The DBCP – GHRSST Pilot Project: an update

1. The DBCP – GHRSST Pilot Project: an update How to make a silk purse out of a sow’s ear (How to make good wine out of bad wine) David Meldrum d.meldrum@unesco.org dtm@sams.ac.uk

2. The DBCP – GHRSST Pilot Project: an update • What is the DBCP? • What does a drifter look like and what does it aim to do? • What is GHRSST? • How can we make drifters perform better for GHRSST? • How can we persuade drifter operators to purchase better drifters? • How do we ensure continuity?

3. What is the Data Buoy Co-operation Panel (DBCP)? • Formed by IOC and WMO in 1985 in response to a crisis: • Ocean Weather Ships had been replaced by drifters • Drifter data quality, quantity and timeliness were very poor • No one understood why • Resolutions etc passed – absolutely no effect on the data! • Solution: full time technical coordinator appointed in 1987: • Based at CLS Argos in Toulouse within JCOMMOPS office • Access to all parts of the data chain and delayed mode QC

4. DBCP – first two technical coordinators

5. What is the Data Buoy Co-operation Panel? • Formed by IOC and WMO in 1985 in response to a crisis: • Ocean Weather Ships had been replaced by drifters • Drifter data quality, quantity and timeliness were very poor • No one understood why • Oceanographers did not release their buoy data • Resolutions etc passed – absolutely no effect on the data! • Solution: full time technical coordinator appointed in 1987: • Based at CLS Argos in Toulouse within JCOMMOPS office • Access to all parts of the data chain and delayed mode QC • Initial problems solved by 2000 • Develop new initiatives: • Outreach to new observing systems under JCOMMOPS umbrella • Capacity development • Pilot Projects

6. DBCP from 1985 to the present day

7. What is a drifter? • ‘Standard’ design • Designed to meet needs of both ocean and met communities • Drogue centred at 15m depth • Pressure ± 1hPa; SST ± 0.2K

8. Some results

9. Some results After Peter Niiler, SIO

10. Drogue presence detection • Drogue adds large inertial mass • Small residual buoyancy • Buoy submerged for much of the time in rough seas • Depth of submersion may be >> 1m • Not a good surface follower • Submersion sensor used to detect presence/absence of drogue Submersion % Drogue loss Courtesy M Pazos, AOML

11. What is GHRSST? • Group for High Resolution Sea Surface Temperature • Formed by GODAE in 2002 • Aims to address an emerging need for accurate HRSST • to constrain the modelled upper ocean circulation thermal structure • for exchange of energy between the ocean and atmosphere • Generate HRSST products from satellite radiometers • Satellite SST validated against in situ observations using ‘matchups’ • Satellite and in situ must be close in space and time • Need to be cloud free • Night-time matchups preferred • Drifter SST matchups found to be best! • Mean discrepancy < 0.1C

12. Monthly global area avg ARC - buoy diff and monthly number of ARC/buoy matchup pairs Distribution of ARC - buoy difference (2002-8) ARC: D3 SST0.2m From early test release Matchup: within 0.1° grid cell, ≤ 3 hrs Smoothed 1° 2002-8 avg ARC - buoy diff (°C) 1° 2002-8 zonal mean ARC - buoy diff.

13. Regional discrepancies between drifter and satellite SST

14. Requirements agreed with GHRSST

15. Issues discussed at DBCP in 2010 • Drifter SST clearly critical to satellite SST retrievals! • GHRSST have expanded a DBCP draft proposal to propose a new joint DBCP-GHRSST pilot project • At least matched funding from GHRSST • 50 drifters upgraded to HRSST reporting • Upgrade cost ~$1k • E-SURFMAR ahead of the game – deploying HRSST-1 drifters • Will require reporting in BUFR to achieve required resolution • Target area to be decided: • General distribution? • Specific area, e.g. Barents Sea? • ToRs and workplan agreed • End date 2014 • PP-HRSST approved

16. Immediate plans (2010) • Need to identify areas that will provide large number of matchups • These areas to be of interest to buoy operators as they will pay most of the cost • PP-HRSST funds being used to help Met Office to purchase high accuracy HRSST-2 drifters • Need to get feedback from GHRSST asap • Problems with receiving BUFR • Need to get funding from GHRSST

17. Cloud probability and target areas

18. Where are we now? • 220 HRSST-1 drifters deployed since end 2010 • Made by MetOcean, Canada • Deployed by ESURFMAR, Meteo France, UKMO, Env Canada • Mostly N Atlantic – poor matchup statistics because of cloud • Iridium comms + GPS • Hourly reports on GTS in BUFR with resolution of 0.01C • Absolute accuracy ~ 0.1C • Now is default MetOcean design • 40 HRSST-2 drifters ordered, 3 deployed • Made by MetOcean, Canada • Some upgraded with DBCP funds • Better design for pre- and post-calibration • Absolute accuracy better than 0.05C • Initially cost $1k more • Still no money from satellite side!

19. Progress on joint funding • Make case at GHRSST and ESA meetings • Message now clearly understood • In situ networks funded in support of NWP, operational oceanography, climate and research: NOT satellite cal/val • In situ community ready to engage in dialogue and collaborate • Needs sustained funding from satellite community • GHRSST now seeking sustained funding for in situ HRSST • Drifter HRSST essential • Argo HRSST will be retained as independent verification • Need to fund drifters in ‘high-value’ areas • ESA Sentinel-3 mission team ready to consider proposal for pre- and post-launch sustained campaign • Draft proposal in prep: target 1 M€ for pre-launch campaign

21. Drogue presence detection • Drogue adds large inertial mass • Small residual buoyancy • Buoy submerged for much of the time in rough seas • Depth of submersion may be >> 1m • Not a good surface follower • Submersion sensor used to detect presence/absence of drogue Submersion % Drogue loss Courtesy M Pazos, AOML

22. The good news for GHRSST • Submersion occurs in higher sea states => well mixed anyway

23. The good news for GHRSST– drogues fall off! • Submersion occurs in higher sea states => well mixed anyway • Drogues eventually fall off • Initially, with drogue on, match-ups may be worse than later on • Recent data suggests that this often happens within 100 days • Drogue retention has got worse over recent years! • Study by M H Rio: derivation of Ekman parameters as fn of time βθ

24. The good news for GHRSST • Submersion occurs in higher sea states => well mixed anyway • Drogues eventually fall off • Initially, with drogue on, match-ups may be worse than later on • Recent data suggests that this often happens within 100 days • Drogue retention has got worse over recent years! • Study by M H Rio: derivation of Ekman parameters as fn of time βθ Red line: first 100 days of data only!

25. The MetOcean HRSST-2 drifter • Demountable SST sensor to aid pre- and post-calibration • High stability thermistor calibrated to better than 0.05K • Cal coeffs embedded in sensor module • Sensor module sends digital SST to buoy controller • Traceable to national standards

26. Early HRSST-2 deployments reveal calibration issue

27. Early HRSST-2 deployments reveal calibration issue

28. Early HRSST-2 deployments reveal calibration issue

29. Summary • Need to identify areas that will provide large number of matchups in shortest possible time • These areas to be of interest to existing buoy operators as they will pay most of the cost • HRSST drifters report ONLY in BUFR • Presence of drogue may make things worse • Need to get feedback from GHRSST asap • If worthwhile, need to win follow-on funding through joint GHRSST/DBCP proposals for: • Further deployments in high value areas • Studies of the near surface temperature profile in the open ocean