Article

Improving extreme hydrologic events forecasting using a new criterion for ANN selection

May 2001
Hydrological Processes 15(8):1533 - 1536

May 2001
15(8):1533 - 1536

DOI:10.1002/hyp.445

Authors:

Paulin Coulibaly

McMaster University

Bernard Bobée

Institut National de la Recherche Scientifique

The issue of selecting appropriate model input parameters is addressed using a peak and low flow criterion (PLC). The optimal artificial neural network (ANN) models selected using the PLC significantly outperform those identified with the classical root-mean-square error (RMSE) or the conventional Nash–Sutcliffe coefficient (NSC) statistics. The comparative forecast results indicate that the PLC can help to design an appropriate ANN model to improve extreme hydrologic events (peak and low flow) forecast accuracy. Copyright © 2001 John Wiley & Sons, Ltd.

The Effects of Climate Change and Extreme Wildfire Events on Runoff Erosion over a Mountain Watershed

Article

Full-text available

Feb 2016
J HYDROL

Increases in wildfire occurrence and severity under an altered climate can substantially impact terrestrial ecosystems through enhancing runoff erosion. Improved prediction tools that provide high resolution spatial information are necessary for location-specific soil conservation and watershed management. However, quantifying the magnitude of soil erosion and its interactions with climate, hydrological processes, and fire occurrences across a large region (>10,000 km2) is challenging because of the large computational requirements needed to capture the fine-scale complexities of the land surface that govern erosion. We apply the physically-based coupled Variable Capacity Infiltration-Water Erosion Prediction Project (VIC-WEPP) model to study how wildfire occurrences can enhance soil erosion in a future climate over a representative watershed in the northern Rocky Mountains - the Salmon River Basin (SRB) in central Idaho. While the VIC model simulates hydrologic processes at larger scales, the WEPP model simulates erosion at the hillslope scale by sampling representative hillslopes.

An Index Used to Evaluate the Applicability of Mid-to-Long-Term Runoff Prediction in a Basin Based on Mutual Information

Article

Full-text available

Jun 2024

Accurate and reliable mid-to-long-term runoff prediction (MLTRP) is of great importance in water resource management. However, the MLTRP is not suitable in each basin, and how to evaluate the applicability of MLTRP is still a question. Therefore, the total mutual information (TMI) index is developed in this study based on the predictor selection method using mutual information (MI) and partial MI (PMI). The relationship between the TMI and the predictive performance of five AI models is analyzed by applying five models to 222 forecasting scenarios in Australia. This results in over 222 forecasting scenarios which demonstrate that, compared with the MI, the developed TMI index can better represent the available information in the predictors and has a more significant negative correlation with the RRMSE, with a correlation coefficient between −0.62 and −0.85. This means that the model’s predictive performance will become better along with the increase in TMI, and therefore, the developed TMI index can be used to evaluate the applicability of MLTRP. When the TMI is more than 0.1, the available information in the predictors can support the construction of MLTRP models. In addition, the TMI can be used to partly explain the differences in predictive performance among five models. In general, the complex models, which can better utilize the contained information, are more sensitive to the TMI and have more significant improvement in terms of predictive performance along with the increase in TMI.

Rational design of high power density “Blue Energy Harvester” pressure retarded osmosis (PRO) membranes using artificial intelligence-based modeling and optimization

Article

Feb 2022
ENERG CONVERS MANAGE

The challenge in harvesting Salinity Gradient Power (SGP) through pressure retarded osmosis (PRO) requires design of high power density (PD) membranes and optimized process for operation. Recent studies show that for a feasible PRO operation the minimum net PD should be around 50 W/m². In this study, a data-driven approach has been adopted for designing optimum membranes as well as operating conditions. 200 papers, from last decade, were extensively reviewed and 34 experimental research articles were shortlisted for possible data mining, to predict water flux (WF) and PD. Comprehensive screened/pre-processed data related to both membrane and process (16 inputs) was obtained from 18 articles amounting to 339 data points. Two artificial neural network (ANN) models were explored (i) Levenberg-Marquardt (ANN-LM), and (ii) Bayesian Regularization (ANN-BR) along with a combination of three different activation functions i.e., hyperbolic tangent sigmoid transfer function (Tan-Sigmoid), logarithmic sigmoid transfer function (Log-Sigmoid) in the input and output layers. Out of the six resulting combinations, the best performing combination was found to be Tan-Sigmoid activation function in both layers with ANN-BR model having an R² value of 0.97 for WF and 0.98 for PD. Membrane properties like the type of membrane, thickness, and water permeability coefficient were found to be the major contributing factors for the prediction of WF while for PD, operating conditions such as applied pressure were found to play the major contributing factor (10–16 %). Optimization results yield a maximum WF of 147 LMH and PD of 87 W/m². These results were compared with the solution diffusion (S-D) model.

Utilization of Low-Frequency Capacitance Sensor Integrated with Artificial Neural Network for Real Time Prediction of Biscuit Shelf-Life

Article

Jan 2013

Taufik Djatna

Capacitance sensors integrated with ANN may possibly to be implemented in solving disadvantages of some shelf-life prediction methods that currently used. The aims of this research are to (1) design a circuit sensors that can measure capacitance value, (2) Design a Artificial Neural Network models to predict the shelf-life of biscuit, (3) Integrate capacitance sensor and ANN that can be applied to predict in real time. In reducing noise occurred commonly the frequency value was not set at single value, therefore the value set up in range is set at 5 kHz-6 kHz. ANN learning algorithm used backpropagation by trial and error the activation function, learning function, the number of nodes per hidden layer, and the number of hidden layer. ANN models are integrated by using a dielectric sensor interface built with MATLAB GUI toolbox through AVR microcontroller ATMega 8535. ANN integrated with capacitance parameters was very good for predicting the shelf-life of biscuit with training performance MSE 0.0001 and R 99.86%. ANN architecture with the best training performance contain 5 hidden layers, 10 nodes per hidden layer, tansig as the hidden layer activation function, purelin as the output layer activation function, trainlm as learning function and 86 epoch. Integration of ANN and capacitance sensors have the ability to predict the shelf-life of biscuit in real time. The performance of intelligent real-time system in predict the shelf-life of biscuit is fairly accurate (≤30%). The results of this study can be used as an alternative method for measuring shelf-life biscuit. This study also showed low frequencies can be used to measure the capacitance as well. Therefore, the integration of ANN and sensor capacitance can minimize time and make cost effective.

An Index Used to Evaluate the Applicability of Mid-long Term Runoff Prediction in a Basin Based on Mutual Information

Preprint

Full-text available

May 2024

Accurate and reliable mid-long runoff prediction (MLTRP) is of great importance in water re-sources management. However the MLTRP is not suitable in each basin and how to evaluate the applicability of MLTRP is still a question. Therefore, the total mutual information (TMI) index is developed in this study based on the predictor selection method using mutual information (MI) and partial MI (PMI). The relationship between the TMI and the predictive performance of five models is analyzed by applying five models in 222 forecasting scenarios in Australia. The results over 222 forecasting scenarios demonstrate that, compared with the MI, the developed TMI index can better represent the available information in the predictors, and has more significant negative correlation with the RRMSE with the correlation coefficient between -0.62 and -0.85. This means the model's predictive performance will become better along with the increase of TMI, and there-fore the developed TMI index can be used to evaluate the applicability of MLTRP. When TMI is more than 0.1, the available information in the predictors can support the construction of MLTRP models. In addition, the TMI is used to partly explain the difference of predictive performance among five models. In general, the complex models, which can better utilize the contained infor-mation, are more sensitive to the TMI, and have more significant improvement in terms of predic-tive performance along with the increase of TMI. This research can provide support for the study of MLTRP.

