A functional framework for flow-duration-curve and daily streamflow estimation at ungauged sites

Requena, Ana I.; Chebana, Fateh; Ouarda, Taha B. M. J.

doi:10.1016/j.advwatres.2018.01.019

Cited by 23 publications

(11 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This could suggest that masking and interaction effects play a role in the estimated effect of basin slope. In Requena, Chebana, and Ouarda (), the coefficient of each basin characteristic was multiplied by the value of the basin characteristic for a given site, obscuring a one‐to‐one comparison with our results.…”

Section: Discussionmentioning

confidence: 96%

“…This is usually mitigated by interpolating a value to replace the violating quantile or simply resorting the predicted quantiles. More recently, work by Poncelet et al () and Requena, Chebana and Ouarda () introduced methods (quantile solidarity and functional regression, respectively) that leverage the physical relation between quantiles and basin characteristics to ensure monotonicity in the quantile estimates. Both quantile solidarity and functional regression offer improvements over methods that ignore the information contained in the relation among quantiles in addition to the relation between each quantile and its associated basin characteristics.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Prediction and Inference of Flow Duration Curves Using Multioutput Neural Networks

Worland

Asquith

Knight

et al. 2019

Water Resources Research

View full text Add to dashboard Cite

We develop multioutput neural network models to predict flow‐duration curves (FDCs) in 9,203 ungaged locations in the Southeastern United States for six decades between 1950 and 2009. The model architecture contains multiple response variables in the output layer that correspond to individual quantiles along the FDC. During training, predictions are made for each quantile, and a combined loss function is used for back propagation and parameter updating. The loss function accounts for the covariance between the quantiles and generates physically consistent outputs (i.e., monotonically increasing quantiles with increasing nonexceedance probabilities). We use neural network dropout to generate posterior‐predictive distributions for FDCs and test model performance under cross validation. Finally, we demonstrate how local surrogate models, via the Local Interpretable Model‐agnostic Explanations method, can be used to infer the relation between basin characteristics and the predicted FDCs. Results suggest that multioutput neural network models can learn the monotonic relations between adjacent quantiles on an FDC; they result in better predictions than single‐output neural network models that predict each quantile independently, and basin characteristics are most useful for predicting smaller quantiles, whereas bias terms from neighboring quantiles are most informative for predicting higher quantiles.

show abstract

Section: Discussionmentioning

confidence: 96%

Section: Introductionmentioning

confidence: 99%

Prediction and Inference of Flow Duration Curves Using Multioutput Neural Networks

Worland

Asquith

Knight

et al. 2019

Water Resources Research

View full text Add to dashboard Cite

show abstract

“…Both the horizontal axis and vertical axis are constructed at a double log scale for viewing purposes. let et al, 2017;Padoan et al, 2010;Russell et al, 2016;Thibaud et al, 2013;Westra and Sisson, 2011), and latentvariable Gaussian models (Bennett et al, 2016b). The ability to simulate rainfall over a region means that hydrological problems need not be confined to individual catchments, but they may cover multiple catchments.…”

Section: Discussionmentioning

confidence: 99%

Spatially dependent flood probabilities to support the design of civil infrastructure systems

Leonard

Westra

2019

Hydrol. Earth Syst. Sci.

View full text Add to dashboard Cite

Abstract. Conventional flood risk methods typically focus on estimation at a single location, which can be inadequate for civil infrastructure systems such as road or railway infrastructure. This is because rainfall extremes are spatially dependent; to understand overall system risk, it is necessary to assess the interconnected elements of the system jointly. For example, when designing evacuation routes it is necessary to understand the risk of one part of the system failing given that another region is flooded or exceeds the level at which evacuation becomes necessary. Similarly, failure of any single part of a road section (e.g., a flooded river crossing) may lead to the wider system's failure (i.e., the entire road becomes inoperable). This study demonstrates a spatially dependent intensity–duration–frequency (IDF) framework that can be used to estimate flood risk across multiple catchments, accounting for dependence both in space and across different critical storm durations. The framework is demonstrated via a case study of a highway upgrade comprising five river crossings. The results show substantial differences in conditional and unconditional design flow estimates, highlighting the importance of taking an integrated approach. There is also a reduction in the estimated failure probability of the overall system compared with the case where each river crossing is treated independently. The results demonstrate the potential uses of spatially dependent intensity–duration–frequency methods and suggest the need for more conservative design estimates to take into account conditional risks.

show abstract

“…Chebana et al (2012) were the first to introduce FDA in the hydrological context, seeing the annual series of streamflow, the hydrograph, as a curve, that is, a functional datum. Many recent studies have applied FDA to the streamflow variable (i.e., Masselot et al 2016;Ternynck et al 2016;Brunner et al 2017;Larabi et al 2017;Requena et al 2018), whereas no attempt has been made to model the stream temperature curve. Similarly to streamflow, where the hydrograph was seen as a curve over a year (Chebana et al 2012), stream temperature can also be adequately represented using a single annual curve, especially given its pronounced seasonal trend.…”

Section: Introductionmentioning

confidence: 99%

Stream Temperature Modeling Using Functional Regression Models

Boudreault

Bergeron

St‐Hilaire

et al. 2019

J American Water Resour Assoc

Self Cite

View full text Add to dashboard Cite

Stream temperature is one of the most important environmental variables in lotic habitats as it has important and direct impacts on the ecosystem. Given the continuous nature of this variable, the aim of this paper was to introduce functional regression for the air‐stream temperature relation, being capable to model an entire seasonal or annual curve of temperatures as one entity, rather than multiple daily or weekly values in classical models. Three types of functional models were explored in the study and compared to two classical models (Generalized Additive Model and Logistic Model) for six rivers from the United States The results show the functional models have the best performance for all the considered rivers. When comparing functional models between them, one variant of the historical functional model performs better than the two other models and is the most parsimonious. Functional regression leads to encouraging results to model the complete annual stream temperature curve as one entity compared to other classical approaches.

show abstract

A functional framework for flow-duration-curve and daily streamflow estimation at ungauged sites

Cited by 23 publications

References 32 publications

Prediction and Inference of Flow Duration Curves Using Multioutput Neural Networks

Prediction and Inference of Flow Duration Curves Using Multioutput Neural Networks

Spatially dependent flood probabilities to support the design of civil infrastructure systems

Stream Temperature Modeling Using Functional Regression Models

Contact Info

Product

Resources

About