Three types of netcdf file are available here:
A given cruise may have up to three sets of files, one corresponding to each of the three possible instruments and ping types, each of which is processed independently.
All files are written in Netcdf version 3 format using the netcdf4-python package by Jeff Whitaker.
These files contain the primary analysis variables: estimated velocity of the water over the ground (u, v), percent good pings in the ensemble (pg), averaged raw signal strength as reported by the instrument on a 0-255 scale (amp), editing flags (pflag) in which any non-zero value may be taken as "bad", and a few time series. These are heading, transducer temperature (tr_temp), and the ship velocity over the ground (uship, vship).
All variables are obtained from the final processed data set with no additional gridding. During processing, pings are averaged in ensembles of a few minutes duration. For historical reasons, the time and position of each such average as given here is the time of the end of each ensemble.
Depth, provided as a 2-dimensional array matching the u and v arrays, is nominal; that is, it has not been adjusted for actual sound speed profiles. The depth of a given cell is usually uniform within a data set, but will vary if the instrument configuration was changed during the cruise. For example, the blanking interval or the pulse length might have been changed.
CODAS processing is built around a data structure that was developed many years ago for this purpose. Data are not stored in the form in which a user typically uses them, but extraction programs are provided. Some of the variables that are stored are used during the processing but are of little later interest; others may be of very occasional interest. The purpose of the long form ADCP netcdf file is to provide access to nearly everything stored in a CODAS ADCP database, without pre-judging its usefulness. That is, the netcdf file is very nearly the CODAS database translated into the standard netcdf format for ease of access without needing the arcane CODAS software.
Some CODAS databases contain either or both of two data types that are not included in these netcdf files: RAW_SPECTRAL_WIDTH and RAW_AMP_SOUND_SCAT. The structure and content of these data types can vary from one database to another, and they are in general not very useful.
There have been many attempts to model barotropic tides; here we provide velocity predictions along cruise tracks, using tidal coefficients from TPXO7.2. Predictions from several other models are also provided where a substantial part of the cruise track is within the model domain. These include the Atlantic, Pacific, and Indian Ocean OTIS regional tidal solutions, the Ross Sea model based on inversion with vessel-mounted ADCP measurements, and the Antarctic regional model CATS2008b (link points to a closely related predecessor; it will be updated when possible.) The zonal and meridional current component predictions were made from the coefficients using pytide.
Bear in mind that these are model-based predictions of the barotropic tide only, and may not be suitable for de-tiding actual ADCP observations. They are provided for exploratory purposes.