Water Quality Technical Note AM-03
US Army Corps
Statistical Verification of Mean-Value
Fixed Water Quality Monitor Sites
in Flowing Waters
by Michael Vorwerk, Joe Carroll, and John Lemons
This technical note describes a method for verifying the representativeness of mean-value and
extreme-value water quality monitoring locations. Recommended techniques are illustrated using
data collected with the total dissolved gas monitoring system on the Columbia and Snake Rivers.
This technical note shows how statistical techniques can be applied to the design of monitoring
systems to ensure that data collected are representative and thus scientifically defensible.
Water quality managers must carefully choose the locations for fixed water quality monitors,
to ensure that the data they collect accurately reflect water quality conditions of the water of
interest. Often, a monitor site will experience some spatial or temporal bias, and data collected
there will not represent the release or river in question.
For rivers and hydroproject releases, bias may be the result of combined spill and generation
releases (Lemons, Vorwerk, and Carroll 1996), releases into lacustrine tailwaters (Vorwerk and
Carroll 1994), generation drawing water from a forebay with heterogeneities (Lemons and others
1996), point sources of pollution, or other processes (Vorwerk, Jabour, and Carroll 1996). A
monitor system intake may be located in some portion of a flow and accurately measure its
water quality, while not reflecting the quality of other portions (Figure 1). Thus, to provide
usable data for operation, regulatory, or background monitoring needs, a manager must verify the
representativeness of monitor sites with regard to the monitoring program goals.
This verification must include quantification of the spatial and temporal similarity between
water quality data gathered at the monitor site and in the stream or river in question. Flowing
water monitor systems can be designed to create temporal records of water quality information
as either means or extreme values (Ward 1979). Different verification techniques are necessary
for each of these designs. This technical note discusses the techniques necessary to verify
mean-value monitor systems.
US Army Engineer Waterways Experiment Station
3909 Halls Ferry Road, Vicksburg, MS 39180-6199