Interesting! - consider boiling water to de-gas as a possibility.  There are ways to test for dissolved oxygen, even probes or sondes.  I dont think I have tested for the other possible gasses.  

Hi Jay,

I am not an expert on this field, but my experience is that dissolved gases do affect significantly to acoustic attenuation. In gas-charged lakes (where the dominant gas is not O2, but usually CO2), ecosounders and other devices working with acoustics normally exhibit bad responses and wrong readings when intercepting water parcels with high concentrations in dissolved gases. 

Dissolved oxygen can be easily measured with optic sensors in modern datasondes or with DO loggers that can measure in situ and automatically. For other gases (CO2, H2S, CH4) you will need water sampling and lab analyses for precise quantification. There are also probes measuring total dissolved gas pressure in waters, which may be convenient for you. Knowing if your tank is oxic or anoxic will help you in determining which gases are being produced in the water column. You may have either DO oxygen production by algae, or other gases released by bacterial activity. Consider sterilization (e.g., by addition of chemicals) as a possible way to reduce the dissolve gas content of the tank.

Regards,

Javier

You must degas your water for good hydrophone calibration. I use a vacuum chamber to do this. As one of your responders says, boiling is also an option but If you have a vacuum chamber that is sufficiently large, use that. The issue is this. When in dissolved state, the gas species have far less effect than when in free gas state (I.e. Bubbles). The problem is that dissolved gas very easily comes out as free bubbles. Consider the glass of water you might leave by your bedside overnight. Every morning there will be visible bubbles on the side of the glass, because as the water got cold overnight, more gas dissolved into the liquid, and in the morning when the temperature in the room rose slightly, it came out of solution and grew to visible size any microscopic bubbles in microscopic cracks in the glass. See page 78 onwards of:

1. Leighton, T.G. (1994) The Acoustic Bubble, Academic Press, 640pages (ISBN 0124419208).

2. Leighton, T.G. (1996) The Acoustic Bubble, (paperback edition), London, San Diego, Academic Press, 640 pages (ISBN 0124419216).

So simple daily oscillations in temperature will cause dissolved gas to come out as bubbles. The bubbles required to cause significant absorption will probably be very small, much smaller than those you see on your overnight glass of water, as you can see from any number of articles, eg the references contained in:

3. Ainslie, M.A. and Leighton, T.G. (2011) Review of scattering and extinction cross-sections, damping factors, and resonance frequencies of a spherical gas bubble, Journal of the Acoustical Society of America, 130(5), 3184-3208 (doi: 10.1121/1.3628321).

4. Leighton, T.G., Meers, S.D. and White, P.R. (2004) Propagation through nonlinear time-dependent bubble clouds and the estimation of bubble populations from measured acoustic characteristics, Proceedings of the Royal Society of London A, 460(2049), 2521-2550 (doi:10.1098/rspa.2004.1298)

It does not matter much what gas is or is not coming out of solution - they will mostly all contribute absorption and scattering. Even if you do take all gas out of the water, there will be small bubbles trapped in crevices when you immerse your hydrophone in the water, so be careful your immersed instruments are properly wetted. 

Cracks and crevices are just one way that microbubbles that be stabilised against dissolving away. Fats, surfactants, and other 'impurities' in the water can also stabilise bubbles, as can microscopic particles. See pages 72-77 of references 1,2 above. 

So, this seems like a difficult problem. The answer is to use clean, degasser water. You might want to do calibrations at night when the water has cooled down (or cool it down with with ice made from degasser water, but be careful not to introduce extra gas in doing so) to make the bubbles in the water shrink, because gas is more soluble in sooner water - do calculate the size of resonance bubbles for your sound field (see equation 4.86 on page 306 of references 1,2 above). You might be friendly advice from your national measurement lab (and you can see why good calibrations cost a lot). 

The effect of the presence of gas on the velocity propagation of sound in liquid is a well known issue, discussed, among Others, by Landau. However its effect on the overall performance of a sensor is difficult to be understood and taken into account (how would a sonar be affected in a real scenarion from the presence of air into the seewater?)

