Considerations on two urban runoff applications in the Canadian context: The modeling of storage facilities with low release rates and the SCS method.

Abstract

A fundamental stormwater management principle is the storage of excessive stormwater runoff at its source and its subsequent release at controlled outflow rates. On-site detention facilities with low release rates are becoming increasingly popular in Canada. To determine the required size of such facilities, proponents advocate methods ranging from the simple rational method to sophisticated continuous simulation computer models. Storage volumes determined by these various methods can produce alarmingly divergent and inconsistent results. The accuracy of such methods has, in many cases, not been adequately tested. Because detention facility design is not well understood, serious misapplication of modeling has occurred in Canada for quite some time. A comparison of two categories of sizing methods---design event methods and continuous simulation techniques---was conducted to determine the relative merits of each method and their respective ranges of application and confidence. Detention facilities sized with single event procedures were subjected to long term continuous simulations using historic rainfall records to determine the actual recurrence frequency of overflows. URBHYD and STORM were selected to represent classes of single event and continuous simulation models, respectively. It was possible to compare the results of both methodologies on the basis of three parameters: storage volumes, release rates and watershed imperviousness. The comparison revealed that for low release rates the single event design approach may significantly underestimate the required storage volume. It also became evident that defining a critical design storm event independently of the drainage basin is impossible. Municipalities accepting release rates as low as 10 to 40 1/s/ha should require continuous simulation in the design process to ensure that adequate storage capacity is provided. Further research is necessary to determine the extent to which snowmelt events effect the design process.