Flood prediction based on climatic signals using wavelet neural network

Article

Jun 2021

Large-scale climatic circulation modulates the weather patterns around the world. Understanding the teleconnections between large-scale circulation and local hydro-climatological variables has been a major thrust area of hydro-climatology research. The large-scale circulation is often quantified in terms of sea surface temperature (SST) and sea-level pressure (SLP). In this paper, we investigate the potential of wavelet neural network (WNN) hybrid model to predict maximum monthly discharge of the Madarsoo watershed, North of Iran considering two large-scale climatic signals like SST and SLP as inputs. Error measures like root-mean-square error (RMSE), and mean absolute error along with the correlation measures like coefficient of correlation (R), and Nash–Sutcliffe coefficient (CNS) were used to quantify the performance of prediction of maximum monthly discharge of three different hydrometry stations of the watershed. In all the cases, the WNN hybrid machine learning model was found to be giving superior performance consistently against the standalone artificial neural network (ANN) model and multiple linear regression model to predict the flood discharges of March and August months. The prediction of flood for August which is more devastating is found to be slightly better than the prediction of floods of March, in the stations served with smaller drainage area. The RMSE, R and CNS of Tamer hydrometry station in August were found to be 0.68, 0.996, and 0.99 m3/s, respectively, for the test period by using WNN model against 1.55, 0.989 and 0.95 by ANN model. Moreover, when evaluated for predicting the maximum monthly discharge in March and August between 2012 and 2013, the wavelet-based neural networks performed remarkably well than the ANN.

Associating Climatic Trends with Stochastic Modelling of Flow Sequences

Article

Full-text available

Jun 2021

Water is essential to all lifeforms including various ecological, geological, hydrological, and climatic processes/activities. With the changing climate, associated El Niño/Southern Oscillation (ENSO) events appear to stimulate highly uncertain patterns of precipitation (P) and evapotranspiration (EV) processes across the globe. Changes in P and EV patterns are highly sensitive to temperature (T) variation and thus also affect natural streamflow processes. This paper presents a novel suite of stochastic modelling approaches for associating streamflow sequences with climatic trends. The present work is built upon a stochastic modelling framework (HMM_GP) that integrates a hidden Markov model (HMM) with a generalised Pareto (GP) distribution for simulating synthetic flow sequences. The GP distribution within the HMM_GP model aims to improve the model’s efficiency in effectively simulating extreme events. This paper further investigated the potential of generalised extreme value distribution (GEV) coupled with an HMM model within a regression-based scheme for associating the impacts of precipitation and evapotranspiration processes on streamflow. The statistical characteristic of the pioneering modelling schematic was thoroughly assessed for its suitability to generate and predict synthetic river flow sequences for a set of future climatic projections, specifically during ENSO events. The new modelling schematic can be adapted for a range of applications in hydrology, agriculture, and climate change.

Associating Stochastic Modelling of Flow Sequences With Climatic Trends

Preprint

Full-text available

May 2021

Water is essential to all life-forms including various ecological, geological, hydrological, and climatic processes/activities. With changing climate, associated El Nino/Southern Oscillation (ENSO) events appear to stimulate highly uncertain patterns of precipitation (P) and evapotranspiration (EV) processes across the globe. Changes in P and EV patterns are highly sensitive to temperature variation and thus also affecting natural streamflow processes. This paper presents a novel suite of stochastic modelling approaches for associating streamflow sequences with climatic trends. The present work is built upon a stochastic modelling framework HMM_GP that integrates a Hidden Markov Model with a Generalised Pareto distribution for simulating synthetic flow sequences. The GP distribution within HMM_GP model is aimed to improve the model's efficiency in effectively simulating extreme events. This paper further investigated the potentials of Generalised Extreme Value Distribution (EVD) coupled with an HMM model within a regression-based scheme for associating impacts of precipitation and evapotranspiration processes on streamflow. The statistical characteristic of the pioneering modelling schematic has been thoroughly assessed for their suitability to generate/predict synthetic river flows sequences for a set of future climatic projections. The new modelling schematic can be adapted for a range of applications in the area of hydrology, agriculture and climate change.

Understanding Uncertainty in Probabilistic Floodplain Mapping in the Time of Climate Change

Article

Full-text available

Apr 2021

An integrated framework was employed to develop probabilistic floodplain maps, taking into account hydrologic and hydraulic uncertainties under climate change impacts. To develop the maps, several scenarios representing the individual and compounding effects of the models’ input and parameters uncertainty were defined. Hydrologic model calibration and validation were performed using a Dynamically Dimensioned Search algorithm. A generalized likelihood uncertainty estimation method was used for quantifying uncertainty. To draw on the potential benefits of the proposed methodology, a flash-flood-prone urban watershed in the Greater Toronto Area, Canada, was selected. The developed floodplain maps were updated considering climate change impacts on the input uncertainty with rainfall Intensity–Duration–Frequency (IDF) projections of RCP8.5. The results indicated that the hydrologic model input poses the most uncertainty to floodplain delineation. Incorporating climate change impacts resulted in the expansion of the potential flood area and an increase in water depth. Comparison between stationary and non-stationary IDFs showed that the flood probability is higher when a non-stationary approach is used. The large inevitable uncertainty associated with floodplain mapping and increased future flood risk under climate change imply a great need for enhanced flood modeling techniques and tools. The probabilistic floodplain maps are beneficial for implementing risk management strategies and land-use planning.

Five centuries of reconstructed streamflow in Athabasca River Basin, Canada: Non-stationarity and teleconnection to climate patterns

Article

Jul 2020
SCI TOTAL ENVIRON

Given the challenge to estimate representative long-term natural variability of streamflow from limited observed data, a hierarchical, multilevel Bayesian regression (HBR) was developed to reconstruct the 1489–2006 annual streamflow data at six Athabasca River Basin (ARB) gauging stations based on 14 tree ring chronologies. Seven nested models were developed to maximize the applications of available tree ring predictors. Based on results of goodness-of-fit tests, the HBR developed was skillful and reliable in reconstructing the streamflow of ARB. From five centuries of reconstructed streamflow for ARB, five or six abrupt change points are detected. The streamflow time series obtained from a backward moving, 46-year window for six gauging sites in ARB vary significantly over five centuries (1489–2006) and at times could exceed the 90% and/or 95% confidence intervals, denoting significant non-stationarities. Apparently changes in the mean state and the lag-1 autocorrelation of reconstructed streamflow across the gauging sites can be similar or radically different from each other. These nonstationary features imply that the default stationary assumption is not applicable in ARB. Further, the reconstructed streamflow shows statistically significant oscillations at interannual, interdecadal and multidecadal time scales and are teleconnected to climate patterns such as El Niño Southern Oscillation (ENSO), Pacific Decadal Oscillation (PDO) and Atlantic Multi-decadal Oscillation (AMO). A composite analysis shows that La Niña (El Niño), cold (warm) PDO, and cold (warm) AMO events are typically associated with increased (decreased) streamflow anomalies of ARB. The reconstructed streamflow data provides us the full range of streamflow variability and recurrence characteristics of extremes spanned over five centuries from which it is useful for us to evaluate and manage the current water systems of ARB more effectively and a better risk analysis of future droughts of ARB.

