The basic approach to verifying the

representativeness of a monitor site is to

compare matched pairs of observations

from the monitor and averaged from the

flow (Figure 2). These pairs must be

taken over as many different times, flow

conditions, and water quality variations

as possible.

The observations in the flow must be distributed so they adequately describe water quality

conditions across the stream. For wider streams and rivers or for more highly variable water

quality conditions, more sample locations are necessary. The sample values from the stream are

averaged with an area-weighted average. If velocities vary greatly in the stream cross section, the

data averaging must also be flow-weighted. The next section provides details on this weighting.

In practice, data are often limited, and the only available option is averaging the transect data

with a simple arithmetic average, and then carrying out the statistical comparison. However, if

the stations are not evenly spaced or if the water column has lateral or vertical heterogeneities in

water quality or velocity, then a flow-weighted average should be calculated.

The following method can be used to

calculate a flow-weighted average. For

each sample station, *i*, and depth, *z*, with

velocity *U*i,z and water quality parameter

value *P*i,z, assign an area *A*i,z that the

information gathered at that location

represents (Figure 3). The area can be

difficult to calculate and is most often

approximated from depth soundings,

maps, surveying techniques, global

positioning equipment, and "best-guess."

The transect flow-weighted average of

the parameter *P *can then be expressed as

∑ ∑A

∑ ∑A

(1)

3