Nortek instruments use the speed of sound at the depth of the transducer to scale the Doppler frequency shift to velocity. The temperature and salinity determines the speed of sound. This is not calculated internally by the instrument, the speed of sound to use is provided by the software, either as a user specified value ("Fixed" option) or as a table of values for different temperatures ("Measured" option). The table is calculated by the software using the user specified salinity. When the "Measured" option is checked, the instrument looks up the speed of sound to use in the table using the measured temperature.
If you need to correct the speed of sound used, as a post processing step, use this formula:
V_new = V_old * (c_new/c_old)
This correction is not available as an option in any of our processing softwares, but almost any numerical analysis package should let you multiply the velocity data by a scaling factor.
If you use manual mode and you enter the wrong salinity, there will be a small error in the velocity, but the sensitivity is not large. At normal temperatures, we are typically talking 1% in velocity error for a 5 C error in the speed of sound. If you use the fixed mode and you enter the wrong speed of sound, the magnitude will depend on how big the difference between the actual speed of sound and the fixed speed of sound is.
If you want to find how large the error is; start out calculating some of the extreme values (in pressure and temperature). Maybe you will find that it is sufficient to apply the same correction for all the data. If you find that you need the full correction you must calculate the true speed of sound for each data point and apply the correction accordingly for each data point.
Sound speed calculator: http://www.es.flinders.edu.au/...
If the salinity is changing rapidly:
1. Use a fixed speed of sound instead of measured and an external conductivity probe to monitor salinity. This would let you calculate speed of sound as a function of time and correct the measurements for the changing conditions.
2. Use an average value for the experiment if the change in speed of sound is small enough it won't really affect uncertainty.
The first is probably preferable, but slightly more work. The second one is simpler and will need some justification to be a viable solution.
For Vectrino users: The Vectrino will be reading the temperature continuously while collecting data, but will not report this nor the speed of sound used. So, if your test facility water temperature increased during an experiment, the Vectrino would know this, and as long as it was not using a fixed speed of sound would adjust its measurements accordingly.