Development of Water Quality Models to Support the European Union Bathing Water Directive in Ireland
Keywords
Bathing Water Directive, Predictive Modelling, Machine Learning, Water Quality, Ireland
Start Date
15-9-2020 12:40 PM
End Date
15-9-2020 1:00 PM
Abstract
The European Bathing Water Directive (BWD; 76/160/EEC 2006) requires the implementation of early warning systems for bathing waters which are subject to short-term pollution events. To this end, the EU SWIM project is developing coastal water quality prediction models and alert systems at nine beach sites in the Republic of Ireland and Northern Ireland, which represent a range of baseline water quality and site conditions. At each site, statistical / machine-learning models are being developed based on their site-specific relationships between faecal indicator bacteria and multiple environmental variables, whose predictions are then transmitted to the public via multiple means (e.g. online, electronic signage). This presentation provides an overview of the end-to-end system under development and a summary of the underlying data and models being used to drive predictions. Following this overview, three key topics related to modelling for support of the EU Bathing Water Directive are discussed: (1) the need for an acceptable “baseline model” to serve as a benchmark for model performance and to improve model development; (2) the importance of utilizing model performance metrics which are interpretable and relevant to stakeholders (i.e. beach managers) to improve model transparency and uptake; and (3) the need to effectively communicate the inherent trade-off between model sensitivity and specificity (i.e. true versus false alarm rates) to establish an acceptable level of model performance on a site-by-site basis. This presentation offers a “model developer perspective” on the challenges and opportunities of these topics and puts forward proposed solutions to better address them.
Development of Water Quality Models to Support the European Union Bathing Water Directive in Ireland
The European Bathing Water Directive (BWD; 76/160/EEC 2006) requires the implementation of early warning systems for bathing waters which are subject to short-term pollution events. To this end, the EU SWIM project is developing coastal water quality prediction models and alert systems at nine beach sites in the Republic of Ireland and Northern Ireland, which represent a range of baseline water quality and site conditions. At each site, statistical / machine-learning models are being developed based on their site-specific relationships between faecal indicator bacteria and multiple environmental variables, whose predictions are then transmitted to the public via multiple means (e.g. online, electronic signage). This presentation provides an overview of the end-to-end system under development and a summary of the underlying data and models being used to drive predictions. Following this overview, three key topics related to modelling for support of the EU Bathing Water Directive are discussed: (1) the need for an acceptable “baseline model” to serve as a benchmark for model performance and to improve model development; (2) the importance of utilizing model performance metrics which are interpretable and relevant to stakeholders (i.e. beach managers) to improve model transparency and uptake; and (3) the need to effectively communicate the inherent trade-off between model sensitivity and specificity (i.e. true versus false alarm rates) to establish an acceptable level of model performance on a site-by-site basis. This presentation offers a “model developer perspective” on the challenges and opportunities of these topics and puts forward proposed solutions to better address them.
Stream and Session
false