Scenario Analysis
An integrated model for management of stormwater micropollutants Luca Vezzaro, Anna Ledin, Peter Steen Mikkelsen
Outline Introduction Model and case study Uncertainty analysis Scenario analysis Conclusion
Stormwater micropollutants
Which substances are we talking about?
33 Substances listed in the EU Directives 2000/60/CE and 2008/105/CE (PP = Priority Pollutants)
+
Any other micropollutant (MP)
Elements of the stormwater system
Catchment
Drainage System
Treatment System
Integrated stormwater quality model
Models in stormwater quality management What can models be used for? Looking at the Driver-Pressure-State-Impact-Response framework (EEA, 1999)
Rainfall Urban activities
Models Micro Pollutants (MP)
Models
Models
Control strategies
Models Models
Water quality
Toxicity
Outline Introduction Model and case study Uncertainty analysis Scenario analysis Conclusion
Source characterization
Lumped characterization
Detailed description Copper roof
Traffic area Commercial
Residential
Highway
Roof
Residential street
Parking lot
Source control End-of-pipe treatment
Pollutant release and transport
STUMP model Stormwater Treatment Unit for MicroPollutants (Vezzaro et al., 2010a) Serial CSTR Number of tanks = same hydraulic behaviour of the treatment unit
Pseudo First order kinetics Fate processes based on substance’s inherent properties = Wide range of substance Easily retrievable data
INPUT
OUTPUT
Integrated model Sources+release+treatment
Emission-based
Dynamic acc/washoff
STUMP
Several source of uncertainty!! GLUE - Generalized Likelihood Uncertainty Estimation (Beven and Binley, 1992)
Case Study The Hersted industripark catchment Industrial-residential area in the Albertslund Kommune (~ 94.7 ha) Measurements 14 months flow data Flow proportional samples (inlet), time proportional (outlet)
Industrial area
Residential area
Outline Introduction Model and case study Uncertainty analysis Scenario analysis Conclusion
Analysis of model performance Pond inlet (hydraulic) Flow predictions are affected by the rainfall input data
Rainfall, but no flow
No rainfall, but flow
Delay in measured rainfall- measured flow
Analysis of model performance Pond inlet (hydraulic) Rainfall corrected according to position of rain gauge (3 km away) Vezzaro et al., 2010b
Delay is removed
Coverage of 40.3% of flow data
Analysis of model performance Pond inlet (quality) 5 rain events (33 samples) One extreme event affects calibration
Coverage of 74.3% of TSS data
Coverage of 82.9% of Cu data
Analysis of model performance Pond outlet (quality) Modelled peaks smoother than measured Pond hydraulic short-circuit higher than expected
Coverage of 53.3% of TSS data
Coverage of 35.7% of Cu data
OUTLET
INLET
Outline Introduction Model and case study Uncertainty analysis Scenario analysis Conclusion
Scenario Analysis Comparison of pollution control strategies The models can be used to compare different MP emission control strategies Two scenario were simulated Scenario A: disconnection of 50% of the roof areas and 30% of the roads and parking areas.
Disconnection of 40% of the catchment impervious area Scenario B: doubling of the pond volume and modification of layout
Doubling of nominal hydraulic residence time (HRT) and increase of effective HRT
Scenario Analysis Long-term results Simulation with a 10-year rainfall series: Scenario A (catchment disconnection) Lower loads to the pond Better settling condition Scenario B (pond improvement) Slight improvement of removal efficiency Dissolved fraction not affected
Source control strategy seems preferable to pond enlargment
Scenario Analysis Impact on downstream environment Effects on downstream aquatic environment Baseline scenario
Emission Limit Value for Cudiss
Dissolved fraction is not affected by the two scenarios Risk for acquatic environment is not decreased
Pond improvement should focus on: Better layout (lower outlet peaks) Removal of dissolved fraction (e.g. filtration systems)
Outline Introduction Model and case study Uncertainty analysis Scenario analysis Conclusion
Conclusions An integrated dynamic model for estimation of MP fluxes in stormwater systems is now available The flexibility of the proposed models can simulate a wide range of substances in various catchments The integrated model can provide a support for scenario analysis and comparison od pollution control strategies Modelling of stormwater MP fluxes requires the use of uncertainty analysis methods
More on this topic at my PhD defence:
Tuesday 29th March at DTU
References Beven, K.J., Binley A. (1992) Future of distributed models: Model calibration and uncertainty prediction, Hydrological Processes, 6(3), 279-298 Vezzaro, L., Eriksson, E., Ledin, A., & Mikkelsen, P. S. (2010a) Dynamic stormwater treatment unit model for micropollutants (STUMP) based on inherent properties. Water Science and Technology, 62(3), 622-629 Vezzaro, L., Ledin, A., Mikkelsen, P.S. (2010b). Integrated modelling of priority pollutants in stormwater systems. In: Proceedings of IDRA 2010. XXXII Italian Conference of Hydraulics and Hydraulic Constructions, Palermo, Italy, 14th-17th September 2010.
