Investigating the appropriate level of complexity for stormwater micropollutants modelling

The increasing importance of stormwater micropollutants (MP) management boosted the use of mathematical models in this field. Several approaches with different level of spatial and temporal resolution have been applied in the last years. This study compared modelling approaches differing for their level of spatial and temporal complexity and investigated their potential for use in stormwater quality management. The methods were tested for total Cu and Zn in two urban catchments with different land usage and number of available measurements. The comparison focused on two approaches for spatial characterization of urban catchments: a lumped and a detailed description of potential MP sources. Also, three modelling approaches with different temporal resolution were compared: a Volume-Concentration approach, an event-based stochastic approach, and a dynamic accumulation-washoff model. The comparison highlighted how the use of a detailed catchment description provided better prediction of MP loads, but was strongly affected by the quality of the used literature data. When a sufficient number of measurements was available, the dynamic model estimated MP loads with narrower uncertainty bounds. When few or no data are available, the stochastic approach is suggested, as it considers inter-event variability with similar result uncertainty as the Volume-Concentration method.