Model based estimation of pollutant loads from wet-weather discharges: something is uncertain in the state of Denmark

Wet weather discharges from urban drainage systems (CSO - Combined Sewer Overflows and outlets from separate systems) can be important contributors to the overall pollutant loads from urban areas into natural surface waters. Given the logistical difficulties in continuously monitoring these discharges, dynamic simulation models have been employed since the 1970s to estimate the discharged pollutant loads. The uncertainty of these model outputs have been widely investigated by the research community during the last decades.

Following the requirement of the EU Water Framework Directive and related legislation, the Danish Environmental Protection Agency (DK-EPA) reports annual discharges of traditional macro pollutants (N, P, Organic matter - NPO) from point sources. There are currently almost 5,000 CSOs and about 14,000 separate outlets in a 43,000 km2 country with 5.8 million inhabitants. Data for all these discharge points are provided by the 98 Danish municipalities and fed into a national database, which has been described as “a good or excellent knowledge base” based on a review of the information on wet-weather discharges in the EU Member States (Moreira et al., 2016). All the reported data are based on the results of model simulations, where calculated discharged volumes are simply multiplied by fixed pollutant concentrations.

To encourage the reduction of pollution from wet-weather discharges, the option of implementing a tax on NPO loads to supplement the taxation of emissions from WWTP is currently evaluated. The purpose is both to decrease the emissions from wet weather emissions in general and to avoid sub-optimization by WWTP operators to reduce inlet flows during peak flows, which in reality increases the direct emissions from the sewer systems via CSOs.

This work presents the outcomes of an analysis that was initiated by the DK-EPA to evaluate the uncertainty of the current model estimations of NPO loads from wet weather discharges. We first analysed the possible sources of uncertainty based on literature review, and then surveyed the modelling practices followed by the municipalities/water utilities leading to their reported emissions. The analysis showed many very unlikely reported results with uncertainties and biases that by far exceeded what should be possible. Based on the outcomes of this analysis, the resulting reactions from the public when realizing how uncertain the reported data were, and the follow-up action taken by the DK-EPA, we propose a roadmap to increase robustness of model results and favour a more widespread and confident inclusion of uncertainty analysis in common practice.