Dave Yake

Rory,

I have copied the comments from our email so that others can view them… thanks again.

dave

The U20-001-04’s (13 ft range) accuracy is 3-6 mm, so that needs to be considered. It also needs to be compensated with barrometric pressure (not mentioned in the report) which adds additional inaccuracies.

With respect to the barometric pressure: a. Pressure sensor for depth: Remember that we are compensating for barometric pressure changes in the Onset logger by having a parallel sensor following barometric pressure. Not sure he understood that.  b. Also, with respect to the ultrasonic sensor, the barometric pressure has a very minimal effect on the speed of sound in air. Marion remembers that it is 2 orders of magnitude below the effect of both temperature and relative humidity (which is much lower than the temp effect.) So, really the only important effect is the temperature effect on speed of sound. 

</li><li>When measuring water in a stilling well with a pressure transducer, and the surface of the water with an ultrasonic sensor, you can’t consider the two data sets as reporting the same values. In high-velocity situations, you can get positive or negative pressure, thus lowering or raising the water level, in your stilling well depending on how it is designed.

We discarded using the ultrasonic sensor on a stilling well due to the signal dropouts which were somewhat erratic as to when they occurred, but clearly were connected to temp changes from high to low which induced some kind of condensation on the walls (that caused deflection of the wave). Similarly, we actually dispensed with the stilling well on the Onset depth logger just for convenience. So, stilling well experience is not something where got comparison data. The Final result comparisons were not data from stilling wells..

</li><li>And most importantly, in your on-ground test, the image shows that the sensor is on a metal boom. This is problematic as long metal objects deform/bend with temperature.

We don’t believe your comments about the metal boom having a significant impact on correct. If you calculate the change in dimension for a uniform metal (ie..
no bimetallic effect) that is heated on only one side (eg. sun) while the other side is cool (shady side) then you can see that the metal is much less than wood. For example, CTE’s for metal will run around 1 * 10**-5 while wood will be at least 10X higher. Still, neither will show much bending with temperature differentials that we see in outside testing. That said, however, it is a concern for planned experiments, and easy to manually measure the separate distance from the boom to the target over the temp range..

We would be interested in your results if/when available…. In particular, it would be helpful to see what the temperature effect (water temp) has on a pressure sensor logger (such as the Onset or the CTD sensor used in the Stroud logger stations..