Does distributed monitoring improve the calibration of urban drainage models?

Abstract

We conclude that while distributed sensing of discharge and water level helps understanding the dynamics of the stormwater runoff, it is not always straightforward to make better predictive models of complex urban drainage networks using this data. This analysis - by delineating the bottlenecks towards informative inference - provides via-negativa recommendations for the placement of distributed sensors, when improved model performance is also one of the motivating factors for such data collection campaigns.