Quality Assurance and Quality Control QAQC Jim Jones, NWS

1
Quality Assurance andQuality Control
(QA/QC)Jim Jones, NWS Cleve Cowles, MMS

• AOOS DMACData Management Workshop
• January 27, 2005

2
AOOS DMAC Data Management Workshop

• Quality Assurance -
• Common thoughts
• QA Fix the Defects
• QA Prevent the Defects

3
AOOS DMAC Data Management Workshop

• Quality Assurance - cont.
• Is a systematic pattern of actions that assures
  all usable data is
• Stable
• Performs as described according to specifications
• Meets customer/user needs and expectations

4
AOOS DMAC Data Management Workshop

• Quality Assurance Framework
• Implement/review/change processes and procedures
• Train and educate on general quality issues
• Notify upper management of systematic
  deficiencies
• Determine, Document, and Promote use of
  established standard methods and procedures

5
AOOS DMAC Data Management Workshop

• Quality Control -
• QC tends to find the defects in a given product
  by a monitoring, auditing, and assessing process

6
AOOS DMAC Data Management Workshop

• Quality Control is
• inherently reactive in nature..
• provides an element of acceptance or rejection
• fits under the idea of QA as a whole

7
AOOS DMAC Data Management Workshop

• NWS QC examples
• Real-time monitoring of various data (ASOS, AWOS,
  Buoys, C-Man, Ship Reports, etc..)
• Near Real-time validation of data and remote
  sensing equipment (Satellite output, NEXRAD,
  Storm spotters)
• Near real-time QC of non-operational
  climatological data

8
AOOS DMAC Data Management Workshop

• Agency Perspective
• QA/QC can be achieved by being sensitive to the
  links of decision making and being responsible to
  prime users.
• If data are used for products that are used
  quickly in peer-reviewed processes and
  applications, then those peer reviews can provide
  the QAQC.
• Accountability is QAQC insurance.
• Decision makers need to be able to determine the
  confidence level of products.
• QAQC is dynamic. What once was an error, when
  fixed, can reveal new errors, and/or free
  resources for identifying and fixing new problems.

9
AOOS DMAC Data Management Workshop

• Agency Perspective (continued)
• Anticipate the worst case scenario early on.
• Protect the organization.
• Require disclaimers and qualifiers for use.
• Inspect products. For example, when projecting
  information on maps, symbols and graticules can
  obscure information.
• Certify and authenticate products.
• Track and acknowledge subordinate data sources in
  product generation and delivery.