Multi-model data fusion for river flow forecasting: an evaluation of six alternative methods based on two contrasting catchments

Abrahart, Robert J.; See, Linda

doi:10.5194/hess-6-655-2002

Cited by 127 publications

(79 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This was achieved in both cases regardless of the number of merged members. This statement is consistent with Abrahart and See (2002) who showed that the most efficient data-fusion methods depended on the particular application case. However, as Georgakakos et al (2004) demonstrated, the simple mean of five model predictions consistently outperformed the best single model prediction in several catchments, it was only true at Sävja in our study.…”

Section: Discussionsupporting

confidence: 79%

Multi-model data fusion as a tool for PUB: example in a Swedish mesoscale catchment

et al. 2011

View full text Add to dashboard Cite

Abstract. Post-processing the output of different rainfallrunoff models allows one to pool strengths of each model to produce more reliable predictions. As a new approach in the frame of the "Prediction in Ungauged Basins" initiative, this study investigates the geographical transferability of different parameter sets and data-fusion methods which were applied to 5 different rainfall-runoff models for a low-land catchment in Central Sweden. After usual calibration, we adopted a proxy-basin validation approach between two similar but non-nested sub-catchments in order to simulate ungauged conditions.Many model combinations outperformed the best single model predictions with improvements of efficiencies from 0.70 for the best single model predictions to 0.77 for the best ensemble predictions. However no "best" data-fusion method could be determined as similar performances were obtained with different merging schemes. In general, poorer model performance, i.e. lower efficiency, was less likely to occur for ensembles which included more individual models.

show abstract

Section: Discussionsupporting

confidence: 79%

Multi-model data fusion as a tool for PUB: example in a Swedish mesoscale catchment

et al. 2011

View full text Add to dashboard Cite

show abstract

“…Since then there have been several more studies which have dealt with multi-model combination of hydrological models (e.g. (Abrahart and See 2002, Ajami, et al 2006, Coulibaly, et al 2005, Hsu, et al 2009, See and Openshaw 2000, Shamseldin, et al 2007, Viney, et al 2009, Xiong, et al 2001). As the nature of the combination function is unknown and no theory exists to analytically derive the combination function from a hydrological or physical point of view, previous studies have used empirical data-driven modeling to derive the combination function and such use is very appropriate.…”

Section: Introductionmentioning

confidence: 99%

Use of Gene Expression Programming for Multimodel Combination of Rainfall-Runoff Models

2012

View full text Add to dashboard Cite

This paper deals with the application of an innovative method for combining estimated outputs from a number of rainfall-runoff models using Gene Expression Programming (GEP) to perform symbolic regression. The GEP multi-model combination method uses the synchronous simulated river flows from four conventional rainfall-runoff models to produce a set of combined river flow estimates for four different catchments.The four selected models for the multi-model combinations are the Linear Perturbation Model (LPM), the Linearly Varying Gain Factor Model (LVGFM), the Soil Moisture Accounting and Routing (SMAR) Model, and the Probability-Distributed Interacting Storage Capacity (PDISC) model. The first two of these models are 'black-box' models, the LPM exploiting seasonality and the LVGFM employing a storage-based coefficient of runoff. The remaining two are conceptual models. The data of four catchments with different geographical location, hydrological and climatic conditions, are used to test the performance of the GEP combination method.The results of the model using GEP method are compared against original forecasts obtained from the individual models that contributed to the development of the combined model by means of a few global statistics. The findings show that a GEP approach can successfully used as a multi-model combination method. In addition, the GEP combination method also has benefit over other hitherto tested approaches such as an artificial neural network combination method in that its formulation is transparent, can be expressed as a simple mathematical function, and therefore is useable by people who are unfamiliar with such advanced techniques. The GEP combination method is able to combine model outcomes from less accurate individual models and produce a superior river flow forecast.

show abstract

“…See and Abrahart [2001] used ANN data fusion strategies for continuous river level forecasting where data fusion is the amalgamation of information from multiple sensors and/or different data sources. Abrahart and See [2002] evaluated six data fusion strategies and found that ANN data fusion provided the best solution for a stable region. Cannon and Whitfield [2002] used a bootstrap aggregated ANN ensemble as a downscaling model to predict streamflow and changes in streamflow conditions in British Columbia, Canada.…”

Section: Introductionmentioning

confidence: 99%

Artificial neural network ensembles and their application in pooled flood frequency analysis

Shu

Burn

2004

Water Resources Research

170

125

View full text Add to dashboard Cite

[1] Recent theoretical and empirical studies show that the generalization ability of artificial neural networks can be improved by combining several artificial neural networks in redundant ensembles. In this paper, a review is given of popular ensemble methods. Six approaches for creating artificial neural network ensembles are applied in pooled flood frequency analysis for estimating the index flood and the 10-year flood quantile. The results show that artificial neural network ensembles generate improved flood estimates and are less sensitive to the choice of initial parameters when compared with a single artificial neural network. Factors that may affect the generalization of an artificial neural network ensemble are analyzed. In terms of the methods for creating ensemble members, the model diversity introduced by varying the initial conditions of the base artificial neural networks to reduce the prediction error is comparable with more sophisticated methods, such as bagging and boosting. When the same method for creating ensemble members is used, combining member networks using stacking is generally better than using simple averaging. An ensemble size of at least 10 artificial neural networks is suggested to achieve sufficient generalization ability. In comparison with parametric regression methods, properly designed artificial neural network ensembles can significantly reduce the prediction error.

show abstract

Multi-model data fusion for river flow forecasting: an evaluation of six alternative methods based on two contrasting catchments

Cited by 127 publications

References 36 publications

Multi-model data fusion as a tool for PUB: example in a Swedish mesoscale catchment

Multi-model data fusion as a tool for PUB: example in a Swedish mesoscale catchment

Use of Gene Expression Programming for Multimodel Combination of Rainfall-Runoff Models

Artificial neural network ensembles and their application in pooled flood frequency analysis

Contact Info

Product

Resources

About