Argo Status M Belbeoch Argo TC inputs from

Argo Status M. Belbeoch, Argo TC (inputs from AST)

Argo Infrastructure Argo is internationally managed by the Argo Steering Team. 2 co-chairs, National Argo programmes representatives, TC • Argo Data Management team coordinates data issues • Argo Data Management is a distributed system – National Data Centres (DACs) feed GTS of WMO and Internet GDACs – REAL TIME ( GTS, GDACs) with standard/automatic QC – DELAYED-MODE files replace RT files (GDACs, after 6 -12 months) • 2 Global Data Centres (USA, FR) - mirrored • Regional Centres being developed • Long term archival centre at US NODC • Argo Information Centre/JCOMMOPS/Argo Project Office 2

Argo Status: what has been achieved ? Argo has achieved the “ 3000 floats milestone” with global distribution and a comprehensive data management system • Argo Core Mission (3200 floats between 60°N and 60°S, no marginal seas) is NOT YET ACHIEVED • All floats are nor performing well. . . • All floats are not operating where they should. . . 3

Argo Status: network density Float density (100%=4 floats): good floats only Challenge in South Indian, and South Atlantic 4

Argo Status: network density 436, 000 Argo profiles were collected during 2004 -2008. Left: profiles per 1 o box, 2004 -mid 2008. (Roemmich, Gilson, Scripps/USA) More floats (600) are needed in the southern hemisphere, especially the South Indian and South Atlantic. Of the ~3300 active floats, >600 are either in marginal seas, high latitudes, over-sampled regions, or greylisted. Argo floats per degree of latitude (black). Argo’s design requirement (red) Equal area sampling (blue). (Freeland, IOS/Canada) 5

Argo Status: network density How effective is Argo in estimating large-scale (10 o x 3 months) variability? Estimates of large-scale signal variance (black, green) and noise (red) are made using (a, b) altimetric height subsampling experiments, (c) subsets of Argo data, and (d) related datasets such as SST. (Roemmich) More floats (600) are needed in the southern hemisphere; signal-to-noise ratio is good in tropics. 6

Argo Status: deployment plans Argo groups are making substantial efforts to plan their deployments and optimize the array coverage taking into account network density /age. Implementing an empty ocean ≠ maintaining a global array. All deployments are registered on line from a draft state to the final confirmation/notification. 7 GODAE => Maps of dispersion to help planning

Argo = logistical challenge 2000 2001 Total 116 295 2002 452 2003 672 2004 869 2005 1008 2006 933 2007 2008 747 848 600 Pacific Ocean 500 400 Atlantic Ocean 300 Indian Ocean 200 Arctic Ocean 100 0 2001 2002 2003 2004 2005 2006 2007 Mediterranean Sea 2008 2009 6344 units deployed ! Ship time is an issue: Sustained funding and cooperation critical 8

Argo = logistical challenge • Most floats are deployed by opportunistic use of commercial ships and RVs. • In some remote regions this is not sufficient. • U. S. /New Zealand collaboration has resulted in deployment of over 600 floats (plus drifters) using a cost effective vessel, RV Kaharoa (28 m length, 5 crew). • Future uncertain (lack of funding) New Zealand’s Minister of Research, Science and Technology studies float deployment. Photo by A. Blacklock, NIWA. Deployment positions of Argo floats from dedicated RV cruises. Kaharoa’s Argo-10 9 voyage to the South Indian Ocean (previous slide) is planned for late 2009.

Argo is the most internationally collaborative program in the history of oceanography USA JAPAN AUSTRALIA GERMANY 2% France 3% KOREA 3% UK 3% CANADA 5% INDIA NETHERLANDS 6% A dozen countries are sustaining the global network, another dozen takes care of regional gaps, and many other are supporting Argo. CHINA 56% 7% EU ARGENTINA CHILE NEW ZEALAND IRELAND 11% Euro Argo contribution growth needs to balance the potential USA contribution decrease. Argo makes a lot of efforts to foster participation by new countries BRAZIL NORWAY ECUADOR MAURITIUS SPAIN RUSSIAN FED. 10

Argo = 100 000+ profiles / year • #All Profiles 562230 • #DM Profiles 289774 • 90% optimal quality (but not for all applications) • 90 % reach the GTS/GDACs within 24 h 11

Argo: QC issues • Salinity Drift: bio-fooling and others reasons DMQC (dedicated working group) Comparison with CTD data, and nearby float data • Accuracy of temperature versus pressure – Applications of Argo data for climate change issues requires highest quality possible (heat content, steric sea level change). – Years required to detect small biases (comparison with CTD) – Free data => educate users • Recent problem on SBE CTD pressure: most of new deployments will be postponed … 12

Argo: QC issues • Salinity drift is estimated and adjusted (Owens and Wong, 2008). • An issue is small systematic errors in pressure. The effect of 1 -dbar time-varying pressure bias on temperature: . 009 o. C on steric height (0/2000): ~ 5 dyn mm. Systematic errors of this magnitude are difficult to detect and correct. Example: SBE 41 changeover from Amatek and Paine p-sensors to Druck in 2003. Global change research requires long time-series and careful standardization. Shipboard CTD data are critical for Argo. The dataset requires several years to achieve best quality. 13 Users need to understand the limitations!

