Overland flow in urban catchments
A three parameter pseudo-linear reservoir model was developed and
tested using data collected from urban catchments in Nottingham. Some
550 individual rainfall-runoff events, from ten different catchments,
were recorded (from May 1983 to June 1985) and available for analysis.
parameters, depression storage and runoff
an effective hyetograph to be obtained from the
The third parameter, KATCH, was a catchment
obtained by an optimisation technique applied to
Two- of the model
constant which was
development data sets.
The rainfall loss parameters were determined using both the
traditional method of linear regression analysis of rainfall-runoff
depths, and by model optimisation of development data sets. Two
versions of the pseudo-linear reservoir model were tested and results
from both versions were compared with the results from the non-linear
reservoir model, which is the overland flow routing model currently
used in the Wallingford Procedure for the design and analysis of urban
storm drainage in the United Kingdom.
When the overall goodness-of-fit between simulated and observed
hydrographs was assessed (using the Biassed Integral Square Error,
BISE), similar model performance results were obtained for both
versions of the pseudo-linear reservoir model and the non-linear
reservoir model applied to paved surfaces.
The pseudo-linear reservoir model accounted for the routing of flows,
both on the surface and through the secondary drainage system, via
the catchment constant, KATCH. For the two sub-catchments with
secondary drains, the non-linear reservoir model over-estimated
peak flows by some 401, since the model ignores routing through
The pseudo-linear reservoir model gave consistently better results
when compared with the non-linear reservoir model for storms monitored
from an uncalibrated sub-catchment of mixed roof and paved surfaces.