for calibration, validation, consistency
  analysis and to derive merged products that
  provide complementary information to modelling
  and data assimilation products.

23
Conclusions and prospects

• There are still a series of advances in data
  processing that are expected to impact
  operational oceanography and its applications
• Continuous improvements are needed so that data
  sets and products evolve according to
  requirements from modeling and data assimilation
  systems (inc. error characterization).
• Prepare the use of new data sets biogeochemical
  in-situ sensors (e.g. oxygen, Chl-a) on floats or
  gliders, new satellite missions for SSS (SMOS,
  Aquarius) and gravity (GOCE) and high resolution
  altimetry (SWOT), etc
• Improve data assembly of key data sets such as
  velocity data (drifters, ADCP, Argo floats,
  altimetry/SST/OC/SAR).
• Better exploit the high resolution information in
  satellite observations.