Outline Introduction Model and case study Uncertainty analysis Scenario analysis Conclusion
Stormwater micropollutants
Which substances are we talking about?
33 Substances listed in the EU Directives 2000/60/CE and 2008/105/CE (PP = Priority Pollutants)
+
Any other micropollutant (MP)
Elements of the stormwater system
Catchment
Drainage System
Treatment System
Integrated stormwater quality model
Models in stormwater quality management What can models be used for? Looking at the Driver-Pressure-State-Impact-Response framework (EEA, 1999)
Rainfall Urban activities
Models Micro Pollutants (MP)
Models
Models
Control strategies
Models Models
Water quality
Toxicity
Outline Introduction Model and case study Uncertainty analysis Scenario analysis Conclusion
Source characterization
Lumped characterization
Detailed description Copper roof
Traffic area Commercial
Residential
Highway
Roof
Residential street
Parking lot
Source control End-of-pipe treatment
Pollutant release and transport
STUMP model Stormwater Treatment Unit for MicroPollutants (Vezzaro et al., 2010a) Serial CSTR Number of tanks = same hydraulic behaviour of the treatment unit
Pseudo First order kinetics Fate processes based on substance’s inherent properties = Wide range of substance Easily retrievable data
INPUT
OUTPUT
Integrated model Sources+release+treatment
Emission-based
Dynamic acc/washoff
STUMP
Several source of uncertainty!! GLUE - Generalized Likelihood Uncertainty Estimation (Beven and Binley, 1992)
Case Study The Hersted industripark catchment Industrial-residential area in the Albertslund Kommune (~ 94.7 ha) Measurements 14 months flow data Flow proportional samples (inlet), time proportional (outlet)
Industrial area
Residential area
Outline Introduction Model and case study Uncertainty analysis Scenario analysis Conclusion
Analysis of model performance Pond inlet (hydraulic) Flow predictions are affected by the rainfall input data
Rainfall, but no flow
No rainfall, but flow
Delay in measured rainfall- measured flow
Analysis of model performance Pond inlet (hydraulic) Rainfall corrected according to position of rain gauge (3 km away) Vezzaro et al., 2010b
Delay is removed
Coverage of 40.3% of flow data
Analysis of model performance Pond inlet (quality) 5 rain events (33 samples) One extreme event affects calibration
Coverage of 74.3% of TSS data
Coverage of 82.9% of Cu data
Analysis of model performance Pond outlet (quality) Modelled peaks smoother than measured Pond hydraulic short-circuit higher than expected
Coverage of 53.3% of TSS data
Coverage of 35.7% of Cu data
OUTLET
INLET
Outline Introduction Model and case study Uncertainty analysis Scenario analysis Conclusion
Scenario Analysis Comparison of pollution control strategies The models can be used to compare different MP emission control strategies Two scenario were simulated Scenario A: disconnection of 50% of the roof areas and 30% of the roads and parking areas.
Disconnection of 40% of the catchment impervious area Scenario B: doubling of the pond volume and modification of layout
Doubling of nominal hydraulic residence time (HRT) and increase of effective HRT
Scenario Analysis Long-term results Simulation with a 10-year rainfall series: Scenario A (catchment disconnection) Lower loads to the pond Better settling condition Scenario B (pond improvement) Slight improvement of removal efficiency Dissolved fraction not affected
Source control strategy seems preferable to pond enlargment
Scenario Analysis Impact on downstream environment Effects on downstream aquatic environment Baseline scenario
Emission Limit Value for Cudiss
Dissolved fraction is not affected by the two scenarios Risk for acquatic environment is not decreased
Pond improvement should focus on: Better layout (lower outlet peaks) Removal of dissolved fraction (e.g. filtration systems)
Outline Introduction Model and case study Uncertainty analysis Scenario analysis Conclusion
Conclusions An integrated dynamic model for estimation of MP fluxes in stormwater systems is now available The flexibility of the proposed models can simulate a wide range of substances in various catchments The integrated model can provide a support for scenario analysis and comparison od pollution control strategies Modelling of stormwater MP fluxes requires the use of uncertainty analysis methods
More on this topic at my PhD defence:
Tuesday 29th March at DTU
References Beven, K.J., Binley A. (1992) Future of distributed models: Model calibration and uncertainty prediction, Hydrological Processes, 6(3), 279-298 Vezzaro, L., Eriksson, E., Ledin, A., & Mikkelsen, P. S. (2010a) Dynamic stormwater treatment unit model for micropollutants (STUMP) based on inherent properties. Water Science and Technology, 62(3), 622-629 Vezzaro, L., Ledin, A., Mikkelsen, P.S. (2010b). Integrated modelling of priority pollutants in stormwater systems. In: Proceedings of IDRA 2010. XXXII Italian Conference of Hydraulics and Hydraulic Constructions, Palermo, Italy, 14th-17th September 2010.
