All of the files in this database have been reviewed and subjected to quality control procedures, after which, data are called the “best basic version” (BBV). The BBV data are in scientific units, have been reviewed by the database manager, and are intended as the starting point for scientific analysis. The individual records are not flagged with data quality words, but inclusion in the database implies the measurements passed all the validation tests.
This database is intended to be useful without requiring detailed knowledge of the particular sensors, loggers, platforms, or field activities that produced the data. Users are encouraged to consult the applicable data reports if detailed information about the instrument configuration is not found in the attributes or metadata.
All measurements in the database were made over some duration, usually at one geographic location. Most of the measurements were made in the ocean or estuaries, but supporting land or buoy-based atmospheric measurements may also be included. The BBV is processed to the following standards:
- Data are calibrated and expressed in standard scientific units.
- Time base has been verified.
- Calibration parameters have been verified.
- Clearly erroneous data values have been removed and replaced with interpolations or pad values (typically “1e35”).
- Data recorded out-of-water or off-location have been removed.
- Appropriate and correct metadata are in the attributes.
- An ISO19115 metadata record for each file is generated by ncISO (the record template was reviewed prior to use).
- Datasets are netCDF files that conform to the CF or EPIC conventions.
- The data files are written by the instrument that collected them; no aggregation is done to combine all the measurements from a platform into one file.
- Data from multiple temporal segments made by the same instruments may be concatenated to make a single time series.
As data standards, conventions, and interoperability best practices evolve, the methods for serving the data may change in order to meet the general objective of describing and serving our time-series data in the most widely used, appropriate manner.
|