Does the ARR2019 ensemble temporal pattern method represent various storm types in NSW?

Accurate design flood estimation is important as they are costly and pose a risk to life. This paper forms part of a larger study investigating the significance of the relationship between storm mechanism and flood magnitude in different regions of NSW, with the goal to provide recommendations on how to consider this impact in design storm estimation. There is a general question about whether the ensemble of 10 temporal patterns recommended by the recent Australian Rainfall and Runoff (ARR2019) are a good representation of various storm patterns, and whether they are the source of the underestimation bias observed in NSW. This paper addresses this question in the East Coast of NSW by running a large sample of storm patterns from different storm types through a catchment hydrology model and comparing the range in peak outflow. Over 2000 storms were extracted from pluviographs located in the East Coast of NSW for events rarer than 20% AEP and with critical durations between 1-72 hours. These storms were broadly classified into Convective, East Coast Lows, and Frontal events. The temporal pattern comparison was completed at over 200 catchments of varying size and proximity to the coast in NSW. The flow estimates at different flow quantiles derived from the ARR ensemble approach were compared to the full suite of event patterns. The results of this comparison suggest that the ARR ensemble patterns are largely unbiased. They sufficiently captured the mean and spread produced by the pseudo-population of storms extracted from the storm database. This was true across all of the storm types tested, and for catchments with differing areas. This study suggests the underestimation bias in NSW is not due to the temporal patterns, meaning preburst, loss mechanism and areal reduction factors are to be investigated. [ABSTRACT FROM AUTHOR]