Impact of climate change on water resource availability in a mountainous catchment: A case study of the Toplica River catchment, Serbia

Article

Apr 2020
J HYDROL

The catchment of the Toplica River, situated in an underdeveloped region of southern Serbia, is studied to examine the potential impact of climate change on the hydrologic regime of mountainous catchments. The study projects precipitation (P), air temperature (T), potential evapotranspiration (PET), and discharge (Q) in the entire catchment, as well as groundwater level (GWL) variation in the lowland part of the catchment, according to scenarios RCP4.5 and RCP8.5. Projections of P and T are based on the results of a multimodel ensemble of seven regional climate models from the EURO-CORDEX project. Runoff is simulated by a calibrated HBV-light model. The correlation between GWL and river discharge was modeled by soft computing techniques of artificial neural networks (ANN). The projections pertain to the period from 2021 to 2100. The Mann-Kendall trend test is used to check for a trend and its statistical significance, and the Mann-Whitney test to examine the statistical significance of a change in the mean ensemble median of time-series for the near future (2021-2050) and distant future (2071-2100), relative to the reference period (1971-2000). No notable changes are expected on an annual scale in the study area. However, the results show that the current non-uniformity of the monthly water distribution is growing. In the winter months at the end of the century, in RCP8.5, P and T are expected to increase, as is Q (by as much as 29-50%). Groundwater responds to increased river discharges by reduced depths to groundwater (increased GWL). A higher Q increases the flood risk in the winter months. In the warm season, RCP8.5 predicts a decrease in Q and increase in the depth to groundwater in the distant future. Reduced quantities of water in the warm period might have an adverse effect on drinking water supply, agriculture, hydropower, fisheries, ecology, and tourism in the study area.

Evaluation of bias correction methods for APHRODITE data to improve hydrologic simulation in a large Himalayan basin

Article

Full-text available

Apr 2020
ATMOS RES

Gridded precipitation products have the advantages of wide spatial coverage and high spatiotemporal resolution compared to conventional rain-gauge data, and have been extensively applied in hydrology-related research. As one of the most representative of the gridded precipitation products, APHRODITE performs well in most regions over Asia except around the Tibetan Plateau. This study implemented four bias correction methods including both mean-based (LS, LOCI) and distribution-based approaches (CDF, LS_CDF) for APHRODITE data over the Yarluzangbu-Brahmaputra River Basin based on different climate zones, and the correction methods were assessed according to the performance of both statistical indices and hydrological simulation. First, it was found that the APHRODITE data underestimated the precipitation amount and overestimated the number of raindays for the entire study region; and the bias is relatively small in upstream (Tibetan Plateau) and large in downstream (floodplains). Second, the result showed that all four correction methods could effectively improve the precipitation estimates of the APHRODITE data and the combined LS_CDF method performed the best because of its greater spatial consistency advantages of bias and closer matching of the wet-day event and cumulative frequency. Hydrological performance also supports that the LS_CDF method is the best due to the fine driving simulation result for all three hydrological stations during long periods and the ability to better force simulation of extreme runoff. Third, this study also found that even though distribution-based approaches performed well in precipitation correction, especially for extreme precipitation, their correction effects also depend on the spatial consistency of the bias, which is even more important than frequency matching. These findings have certain reference values for the evaluation, correction, and application of grid precipitation data, not only for APHRODITE data, but also satellite and reanalysis precipitation data.

Optimized fuzzy inference system to enhance prediction accuracy for influent characteristics of a sewage treatment plant

Article

Mar 2020
SCI TOTAL ENVIRON

Sewage treatment plants (STPs) keep sewage contamination within safe levels and minimize the risk of environmental disasters. To achieve optimum operation of an STP, it is necessary for influent parameters to be measured or estimated precisely. In this research, six well-known influent chemical and biological characteristics, i.e., biochemical oxygen demand (BOD), chemical oxygen demand (COD), Ammoniacal Nitrogen (NH3-N), pH, oil and grease (OG) and suspended solids (SS), were modeled and predicted using the Sugeno fuzzy logic model. The membership function range of the fuzzy model was optimized by ANFIS, the integrated Genetic algorithms (GA), and the integrated particle swarm optimization (PSO) algorithms. The results were evaluated by different indices to find the accuracy of each algorithm. To ensure prediction accuracy, outliers in the predicted data were found and replaced with reasonable values. The results showed that both integrated GA-FIS and PSO-FIS algorithms performed at almost the same level and both had fewer errors than ANFIS. As the GA-FIS algorithm predicts BOD with fewer errors than PSO-FIS and the aim of this study is to provide an accurate prediction of missing data, GA-FIS was only used to predict the BOD parameter; the other parameters were predicted by PSO-FIS algorithm. As a result, the model successfully could provide outstanding performance for predicting the BOD, COD, NH3-N, OG, pH and SS with MAE equal to 3.79, 5.14, 0.4, 0.27, 0.02, and 3.16, respectively.

comparison of artificial neural network and combined models in estimating spatial distribution of snow depth and snow water equivalent in Samsami basin of Iran

Article

Full-text available

Nov 2010
NEURAL COMPUT APPL

Mohammad Reza Sharifi

Snow water equivalent (SWE) is a key parameter in hydrological cycle, and information on regional SWE is required for various hydrological and meteorological applications, as well as for hydropower production and flood forecasting. This study compares the snow depth and SWE estimated by multivariate linear regression (MLR), discriminant function analysis, ordinary kriging, ordinary kriging-multivariate linear regression, ordinary krigingdiscriminant function analysis, artificial neural network (ANN) and neural network-genetic algorithm (NNGA) models. The analysis was performed in the 5.2 km2 area of Samsami basin, located in the southwest of Iran. Statistical criteria were used to measure the models’ performances. The results indicated that NNGA, ANN and MLR methods were able to predict SWE at the desirable level of accuracy. However, the NNGA model with the highest coefficient of determination (R2 = 0.70, P value\0.05) and minimum root mean square error (RMSE = 0.202 cm) provided the best results among the other models. The lower SWE values were registered in the east of study area and higher SWE values appeared in the west of study area where altitude was higher.

Analysing the accuracy of machine learning techniques to develop an integrated influent time series model: case study of a sewage treatment plant, Malaysia

Article

Full-text available

Apr 2018
ENVIRON SCI POLLUT R

The function of a sewage treatment plant is to treat the sewage to acceptable standards before being discharged into the receiving waters. To design and operate such plants, it is necessary to measure and predict the influent flow rate. In this research, the influent flow rate of a sewage treatment plant (STP) was modelled and predicted by autoregressive integrated moving average (ARIMA), nonlinear autoregressive network (NAR) and support vector machine (SVM) regression time series algorithms. To evaluate the models’ accuracy, the root mean square error (RMSE) and coefficient of determination (R2) were calculated as initial assessment measures, while relative error (RE), peak flow criterion (PFC) and low flow criterion (LFC) were calculated as final evaluation measures to demonstrate the detailed accuracy of the selected models. An integrated model was developed based on the individual models’ prediction ability for low, average and peak flow. An initial assessment of the results showed that the ARIMA model was the least accurate and the NAR model was the most accurate. The RE results also prove that the SVM model’s frequency of errors above 10% or below − 10% was greater than the NAR model’s. The influent was also forecasted up to 44 weeks ahead by both models. The graphical results indicate that the NAR model made better predictions than the SVM model. The final evaluation of NAR and SVM demonstrated that SVM made better predictions at peak flow and NAR fit well for low and average inflow ranges. The integrated model developed includes the NAR model for low and average influent and the SVM model for peak inflow.

