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

I
I
I
I

be provided for the tropical regions of Australia with the
foreshadowed review of the GTSM.

3.9.2 Specific Recommendations

(a) Selection of appropriate patterns

Table 8 summarises the different temporal patterns that
can be applied to design rainfall bursts. There are four main
sources of design information as follows:

. temporal patterns for use with the more frequent design
events (Book II, Section 2);

. long duration patterns for use in southeastern Australia
(Nathan, 1992; Bureau of Meteorology, 1998);

. short duration patterns from Bulletin 53 (Bureau of
Meteorology. 1994); and,

. tropical storm patterns (Kennedy and Hart, 1984).

The first two sets of patterns involve application of the
Average Variability Method (Pilgrim et at, 1969), and as
such are specifically suited to the AEP-neutral design
objective. The last three sets of patterns were originally
derived for application to PMP events, but in the absence of
any more relevant information they are applied to events
with AEPs rarer than 1 in 100. The one exception to this is
the use of unsmoothed GSAM patterns for long duration
storms with AEPs to the credible limit of extrapolation. The
unsmoothed patterns were derived using the Average
Variability Method, but they differ from the patterns
provided in Book II Section 2 as they were based on a
sample of larger (areal) storms. These patterns were then
smoothed for use with PMP estimates, as mentioned
above.

Another issue requiring mention is the absence of
temporal patterns in southeastern Australia for use with
storm durations between the upper limit of the Bulletin 53
method (3 or 6 hours) and the lower limit of the GSAM
method (generally 24 hours). A pragmatic solution to this
problem is to apply both sets of temporal patterns and to
adopt a weighted average peak flow, where the weighting is
based on storm duration (as indicated in Table 8). The
weighted average peak flow is then used to scale the
hydrograph obtained using the most relevant generalised
method; weighting all the ordinates of the hydrograph is not
recommended as the resulting hydrograph may exhibit a
lower peak than etther of the individual hydrographs.

In the GTSM region, the GSOM temporal pattern should
be used for the 6 hour event In the transition zone between
the GTSM and GSAM regions, temporal patterns from both
the GSAM and GTSM methods should be applied
separately (in conjunction with the corresponding spatial
patterns), and the largest flood adopted.

(b) Dealing with inconsistencies and
smoothing of results

In practice, the simplistic use of single design patterns
for different durations and AEPs can yield flood estimates
that do not vary in a consistent manner. In extreme cases,
this can result in design flood magnitudes that decrease
with decreasing AEP. M ore typically, the patterns may
result in crttical storm durations that vary inconsistently with
AEP; such . variation wHl impa(:t up.on the volulll!> Df
design hydrographs, which when routed through a
reservoir, may produce inconsistent results. Problems are
more likely to occur with the transition between Book II
Section 2 patterns and those derived for PMP events. The
use of unsmoothed GSAM patterns for longer duration
events will largely obVIate the problem In southeastern
Australia, though for other applications consideration
should be given to filtering out embedded bursts of lower

I

I

I
1
I
1
I
I

I
1

I
I

I
I

I

..........." '"~ - ...._n...'...~.......... .....~" ,.... ....^.......,... I I...."...".

AEP intensities (using, for example, the simple algorithm
for re-distribution provided in the RORB model; Laurenson
and Mein, 1994). Where significant inconsistencies remain,
practitioners will need to apply judicious smoothing of
results for different durations and AEPs.

3.10 Spatial Patterns

(a) Basis of adopted patterns

Spatial patterns are also required for the design
rainfalls. The source of spatial patterns as a function of
burst duration and AEP is broadly similar to that adopted
for temporal patterns, and is summarised in Table 9. It is
seen that there are four main sources of design
information: uniform patterns for use with the more frequent
design events, spatial patterns for use in southeastern
Australia (Minty et at, 1996), thunderstorm patterns from
Bulletin 53 (Bureau of Meteorology, 1994), and tropical
storm patterns (Kennedy and Hart, 1984). As with temporal
patterns, the last three sets of patterns were originally
derived for application to PMP events, but in the absence of
any more relevant information they are applied to events
with AEPs rarer than 1 in 100. .

(b) Issues for consideration

. Catchments with large inundated areas. Except for
catchments with marked rainfalt gradients, the spatial
distribution of rainfall has generally less influence on
the shape and size of the resulting hydrograph than
temporal patterns, however the thunderstorm and
tropical patterns can have an appreciable effect on
flood magnitude, particularly if the catchment contains
extensive drowned reaches resulting from reservoir
inundation. For such catchments, the largest flood
may result from location of the eye of the areal pattern
close to the storage, even thDUgh this does not result
in the greatest average depth of rainfall. Where
relevant, this possibilily should be checked.

. Large design rainfalls. Design rainfalts should
generalty be distributed uniformly over the catchment,
unless there is clear evidence from either observed
storm events or the design rainfall isopleths in
Volume 2 ARR of a marked spatial trend in storm
rainfall. For such catchments, significant variation in
design rainfall characteristics can be broadly captured
by dividing the catchment into two or more sub-
catchments; design rainfall information can be derived
separately for each region and then applied uniformly
over each sub-catchment If the catchment is sub-
divided into several uniform zones, the areal reductiDn
factors applied must be applicable to the total
catchment size, and not each sub-catchment area.

. Rare to Extreme short duration design rainfalis. The
thunderstorm patterns provided in Bulletin 53 (Bureau
of Meteorology, 1994) should be used. The
supplement to Bulletin 53 issued in December 1996
by the Bureau contains a substantially revised method
of spatially distributing rainfall and it is important that
practitioners no longer use the method described in
the original Bulletin 53. The spatial pattern should
generafly be centred over the catchment and
orientated in such a way as to overlap the catchment
boundary with the smallest possible ellipse.

. Rare to Extreme long duration design rainfalls in
southeastern Australia. The spatial patterns provided
with GSAM estimates (Minty et aI., 19$6) 5l1ou4d be
applied to all Rare to Extreme events. The spatial
patterns are based on modified 72 hour 50 year ARI
intensity fields of design rainfalls from Book II,