Argo: QC issues, pressure sensor severe pb. • “micro leaks” in the Druck sensor used in SBE CTD, on all floats • Recent study (by UW & Scripps / USA) – Increase in the occurrence rate of floats showing negative surface pressure – Before 2007: 3% – 2007: 12% – 2008: 10 -30 % … • • most of new deployments will be postponed … Manufacturer doing tests but still in phase of understanding Equip some CTD with new pressure sensor (Paine) Return floats and repair CTDs • It takes more than a year for a float to exhibit such negative pressure offset • In general to have feedback on a float generation and impact of new features we need to wait years 14

Argo: QC issues • DM QC time consuming activity (resources required): 67% achieved 15

Argo is revolutionizing global oceanography. . . 2000 -2008 August XBT profiles (> 300 m, source: WOD) 1951 -2000 August hydrographic T/S stations (> 1000 m, source: WOD) 2004 -2008 August Argo T, S profiles . . . and its impact will be greatest in the southern hemisphere where there are large climate signals and there is little historical data. 16

Argo benefits • Operational use: requirements for long term 14 Operational Centres using Argo data • Research applications growing • Training Workshops Capacity building initiatives on data use • Argo has enormous potential value in education applications. • Google Ocean & Argo partnership network status (gateway) products (T, S, anomalies) stories by oceanographers climate change focus • • To be extended to other networks monitored by JCOMMOPS. GODAE, my. Ocean: prepare KML … 17

Capacity Building: cooperation with GODAE ? • Raising capacity in Argo national programs (3 DMQC, 1 float technology workshop, one-on-one visits and training) • Increasing Argo’s international user community (e. g. PI-GOOS Marine Data Workshop). All nations should benefit from global ocean observations. • Secondary and tertiary education (teacher training workshops). PI-GOOS Marine Data Workshop in Fiji to acquaint meteorological service and fisheries scientists with Argo and other ocean datasets. South Atlantic Argo training workshop on US Navy vessel HSV-2 Swift. SEREAD (Argo-sponsored) teacher training workshop in Samoa. 18

Argo Float Technology: float reliability improving 45% of floats deployed in 2004 are still active at age 4+ years. Deployment failure rate: 2. 5 % • Still room for improvement: all floats do not reach yet the 4 years lifetime • This target will likely be reached and exceeded 19 • This will help to fill gaps … without deploying more floats

Argo Float Technology: Telecom. slowly improving ~8% of deployments with Iridium (2007, 2008) – Argos 3 pilot projects started 20 More anticipated in 2009 (Australia)

Argo Float Technology: Cycles rather homogeneous 80% of the fleet is set up to drift at 1000 dbar as decided by the AST 70 % to profile at a depth > 1500 dbar 80 % to cycle on a 9 -11 days basis 21

Argo Float Technology: new designs, new sensors • New generation floats are longer-lived, smaller, and more capable. • Extended domains are being explored or considered: under ice, marginal seas, boundary currents, abyssal ocean. • New sensors are being developed and tested (oxygen, bio-optical, surface layer, …) • At present ~150 Argo floats carry dissolved oxygen sensors. • First floats are providing SST (at no cost). This is likely to be extended to all floats. 22 Left to right: SOLO/SOLO-II w/Iridium, ARVOR, PROVOR w/optical sensor, APEX w/SBE oxygen sensor.

Argo & GODAE / OSE • Feedback from data users on requirements (networks design, data issues, . . . ) – Argo Science Workshops , Ocean. Obs’ 09, (next ADMT in Toulouse. . . ) – 3°x 3° everywhere ? – How many T/S levels are assimilated by models ? (50 or 1000 needed? ) – Metadata required? • Routine feedback from operational centres on data quality: tools to be developed and maintained – – – DBCP/SOT operational system Argo being developed (e. g. Altimetry QC by CLS/Coriolis) JCOMMOPS acts as a centralized relay tool: data users – data producers UKMO maintain blacklists Integrated QC from data centres (GODAE Server) • JCOMMOPS can help to design, finalize and promote products • Continuous demonstration of the value of the Argo array – – Specific impact studies E. g. : What would give the models with half Argo array ? 23

Argo Objectives for the Argo Program in the coming years related to array performance are: • Achieve mean float lifetimes of 4 years or longer, needed to sustain the core Argo array with 800 floats deployed per year. • Deploy more floats in the southern hemisphere to achieve the array’s design requirements. • Extend instrument capabilities for profiling to 2000 m everywhere in the oceans. • Sustain funding (Argo is 20% underfunded) • GODAE/Argo: – What should be Argo‘s sampling plan for high latitudes or marginal seas ? – Should Argo sample the deep ocean ? – Should Argo be denser in all WBC regions ? – How the data distribution could be improved ? 24