Identifying Interference with the StreamPro in TRDI SxS Pro

Video Transcript
Download Video
Right-click and save to download

Detailed Description

This video discusses the process of identifying potential site interference with the StreamPro when using TRDI SxS Pro.


Date Taken:

Length: 00:04:18

Location Taken: Augusta, ME, US


Hi, my name is Nick Stasulis and I am a hydrologic
technician with the Maine office of the New

England Water Science Center.

In this screencast I will discuss how you
can identify site interference with the StreamPro

in SxS Pro.

A recent evaluation of StreamPro data collected
throughout the USGS indicated that the StreamPro

is susceptible to electromagnetic interference
at some sites, more than other ADCPs.

The StreamPro ADCP is more susceptible to
EMI because the transducer is separated from

the electronics by a cable.

Think of this cable as an antenna, and the
longer the cable is, the more potentially

susceptible the unit is to interference.

While we believe that this problem is mostly
confined to StreamPro ADCPs, it is conceivable

that other ADCPs may experience the problem.

First, let’s discuss the interference itself.

This interference is due to sources specific
to a site or measuring section.

So, just because a unit experiences interference
at one site or one measurement cross section,

doesn’t mean it would absolutely experience
interference at another site or in a different

cross section at the same site.

Also, sites with a low backscatter condition
are more susceptible to interference as the

EMI can more easily overwhelm the returns
of the true signal transmitted by the ADCP.

The cause of this interference is site-specific
electromagnetic interference, or EMI.

EMI sources could include television, AM,
FM and satellite transmissions, solar magnetic

storms, lightning, power transmission lines,
airport radar, railroad and mass transit infrastructure,

and several others not listed here.

Now, let’s talk about how to identify interference
in the SxS Pro software.

First, interference usually causes erroneous
or very odd looking data, particularly unusual

patterns that change with depth, and using
the simple review steps outlined in an earlier

screencast should help identify situations
where interference exists.

This is a pretty classic example of a measurement
with interference.

First, notice how the there is a very clear
increase in velocities vertically in the contour

plot, velocities at the bottom are much slower
than the velocities at the surface.

Next, notice the velocity profile for this
ice measurement appears to be non-standard.

Intensity values look reasonable, but switching
to correlation shows an increase with depth,

which is not what we’d expect in a normal

Another sign of interference could be high
error velocities, and in this case, they are

fairly consistent in the contour plot, with
only a few outliers.

The most telling sign that interference is
possible is the PT3 test, collected as part

of the ADCP test.

In this case, viewing the ADCP test shows
correlation values that stay nearly the same

over some number of lags.

In normal conditions, the correlation values
in all four beams would decrease to 15 or

so by about lag 3 and remain low.

If interference is expected, you can’t rely
on one of these checks alone, you have to

evaluate each and determine if interference
is likely.

In this case, we have an oddly shape velocity
profile vertically, unexpected correlation

values, and a non-standard PT3 test result,
though our error velocities look reasonable.

If you identify interference, you may be able
to use water up and water error thresholds

to remove the erroneous data, but this may
screen out too much data to allow the measurement

to be used.

Your best option is likely to try moving a
substantial distance to a different measurement

section, or switch instruments.

As of Spring 2015, TRDI is working on a hardware
solution to help minimize this issue.

If you have questionable data, consider consulting
with the USGS Hydroacoustics Workgroup using

the email address shown.