Evaluating Extrapolation in SonTek RiverSurveyor Stat. Live Data

Video Transcript
Download Video
Right-click and save to download

Detailed Description

This video discusses the process of evaluating extrapolation methods in the SonTek RiverSurveyor Stationary Live software. Note: Use of trade names is for descriptive purposes only, and does not imply endorsement by the USGS.

Details

Image Dimensions: 480 x 360

Date Taken:

Length: 00:06:00

Location Taken: Augusta, ME, US

Transcript

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 the basics
of reviewing extrapolation methods for midsection

measurement data collected in RiverSurveyor
Stationary Live.

It’s important to note that the USGS extrap
program cannot be used to evaluate extrapolation

methods in midsection software, it is a manual
process, as we’ll discuss.

-Before reviewing extrapolation in RSSL, be
sure the screening distance is set appropriately

in the software, as this will cause data near
the surface to be removed, and might change

your evaluation of the methods used.

As a reminder, the screening distance is set
to 0.52 plus the transducer depth for open

water measurements.

For ice measurements, the screening distance
is still based on the water surface, but must

include ice information.

So, you would add the 0.52 to the transducer
depth below water surface, which would vary

for each station if ice information was entered
into the software.

-When evaluating extrapolation methods for
any ADCP data, it is useful to have some knowledge

of the site conditions to aid in your understanding
of an appropriate extrapolation method.

For example, it’s useful to know about the
bottom substrate type and size in relation

to the channel depth, the channel shape and
hydraulic features upstream and downstream

of the cross section.

Certain channel shapes can tell us something
about what the expected velocity profile might

be, as we see in this diagram from Chow.

It shows that some site conditions might cause
us to expect a bend back at the surface, indicated

by the blue arrows, while some might cause
to expect an increase in velocity towards

the surface, indicated by the red arrows.

Also, the roughness of the bottom could tell
us something about the expected power exponent.

If the average depth is 4 ft and there were
1 ft boulders on the channel bottom, you would

expect the profile to trend towards zero higher
in the water column than if the channel were

a sand bottom.

The rougher the bottom, the higher you would
expect the extrapolation power fit exponent

to be.

-Start your evaluation of the extrapolation
method by scrolling through each vertical

and looking at the velocity profile plot.

This is easily done using the left and right
arrow keys.

Remember that the blue measured data is always
used, and we want to evaluate how the green

top fit and orange bottom fit line up with
the measured data and extend beyond that data

into the unmeasured areas.

-If you start noticing a consistent trend
of measured data and the extrapolated areas

not agreeing, it’s likely a change is needed
from the power/power default.

If a change in slope to the measured data
will help align the extrapolation, it may

be as simple as increasing or decreasing the
exponent, RSSL software labels the exponent

as coefficient.

A lower coefficient will cause a more vertical
profile (due to a smoother bed), while a higher

coefficient will flatten the profile (due
to a rough bed).

If the coefficient alone doesn’t allow a
proper fit, you might need to change the method,

as well.

Note that when changing the exponent for a
power fit, the coefficient value needs to

be changed for both the top and bottom method.

-Constant/no slip will take the upper most
bin and carry that velocity to the surface

and fit a power curve through the lower 20%
of the data.

The cells to use for the constant method is
typically not changed, left at 1, and the

% of the cells used for the no slip method
should typically not be changed, either.

It is important to note that you can, and
should, change the coefficient for the no

slip method if the standard exponent doesn’t
follow the trend of the measured data.

-Let’s make a few other points on the available
methods.

No slip should only be used on the top for
ice measurements, never for open water data.

Also, constant should never be used on the
bottom, for any measurement, as we know the

velocity eventually goes to zero.

3-point will fit a line through the top three
bins, and is useful in conditions where strong

winds cause a severe bend towards zero on
the surface.

Use of this method should be confirmed with
detailed field notes.

-While many measurements will have the same
extrapolation method used for the entire measurement,

some cross sections may require different
methods for each station.

If the portions of your cross section vary
significantly in bed roughness, bed type,

and/or flow direction the velocity profile
for some sections may also vary.

In RiverSurveyor Stationary Live you can specify
extrapolation methods, including the power-fit

exponent, for individual stations when the
shape of the velocity profile varies from

station to station.

-After your evaluation is complete, and you’ve
selected the method that you feel is appropriate,

there are a couple steps left.

First, apply the newly determined extrapolation
settings to each station, or the entire measurement,

as appropriate.

Take a look at the change that resulted to
the final discharge, as this will help determine

the uncertainty associated with your new extrapolation
method.

If you are confident in the selected method
and the change in final discharge is quite

small, it’s likely the change in extrapolation
method will have little influence in how you

rate the measurement.

If the change is quite large or uncertain,
you should likely down-rate your measurement

due to the increased uncertainty with the
extrapolation method.

Also, any time you apply a change to the extrapolation
methods, make sure that the change is documented

in field notes.