Embedded Files
Modelling SuDS from Device-Scale to Street-Scale
Abstract:
As Sustainable Drainage Systems (SuDS) / Low Impact Development (LID) methods are increasingly utilised to manage stormwater in urban areas, it becomes crucial for drainage engineers to accurately represent their hydrological and hydraulic impacts within drainage modelling tools. However, not all tools include explicit SuDS modelling capabilities or their utilisation may be considered too computationally expensive in some practical contexts. This paper examines the potential for representing two common SuDS devices – a green roof and a bioretention cell – using generic hydrological/hydraulic model components. Various approaches, including initial losses, detention storage, catchment area disconnection, and reassignment to permeable surfaces, were compared against outputs from the SWMM LID module.
The main findings indicate that all approximation approaches have limitations. Methods involving complete disconnection or transferring catchment areas to pervious surfaces failed to accurately simulate the hydrological dynamics at the device scale. In contrast, two approaches based on initial losses and detention storage showed greater potential, with continuous losses based on evapotranspiration (ET) providing more realistic responses compared with a daily fixed recharge depth. In scenarios where bioretention drainage is controlled via an orifice, these two approaches produced outflow profiles similar to those of the SWMM’s LID module in continuous simulation (both NSE > 0.8). However, neither model performed acceptably in scenarios where detention effects were dominated by substrate percolation.
The aim of the paper is to not to propose a better model; rather to understand how further simplifications to key hydrological processes impact on the accuracy and usefulness of model outputs, specifically in the context of vegetated SuDS devices.
Results so far
This study compared the predicted runoff responses of the SWMM LID module (LID model) with four simplified model formulations. These formulations varied in complexity, ranging from simple approaches such as subcatchment disconnection (Disconnect model) and transfer to pervious areas (TransferP model) to more detailed representations of detention processes through flow controls. Retention processes were modelled either with fixed daily losses (Daily Loss model) or continuous losses using dynamic evapotranspiration estimates (Cont. Loss model).
The study reveals that all simplified model formulations have inherent limitations in accurately simulating hydrological dynamics at the device scale. Methods that involve complete disconnection (Disconnect) or transferring catchment areas to pervious surfaces (TransferP) do not effectively capture these dynamics, and overestimate SuDS hydraulic performance. In contrast, approaches utilizing initial losses and storage, particularly those accounting for continuous losses based on evapotranspiration (Cont. Loss), yield more realistic responses compared to those using a daily fixed recharge depth (Daily Loss) in continuous rainfall simulation. In continuous simulations, outflow simulations from Daily Loss and Cont. Loss were able to approximately capture peak flows and changes in hydrologic dynamics during rainfall events with high rainfall amounts. In particular, in a SuDS device with orifice control, the Daily Loss and Cont. Loss models perform reasonably well. However, neither model performed acceptably in scenarios where detention effects were dominated by substrate percolation.
Shuxin Ren | The University of Sheffield | 2025
Page updated
Report abuse