A new version of DRAINS is now available that incorporates ARR2016 procedures. It includes a new Initial Loss - Continuing Loss Model, use of IFD 2016 data from the BOM web site, and use of ARR 2016 rainfall pattterns from the ARR data hub. Please refer to the built-in DRAINS help system for more detail (start with the Recently-added features topic).
The ARR data hub now provides temporal patterns for rainfall design bursts of various durations. But there is a problem with the data now available.
The temporal patterns are for design bursts only (not for complete storms). ARR 2016 (Book 5, Section 3.3.2) states:
However, if design bursts, rather than complete storms, are used in design then the burst initial loss needs to be reduced to account for the pre-burst rainfall.
If you look at Section 3.3.2 the meaning of the above quote is clearer. It means the burst initial loss (ILb) is less than the complete storm initial loss (ILs) by the depth of pre-burst rainfall depth, or
ILb = ILs – pre-burst depth
ARR 2016 recommends use of an initial loss - continuing loss model. The data hub provides Initial Loss data for complete storms. Since it only provide patterns for design bursts (not complete storms) it is clear we need to reduce the Initial Loss data values to account for the pre-burst rainfall. The problem is that the data hub does not provide the pre-burst rainfall depth.
Bob Stack pointed out this problem to the ARR team and requested the data hub provide suitable pre-burst data (emails to the ARR team in June 2016, to Mark Babister in July 2016 and to James Ball in December 2016). These requests have been ignored.
The new version of DRAINS contains a new data item for each storm (Pre-burst rainfall depth). The default value used in DRAINS is 0. The data hub does provide median values of pre-burst depth but only for bursts of 60 minutes and longer. There is no guidance for pre-burst depths for burst durations less than 60 minutes, and these shorter burst durations are often critical in urban situations. So how can we proceed?
We do know that if the pre-burst depth exceeds the Initial Loss for the storm then the initial loss for the burst (ILb) will be 0. At the other extreme, if there is no pre-burst rainfall, then ILb will equal ILs. So we could use these two values of pre-burst depth (0 and ILs) to do sensitivity testing to get a range of answers. Hint: Rather than change the pre-burst depth for a large number of storms in DRAINS, you can get the same effect by leaving the pre-burst depths at 0 and changing the initial losses in the loss model to 0.
Click here for a link to a more detailed discussion on the Sensitivity of Results to Pre-Burst Depth.
This sensitivity testing (doubling the amount of analysis required) is undesirable and we would hope that the ARR team will get the message and provide the data we all so obviously need. The information is available in the original rainfall data used to develop the temporal patterns. It could be provided either as a single number for each pattern or by publishing longer duration patterns that include the pre-burst rainfall.
The latest version of DRAINS continues to provide the option of using a Horton Loss Model (ILSAX hydrology). Section 3.3.2 correctly points out that we still have the same problem:
For the same reason, the initial moisture content for storage capacity models (such as Horton and SWMOD) need to be increased to account for this pre-burst rainfall.
But we have been living with this problem for many years with the patterns from ARR 1987. We were hoping that ARR 2016 would solve this problem for both Initial-Continuing and Horton loss models. Until there is better information on pre-burst depth we prefer to continue using ILSAX hydrology (which uses Horton losses). Our preliminary work suggests this is less sensitive to pre-burst rainfall.