Building a Global Ocean Profile Database

GSOP Quality Control WorkshopJune 12, 2013

Outline

• The size and shape of a global ocean profile database

• Data Archeology and Rescue

• Data flow into a global ocean profile database

3

World Ocean Database (WOD) A quality controlled collection of ocean profiles,

plankton tow , and ship-based surface measurements from 1772-present.

4

Variables in WOD1) temperature2) salinity3) oxygen4) phosphate5) nitrate6) nitrate + nitrite6) silicate5) chlorophyll6) pH7) alkalinity8) pCO2

9) TCO2

10) Plankton11) CFCs 11, 12, and 11312) Tritium13) Helium (noble gas)14) ΔHe-3 (isotope)15) ΔC-13 (isotope)16) ΔC-14 (isotope)17) Argon (noble gas)18) Neon (noble gas)19) O-18 (isotope)20) Beam Attenuation Coefficient (transmissivity)

Instrument types in WOD

1) Station Data (bottles, thermometers)2) CTD3) XBT4) MBT5) Towed CTD6) Profiling Floats7) Drifting buoy (thermistor chains)8) Moored buoy (e.g., TAO, PIRATA TRITON)9) Autonomous Pinniped (instrumented

elephant seals)10) Gliders

Blue: Quality control beyond automatic checks

Growth of the World Ocean Database

51970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 20140

2,000,000

4,000,000

6,000,000

8,000,000

10,000,000

12,000,000

14,000,000

0.4

1.31.5

2.1

2.6

3.5

# of Temperature profiles

# of Salinity profiles

Year

# of

Pro

files

(mill

ions

)

WOD13 12.9

5.4

WOD by Instrument Type

6

1934 : Nansen Cast

Temperature Data During Peek of DifferentObserving Systems

Red=Nansen Cast /CTD[1890s/1964]Light Blue=MBT [1939]Dark Blue=XBT [1967]Green=Argo float [2001]Orange=Tropical buoy [1984]

7

1934 : Nansen Cast

1960 : MBT

Temperature Data During Peak of DifferentObserving Systems

Red=Nansen Cast /CTD[1890s/1964]Light Blue=MBT [1939]Dark Blue=XBT [1967]Green=Argo float [2001]Orange=Tropical buoy [1984]

8

1934 : Nansen Cast

1960 : MBT

1985 : XBT

Temperature Data During Peak of DifferentObserving Systems

Red=Nansen Cast /CTD[1890s/1964]Light Blue=MBT [1939]Dark Blue=XBT [1967]Green=Argo float [2001]Orange=Tropical buoy [1984]

9

1934 : Nansen Cast

1960 : MBT

1985 : XBT

2009 :Argo

Temperature Data During Peak of DifferentObserving Systems

Red=Nansen Cast /CTD[1890s/1964]Light Blue=MBT [1939]Dark Blue=XBT [1967]Green=Argo float [2001]Orange=Tropical buoy [1984]

10

11

Measurements vs. Depth WOD05Measurements (X105) Measurements (X104)

SilicateChlorophyll

12

Measurements vs. Depth WOD13Measurements (X105) Measurements (X104)

Dep

th (m

)

Dep

th (m

)

13

World Ocean Database• Contains contributions from national data centers, universities, special projects, fisheries, navies, government agencies, individual

scientists, merchant ships etc.

• Data found on internet, index cards, glass slides, outdated computer media, published cruise reports, etc.

• Original data must be accessible in the NODC archive/permission must be obtained for all data

14

Data from universities, fisheries

Left: Bottle data from Hokkaido University (Japan)(1962-1994)

Right: Temperature Data from US National Marine Fisheries (1964-1997)

15

Data from international projects

Left: GEOSECS bottle and CTD data (1972-1979)

Right: WOCE CTD data (1990-1998)

16

Data Archeology and Rescue

Cruise Reports found:

Red/Yellow: Library of University of Tromso, Norway

Blue: Public Library, St. Petersburg, Russia

Green: Public Library, New York City, USA (in part)

Are there still data to be rescued?