I personally agree with TG Leighton, although eliminating the air trapped into the water is not a simple task. I worked for years on such topic when dealt with space experiment with liquids: we "Always" had the need of degassing the liquid under test to avoid problems due to bubbles during the experiment.

Boiling the water is a good idea, although the best solution is that of degassing it using a suitably designed apparatus (vacuum pump, reservoir, lung for trapping vapours, etc., not a simple experimental arrangement). It is difficult to judge from here which grade of degassing You need. Anyway You should consider that, also after degassing, the air shall be again dissolved into the water if You expose the reservoir to free air. Mixing and/or flushing the water after the degassing procedure is the "best way" to regas it again, unless these operations are performed under vacuum.

Anyway, before starting to put on the shelf the intricated and expensive (especially in terms of time) degassing apparatus, I suggest to follow the indications provided by TG Leighton at the end of his answer.

Do not hesitate to ask more about this issue.

Hi Jay,

It affects more things than I first thought it could affect and yes, it is all the more important at high frequency and in situtations with low line pressure.

Not only is wave speed greatly affected but the formation of shock waves through wave steepening arises as well as soon as pulsation amplitude starts to become significant.

You might want to take a look here

http://qringtech.com/2010/09/15/wave-steepening-increase-peak-pressure-piping-pumps/

and read this excellent book on multiphase flows

http://authors.library.caltech.edu/25021/2/cabook.pdf

As a comment - I was working on noise from an automotive servo steering system many years ago. Making tests was laborious as filling the system with oil required the car to run for 24 hrs before oil charcateristics started to resemble those of normal operation. Hydrualics oil should have about 1500 m/s wave speed when perfectly degassed. As a rule of thumb, it usually lands in the ball park 800-900 m/s.

Have fun

Claes

Did you take into account the reverberation.

You can see : ] N.COCHARD, J.L. LACOUME, P. ARZELIES and Y. GABILLET, Underwater acoustic noise measurement in test tanks, IEEE Journal of Oceanic Engineering, Vol 25, n° 4, pp. 516,522, october  2000.

Thank you all for your valuable responses.  I can address a few of the issues above for clarity.  In our case, the hydrophones are wetted with a 50% solution of water and dishwashing detergent, though I couldn't tell you the brand.  I've been looking for a brand-neutral solution but haven't found any definitive reference.

We are accounting for reverb from the bottom, sides and surface of the tank, but have not considered more complex reflections from the hydrophone cases or the apparatus which holds them, nor have we made any attempt to measure and characterise any scattering effect (dispersion) along the direct path.

My working theory is that absorption and/or dispersion effects due to scattering results in unstable readings for the first few weeks.  After a settling period, we see a consistent result between tests, which we naively assume to be a correct result, in the absence of any information to the contrary.

Because the tank is indoors the temperature doesn't vary by more than a degree, if that, from day to night; the source of our water is a turbulent groundwater source fed by an underground spring.  The water is filtered and UV treated, but not degassed.  My assumption is that the scattering effect is initially much worse because the water source is very cold, is aerated during the pumping and filling process, but that the active bubble content is reduced as the dissolved gases reach an equilibrium concentration.

We are due to refill the tank soon, so I'll make some before and after measurements and post the results here, some time in the next few months.

Update.  We never did refill the tank, opting instead to control algae with very light chlorine, pH was stabilized to just below 7 (Tank is fiberglass, all other fittings are plastic).  Instead of draining and refilling the tank, we decant a few dozen liters of water, let it sit over the weekend, then carefully add it.  This regime has kept the tank full (and clean) without needing to shut down for weeks at a time.  I should have mentioned that the tank capacity is 6000 liters.  Even though the temperature variation is not great, the tank is slightly colder in the morning so calibrations are performed then.  We avoid using the tank in the afternoon as it warms up a few degrees then (in summer).  This procedure falls short of a laboratory standard, but does appear to be effective in reducing the problem to an acceptable level.

We are currently designing an apparatus for our new facility to explore some of the options presented above.