Regional flood frequency estimation with canonical correlation analysis

Regional flood frequency analysis at ungauged sites using the adaptive neuro-fuzzy inference system

Non-stationary regional flood frequency analysis at ungauged sites

Estimation of water quality characteristics at ungauged sites using artificial neural networks and canonical correlation analysis

[1] Models based on canonical correlation analysis (CCA) and artificial neural networks (ANNs) are developed to obtain improved flood quantile estimates at ungauged sites. CCA is used to form a canonical physiographic space using the site characteristics from gauged sites. Then ANN models are applied to identify the functional relationships between flood quantiles and the physiographic variables in the CCA space. Two ANN models, the single ANN model and the ensemble ANN model, are developed. The proposed approaches are applied to 151 catchments in the province of Quebec, Canada. Two evaluation procedures, the jackknife validation procedure and the split sample validation procedure, are used to evaluate the performance of the proposed models. Results of the proposed models are compared with the original CCA model, the canonical kriging model, and the original ANN models. The results indicate that the CCA-based ANN models provide superior estimation than the original ANN models. The ANN ensemble approaches provide better generalization ability than the single ANN models. The CCA-based ensemble ANN model has the best performance among all models in terms of prediction accuracy.

https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/2006WR005142

Flood frequency analysis at ungauged sites using artificial neural networks in canonical correlation analysis physiographic space

It is often necessary to estimate extreme events at sites where little or no hydrometric data are available. In such cases, one may use a regional estimation procedure, utilizing data available from other stations in the same hydrologic region. In general, a regional flood frequency procedure consists of two steps, delineation of hydrologically homogeneous regions and regional estimation. This paper focuses on the development of a regional flood frequency procedure based on canonical correlation analysis (CCA) and its application to data from a northern Canadian basin in which floods are dominated by spring snowmelt. This CCA-based procedure allows the joint regional estimation of spring flood peaks and volumes. The CCA method allows the determination of pairs of canonical variables such that the correlation between the canonical variables of one pair is maximized and between the variables of different pairs is equal to zero. Therefore, it is possible to infer hydrological canonical variables, knowing the physiographical-meteorological canonical variables. The methodology developed was applied to the St. Maurice River basin system, which is operated by Hydro-Quebec and characterized by the relatively low precision of flow data available. Results show that the proposed method allows for a significant reduction in the 100-year spring flood and volume quantile estimation bias and mean square error. The study also shows that, in 60% of cases, the method that was previously used overestimates quantile values, which leads to an overdesign of retention structures.

Mario Haché

Papers

Regional flood peak and volume estimation in northern Canadian basin