Application of artificial neural network in horizontal subsurface flow constructed wetland for nutrient removal prediction

Article

Full-text available

Jan 2016

Nihan Özengin

The aim of this study is to determine the appropriateness of the field measurements for the effectiveness of nutrients removal of Phragmites australis (Cav.) Trin. Ex. Steudel by applying artificial neural network (ANN) and also evaluate the removal capacity of LECA (light expanded clay aggregate) in a horizontal subsurface flow constructed wetland (SSFW). Two laboratory scale reactors were operated with weak and strong synthetic domestic wastewater continuously. One unit was planted with P. australis and the other unit remained unplanted (control reactor). The best performance was achieved with strong domestic wastewater treatment, the average removal efficiencies obtained from the evaluation of the system were 70.15% and 65.29% for TN, 66% and 57.4% for NH4-N, 61.64% and 67.37% for TP and, 66.52% and 51.7% for OP in planted and unplanted reactors, respectively. The average NO3-concentration was 0.90 mg l⁻¹ in the influent and 0.47 mg l⁻¹ and 0.60 mg l⁻¹ from planted and unplanted reactors, respectively. The average NO2⁻ concentration was 0.80 mg l⁻¹ in the influent and 0.56 mg l-1 and 0.88 mg l⁻¹ from planted and unplanted reactors, respectively. Based on the obtained results, this system has potential to be an applicable system to treat strong domestic wastewater. The data obtained in this study was assessed using NeuroSolutions 5.06 model. Each sample was characterized using eight independent variables (hydraulic retention time (HRT), dissolved oxygen (DO), pH, temperature (T), ammonium- nitrogen (NH4-N), nitrate (NO3⁻), nitrite (NO2⁻), ortho-phosphate (OP), and two dependent variable (total nitrogen (TN) and total phosphorus (TP)). The correlation coefficients between the neural network estimates and field measurements were as high as 0.9463 and 0.9161 for TN and TP, respectively. The results indicated that the adopted Levenberg-Marquardt back-propagation algorithm yields satisfactory estimates with acceptably low MSE values. Besides, the support matrix may play an important role in the system. The constructed wetland planted with P. australis and with LECA as a support matrix may be a good option to encourage and promote the prevention of environmental pollution.

Multi-metric calibration of hydrological model to capture overall flow regimes

Article

Jun 2016
J HYDROL

Flow regimes (e.g., magnitude, frequency, variation, duration, timing and rating of change) play a critical role in water supply and flood control, environmental processes, as well as biodiversity and life history patterns in the aquatic ecosystem. The traditional flow magnitude-oriented calibration of hydrological model was usually inadequate to well capture all the characteristics of observed flow regimes. In this study, we simulated multiple flow regime metrics simultaneously by coupling a distributed hydrological model with an equally weighted multi-objective optimization algorithm. Two headwater watersheds in the arid Hexi Corridor were selected for the case study. Sixteen metrics were selected as optimization objectives, which could represent the major characteristics of flow regimes. Model performance was compared with that of the single objective calibration. Results showed that most metrics were better simulated by the multi-objective approach than those of the single objective calibration, especially the low and high flow magnitudes, frequency and variation, duration, maximum flow timing and rating. However, the model performance of middle flow magnitude was not significantly improved because this metric was usually well captured by single objective calibration. The timing of minimum flow was poorly predicted by both the multi-metric and single calibrations due to the uncertainties in model structure and input data. The sensitive parameter values of the hydrological model changed remarkably and the simulated hydrological processes by the multi-metric calibration became more reliable, because more flow characteristics were considered. The study is expected to provide more detailed flow information by hydrological simulation for the integrated water resources management, and to improve the simulation performances of overall flow regimes.

Optimal design of BP algorithm by ACOR model for groundwater-level forecasting: A case study on Shabestar plain, Iran

Article

Full-text available

May 2016

Groundwater always has been considered as one of the major sources of drinking and agricultural water supply, especially in arid and semi-arid zones. Thus, there is a need to simulate (i.e., forecast) groundwater levels with an acceptable accuracy. In this paper, we present two applications of intelligent optimization algorithms for simulations of monthly groundwater levels in an unconfined coastal aquifer sited in the Shabestar plain, Iran. First, the backpropagation neural network (ANN-BP) with seven neurons in its hidden layer is utilized to reproduce groundwater-level variations using the external input variables including the following: rainfall, average discharge, temperature, evaporation, and annual time series. In the next application, ant colony optimization is used to optimize and find initial connection weights and biases of a BP algorithm during the training phase (ACOR-BP). The results were found to be acceptable in terms of accuracy and demonstrated that a hybrid ACOR-BP model is a much more rigorous fitting prediction tool for groundwater-level forecasting. This study has shown that such a hybrid network can be used as viable alternative to physical-based models for simulating the reactions of the aquifer under conceivable future scenarios. In addition, it may be useful for reconstructing long periods of missing historical observations of the influencing variables.

Improving Hydrological Process Modeling Using Optimized Threshold-Based Wavelet De-Noising Technique

Article

Full-text available

Mar 2016
WATER RESOUR MANAG

Peyman Abbaszadeh

The extent of the noise on hydrological data is inevitable, which reduces the efficiency of Data-Driven Models (DDMs). Despite of this fact that the DDMs such as Artificial Neural Network (ANN) are capable of nonlinear functional mapping between a set of input and output variables, but refining of the time series through data pre-processing methods can provide with the possibility to increase the performance of these set of models. The main objective of this study is to propose a new method called Optimized Threshold- Based Wavelet De-noising technique (OTWD) to de-noise hydrological time series and improve the prediction accuracy while the DDM is being used. For this purpose, in the first step, Wavelet-ANN (WNN) model was developed for identifying suitable wavelet function and maximum decomposition level. Afterward, sub-signals of original precipitation time series which were determined in the first step were de-noised by using of OTWD technique. Therefore, these clean sub-signals of precipitation time series were imposed as input data to the ANN to predict the precipitation one time step ahead. The results showed that OTWD technique could improve the efficiency of WNN model dramatically; this outcome was reported by the different efficiency criterions such as Nash-Sutcliffe Efficiency (NSE= 0.92), Root Mean Squared Error (RMSE = 0.0103), coefficient correlation of linear regression (R= 0.93), Peak Value Criterion (PVC= 0.021) and Low Value Criterion (LVC= 0.026). The best fitted WNN model in comparison by proposed model showed weaker performance by the NSE, RMSE, R, PVC and LVC values of 0.86, 0.043, 0.87, 0.034 and 0.045, respectively.

Short-Term Forecasting of Water Yield from Forested Catchments after Bushfire: A Case Study from Southeast Australia

Article

Full-text available

Feb 2015

Forested catchments in southeast Australia play an important role in supplying water to major cities. Over the past decades, vegetation cover in this area has been affected by major bushfires that in return influence water yield. This study tests methods for forecasting water yield after bushfire, in a forested catchment in southeast Australia. Precipitation and remotely sensed Normalized Difference Vegetation Index (NDVI) were selected as the main predictor variables. Cross-correlation results show that water yield with time lag equal to 1 can be used as an additional predictor variable. Input variables and water yield observations were set based on 16-day time series, from 20 January 2003 to 20 January 2012. Four data-driven models namely Non-Linear Multivariate Regression (NLMR), K-Nearest Neighbor (KNN), non-linear Autoregressive with External Input based Artificial Neural Networks (NARX-ANN), and Symbolic Regression (SR) were employed for this study. Results showed that NARX-ANN outperforms other models across all goodness-of-fit criteria. The Nash-Sutcliffe efficiency (NSE) of 0.90 and correlation coefficient of 0.96 at the training-validation stage, as well as NSE of 0.89 and correlation coefficient of 0.95 at the testing stage, are indicative of potentials of this model for capturing ecological dynamics in predicting catchment hydrology, at an operational level.

