From the MIDUSS Version 2
Reference Manual - Chapter 7
(c) Copyright Alan A. Smith Inc.
Figure 7-10 - Alternative
definitions of catchment shape.
catchment is assumed to be represented by two idealized, rectangular
inclined planes - one for the pervious surface and the second for the
impervious fraction. The two planes are commonly assumed to be
inclined at the same gradient, but MIDUSS lets you define this and all
other characteristics to be different.
For each catchment you
must first specify the total catchment area and the percentage of that
area which is impervious. MIDUSS then provides three options to
define the shape of the two surfaces. Each of these is illustrated in
a rather idealized way in Figure 7-10.
The default assumption
is that the length of overland flow on the impervious surface is the
same as that specified for the pervious fraction. This case is shown
in Figure 7-10(a).
Alternatively, you may
choose an option which assumes that the width of both rectangles is
the same. This is equivalent to assuming that the overland flow
lengths are in the same proportion as the areas of the two fractions,
and is illustrated in Figure 7-10(b).
The third option
allows you to define a specific length of overland flow for each of
the two rectangles, so that neither length nor width need be the
same. This case is shown in Figure 7-10(c).
From the sketches of
Figure 7-10 it should be clear that the overland flow length is the
distance from the boundary of the idealized rectangle to the drainage
conduit (pipe or channel). It is along the overland flow length that
the surface gradient should be estimated. The idealized catchments of
Figure 7-10 are shown as non-symmetrical (i.e. with all the pervious
or impervious area on one side of the drainage conduit) only to
illustrate the concept.
In practice, it is
usual for both pervious and impervious surfaces to be distributed more
or less symmetrically about the drainage conduit.
Avoid the mistake of
estimating overland flow length and slope between the outflow point
and the point on the catchment boundary which is furthest from the
outlet. This overestimates the time of concentration and
underestimates the peak outflow.
If the catchment area
is symmetrically distributed around the drainage network, an
approximate value for the overland flow length can be found by
dividing the area by twice the length of the drainage channel. If the
catchment is unsymmetrical so that the drainage channel is along one
edge of the catchment, the overland flow length can be approximated as
(Area/Channel length). The two cases of symmetrical and one-sided
catchments are illustrated in Figures 7-11(a) and (b) respectively.
If neither of these cases applies then you must either make a
subjective judgment or simulate the area as two separate
Figure 7-11 - Estimating
overland flow length in symmetrical and one-sided catchments.
Another point to note
is that in MIDUSS the impervious fraction is assumed to be directly
connected to the drainage network. This means that flow from the
impervious areas does not pass over a pervious area before reaching
the drainage channel. In some urban drainage models the impervious
area is further subdivided into directly and indirectly connected
fractions but these methods assume that runoff from the indirectly
connected impervious area is uniformly distributed over the pervious
fraction. In practice, such runoff is usually concentrated over a
relatively small pervious area thus reducing the potential for
infiltration. The assumption in MIDUSS therefore leads to a
conservative estimate of the total runoff from the catchment.
The Manning 'n' value
is used to estimate the time of concentration (see equation [7.41])
for any specific intensity of effective rainfall. Typical values of
‘n’ for overland flow on pervious surfaces should be in the range 0.2
‑ 0.35 and do not represent realistic values of 'n' that might be
used in channel flow calculations.
In addition to the
above description, parameters must be defined which describe the
infiltration process and rainfall abstractions on the pervious area.
These will depend on the infiltration model selected and are as
described in the section Calculating Effective Rainfall .