Environmental Engineering Reference
In-Depth Information
of the nature of the drainage pathways and the
definition of the appropriate weir outlets from
storage ponds is complex.
The 1D modelling approach was demonstrated
to be an effective approach for the modelling of
relatively large surface areas.
The Cowes catchment was satisfactorily mod-
elled using both the Infoworks CS 2Dprogramand
the SIPSON/UIM 1D/2D program. Both models
were able to simulate the response corresponding
to a historical flood event.
The 1D surface modelling approach should be
seen as complementary to the 2Dapproach, which
has been developed by several software houses
since the inception of FRMRC. Although the 1D
surface modelling approach is marginally more
resource demanding at the data preparation stage,
equivalent file sizes and simulation times are
significantly less than for the 2D.
Two-dimensional modelling provides better
graphical representation of the flooding process
than the 1D model, but 1D is currently much
more effective at modelling adaptive surface re-
sponses to flooding problems.
The 1D approach uses all the functionality of
current software packages for water quality
modelling. This means that it can be used to
demonstrate the impact of adaptive responses on
receiving water quality and also to quantify the
impact of flooding from combined sewers on pub-
lic health
evaluate the new1Dsurface flowmodel developed
by FRMRC.
Data collection and clean-upwere a significant
part of each study and the UKWIR contractors
spent considerable time within the catchments
to ensure that the elevation models were repre-
sentative of reality. The base data for the three
studies were 1-m horizontal resolution LiDAR
with a vertical resolution of
150mm. It is im-
portant that surface features that act as urban
flood pathways, for example roads with kerbs and
drop kerbs, should be accurately represented.
The use of LiDAR data to develop the DTM/
DEM should be supported by other forms of data
collection such as land and GPS surveys and site
visits (walkovers). Specific issues are associated
with flowpaths at bridges, embankments, walls
and walls with gates, hedges and narrow gaps
between buildings. Overall therefore the study
reinforced the need for sufficient time and re-
sources to be devoted to catchment familiariza-
tion and undertaking sufficient field observations
so that all the necessary small-scale surface fea-
tures that can divert or constrain flows can be
incorporated into the model.
This study highlighted that within urban areas
the definition of overland flow routes needs to be
considered at both the 'macro' scale and also at the
'micro' scale. The vertical accuracy of the Digital
Terrain Model was particularly important in this
regard, and it was concluded that the vertical
accuracy needs to be better than a typical 125-mm
height kerb face.
The FRMRC overland flow software proved to
be a relatively quick means of adding overland
pathways to existing sewer models and, further
to teething troubles with the data structure, the
data could be quickly imported into the Infoworks
CS software. It was recommended that the data
clean-up operation be made automated.
The studies identified some key issues for urban
drainagemodellers. Flat areas create problems and
the accurate identification of storage nodes and
drainage pathways is paramount, especiallywhere
there are buildings across potential flow routes. In
addition, in flatter areas some storage ponds lie
within larger storage ponds and the interpretation
Recommendations
The outputs from the three UKWIR-funded stud-
ies yielded several recommendations for improved
urban flood risk modelling:
.
Further research be undertaken to consider the
benefits to be gained from using 1D/1Dmodelling
and 1D/2Dmodelling indifferent parts of the same
catchment. In theory the upper parts of catch-
ments, where the flooding is predominantly
'conveyance' flooding, can be adequately mod-
elled with a 1D/1D approach, whilst the lower
parts of catchments or where there is flattening of
the ground slope the flooding is predominantly
'ponding' flooding, which requires
1D/2D