Selection of smoothing parameter estimators for general regression neural networks – Applications to hydrological and water resources modelling

Article

Sep 2014
ENVIRON MODELL SOFTW

Multi-layer perceptron artificial neural networks are used extensively in hydrological and water resources modelling. However, a significant limitation with their application is that it is difficult to determine the optimal model structure. General regression neural networks (GRNNs) overcome this limitation, as their model structure is fixed. However, there has been limited investigation into the best way to estimate the parameters of GRNNs within water resources applications. In order to address this shortcoming, the performance of nine different estimation methods for the GRNN smoothing parameter is assessed in terms of accuracy and computational efficiency for a number of synthetic and measured data sets with distinct properties. Of these methods, five are based on bandwidth estimators used in kernel density estimation, and four are based on single and multivariable calibration strategies. In total, 5674 GRNN models are developed and preliminary guidelines for the selection of GRNN parameter estimation methods are provided and tested.

Predicción de caudales en río Colorado, Argentina

Article

Full-text available

Nov 2012

The identification of suitable models for predicting daily water flow is important for planning and management of water storage in reservoirs of Argentina. Long-term prediction of water flow is crucial for regulating reservoirs and hydroelectric plants, for assessing environmental protection and sustainable development, for guaranteeing correct operation of public water supply in cities like Catriel, 25 de Mayo, Colorado River and potentially also Bahía Blanca. In this paper, we analyze in Buta Ranquil flow time series upstream reservoir and hydroelectric plant in order to model and predict daily fluctuations. We compare results obtained by using a three-layer artificial neural network (ANN), and an autoregressive (AR) model, using 18 years of data, of which the last 3 years are used for model validation by means of the root mean square error (RMSE), and measure of certainty (Skill). Our results point out to the better performance to predict daily water flow or refill them of the ANN model performance respect to the AR model.

Snow cover thickness estimation using radial basis function networks

Article

Full-text available

May 2013
TCD

This paper reports an experimental study designed for the in-depth investigation of how the radial basis function network (RBFN) estimates snow cover thickness as a function of climate and topographic parameters. The estimation problem is modeled in terms of both function regression and classification, obtaining continuous and discrete thickness values, respectively. The model is based on a minimal set of climatic and topographic data collected from a limited number of stations located in the Italian Central Alps. Several experiments have been conceived and conducted adopting different evaluation indexes. A comparison analysis was also developed for a quantitative evaluation of the advantages of the RBFN method over to conventional widely used spatial interpolation techniques when dealing with critical situations originated by lack of data and limited n-homogeneously distributed instrumented sites. The RBFN model proved competitive behavior and a valuable tool in critical situations in which conventional techniques suffer from a lack of representative data.

Prediction of water flows in Colorado River, Argentina

Article

Full-text available

Nov 2012

Two Decades of Anarchy? Emerging Themes and Outstanding Challenges for Neural Network River Forecasting

Article

Full-text available

Jul 2012

This paper traces two decades of neural network rainfall-runoff and streamflow modelling, collectively termed 'river forecasting'. The field is now firmly established and the research community involved has much to offer hydrological science. First, however, it will be necessary to converge on more objective and consistent protocols for: selecting and treating inputs prior to model development; extracting physically meaningful insights from each proposed solution; and improving transparency in the benchmarking and reporting of experimental case studies. It is also clear that neural network river forecasting solutions will have limited appeal for operational purposes until confidence intervals can be attached to forecasts. Modular design, ensemble experiments, and hybridization with conventional hydrological models are yielding new tools for decision-making. The full potential for modelling complex hydrological systems, and for characterizing uncertainty, has yet to be realized. Further gains could also emerge from the provision of an agreed set of benchmark data sets and associated development of superior diagnostics for more rigorous intermodel evaluation. To achieve these goals will require a paradigm shift, such that the mass of individual isolated activities, focused on incremental technical refinement, is replaced by a more coordinated, problem-solving international research body.

Can Machine Learning Accelerate Process Understanding and Decision‐Relevant Predictions of River Water Quality?

Article

Mar 2022

The global decline of water quality in rivers and streams has resulted in a pressing need to design new watershed management strategies. Water quality can be affected by multiple stressors including population growth, land use change, global warming, and extreme events, with repercussions on human and ecosystem health. A scientific understanding of factors affecting riverine water quality and predictions at local to regional scales, and at sub‐daily to decadal timescales are needed for optimal management of watersheds and river basins. Here, we discuss how machine learning (ML) can enable development of more accurate, computationally tractable, and scalable models for analysis and predictions of river water quality. We review relevant state‐of‐the art applications of ML for water quality models and discuss opportunities to improve the use of ML for emerging computational and mathematical methods for model selection, hyperparameter optimization, incorporating process knowledge into ML models, improving explainablity, uncertainty quantification, and model‐data integration. We then present considerations for using ML to address water quality problems given their scale and complexity, available data and computational resources, and stakeholder needs. When combined with decades of process understanding, interdisciplinary advances in knowledge‐guided ML, information theory, data integration, and analytics can help address fundamental science questions and enable decision‐relevant predictions of riverine water quality. This article is protected by copyright. All rights reserved.

Enhancement of nitrogen prediction accuracy through a new hybrid model using ant colony optimization and an Elman neural network

Article

Nov 2021

Advanced human activities, including modern agricultural practices, are responsible for alteration of natural concentration of nitrogen compounds in rivers. Future prediction of nitrogen compound concentrations (especially nitrate-nitrogen and ammonia-nitrogen) are important for countries where household water is obtained from rivers after treatment. Increased concentrations of nitrogen compounds result in the suspension of household water supplies. Artificial Neural Networks (ANNs) have already been deployed for the prediction of nitrogen compounds in various countries. But standalone ANN have several limitations. However, the limitations of ANNs can be resolved using hybrid models. This study proposes a new ACO-ENN hybrid model developed by integrating Ant Colony Optimization (ACO) with an Elman Neural Network (ENN). The developed ACO-ENN hybrid model was used to improve the prediction results of nitrate-nitrogen and ammonia-nitrogen prediction models. The results of new hybrid models were compared with multilayer ANN models and standalone ENN models. There was a significant improvement in the mean square errors (MSE) (0.196→0.049→0.012, i.e. ANN→ENN→Hybrid), mean absolute errors (MAE) (0.271→0.094→0.069) and Nash–Sutcliffe efficiencies (NSE) (0.7255→0.9321→0.984). The hybrid model had outstanding performance compared with the ANN and ENN models. Hence, the prediction accuracy of nitrate-nitrogen and ammonia-nitrogen has been improved using new ACO-ENN hybrid model.

Prediction of the karstic spring flow rates under climate change by climatic variables based on the artificial neural network: a case study of Iran

Article

Full-text available

May 2020
ENVIRON MONIT ASSESS

Few studies have evaluated the impact of climate change on groundwater resources for a region with no pumping well. Indeed, the uncertainty of pumping wells may undesirably influence the results. Therefore, a region without any pumping well was selected to assess the impact of climate change on the karstic spring flow rates. NASA Earth Exchange Global Daily Downscaled Projections (NEX-GDDP) dataset was used to extract the climatic variables for the present (1961–1990) and future (2021–2050) time periods by two Representative Concentration Pathways (RCPs), i.e., RCP4.5 and RCP8.5, in Lali region, southwest Iran. Although this dataset has been already verified, its output was evaluated for Lali region. Then, the impact of climate change on the discharge of Bibitarkhoun karstic spring was examined by the Artificial Neural Network (ANN). In this regard, if considering the daily data, ANN is not trained satisfactorily, because of the spring’s lag time response to the precipitation; if monthly time step is considered, the data would not be adequate. Therefore, the average of some previous days was considered to calculate the variables. The average precipitation is 344, 329, and 324 mm/year and the average temperature is 14.18, 15.98, and 16.3 °C both for the present, future time period under RCP4.5 and future time period under RCP8.5, respectively. The network selected demonstrated no climate change impact on the average of spring discharge. However, the discharge increased by about + 8% in spring and summer and decreased by about − 7% in autumn and winter in the future time period.

