Delayed Mode Quality Control (DMQC)

Delayed Mode Quality Control (DMQC) is a detailed scientific assessment of Argo data that ensures observations are accurate, consistent, and ready for scientific use. It uses internationally agreed, parameter-specific methods developed by the Argo community.

Take a look at the Delayed Mode Quality Control of Argo Floats presentation

 

From real-time observations to science-ready data

Argo data undergo several stages of quality control before they become fully quality-assured, science-ready observations.

As soon as data are received from an Argo float, they undergo Real-Time Quality Control (RTQC) tests. These automated checks identify potential problems and assign quality flags to the observations. The resulting flags provide important information about the reliability and context of the data and are used as an essential input to the subsequent DMQC process.

An important next step is visual inspection of the observations. Expert review can identify unusual or unexpected features that may not be detected by automated real-time tests. Additional checks, including Objective Analysis and Min/Max tests performed by Coriolis, can also highlight potential issues. Alerts are communicated to the relevant Data Assembly Centres (DACs) for investigation, review and, where appropriate, flagging of the affected data.

Scientific assessment through DMQC

DMQC builds on these earlier quality-control steps and provides a more detailed scientific assessment using internationally agreed, parameter-specific methods.

DMQC is applied across a range of Argo observations, including:

  • Core parameters: temperature, salinity and pressure
  • Deep Argo: observations below 2,000 dbar, down to 6, 000 dbar
  • Biogeochemical (BGC) parameters: oxygen, chlorophyll-a, pH, nitrate, suspended particles and irradiance
  • Trajectory data

During DMQC, float observations are carefully compared with independent, high-quality reference datasets from the same region and a similar time period. Depending on the parameter, these may include ship-based CTD observations, historical high-quality Argo profiles, in-situ bottle measurements, climatologies and reanalysis products.

DMQC combines data analysis with expert scientific judgement. Understanding the characteristics of a particular ocean region and parameter is essential for identifying anomalies, assessing measurement uncertainty and determining the quality of the observations.

For Core parameters, DMQC is normally performed approximately one year after deployment and repeated annually for each parameter. For BGC observations, quality-controlled data can become available much sooner, typically within 5–6 cycles after data transmission.

Where possible, identified inconsistencies or systematic errors are corrected. Data that cannot be reliably corrected are appropriately flagged, ensuring that the final dataset provides the highest possible quality for scientific applications.

Each Argo parameter has a dedicated DMQC methodology developed and agreed by the international Argo community, supported by open-source software and tools.

Argo Data Management Documentation | Euro-Argo open-source software | Argo DMQC software

What Argo data looks like 

Argo profiles contain several versions of each parameter to record the progression from the original measurement through quality control and scientific adjustment (Argo user's Manual ).

For each parameter (<PARAM>), associated variables include:

  • <PARAM>_QC — quality-control flag for the original data, e.g. PRES_QC
  • <PARAM>_ADJUSTED — adjusted parameter value, e.g. PRES_ADJUSTED
  • <PARAM>_ADJUSTED_QC — quality-control flag for the adjusted data, e.g. PRES_ADJUSTED_QC
  • <PARAM>_ADJUSTED_ERROR — estimated error associated with the adjusted value, e.g. PRES_ADJUSTED_ERROR

The adjusted data provide an important link between real-time and delayed-mode processing. Where scientific calibration or other adjustments can be applied during real-time processing, the resulting values are recorded in the profile as an additional <PARAM>_ADJUSTED variable. These adjustments can be particularly important for BGC parameters, where calibration and correction are often required before the observations can be used reliably for scientific applications.

The adjusted values and their associated quality flags and estimated errors are subsequently reviewed and refined through the DMQC process as more reference data and scientific information become available.

At the Argo Global Data Assembly Centres (GDACs), profiles containing delayed-mode data can be identified by the “D” included before the WMO number in the filename, for example D5900400_001.nc or BD5904179_001.nc.

BODC DMQC capabilities

The BODC Argo team plays an active role in advancing Argo data management and quality control at both European and international levels. Our expertise in DMQC supports the production of consistent, high-quality, science-ready Argo data and contributes to the continued development of the global Argo observing programme.

At BODC, we currently provide expert DMQC for Core parameters, oxygen, pH, nitrate and irradiance, delivering quality-controlled data to the Argo Global Data Assembly Centres (GDACs). We are also actively implementing and evaluating new and improved quality-control methods, helping to strengthen consistency and best practice across the international Argo network.

Supporting the international Argo community

The BODC Argo DMQC team contributes specialist expertise to European and international Argo data-management activities. Through our involvement in European and international projects, we contribute to the development of new DMQC methods and procedures, organise and deliver training workshops and technical sessions, contribute to expert discussions and working groups, and provide advice and support to other Argo programmes around the world.

Our work helps to promote the harmonisation of data-management practices across national programmes and supports the wider Argo community in applying robust, consistent and scientifically validated quality-control approaches.

BODC also contributes to the design and development of new data-management procedures, methods and software. We develop and share open-source tools based on community frameworks, supporting transparent development, community collaboration and sustainable governance of Argo data-management software. This helps the Argo community adopt improved practices while ensuring the long-term delivery of high-quality, science-ready observations.