Robustness-Optimality Tradeoff for Watershed Load Reduction Decision Making under Deep Uncertainty

Su, Han; Dong, Feifei; Liu, Yong; Zou, Rui; Guo, Hao

doi:10.1007/s11269-017-1689-3

Cited by 2 publications

(2 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Deep learning, a new generation of artificial neural network algorithms, has the potential to help us unearth new feedback loops and processes that govern BMP effectiveness and diffuse nutrient pollution . Similarly, deep learning could be used to consider and assess multiple possible futures. , There are new suites of machine learning approaches that use deep equifinality to unearth new understandings of processes and feedback loops within catchment-scale water quality processes . However, as Lakkaraju et al highlight, it is critical to be aware that training data will bias and influence outcomes.…”

Section: Discussionmentioning

confidence: 99%

“…184 Similarly, deep learning could be used to consider and assess multiple possible futures. 185,186 There are new suites of machine learning approaches that use deep equifinality to unearth new understandings of processes and feedback loops within catchment-scale water quality processes. 187 However, as Lakkaraju et al 183 highlight, it is critical to be aware that training data will bias and influence outcomes.…”

Section: Lack Of Knowledge About Bmp Functionmentioning

confidence: 99%

See 1 more Smart Citation

Best Management Practices for Diffuse Nutrient Pollution: Wicked Problems Across Urban and Agricultural Watersheds

Lintern

McPhillips

Winfrey

et al. 2020

Environ. Sci. Technol.

125

View full text Add to dashboard Cite

Extensive time and financial resources have been dedicated to address nonpoint sources of nitrogen and phosphorus in watersheds. Despite these efforts, many watersheds have not seen substantial improvement in water quality. The objective of this study is to review the literature and investigate key factors affecting the lack of improvement in nutrient levels in waterways in urban and agricultural regions. From 94 studies identified in the academic literature, we found that, although 60% of studies found improvements in water quality after implementation of Best Management Practices (BMPs) within the watershed, these studies were mostly modeling studies rather than field monitoring studies. For studies that were unable to find improvements in water quality after the implementation of BMPs, the lack of improvement was attributed to lack of knowledge about BMP functioning, lag times, nonoptimal placement and distribution of BMPs in the watershed, postimplementation BMP failure, and socio-political and economic challenges. We refer to these limiting factors as known unknowns. We also acknowledge the existence of unknown unknowns that hinder further improvement in BMP effectiveness and suggest that machine learning, approaches from the field of business and operations management, and long-term convergent studies could be used to resolve these unknown unknowns.

show abstract