1.4 Data levels and Best Practice Recommendations

You are here:

Data Levels

Modified from GLEON Best Practice Recommendations for Data Management and ODM2 core: Processing Levels

Level 0 – Raw data: unprocessed data and data products that have not undergone quality control. Depending on the data type and data transmission system, raw data may be available within seconds or minutes after real-time. Examples include real-time precipitation, streamflow, and water quality measurements

Level 0.1 – First pass QC: A first quality control pass has been performed to remove out of range and obviously erroneous values. These values are deleted from the record. E.g: Online Environment Canada stream-flow data, laboratory data

Level 1Quality Controlled Data: Data that have passed quality assurance procedures such as Level 0.1 and have been further quality controlled by data provider before being submitted to CanWIN (e.g. Idronaut data with only downwelling (upwelling data removed) data included.

Level 1.5Advanced Quality Controlled Data: Data have undergone complete data provenance (i.e. standardized) in CanWIN. Metadata includes links to protocols and methods, sample collection details, incorporates CanWIN’s or another standardized vocabulary, and has analytical units standardized. Note: Process still under development in CanWIN (as of May 13, 2020).

Level 2Derived Products: Derived products require scientific and technical interpretation and can include multiple data types. E.g.: watershed average stream runoff derived from stream-flow gauges using an interpolation procedure.

Level 3 – Interpreted Products: These products require researcher (PI) driven analysis and interpretation and/or model-based interpretation using other data and/or strong prior assumptions. E.g.: watershed average stream runoff and flow using streamflow gauges and radarsat imagery

Level 4Knowledge Products: These products require researcher (PI) driven scientific interpretation and multidisciplinary data integration and include model-based interpretation using other data and/or strong prior assumptions. E.g.: watershed average nutrient runoff concentrations derived from the combination of stream-flow gauges and nutrient values.

Data Processing Best Management Practices

  1. Get Level 0 (raw) data (e.g. downloaded instrument data ), Level 0.1 OR Level 1 data (e.g. laboratory results).
  2. Decide which level of data you will be publicly sharing and which, if any,  you will be archiving.
  3. If you are archiving data, and need assistance, discuss options with the CanWIN data curator. Depending on data size, options can include CanWIN’s GitLab or Alfresco servers or offsite archival storage.
  4. If you produce additional results from the original dataset you can also add this to CanWIN under the same project as Levels 2,3 or 4 products. This helps showcase your work

 

CanWIN DataMgmt