In regulated rivers, the relationship between wetted perimeter and dis
charge is sometimes used as an expedient technique for determining the
minimum flow allowable for environmental purposes. The critical minim
um discharge is supposed to correspond to the point where there is a b
reak in the shape of the curve (usually a logarithmic or power functio
n). Below this discharge, wetted perimeter declines rapidly. This crit
ical point on the curve is almost universally, but incorrectly, termed
an 'inflection' point, and is usually determined subjectively by eye
from a graph. The appearance of a break in the shape of the curve is s
trongly dependent on the relative scaling of the axes of the graph. Th
is subjectivity can be overcome by defining the break in shape using m
athematical techniques. The important break in the shape of the curve
can be systematically defined by the point where the slope equals 1, o
r where the curvature is maximized. The technique can be applied to ot
her habitat-discharge relationships, provided the habitat variable inc
reases with discharge. These techniques were applied to two regulated
headwater streams located in the Melbourne catchment area. Channel sur
vey data were used to model the relationship between discharge and wet
ted perimeter, flowing water perimeter and blackfish habitat area. A l
ogarithmic function could be fitted to the wetted perimeter data for S
tarvation Creek, but the relationship for Armstrong Creek was linear.
Both streams showed logarithmic relationships between discharge and fl
owing water perimeter. For these streams, the wetted perimeter relatio
nships did not suggest an optimum environmental flow, nor did they sug
gest a flow level that would maintain the macroinvertebrate community
in its unregulated state if it was applied for a long period of time.
Fish habitat area does not necessarily increase with discharge, so the
method of curve analysis suggested here for wetted perimeter may not
be applicable to some fish habitat area data. Flowing water perimeter
is preferable over wetted perimeter as a variable to define habitat su
itable for macroinvertebrates. (C) 1998 John Wiley & Sans, Ltd.