Keywords
Candidate models, environmental impact assessment, model structure sensitivity, urban drainage impacts, river water quality modeling
Location
Session C1: Compexity, Sensitivity, and Uncertainty Issues in Integrated Environmental Models
Start Date
16-6-2014 2:00 PM
End Date
16-6-2014 3:20 PM
Abstract
Numerical modeling of physicochemical conditions in rivers influenced by urban drainage and other pressures is increasingly used as supportive method for an integrative environmental impact assessment. In various European countries, protocols for water quality based impact assessment (WQA protocols) are in use, of which many propse the use of water quality models. Despite the increased effort to study uncertainty issues in river water quality modeling in recent years, identifying and differentiating model structure uncertainty remains a challenging task. This study elaborates upon a key conflict in model development: the need to simplify and to still ensure structural adequacy to obtain reliable modeling results. The paper evaluates the adequacy of diverse river water quality modeling approaches when subjected to a changing model structure. This evaluation is achieved by applying a set of calibrated candidate models to two different river case studies under varying pollution conditions. The term 'model structure sensitivity' is introduced to quantify the output variation as a result of a changing model structure. Sensitivity is here, in contrast to previous works, positively interpreted as model flexibility. The study illustrates that the interdependence between model sensitivity and model error is conditional upon complexity, but model adequacy differs depending on the pollution dynamics in the modeled system. Results show that model structure uncertainty has a composite nature: effects related to transport and conversion model contribute with varying shares. Two actions should be considered to improve model structure characterization: i) use of a variety of model structures combined with an analysis of structure and parameter sensitivity, and ii) acquisition of high-resolution reference data to capture varying pollution load dynamics.
Included in
Civil Engineering Commons, Data Storage Systems Commons, Environmental Engineering Commons, Hydraulic Engineering Commons, Other Civil and Environmental Engineering Commons
Model structure sensitivity of river water quality models for urban drainage impact assessment
Session C1: Compexity, Sensitivity, and Uncertainty Issues in Integrated Environmental Models
Numerical modeling of physicochemical conditions in rivers influenced by urban drainage and other pressures is increasingly used as supportive method for an integrative environmental impact assessment. In various European countries, protocols for water quality based impact assessment (WQA protocols) are in use, of which many propse the use of water quality models. Despite the increased effort to study uncertainty issues in river water quality modeling in recent years, identifying and differentiating model structure uncertainty remains a challenging task. This study elaborates upon a key conflict in model development: the need to simplify and to still ensure structural adequacy to obtain reliable modeling results. The paper evaluates the adequacy of diverse river water quality modeling approaches when subjected to a changing model structure. This evaluation is achieved by applying a set of calibrated candidate models to two different river case studies under varying pollution conditions. The term 'model structure sensitivity' is introduced to quantify the output variation as a result of a changing model structure. Sensitivity is here, in contrast to previous works, positively interpreted as model flexibility. The study illustrates that the interdependence between model sensitivity and model error is conditional upon complexity, but model adequacy differs depending on the pollution dynamics in the modeled system. Results show that model structure uncertainty has a composite nature: effects related to transport and conversion model contribute with varying shares. Two actions should be considered to improve model structure characterization: i) use of a variety of model structures combined with an analysis of structure and parameter sensitivity, and ii) acquisition of high-resolution reference data to capture varying pollution load dynamics.
