internal vs. post processing

Hello,

Can someone inform about the pros and cons, if any, regarding

sampling at small profile intervals and post-averaging vs. sampling

at larger intervals and have the AquaDopp do the averaging internally.

For example:

Profile interval 1 sec & Average interval 1 sec; then post-average for 60 sec

vs.

Profile interval 60 sec & Average interval 60 sec

Assume, of course, that there is enough memory/power for both

options.

Best,

Ayal

The only difference I know is the ability to do QA/QC at the shorter intervals.  And then of course there is the issue of instruments that are moving -- you have to resolve the data in ENU at the shorter interval if you are going to average data from a moving system.  For instruments that are sitting still, I would think QC is the big one and there were several studies on this back in the 80s/90s for vessel mounted systems that showed how the quality of the data could be improved using different schemes.   Len Zedel also gave an interesting paper on "fish removal" at the latest CMTC conference that would require the collection of "raw data".

Best, Atle Lohrmann 

Atle,

Can you elaborate on what you mean regarding the QC problem.

How is QC done, and what trouble one may get into when sampling

with short intervals. I plan to have the instrument on a fixed bottom pod,

looking up, and intervals of e.g. 5-10 sec. At least no worries of motion

(unless something really bad happens...)

Best,

Ayal

Previously Atle Lohrmann wrote:

The only difference I know is the ability to do QA/QC at the shorter intervals.  And then of course there is the issue of instruments that are moving -- you have to resolve the data in ENU at the shorter interval if you are going to average data from a moving system.  For instruments that are sitting still, I would think QC is the big one and there were several studies on this back in the 80s/90s for vessel mounted systems that showed how the quality of the data could be improved using different schemes.   Len Zedel also gave an interesting paper on "fish removal" at the latest CMTC conference that would require the collection of "raw data".

Best, Atle Lohrmann 

QC procedures depend on the issues that may be relevant for a particular site.  A typical example would be fish swimming through the beams -- if you want to remove the individual fish you have to run the QC on the raw data and not wait for the effect to end up as a bias in the mean data.  This is particularly a problem in areas around fish farms, where the excess feed is eaten by the wild stock.  Another example could be plants swinging in and out of the Vector sampling volume when studying the bottom boundary layer. 

Qartod in the US has made an effort to look at the QA/QC problem in some detail.  A link to their WEB site is here:  http://www.qartod.org

- Atle