[Acoustics] index velocity rating question

[James J Duncker]

We're in the middle of an error analysis on a number of AVM gaging stations
and have a question regarding the grouping of ADCP transects for
development of an index-velocity rating.  We've always run the AVM site at
a 5-minute interval, generating 288 unit-value velocity readings per day.
Initially, we used each transect as a seperate measurement and plotted all
of the transects on the index-rating curve.  Later, to avoid serial
correlation, we moved to grouping the measurements by date and flow
condition, and then changed our procedures again to always making the same
number of transects at the site (8-10).  So now we have an index-velocity
rating where each point represents the average of ten ADCP transects (that
collectively took 1 hour to collect) and 1 hour's worth of AVM velocity
data.  The regression equation from this "grouped" transect data is then
applied to all of the 5-minute AVM unit-value velocity time series to
compute discharge throughout the year.  This is where we are starting to
have some questions...our error analysis is leading us to think that the
regression developed should only be applied to hourly averaged avm
time-series data, rather than applying it to 5-minute velocity data.
Anyone have any thoughts on this?  Is this addressed in the index-velocity
rating class?

Thanks;
Jim


James J. Duncker
Hydrologist
U.S. Geological Survey