Methodology for spatio‐temporal predictions of traffic counts across an urban road network and generation of an on‐road greenhouse gas emission inventory

Article

Nov 2019

On‐road emission inventories in urban areas have typically been developed using traffic data derived from travel demand models. These approaches tend to underestimate emissions because they often only incorporate data on household travel, not including commercial vehicle movements, taxis, ride hailing services, and other trips typically underreported within travel surveys. In contrast, traffic counts embed all types of on‐road vehicles; however, they are only conducted at selected locations in an urban area. Traffic counts are typically spatially correlated, which enables the development of methods that can interpolate traffic data at selected monitoring stations across an urban road network and in turn develop emission estimates. This paper presents a new and universal methodology designed to use traffic count data for the prediction of periodic and annual volumes as well as greenhouse gas emissions at the level of each individual roadway and for multiple years across a large road network. The methodology relies on the data collected and the spatio‐temporal relationships between traffic counts at various stations; it recognizes patterns in the data and identifies locations with similar trends. Traffic volumes and emissions prediction can be made even on roads where no count data exist. Data from the City of Toronto traffic count program were used to validate the output of various algorithms, indicating robust model performance, even in areas with limited data.

Forecasting Groundwater Levels using Nonlinear Autoregressive Networks with Exogenous Input (NARX)

Article

Feb 2018
J HYDROL

While the application of neural networks for groundwater level forecasting in general has been investigated by many authors, the use of nonlinear autoregressive networks with exogenous inputs (NARX) is relatively new. For this work NARX were applied to obtain groundwater level forecasts for several wells in southwest Germany. Wells in porous, fractured and karst aquifers were investigated and forecasts of lead times up to half a year were conducted for both influenced (e.g. nearby pumping) and uninfluenced wells. Precipitation and temperature were chosen as predictors, which makes the selected approach easily transferable, since both parameters are widely available and simple to measure. Input and feedback delays were determined by applying STL time series decomposition on the data and using auto- and cross-correlation functions on the remainders to determine significant time lags. Coefficient of determination, (relative) root mean squared error and Nash-Sutcliffe efficiency were used to evaluate forecasts, the model selection was based on an out-of-sample validation on rolling basis. The results are promising and indicate an outstanding suitability of NARX for groundwater level predictions with such a small set of inputs in all three aquifer types.

Simulating Water Resources within Biomass Structure of Trees Storey in the Natural and Mixed-Beech Forests in North of Iran using Artificial Neural Network

Article

Jan 2017

علی‌اصغر واحدی

The Future

Chapter

Nov 2017

B. Sivakumar

The tremendous progress that has been achieved, through three decades of research, in the applications of chaos theory in hydrology inevitably leads to questions regarding the future of chaos theory in hydrology. Of particular interest is to identify potential areas for further applications and advancement of the theory and possible ways to achieve fruitful outcomes. This chapter addresses these questions. In light of some of the research questions at the forefront of hydrology at the current time and will be in the future, and also looking at some studies that have already addressed these questions from the perspective of chaos theory (albeit rudimentary), several different areas are identified to further advance chaos theory in hydrology. These are: parameter estimation in hydrologic models, simplification in hydrologic model development, integration of different concepts in hydrology, development of catchment classification framework, extensions of chaos studies using multiple hydrologic variables, reconstruction of hydrologic system equations, and downscaling of global climate models. Finally, the need and the potential to establish reliable links between chaos theory, hydrologic data, and hydrologic system physics are also discussed.

Estimation of groundwater level using a hybrid genetic algorithm-neural network

Article

Full-text available

Jan 2015

A Novel Solution to Define the Optimum Number and Location of New Wells to Improve Groundwater Level Map

Article

Full-text available

Dec 2016

Groundwater is one of the major sources of water. Maintenance and management of this vital resource is so important in arid and semi-arid region. Thus, reliable and accurate groundwater measurement is essential to introduce as a basic data for any groundwater management study. Observation wells are adequate tools for monitoring groundwater level. It is clear that more wells lead to better illustration the groundwater map. The aim of this study is to find the appropriate number and location of wells, having maximum effect on improvement of accuracy of groundwater level map. The number of the added wells should be determined by considering both economic restrictions and the amount of improvement that caused by the added well simultaneously. Hence, genetic algorithm was employed to find the optimum number and location of the wells to be added. The objective function of this algorithm is to maximize the gradient of the inverse mean square error per number of added wells.

Comparison of an artificial neural network and a conceptual rainfall-runoff model in the simulation of ephemeral streamflow

Article

Feb 2016

The rainfall–runoff process is governed by parameters that can seldom be measured directly for use with distributed models, but are rather inferred by expert judgment and calibrated against historical records. Here, a comparison is made between a conceptual model (CM) and an artificial neural network (ANN) for their ability to efficiently model complex hydrological processes. The Sacramento soil moisture accounting model (SAC-SMA) is calibrated using a scheme based on genetic algorithms and an input delay neural network (IDNN) is trained for variable delays and hidden layer neurons which are thoroughly discussed. The models are tested for 15 ephemeral catchments in Crete, Greece, using monthly rainfall, streamflow and potential evapotranspiration input. SAC-SMA performs well for most basins and acceptably for the entire sample with R2 of 0.59–0.92, while scoring better for high than low flows. For the entire dataset, the IDNN improves simulation fit to R2 of 0.70–0.96 and performs better for high flows while being outmatched in low flows. Results show that the ANN models can be superior to the conventional CMs, as parameter sensitivity is unclear, but CMs may be more robust in extrapolating beyond historical record limits and scenario building. EDITOR M.C. Acreman; ASSOCIATE EDITOR not assigned

Groundwater salt content change and its simulation based on machine learning model in hinterlands of Taklimakan Desert

Article

Full-text available

Jan 2013

The Taklimakan Desert, in the center of the Tarim Basin, Northwest China, is near the arid center of the Eurasian Continent. It is an ideal place for carrying out research on circulation systems of ocean-continent and continent- continent, as well as on the environmental effects of the Tibetan Plateau uplift. Oil and natural gas resources are rich in the desert hinterlands, where weather is extremely dry and the ecological environment is extremely frail. How to achieve sustainable development of oil gas energy exploitation and water resources will become a key issue of rapid and sustained energy development in the exploitation area. Since 2000, with the increase of oil and gas production, as well as construction of ecological engineering in oil field areas, regional water use, i.e., water consumption by oil exploitation, life-service water, and ecological engineering water has been increasing. Under this condition, the response of regional groundwater to water-use behavior, and sustainable use of groundwater under scaled development of future oil gas energy, is a focus of discussion in academic circles. All the water used for oil field production, living, and an ecological shelterbelt program in the Tarim oil field was supplied by 19 local water wells. The main groundwater was from surrounding snow cover and glacier meltwater that leak into the groundwater system at a mountain pass, which then inflows slowly as runoff with terrain change. There is a single chemical type of groundwater in the study area, which is mainly alkali and a strong acid ion. Groundwater is an important resource within desert areas, for both human use and maintenance of the ecological environment. However, little information is available on the long-term change and evolutionary trends of groundwater salt content in the region. The aim of this study is to increase knowledge of such content and its influences there. We simulated this content, studied its change, and evaluated the effects and applicability of machine learning models for the hinterland of Taklimakan Desert. For these purposes, three machine learning models and an appropriate integrated model were designed to simulate the variation of groundwater salt content. Based on statistical analysis of evaporation, precipitation, air temperature, air pressure, groundwater level and conductivity, this paper reports some statistical features of that content and its influences. The temporal change of content was simulated using Gaussian Process (GP), Gaussian Process Latent Variable Model (GPLVM), Back Propagation Artificial Neural Network (BP-ANN), and an integrated machine learning model, under the impacts of climate change and human activities. The modeling was evaluated using root mean square error (RMSE) and uncentered R-squared (uncentered R2).The results show that the groundwater flow system is affected by climate change and human activities, and groundwater level appears steady with small fluctuation in the study area. The change of groundwater level was concordant with air pressure, and the change of air temperature was consistent with evaporation. There was an overall uptrend of groundwater salt content. Accurate prediction can be achieved with GP, whereas GPLVM was the better multiple regression method for a known groundwater salt content among the three models. The integrated model of GP, GPLVM, and BP-ANN gave the best results for all observation datasets, including training and prediction data. The results not only help clarify regional hydrologic cycle processes under extremely dry conditions, but can also provide appropriate theoretical evidence for sustainable utilization of water resources and safe production in desert oilfields.

