QARTOD

1
QARTOD

• Quality Assurance of Real-Time Ocean Data

Matt Howard Texas AM University NFRA Rep. to
DMAC-ST
2
QARTOD Workshops

• QARTOD I December 3-5 2003
• QARTOD II February 28 - March 2, 2005

3
QARTOD - I

• Sponsored by NDBC NWS
• Held 3-5 December 2003 in Stennis, MS
• 80 Participants
• Most participants already running observing
  systems
• Primary Tasks for Breakout Groups
• Develop minimum standards for calibration
• QA/QC Methods
• Metadata

4
Introduction
Introduction The dawn of the Integrated Ocean
Observing System (IOOS) era brings with it many
challenges related to the distribution and
description of real-time ocean data. One of the
primary challenges facing the ocean community
will be the fast and accurate assessment of the
quality of the data streaming from the IOOS
measurement systems. Operational data merging and
assimilation from multiple data sources will be
essential to the ability to adequately describe
and predict the physical, chemical, and
biological state of the coastal ocean. These
activities demand a simple, trustworthy, and
consistent quality description for every
observation distributed as part of the IOOS
system.
- QARTOD I Final Report
5
Agenda
1 Day of show and tell 1 Day Breakout
groups Calibration Issues QA/QC
Methods Metadata Requirements 0.5 Day Report
Out
6
Presenters

• NOS CO-OPS NWLOP, PORTS, OSTEP, HF-RADAR
• NOAA/AOML
• USF/COMPS
• LSU/CSI/WAVCIS
• FSU/COAPS
• NOAA/NESDIS/GTSPP
• NOAA/PMEL/TAO/TRITON
• UAlaska/GINA
• Oregon HealthScienceU

• UCSDC/SCRIPPS/CDIP
• MarylandDNR
• MBARI/Oasis
• NOAA/NESDIS/NCDC
• UM/RSMAS/Shipboard/VOS/MODIS
• NOAA/NDBC
• NOAA/CSC
• USACE/FRFM
• NRL/SSH
• NOAA/NCDDC

7
Parameters Considered by QARTOD-I Teams(GOOS
Report 125)

• Water level
• Temperature
• Salinity
• Oxygen
• Nutrients
• Chlorophyll
• CDOM
• Others with similar scalar character

8
Calibration

• Calibrate your sensors
• At your own institution
• At the factory
• At outside facility
• Standards NIST or other
• Higher standards for low variability environments

9
RT Quality-Control Tests

• Transmission validity (checksum, timestamp)
• Gross error, plausibility, sanity checks
• Range exceeds (climatological, local) values
• Sensor comparisons (sensor, neighbor, model)
• Rate of change (temporal-spatial)
• Parameter-parameter (TS, Nutrient, density)
• Pass-fail results carried along with data

10
How to come to grips with all possible standards
for metadata and metadata storage schemes

• Flat files (ASCII)
• Flat files w/imbedded metadata
• Defined binary
• NetCDF, BUFR, HDF
• Relational Databases
• OO Databases
• ESRI Other Vendor formats

• FGDC, DIF, ISO
• Dublin Core, MIF
• OGC, OAI
• DODS/OPeNDAP
• XML, EML, ESML
• MarineML, GML
• ADN, IMS, Z39.50

11
Final Recommendations

• Calibrate sensors quantify accuracy and error
• All data should be subjected to automated QC with
  appropriate levels of manual checks.
• Quality descriptors must accompany all data
• Metadata (calibrations, tests, flags)
• Aggregate (overall pass/fail) and individual
  tests
• Flags
• -9 missing data
• 0 quality not evaluated
• 1 bad
• 2 questionable/suspect
• 3 good

12
Standards?

• The Participants felt strongly that
• Data provider is free to decide description,
  delivery, and testing with minimum of mandatory
  requirements.
• No particular storage format (ascii,netcdf,db),
  data delivery method (OPeNDAP2), or type of
  metadata standard were adopted.
• (assumed translational gateways would handle)

13
QARTOD II

• Hosted by NDBC
• February 28 - 02 March 2005
• Nauticus Auditorium Hotel
• Norfolk, VA
• Waves and Currents

14
Agenda

• 1.5 days Intro Show and Tell
• 2.5 hours breakout groups
• 0.5 days Report and Synthesis, and planning of
  next QARTOD (oxygen?)

