FLOOD01798 - Laserfiche WebLink

Home Browse Search

DRAFT D I 1 I 1 1 (c} CRC-FORGl'! method More recently in Australia, the concepts first proposed by Reed and Stewart (1989) have been further developed into the GRG-FORGE method by Nandakumar et aL (1997). The GRG-FORGE method uses the concept of an expanding region fcx:wssed at the site of interest Design rainfall estimates for frequent events (e.g. 1 in 50 AEP) are based on pooled data from a few gauges around the focal point, while design rainfall estimates at the AEP limit of extrapDlation are based on pooled rainfall data from up to several hundred gauges. Before the data from different sites can be pooled, maximum annual rainfalls from each site need to be standardised by dividing by an "index variable" (which may be the mean annual maximum for the site, or a rainfall of any specffied AEP that is reasonably accurately determined from short records). Details of how to undertake appropriate regional analyses are outside the SCDpe of ARR, though practitioners are encouraged tD understand the assumptions underlying the estimates derived for their region of interest, and their limitations. I I I (d) Derivation of design rainfall estimates from regional information Specific details on deriving rainfall estimates from regional information will depend on the nature of the information available, thDUgh some general recommendations are presented below. At the time of writing (1998), design estimates have been prepared for Victoria and Tasmania, and work is underway to prepare similar information for New South Wales and Queensland. It must be stressed that this area of design flood estimation is undergoing rapid advances, and practitioners are encouraged to seek out the most accurate information relevant to their region of interest It is expected that over the next few years a number of different approaches will be develDped and that the spatial coverage of relevant design information will increase as funds to undertake the necessary wDrk becDme available. The generic sequence of steps required to derive Rare design rainfalls can be described as follows: 1. locate the catchment boundary on a map; 2. derive a representative index variable for the catchment; 3. derive representative growth factors for the catchment corresponding to the selected AEPs of interest; 4. multiply the index variable by the growth factDrs to obtain representative design point rainfalls; 5. reduce the point rainfall depths to a representative areal value using the apprDpriate areal reduction factor for the total catchment area. The actual steps required to derive rainfall depths up to the credible limit of extrapolation will vary with the method adopted for the region of interest At present, a data base of design estimates (based on the GRG-FORGE method) is under preparation for the eastern states, and efforts should be made to ensure that the most recent information relevant to the SITe of interest is used. I I I I I I I I I I I 3.4 Estimation of the PMP Depth 3.4.1 General The theoretical definition of the PMP is "the greatest depth of precipitation fDr a given duration that is physically possible over a given size storm area at a particular Book VI - ~stlmatlon of Large to Extreme "'Ioads 9J'ographical location at a certain time 01 year" (Wotld Meteorological Organisation, 1986'). Estimates are derived using generalised methods that are based on the analysis of data over a wide region. The PMP design rainfall estimates proeuced by these generalised methods are based on a set of simplifying assumptions applied when extrapolating from observed iarge events to "maximised conditions". They thus represent operational estimates of the PMP and are not necessarily equivalent to a theoretical upper limit on rainfall fDr that location i.e., there is a very small, but finite, probability that the estimates may be exceeded (Section 3.5). The procedures applied in Australia are generally consistent with procedures applied elsewhere and, in the case of the short duration estimates, are partly based on overseas data. Generalised methods for estimating PMP were introduced from the mid-1970s onwards, following the recommendations of the Wortd Meteorological Organisation (World Meteorological Organisation, 1986'). These methods use data from a large region and make adjustments for moisture availability and topographic effects on extreme rainfall depths. Tl1e generalised methods currently available in Australia are: (i) the Generalised Short Duration MethDd (GSDM, "Bulletin 53"), suitable for short duration, small area storms over any part of Australia (Bureau of Meteorology, 1994); (Ii) the Generalised Southeast Australia Method (GSAM), suitable fDr long duration storms over catchments of any size in southeast Australia (Minty et aI., 1996); and, (Iii) the Generalised Tropical Storm Method (GTSM), for long duration storms applicable to those parts of Australia affected by tropical storms (Kennedy, 1982; Kennedy and Hart, 1984). The GSAM region is sub-divided into two zones, Goastal and Inland, separated by the Great Dividing Range. The GTSM region is also divided into two zones; the East Goast Tropical Zone, where a persistent low pressure trough can enhance the rainfall from tropical cyclones and other tropical low pressure systems, and the Remaining Tropical Zone. Figure 3 illustrates the boundaries for application of the longer duration generalised methods and the subdivision into zones. It is seen that the longer duration generalised methods combine to cover all areas of Australia except western Tasmania, thDUgh IT should be noted that there are some concerns with estimates for some regions (most notably the southwest of Western Australia). The following sub-sections provide further information about these methods, and give recommendatiDns for durations not covered by the generalised methods. Details of the estimation of PMP are outside the scope of ARR, and Dnly general recommendatiDns are given herein. Descriptions of current Australian methods are given by Kennedy (1982), Kennedy and Hart (1984), Pearce and Kennedy (1993), the Bureau of Meteorology (1994), and Minty et aJ. (1996). The details of the methods which can have appreciable effects on estimates, such as assumptions regarding the time span or season within which a large Dbserved storm could have occurred, transposition zones, and the recommended limit of the maximisation factor, require judgement by a meteorologist experienced in PMP studies. It should be noted that the PMP design rainfall estimates relate tD design bursfs. The relationship between design bursts and compiete design storms is discussed in Section 4.2.1 (c).