Quality Control Standards and Data Levels
There are four data levels corresponding to different degrees of quality control measures.
Level 0 Data
Level 0 data is the raw data that is downloaded from the dataloggers. It is permanently archived in it's original format, and as such is mostly uncalibrated, not converted to meaningful units, and unfiltered. It is 10-minute data that is downloaded weekly, and contains both 10-minute means and standard deviations, except for totalizing measurements such as tipping bucket gauges. Interpretation of this data requires quite a bit of ancillary information, such as calibration coefficients, etc. As such, we do not widely distribute the raw data, and consider it to be of proprietary status.
Level 1 Data
Level 1 data is the result of our first examination of the raw data. This examination is usually the same day that the datalogger storage module is retrieved from the Subnivean Lab. We pass the raw data through a program that applies the necessary conversions, calibration coefficients, bad value filters, and so forth in order to bring the raw data into meaningful numbers. At this level, filters are very conservative, and values are replaced with missing value indicators only when it is absolutely clear that they are out of range.
Once the data are converted, a set of preliminary turbulent fluxes are computed.
The fluxes are corrected for atmospheric stability using stability functions based on the Richardson number (Ri) developed by ETH in Switzerland.
After completion of the program, an analyst begins interpreting the data. All the data variables are plotted in sequence, and the analyst makes note of anything out of the ordinary that may be noticed. The Level 1 data are not further modified, but any notes from the analyst are recorded in a metadata file that accompanies the data distribution.
Level 2 Data
Level 2 data is the first descendant of Level 1 data. The 10-minute Level 1 data is averaged over hourly and daily time periods, and it is the hourly and daily means of the Level 1 data that are available as Level 2.
The full aerodynamic profile dataset is not provided in Level 2; only the 2.0 m Height air temperature, relative humidity, and wind speed is provided.
Level 2 data is the first level of "smoothed" data, however it is EXTREMELY IMPORTANT that the user is aware that THESE ARE PRELIMINARY ONLY!!! They are acceptable for rough comparisons to other sites, etc., but should not be considered to be especially accurate for purposes of energy budget analysis, etc.
Level 3 Data
Level 3 data are segments of the database that have undergone rigorous analysis by one or more of the project scientists.
The data may have been manipulated in any variety of ways, but are the "finished product" that is discussed in publications pertaining to the data. As such, the content of these databases segments may vary considerably, depending on the nature of the investigation. For this reason, the project scientist performing the analysis determines what to include and disclude from these data files, and the files are made available to the public as is (i.e. you get the whole thing, no options!). In most cases, reference to relevant publications is included with the data instead of the full metadata files that are bundled with the Level 1 and Level 2 data.
If any problems arise please e-mail
Todd Ackerman
(todda@culter.colorado.edu)