Background

SWMM (Storm Water Management Model) is one of the most widely used software for storm water modeling. SWMM is released by the US Environmental Protection Agency (EPA) but the development of the official releases has been made with contributions from both EPA and academic institutions as well as private corporations. The first version, SWMM1, was released in 1971 and the latest release, SWMM5, was released in 2005 (Rossman & Huber, 2016a). During recent years the application of LID-solutions (Low Impact Development) have seen a sharp rise (Fletcher, et al., 2014). SWMM5 was updated with a discrete tool for LID-modeling in 2009, in order to allow further assessment of the impacts from implementing LIDs (McCutcheon & Wride, 2013). The US EPA defines LID as a set of practices engineered to mitigate the effects of storm water run off (US EPA, 2012).

The LID-control in SWMM5 includes an option to add an underdrain for several types of LID’s. The underdrain is stated to have the intention of preventing flooding by controlling the outflow from the storage (Rossman & Huber, 2016b). Correct design of the underdrain for a specific LID-unit will help towards keeping ponding time, CSO (combined sewer overflow) and other common storm water system design criteria within its limits (New York State, 2015).

The complexity of modeling underdrain flow leads to vastly different approaches being used in available computer models depending on their usage. In several models, such as SWMM and winSLAMM (windows Source Loading and Management Model), it is assumed that the flow is limited by either the capacity of the pipe or an outlet orifice, thus applying the orifice equation or Manning’s formula (Rossman & Huber, 2016b); (PV & Associates, 2015). Models with head driven flow, such as SWMM, are typically build for LID-modeling, while DRAINMOD is an example of a modeling software designed for agricultural land. Thus, it is originally made for drainpipes laid in finer soils where the low conductivity of the soil causes the water table to slope towards the drain, which leads to a more complex modeling approach of underdrain flow (Skaggs, 1980). Although not explicitly developed for LID’s, DRAINMOD has been successfully implemented for continuous modeling of several sites, including bioretention and permeable pavement (Hunt, et al., 2013); (Smolek, et al., 2015). A limitation in DRAINMOD is that it only reports output at the end of each day (Hunt, et al., 2013).

Problem description

Although LID-controls are still a fairly recent addition to SWMM they have undergone several updates in the last years (CHI WATER, 2017). For this reason, many of the recommendations made in published reports and online discussions on improved LID-modeling in SWMM have quickly been made obsolete, simply because they have already been added in subsequent updates to the software. Updates include better accounting for flow through LIDs under flooded conditions in 5.1.007 and allowing for separate routing of the underdrain flow in 5.1.008. However, some concerns that have been raised are still highly relevant, such as the difficulties to represent conduit outlets from LID’s in SWMM (Winston, 2015). As research on DRAINMOD is indicating advantages of its more complex modeling of soil water conditions there is reason to believe that the soil layer in SWMM could be exceedingly simplified for some LID-designs, which in turn impacts the flow through the underdrain.