FLOOD01798

• County: Statewide
• Title: Australian Rainfall and Runoff 1998, Revision of Book VI - Estimation of Large to Extreme Floods
• Date: 11/28/1998
• Prepared By: Rory Nathan, Sinclair Knight Merz
• Floodplain - Doc Type: Educational/Technical/Reference Information

Ul'V\r I U

I

I
I

:c
C ~.C
Pre-burst rainfall :

1

~
o
t;::
E
III
~
-
Ul
-

:c

I

~
c
'(ij
0::

Storm Initial Loss (IL,,)

I

Pre-storm
rainfall

1
I

DOOK VI - t:.StlmatlOn or Large to t:.xtreme rlooas

Design Storm

)':

Design burst

~

).:
Burst Initial
Loss (ILs)

Start of
hydrograph
rise

Figure 7 Schematic diagram illustrating concepts used in design loss formulation (adapted from Hill et al., 1997).

Time

I
1

initial loss are therefore treated separately in (i) and (ii)
below.
The interpretation of proportional loss as the
unsaturated proportion of the catchment means that with
larger storm events the unsaturated proportion of the
catchment is reducing and thus the proportional loss also
reduces. As it is difficu~ to extrapolate the rate of this
reduction to extreme events, the proportional loss model is
generally considered less appropriate for estimating Large
to Extreme floods. On the other hand. the continuing loss
rate (CL) is expected to approach a limiting value for
saturated catchment conditions, and this limiting value is
the appropriate design loss rate for Extreme events. More
detailed discussion of the variation of CL with event
magnitude is given in (iii) below.

(/) Storm /nit/a//oss (/L.)

The available evidence to support the conceptual
interpretation of loss variation includes the resu~s obtained
by HiD et aL (1996'); these indicated little or no variation of
design losses with rainfall severity for events more frequent
than 1 in 100 AEP. For IL" this finding implies little or no
correlation between the magnitudes of pre-storm rainfan
(producing the storm antecedent conditions) and storm
event rainfall for events more frequent than 1 in 100 AEP.
An analysis of the rainfall conditions prior to the largest
storms on record in the GSAM region of southeastern
Australia (Bureau of Meteorology, 1998') indicated
qualitatively no propensity for "greater than normal" rainfall
in the 15 days immediately preceding these storms. The
analysis by the Bureau of Meteorology shows that about
15% of stonns 15 daylt immediaboly prece.ding the largest
storms on record in southeastern Australia were preceded
by rainfall totals of less than 10% of the depth of the storm.
Further, the analysis revealed that the average length of
lt1e dry period between pre-storm rainfall and the storms
was about 8 days._
The avaltable evidence thus suggests that there Is no
need to vary IL, with event magnitudes up to the largest
event on record. Further research is desirable to confirm

I
I

1

1

1
1

I

I

I

the applicability of these findings of little or no variation of
IL. with event magnitude to regions outside south-eastern
Australia.

(ii) Burst initia//oss (/L.):

The pre-burst rainfall (the rainfall from the beginning of
the complete storm to the start of the rainfall burst), rather
than the pre-storm rainfall, is the key determinant of ILo.
This results in IL. being systematically smaller than IL.; the
difference decreasing with increasing burst duration as
there is a greater tendency for long duration bursts to
represent complete storms. As an increasing storm
magnitude is generally also associated with larger pre-burst
rainfall, IL. tends to further decrease with increasing event
magnitude.

(Iii) Continuing loss (CL):

For events of increasing duration and intensity of
rainfall, an increasing proportion of the catchment is
expected to become saturated, resulting in a reduced
catchment average value of CL However, the available
evidence from Hill et aL (1996') based on catchments
located in Victoria and the ACT indicates no systematic
differences in CL for observed events in the AEP range
from 1 in 2 to 1 in 100. This can be interpreted to mean
that, except in catchments with highly pervious soils,
catchment saturation is approached already during
moderate to large storm events. Nevertheless, it should be
conservatively assumed that only the CL values associated
with the largest observed events are representative of
design loss rates for Large to Extreme floods.

(d) Variation of design losses with season

There is clear evidence that initial loss values vary
seasonally in some regions of Australia (e.g. Laurenson
and Pilgrim, 1963; Hill et aL 1998), and this can be readily
explained by differences in the liI<elihaod of pre"storm
rainfall for different seasons. However, the interpretation of
the observed seasonal differences in continuing losses is
more difficult Little published information is available on