Three types of functionally different artificial neural network (ANN) models are calibrated using a relatively short length groundwater level records and related hydrometeorological data to simulate water table fluctuations in the Gondo aquifer, Burkina Faso. Input delay (IDNN) with static memory structure globally recurrent (RNN) inherent dynamical proposed for monthly modeling. The simulation performance IDNN RNN is compared results obtained from two variants radial basis function (RBF)...

This paper presents a comprehensive review of fundamental and challenging issue in hydrology: the regionalization streamflow its advances over last two decades, specifically 1990–2011. includes discussion developments continuous regionalization, model parameter optimization methods, application uncertainty analysis procedures, limitations challenges, future research directions. Here, refers to process transferring hydrological information from gauged ungauged or poorly basins estimate...

This paper traces two decades of neural network rainfall-runoff and streamflow modelling, collectively termed ‘river forecasting’. The field is now firmly established the research community involved has much to offer hydrological science. First, however, it will be necessary converge on more objective consistent protocols for: selecting treating inputs prior model development; extracting physically meaningful insights from each proposed solution; improving transparency in benchmarking...

Regionalization, a process of transferring hydrological information [i.e., parameters conceptual rainfall-runoff model, namely, the McMaster University-Hydrologiska Byråns Vattenbalansavdelning (MAC-HBV)] from gauged to ungauged basins, was applied estimate continuous flows in basins across Ontario, Canada. To identify appropriate regionalization methods, different methods were applied, including spatial proximity kriging, inverse distance weighted (IDW), and mean parameters], physical...

This paper examines the potential of support vector machine (SVM) in long-term prediction lake water levels. Lake Erie mean monthly levels from 1918 to 2001 are used predict future up 12months ahead. The results compared with a widely neural network model called multilayer perceptron (MLP) and conventional multiplicative seasonal autoregressive (SAR). Overall, SVM showed good performance is proved be competitive MLP SAR models. For 3- 12-month-ahead prediction, outperforms two other models...

Wavelet analysis is used to identify and describe variability in annual Canadian streamflows gain insights into the dynamical link between dominant modes of climate Northern Hemisphere. Results from applying continuous wavelet transform mean 79 rivers selected Reference Hydrometric Basin Network (RHBN) reveal striking climate‐related features before 1950s after 1970s. The span available observations, 1911–1999, allows for depicting variance periods up 12 years. Scale‐averaged power spectra...

Abstract The issues of downscaling the outputs a global climate model (GCM) to scale that is appropriate hydrological impact studies are investigated using temporal neural network approach. time-lagged feed-forward (TLFN) proposed for daily total precipitation and maximum minimum temperature series Serpent River watershed in northern Quebec (Canada). models developed validated large-scale predictor variables derived from National Centers Environmental Prediction–National Center Atmospheric...

Abstract. In Canada, risk of flooding due to heavy rainfall has risen in recent decades; the most notable examples include July 2013 storm Greater Toronto region and May 2017 flood Islands. We investigate nonstationarity trends short-duration precipitation extremes selected urbanized locations Southern Ontario, evaluate potential nonstationary intensity–duration–frequency (IDF) curves, which form an input civil infrastructural design. Despite apparent signals all locations, stationary vs....

Recent years have witnessed considerable developments in multiple fields with the potential to enhance our capability of forecasting pluvial flash floods, one most costly environmental hazards terms both property damage and loss life. This work provides a summary description recent advances related insights on atmospheric conditions that precede extreme rainfall events, development monitoring systems relevant hydrometeorological parameters, operational adoption weather hydrological models...

Waterford River watershed, St. John's, Newfoundland and Labrador (NL), Canada. This study investigates five hydrological models to identify adequate model(s) for operational flood forecasting at watershed. These included three lumped conceptual (SAC-SMA: Sacramento Soil Moisture Accounting, GR4J: modèle du Génie Rural à 4 paramètres Journalier, MAC-HBV: McMaster University Hydrologiska Byråns Vattenbalansavdelning), a semi-distributed model (HEC-HMS: Hydrologic Engineering Center's Modeling...

Accurate forecasting in hydrologic modeling is crucial for sustainable water resource management across various sectors, where predicting extreme flow phases holds particular significance due to their severe impact on the territory. Due inherent uncertainties forecasting, relying solely a single rainfall–runoff model may not provide reliable predictions. To address this challenge, over years, researchers have developed and applied forecast merging (HFM) techniques that combine multiple...

An experiment on predicting multivariate water resource time series, specifically the prediction of hydropower reservoir inflow using temporal neural networks, is presented. This paper focuses dynamic networks to address relationships hydrological series. Three types network architectures with different inherent representations information are investigated. input delayed (IDNN) and a recurrent (RNN) without delays proposed for forecasting. The forecast results indicate that, overall, RNN...

In this paper, a Bayesian learning approach is introduced to train multilayer feed‐forward network for daily river flow and reservoir inflow simulation in cold region basin Canada. approach, uncertainty about the relationship between inputs outputs initially taken care of by an assumed prior distribution parameters (weights biases). This updated posterior using likelihood function following Bayes' theorem while data are observed. called objective approach. The maximized suitable optimization...

The objectives of this study are to describe the local interannual variability in southern Québec, Canada, streamflow, based on wavelet analysis, and identify plausible climatic teleconnections that could explain these variations. Scale-averaged power spectra used simultaneously assess spatial 18 contiguous annual streamflow time series. span available observations, 1938–2000, allows depicting variance for periods up about 12 yr. most striking feature, 2–3-yr band 3–6-yr band—the 6–12-yr is...