Ultramoor Data inter-comparisons for the Ultramoor Buoy Farm test mooring.  

1.     The basic inter-comparisons were accomplished by comparing the instruments at the 1 second sampling rate.  This required data interpolation of the Aanderaa instrument from 2 minutes to 1 minute and of the VACM from 56.25 seconds to 1 minute.

2.     One-minute samples of the data from the different instruments show the Nobska Mavs has much more variability than the other instruments.  This appears to be an artifact of the Nobska Mavs sampling rate (5 samples @ 2 second sampling each minute).  A different sampling scheme for the instrument might eliminate the variability.  Note that the 15-minute averages show that Nobska Mavs has a good comparison with other instruments.

3.     It should be noted that the FSI instrument overwrote part of it’s at sea data.  This occurred in the lab during the download process by an experienced laboratory engineer.

4.     The ‘up’ variable has a strong signal (+/- 6 cm) in the FSI.  The ‘up’ variable in the Nobska Mavs and Sontek Argonaut instruments appear to have random fluctuations of +/- 2 cm/sec.

5.     The VACM and the Nobska Mavs had short time series due to running out of storage space.  The FSI instrument is short at the beginning in order to start the instrument after launch and stops between 15 and 20 minutes short of recovery, reason unknown.

6.     It was not possible to totally evaluate consistency in clock function for all instruments.  There were no ‘first’ or ‘last’ events, nor instrument recovery times.  Temperature and pressure agree on the launch times for all the instruments except the FSI which did not record the launch. However the two instruments with  pressure data at the end (Sontek and Aanderaa) show a 4-minute difference in the mooring release time for the 34 day recording period.  This time discrepancy is outside our normal time specs of +/- 1 second a day and would translate to a 45 minute drift per year. The release time as shown by the temperatures show the same 4-minute difference between the Sontek and the Aanderaa.   The VMCM temperature release time was in the middle, that is, 2 minutes after the Aanderaa and 2 minutes before the Sontek.  As temperature values were not available for the other four instruments approximate clock checks were not possible.

7.     Thermistor stabilization at instrument depth occurred for the VMCM and the Aanderaa within 3 minutes; the VACM and the Sontek within 15 minutes.  The Nobska Mavs temperature sensor took about 66 minutes to stabilize.  Also the Nobska Mavs showed a one degree higher temperature reading than the other instruments prior to launch.  Note: The Nobska Mavs temperature sensor is an internal sensor and not coupled to the instrument case either mechanically or in direct physical contact.

8.     15 minute and one day averages from all instruments show reasonable correspondence between the main variables (U, V, T)             allowing for the ‘real’ differences in the water.

9.     In conclusion:

• Close comparisons between the U, V and T data from the various instruments are not possible because the mooring placement resulted in the instruments being in different water regimes.  See the temperature plots available on the web site http://www.whoi.edu/science/ultramoor.

• The time discrepancy needs more careful attention.  Checking the time base using launch and release data is not the most accurate method, but does give an indication of a possible problem.

10.     Recommended:

• A near bottom mooring in deeper water to enable a better comparison of the main data components.
• Also pre and post deployment events placed in the data stream of each instrument to establish clock accuracy.

 
Susan Tarbell
Dec. 11, 1998