Velocity precision

Hi I do hydrographic surveys for finfish farms in Scotland and there are criteria for this surveys set out by a Scottish goverment department. In Shetland (where I am based) we have been using varoius models of nortek ADPs for around 7 years. Previously there was an velocity presiion criteria for our surveys of 2 cm/s however they have recently changed the criteria so that the velocity prescion must be better than +/- 10% of the mean current speed. For the 400 khz Aquadopp the maximum possible velocity presicion (with their other requirements such as cells less than 2m) is 0.7cm/s and for the 600 khz aquadopp and the AWAC the maximum velocity presion is 0.4 cm/s to 0.5 cm/s (depending on the exact site set up). The older 500khz NDP can only achieve around 0.8 - 0.9cm/s.

The current speeds we are tyring to measure tend to be in the range of 3cm/s to 6.5 cm/s. This means it will never be possible for the 400khz or NDP profiler to achive this presion and the 600khz meters will not be able to achieve this pression for many locations.

I don't have a great understanding of the velecoty pression. When we do repeat 15 day surveys (the minium survey length) the results tend to be within +/- 0.5 cm/s this is when the surveys are carried out at completly different time of year, sometimes many years apart and locations up to 500m apart. Some variation would be expected as the tidal strength will vary from month to month. Therefore I don't really understand what the velocity precsion is an indication of and whether this new criteria is necassary. Their new guidance suggests using an alterative form of measuing the tide. This includes using the mechanical meters which turn in the current, some which won't even turn below 3 cm/s!

Anyway, I was wondering what Norteks views were on the required velocity precision for surveys.

Thanks

Rachel

Hello

Just to add to the previous inquiry as this "+/- 10%" criteria could have serious implications to our business activities as well.

The estimated horizontal velocity precision is being treated as a standard deviation. To quote the example given a current speed of 20 cm/s with a recorded precision of 2.5 cm/s will mean that it is highly unlikely that the data will be incorrect by more than 7.5 cm/s (three standard deviations). Is this a fair application of the velocity precision? My understanding of it was that the velocity precision would give a more absolute range i.e. a speed of 20 cm/s with an estimated precision of 2.5 cm/s will mean that it is highly unlikely that the actual value will be outwith 17.5 to 22.5 cm/s

Can you also tell me how the estimated velocity precision is calculated, and if actual velocity precision can be calculated from the data collected.

Thanks in advance
Alan Harpin

I refer you to pages 10 and 11 of this document http://www.sepa.org.uk/water/aquaculture/marine_aquaculture/idoc.ashx?docid=38420d21-0875-4de5-93f7-7295ee8c65f7&version=-1 (note this link takes you to a pdf file)

The concept of "precision" is related to idea of "repeatability" as it is being used for acoustic current profilers and current meters.  Because of issues that are inherent to the estimation technique that is being used, the velocity estimate will include some random noise.  The magnitude of this noise is often described through its standard deviation, which is really the square root of the variance of the noise.  If, for example, you could remove the true signal from the data, the "precision" would be equal to the standard deviation of the noise time series. Since you cannot really remove the signal (you don't know what it is), the variance of the noise is calculated using spectral techniques and the square root of this noise is provided in the software as an estimate of the standard deviation.

Because the precision is independent of the velocity magnitude, you can potentially run into problems if the currents are very weak.  There are two ways around this problem - average over longer periods of time (i.e. spend more battery using Lithium with 3-4 the capacity of Alkaline) or use a different processing technique like the Nortek HR profilers (firmware upgrade).

Best regards, Atle Lohrmann

Hello,

Just wanted to add a few words about site studypractice in Norway....    

The 400kHz Aquadopp Profiler has become the business standard Doppler instrument within salmon farm site studies in Norway. We are still up against mechanical propeller type devices which of couse are well established in the business.

The requirements presented from Norwegian fishery authorities specify measurement intervals and averaging time. As far as I have been able to verify there is no specified requirement in terms of measurement accuracy or precision. The present requiremens were written with mechanical instruments in mind. Actually the business is very conservative and Doppler instruments are still judged by how close we can get to values previously acquired with propeller type instruments ! 

My clients indicate that even a measurement accuracy of e.g. +/- 2cm/s would be ok and not a big issue. Basically what they are looking for is to detect periods of zero current, mean values and max currents. The objective is to determine how well the location is suited in terms of fish health/growth, water column environmental impact (spreading current), bottom recovery and of course cage breakdown (fish escape).

The are local variations to the regulations, but typically the requirements are measurements at three levels - in some regions four levels. Measurements are required every 10 minutes and over a four week period. 

By the way, we are working on a SW tool especially designed for standardised reporting of fish farming site surveys (Aquadopp profiler 400kHz). The objective is to present a user friendly SW that takes data directly from the Aquapro, filters out bad data and presents the results in a variety of relevant modes, e.g progressive vectors, water flux rose plots and statistics.

Best Regards

Ole-Erik Gunnulfsen