Discharge is most often directly measured using the velocity area method. A flow meter is used measure the velocity and water depth at successive locations along a stream cross section. These data are then integrated to give the total discharge.
The more locations along the cross section, the more accurate the measured discharge. Figure F-5 of Dingman (Physical Hydrology, Prentice Hall, 2002) gives an interesting plot of error in discharge measurement versus number of measurement points along a cross section. Usually at least 20 locations are recommended.
When working on small streams its often easiest to begin by stretching a tape measure across the stream. Then select a measuring distance interval that will give about 20 points Xi across the stream.
At each distance Xi , measure the stream depth Yi and the stream velocity vi. In doing this, follow the USGS “Six-Tenths-Depth Method” whereby you measure the velocity at a depth below the stream surface of 0.6*Yi (or 0.4*Yi above its bottom).
The total discharge of the stream is given by

where the area Ai is defined as

When selecting a location to measure discharge, the USGS seeks out these characteristics (from Dingman, page 609):
1) Converging
flow (i.e., cross-sectional area decreasing downstream) without areas of
near
zero velocity or eddies.
2) Absence of
backwater conditions (due, for example, to high water levels in a stream or lake to
which the gaged stream is a
tributary.
3) Smooth cross-section with minimal flow obstructions upstream or downstream
4) Velocities
and depths not exceeding the range for which the velocity- and depth-measuring
devices give accurate results and for which one can safely negotiate the
section.