A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Deriving a Bayesian Network to Assess the Retention Efficacy of Riparian Buffer Zones
Bayesian networks (BN) have increasingly been applied in water management but not to estimate the efficacy of riparian buffer zones (RBZ). Our methodical study aims at evaluating the first BN to predict the RBZ efficacy to retain sediment and nutrients (dissolved, total, and particulate nitrogen and phosphorus) from widely available variables (width, vegetation, slope, soil texture, flow pathway, nutrient form). To evaluate the influence of parent nodes and how the number of states affects prediction errors, we used a predefined general BN structure, collected 580 published datasets from North America and Europe, and performed classification tree analyses and multiple 10-fold cross-validations of different BNs. These errors ranged from 0.31 (two output states) to 0.66 (five states). The outcome remained unchanged without the least influential nodes (flow pathway, vegetation). Lower errors were achieved when parent nodes had more than two states. The number of efficacy states influenced most strongly the prediction error as its lowest and highest states were better predicted than intermediate states. While the derived BNs could support or replace simple design guidelines, they are limited for more detailed predictions. More representative data on vegetation or additional nodes like preferential flow will probably improve the predictive power.
Deriving a Bayesian Network to Assess the Retention Efficacy of Riparian Buffer Zones
Bayesian networks (BN) have increasingly been applied in water management but not to estimate the efficacy of riparian buffer zones (RBZ). Our methodical study aims at evaluating the first BN to predict the RBZ efficacy to retain sediment and nutrients (dissolved, total, and particulate nitrogen and phosphorus) from widely available variables (width, vegetation, slope, soil texture, flow pathway, nutrient form). To evaluate the influence of parent nodes and how the number of states affects prediction errors, we used a predefined general BN structure, collected 580 published datasets from North America and Europe, and performed classification tree analyses and multiple 10-fold cross-validations of different BNs. These errors ranged from 0.31 (two output states) to 0.66 (five states). The outcome remained unchanged without the least influential nodes (flow pathway, vegetation). Lower errors were achieved when parent nodes had more than two states. The number of efficacy states influenced most strongly the prediction error as its lowest and highest states were better predicted than intermediate states. While the derived BNs could support or replace simple design guidelines, they are limited for more detailed predictions. More representative data on vegetation or additional nodes like preferential flow will probably improve the predictive power.
Deriving a Bayesian Network to Assess the Retention Efficacy of Riparian Buffer Zones
Andreas Gericke (author) / Hong Hanh Nguyen (author) / Peter Fischer (author) / Jochem Kail (author) / Markus Venohr (author)
2020
Article (Journal)
Electronic Resource
Unknown
Metadata by DOAJ is licensed under CC BY-SA 1.0
Deriving a Bayesian Network to Assess the Retention Efficacy of Riparian Buffer Zones
| BASE | 2020
LiDAR-based predictions of flow channels through riparian buffer zones
| Taylor & Francis Verlag | 2015
Establishment and management of riparian buffer zones in Han River basin, Korea
| British Library Conference Proceedings | 2001
Urban Riparian Zones: Designing Mitigation Buffers
| British Library Conference Proceedings | 2014
Riparian Buffer Widths at Rocky Mountain Resorts
| British Library Conference Proceedings | 1997