A comparative study of shallow groundwater level simulation with WA–ANN and ITS model in a coastal island of south China

Article

Nov 2014

Accurate and reliable prediction of shallow groundwater level is a critical component in water resources management. Two nonlinear models, WA–ANN method based on discrete wavelet transform (WA) and artificial neural network (ANN) and integrated time series (ITS) model, were developed to predict groundwater level fluctuations of a shallow coastal aquifer (Fujian Province, China). The two models were testified with the monitored groundwater level from 2000 to 2011. Two representative wells are selected with different locations within the study area. The error criteria were estimated using the coefficient of determination (R 2), Nash–Sutcliffe model efficiency coefficient (E), and root-mean-square error (RMSE). The best model was determined based on the RMSE of prediction using independent test data set. The WA–ANN models were found to provide more accurate monthly average groundwater level forecasts compared to the ITS models. The results of the study indicate the potential of WA–ANN models in forecasting groundwater levels. It is recommended that additional studies explore this proposed method, which can be used in turn to facilitate the development and implementation of more effective and sustainable groundwater management strategies.

Model Calibration and Parameter Estimation: For Environmental and Water Resource Systems

Book

Jan 2015

This three-part book provides a comprehensive and systematic introduction to these challenging topics such as model calibration, parameter estimation, reliability assessment, and data collection design. Part 1 covers the classical inverse problem for parameter estimation in both deterministic and statistical frameworks, Part 2 is dedicated to system identification, hyperparameter estimation, and model dimension reduction, and Part 3 considers how to collect data and construct reliable models for prediction and decision-making. For the first time, topics such as multiscale inversion, stochastic field parameterization, level set method, machine learning, global sensitivity analysis, data assimilation, model uncertainty quantification, robust design, and goal-oriented modeling, are systematically described and summarized in a single book from the perspective of model inversion, and elucidated with numerical examples from environmental and water resources modeling. Readers of this book will not only learn basic concepts and methods for simple parameter estimation, but also get familiar with advanced methods for modeling complex systems. Algorithms for mathematical tools used in this book, such as numerical optimization, automatic differentiation, adaptive parameterization, hierarchical Bayesian, metamodeling, Markov chain Monte Carlo, are covered in details. This book can be used as a reference for graduate and upper level undergraduate students majoring in environmental engineering, hydrology, and geosciences. It also serves as an essential reference book for professionals such as petroleum engineers, mining engineers, chemists, mechanical engineers, biologists, biology and medical engineering, applied mathematicians, and others who perform mathematical modeling.

Estimation of groundwater level using a hybrid genetic algorithm-neural network

Article

Jan 2015

Ziba Hosseini

In this paper, we present an application of evolved neural networks using a real coded genetic algorithm for simulations of monthly groundwater levels in a coastal aquifer located in the Shabestar Plain, Iran. After initializing the model with groundwater elevations observed at a given time, the developed hybrid genetic algorithm-back propagation (GA-BP) should be able to reproduce groundwater level variations using the external input variables, including rainfall, average discharge, temperature, evaporation and annual time series. To achieve this purpose, the hybrid GA-BP algorithm is first calibrated on a training dataset to perform monthly predictions of future groundwater levels using past observed groundwater levels and additional inputs. Simulations are then produced on another data set by iteratively feeding back the predicted groundwater levels, along with real external data. This modelling algorithm has been compared with the individual back propagation model (ANN-BP), which demonstrates the capability of the hybrid GA-BP model. The later provides better results in estimation of groundwater levels compared to the individual one. The study suggests that such a network can be used as a viable alternative to physical-based models in order to simulate the responses of the aquifer under plausible future scenarios, or to reconstruct long periods of missing observations provided past data for the influencing variables is available.

Modeling soil water content in extreme arid area using an adaptive neuro-fuzzy inference system

Article

May 2015
J HYDROL

Modeling of soil water content (SWC) is one of the most studied topics in hydrology due to its essential application to water resources management. In this study, an adaptive neuro fuzzy inference system (ANFIS) method is used to simulate SWC in the extreme arid area. In-situ SWC datasets for soil layers, with depths of 40. cm (layer 1), 60. cm (layer 2) below surface was taken for the present study. The models analyzed different combinations of antecedent SWC values and the appropriate input vector has been selected based on the analysis of residuals. The performance of the ANFIS models in training and validation sets are compared with the observed data. In layer 1, the model which consists of six antecedent values of SWC, has been selected as the best fit model for SWC modeling. On the other hand, which includes two antecedent values of SWC, has been selected as the best fit model for SWC modeling at layer 2. In order to assess the ability of ANFIS model relative to that of the ANN model, the best fit of ANFIS model of layer 1 and layer 2 structures are also tested by two artificial neural networks (ANN), namely, Levenberg-Marquardt feedforward neural network (ANN-1) and Bayesian regularization feedforward neural network (ANN-2). The comparison was made according to the various statistical measures. A detailed comparison of the overall performance indicated that the ANFIS model performed better than both the ANN-1 and ANN-2 in SWC modeling for the validation data sets in this study.

Spatiotemporal Groundwater Level Forecasting in Coastal Aquifers by Hybrid Artificial Neural Network-Geostatistics Model: A Case Study

Article

Mar 2011

Prediction of groundwater level in a basin is of immense importance for the management of groundwater resources, especially in coastal regions where the water table fluctuations are to be limited to avoid sea water intrusion. Lack of strong predictive tools, or perhaps the lack of experienced users of those tools, may contribute to problems in data interpretation and failure to reach consensus about the need for key water management actions. Therefore, it is extremely important to comprehend the spatiotemporal variations of the water level for the management of groundwater in the coastal areas. In this article, a hybrid, artificial neural network-geostatistics methodology is presented for spatiotemporal prediction of groundwater levels. The proposed model contains two separated stages. At the first stage, an artificial neural network is trained for each piezometer for time-series modeling of the water level, so that the model can predict the groundwater level the next month. At the second stage, predicted values of water levels at different piezometers are imposed to a calibrated geostatistics model in order to estimate groundwater level at any desired point in the plain. This methodology is applied for the Shabestar plain, which adjoins to Urmieh Lake as a coastal aquifer in East Azerbaijan Province, Iran. The most appropriate set of input variables to the model are selected through a combination of domain knowledge and available data series. Results suggest that the feed-forward neural network trained with Levenberg-Marquardt algorithm for temporal and Kriging scheme for spatial modeling are good choices for predicting groundwater levels in the coastal aquifer.