15
WG Tasks

• Charge to the Breakout Groups
• For sensor specific data, can the group
• Identify/standardize real-time quality tests?
• Provide quality descriptor flags based upon test
  results
• Identify metadata requirements?
• Identify calibration requirements?
• Identify common data formats?
• Identify any additional requirements associated
  with DMAC compatibility?
• Describe the roadmap and current roadblocks to
  developing an operational QA/QC capability?

16
Presenters Breakout Groups

• NOAA/NOS
• NOAA/NDBC/ADCP
• ATON/ADCP
• HF RADAR
• Data Integration
• ADCP-based Waves
• DMAC/MMI/Metadata
• TWiki

• In-situ (and remote) waves
• Remote Currents
• In-situ Currents

17
In-Situ CurrentsBreakout Group Report Out
QARTOD IIFebruary 28 March 2, 2005
18
Technology Selection

• 25 participants
• Three technology groups
• Current meters
• ADCPs technology selected by group
• Drifters
• In general, did not want to be vendor specific

19
(No Transcript)
20
Real-Time Quality Descriptor Flags

• Each data record goes out with quality descriptor
  flags
• Group Discussion
• Two principle customer groups
• Customer interested in observation only (e.g.,
  maritime community)
• Ensemble flag linked to release of data
• Customer interested in full record (e.g.,
  academia, oil and gas industry)
• Flags for each hard parameter
• Ensemble flag linked to release of data
• Flags for soft flags, if affordable
• Two data sets
• Real-time observations (only Level 2 and 3 data)
• Archived observations (all data)

21
Real-Time Quality Descriptor Flags

• Categories agree with QARTOD I recommendations
• -9 missing value
• 0 quality not evaluated
• 1 bad
• 2questionable/suspect
• 3good

22
Real-Time Metadata Descriptors

• Group Discussion
• Developed a list of recommended fields to send
  with the real-time observation
• Instrument-level description
• Metadata should be in the header of the record

23
Metadata to be Delivered in Real-Time

• Latitude (with a designator fixed or moving)
• Longitude (with a designator fixed or moving)
• Horizontal datum / reference
• Depth of water
• Depth of instrument
• Vertical datum / reference
• Time (with a designator for time zone, e.g., UTC)
• Data collected at beginning, middle, or end of
  sample
• Compass reference
• Serial number
• Units for data reporting
• Orientation

24
Metadata to be Delivered if Affordable

• Sensor type
• Deployment date
• Calibration procedures
• Date of calibration
• System frequency
• Bin size
• Number of bins
• Sampling interval
• Bad beam indicator
• Average interval
• Pings per ensemble
• Compass update rate
• Quality checks descriptions and thresholds
• Blanking distance
• Coordinate system for velocity measurements
• Joyce parameters
• Platform description, including fixed or moving

25
Real-Time Calibration Flags Descriptors

• Place reference to date of calibration in metadata

26
Common Data Formats

• Group Discussion
• Many data formats represented in group
• ASCII is universally understood
• Future holds sensor sending data in XML
• Moving towards automation and interoperability
• Recommendations
• Community should strive to be DMAC compliant
• Recommended format NetCDF, along with data
  dictionary and convention

27
Next Steps and Roadblocks

• Key Next Steps
• Define thresholds for quality control tests
• Develop open source environment for collaboration
  (e.g., TWiki)
• Define specifications for metadata descriptors
• Roadblocks
• Disparity in requirements for different user
  groups could lead to two or more datasets

28
Website
http//nautilus.baruch.sc.edu/twiki/bin/view/Main/
WebHome