In a letter dated 11 May, 1990 from Lawrence Hall (Director of Ocean Programs, Sippican Inc.) to Greg Withee (director US NODC), it was noted that Sippican had manufactured > 4 million XBTs to date.

In the World Ocean Database, there are 1.3 million XBT profiles which are from drops prior to the end of 1990.

Are there still data to be rescued?

WOD09: 7,955 bottle/CTD casts in the Sea of Okhotsk

WOD13: 14,264 bottle/CTD casts in the Sea of Okhotsk

6,309 casts added, mostly from the Russian Far East Institute

Years for which data were added: 1907-1990

Gap 1930-1940

Gap 1960-1986

1939 NODC database

1939 observations

Figure 1

Data in the Barent Sea: Available in WOD vs. known observations [Figures courtesy of Igor Smolyar.]

Source: Proceedings of the Polar Institute of Fisheries and Oceanography from the 1950s-1960s.

Are there still data to be rescued?

20

Argo/CoriolisGTSPP

Delayed ModeData

NODCPermanent

Archive[non-uniform

Format]

Near Real-Time Availability

World Ocean DatabaseAdditional QC

Uniform format

BuoysPMEL/JAM

Data quality inform

ation

Not AlwaysPublicly Accessible

NODC Information Stream (Profile Data)Quarterly Updates of WOD

Argo data 2012: 132,244 profiles[Coriolis GDAC]

Tropical Moored Buoy ArraysTAO/TRITON – PIRATA – RAMA33,012 daily means [PMEL]

Near-real time data from GTSPP: 2012Green – XBT (14,775 casts)Red – CTD (12,549 casts)Blue - Pinniped (21,567 dives)Turquoise – Glider (7,430 half cycles)

Red: CCHDO – 164 CTDsGreen: ICES – 7,254 CTD/bottleBlack: TAO CTD – 59 CTDsOrange: SISMER (France) – 194 CTDsBlue [Atlantic/Medit]: AOML – 386 XBTs

Blue [Pacific/Indian]: CSIRO – 162 CTDsMagenta: ISDM (Canada) – 181 CTDsTurquoise: CalCOFI – 75 CTDsGrey: WHOI Ice Buoy – 6,202 cycles

Delayed-mode data for 2012 to date: 14,677 profiles

25

Examples of Delayed-mode data Updates to WOD

World Ocean Database Quarterly

Updates

NOAA

Northeast Fisheries

CCHDOWHOI Ice

Tethered Buoys

ICES

CalCOFI

Japan Ocean Data Center

CSIROLine W

INIDEP (Argentina)

TAO maintenance cruises

Blue – QuarterlyOrange- YearlyGreen - Irregular

26

Example of already aggregated data passed to WOD

CLIVAR and Carbon Hydrographic Data Office

(CCHDO) GO-SHIP

WOCE

BATS

HOT

CDIAC

CARINA

PACIFICAOceanSITES

Other

CLIVAR

27

Science andData Community

Convert Data to CommonFormat/Initial Quality Control

Calculate ClimatologiesSecondary Quality Control

Release DatabaseWith Quality Control Flags

Scientific ResearchPost-release Quality ControlMonthly database updates

OCL/NODCMAKE DATA AVAILABLE

Publish Results

FEEDBACK LOOP

Data Dissemination:Yearly by instrumentGeographically by instrumentWODselectAll updated quarterly

Format:Native WOD format(compact ASCII) with conversionroutines

netCDF -> feeds intoNODC Geoportal/THREDDS server

Working now on subset dissemination through Geoportal

Relevant to Workshop

• Always more historical/recent data – are we setting up a framework for continued quality control or working with a static historical database?

• If continuous, at which stage of the data flow do we insert ourselves and in what manner?

• What data sets are we dealing with – how do we handle each data set (pinniped? moored buoy?)

• Is the WOD a proper vessel for this project, or is a different system necessary?

• How will data be disseminated?