Erna NN Wafer Sensor2012

Data

Mar 2015

PREDIKSI MASA KEDALUWARSA WAFER DENGAN ARTIFICIAL NEURAL NETWORK (ANN) BERDASARKAN PARAMETER NILAI KAPASITANSI Prediction of Wafer Shelf Life using Artificial Neural Network based on Capacitance Parameter

Article

Nov 2013

ABSTRAK Wafer adalah jenis makanan kering yang sering ditemukan kedaluwarsa. Penentuan masa kedaluwarsa dengan observasi laboratorium memiliki beberapa kelemahan, diantaranya memakan waktu, panelis terlatih, suasana yang tepat, biaya dan alat uji yang kompleks. alternatif solusinya adalah penggunaan artificial Neural Network (aNN) berbasiskan parameter kapasitansi. Tujuan kerja ilmiah ini adalah untuk memprediksi masa kedaluwarsa wafer menggunakan aNN berbasiskan parameter kapasitansi. algoritma pembelajaran yang digunakan adalah Backpropagation dengan trial and error variasi jumlah node per hidden layer, jumlah hidden layer, fungsi aktivasi, fungsi pembelajaran dan epoch. Hasil prediksi menunjukkan bahwa aNN hasil pelatihan yang dikombinasikan dengan parameter kapasitansi mampu memprediksi masa kedaluwarsa wafer dengan MSE terendah 0,01 dan R tertinggi 89,25%. ABSTRACT Wafer is type of biscuit frequently found on expired condition in market, therefore prediction method should be implemented to avoid this condition. apart from the prediction of shelf-life of wafer done by laboratory test, which were time-consuming, expensive, required trained panelists, complex equipment and suitable ambience, artificial neural network (aNN) based dielectric parameters was proposed in nthis study. The aim of study was to develop model to predict shelf-life employing aNN based capacitance parameter. Back propagation algorithm with trial and error was applied in variations of nodes per hidden layer, number of hidden layers, activation functions, the function of learnings and epochs. The result of study was the model was able to predict wafer shelf-life. The accuracy level was shown by low MSE value (0.01) and high coefficient correlation value (89.25%).

Prediction of groundwater levels using different artificial neural network architectures and algorithms

Article

Full-text available

Apr 2012

Performance of four types of functionally different artificial neural network (ANN) models, namely Feed forward neural network, Elman type recurrent neural network, Input delay neural network and Radial basis function network and fourteen types of algorithms, namely Batch gradient descent (traingd), Batch gradient descent with momentum (traingdm), Adaptive learning rate (traingda), Adaptive learning rate with momentum (traingdx), Resilient backpropagation (trainrp), Fletcher-Reeves update (traincgf), Polak-Ribiere update (traincgp), Powell-Beale restarts (traincgb), Scaled conjugate gradient (trainscg), BFGS algorithm (trainbfg), One step secant algorithm (trainoss), Levenberg-Marquardt (trainlm), Automated regularization (trainbr) and Random order incremental training (trainr) were examined in order to identify an efficient ANN architecture and algorithm that can simulate the water table fluctuations using a relatively short length of groundwater level records. Tirupati, located in Chittoor district of the drought-prone Rayalaseema region in India, having resident population of over 3.0 lakhs and pilgrims of over 50,000 per day was chosen as the study area. As its groundwater levels showed a rapid decline in the last decade due to the overexploitation for the domestic, agricultural and industrial needs, accurate prediction is very essential to plan better conservation of groundwater resources. Results showed that Feed forward neural network trained with training algorithm Levenberg-Marquardt is suitable for accurate prediction of groundwater levels.

Flood forecasting using artificial neural networks

Conference Paper

Full-text available

Jan 2005

Stream Flow Forecasting using ANN and Fuzzy Logic

Article

Full-text available

Jan 2012

Multi model data fusion for hydrological forecasting using K-nearest neighbor method

Article

Full-text available

Feb 2010
IRAN J SCI TECHNOL B

Hydrological forecasting is one the most important issues in water resources systems which helps in dealing with the real time operation, flood and drought warning, and irrigation scheduling. Recent studies have suggested that the use of data fusion approach instead of using a single forecast approach may improve the hydrological forecast skill. This paper presents a comparative assessment of five different methods of data fusions including simple and weighted averaging; relying on the user's experience; artificial neural networks; and error analysis, by applying them in two real case studies. Multiple linear regression, non-parametric K-nearest neighbour regression, conventional multilayer perceptron, and an artificial neural network improved for extreme value forecasting are used as individual forecasting methods at each case study. Conventional data fusing methods as well as a new proposed statistical method based on the non-parametric K-nearest neighbor model are used for hydrological forecasting. Results of data fusion approach in two contrasting case studies are thoroughly analyzed and discussed. The results demonstrate that the use of data fusion could significantly improve forecasts in comparison with the use of singe models. As a result of this study, it is concluded that time-varying combining methods which benefit from the use of real-time predictors in their fusion procedure could be more promising than others. Also, data fusion by K-NN method outperforms conventional methods by improving forecasts through decreasing the bandwidth of ensemble forecast and error of point forecast in both case studies.

Application of ANN in sketching spatial nonlinearity of unconfined aquifer in agricultural basin

Article

Feb 2014
AGR WATER MANAGE

This paper endeavors the growing challenges of groundwater economy in agriculture with information and analysis of the spatial nonlinearity in groundwater depletion due to anthropogenic abstraction and proposes a way to find the water table imprints by judicious application of artificial neural networks (ANN). The results exhibit that groundwater problems and their agricultural consequences are heterogeneous across space and time. While the problems are contemplative and impressionistic, the severity scales varying dimensions. It is found that ANN models are realistic and viable due to their inherent stochastic nature of neural computation using artificial intelligence decoding ingrained nonlinearity and strong synchronicity. The result demonstrates that ANN is capable of recognizing local optimal in a time series analyses and can successfully forecast seasonal variability. It can be used to closely monitor the water variables to meet and anticipate the growing challenges of groundwater resource sustainability and precision irrigation. The model can be leveraged in devising water economy policy and seasonal cropping practices which in turn can aid policies to be tailored to local hydrogeological settings and agro economic realities. While market forces and economic incentive policy can change water use, public initiatives for agricultural groundwater regulation to balance short term economic efficiency with long resource sustainability are urgently needed.

Artificial neural network modeling of dissolved oxygen in the Heihe River, Northwestern China

Article

Sep 2012

Identification and quantification of dissolved oxygen (DO) profiles of river is one of the primary concerns for water resources managers. In this research, an artificial neural network (ANN) was developed to simulate the DO concentrations in the Heihe River, Northwestern China. A three-layer back-propagation ANN was used with the Bayesian regularization training algorithm. The input variables of the neural network were pH, electrical conductivity, chloride (Cl−), calcium (Ca2+), total alkalinity, total hardness, nitrate nitrogen (NO3-N), and ammonical nitrogen (NH4-N). The ANN structure with 14 hidden neurons obtained the best selection. By making comparison between the results of the ANN model and the measured data on the basis of correlation coefficient (r) and root mean square error (RMSE), a good model-fitting DO values indicated the effectiveness of neural network model. It is found that the coefficient of correlation (r) values for the training, validation, and test sets were 0.9654, 0.9841, and 0.9680, respectively, and the respective values of RMSE for the training, validation, and test sets were 0.4272, 0.3667, and 0.4570, respectively. Sensitivity analysis was used to determine the influence of input variables on the dependent variable. The most effective inputs were determined as pH, NO3-N, NH4-N, and Ca2+. Cl− was found to be least effective variables on the proposed model. The identified ANN model can be used to simulate the water quality parameters.

Sensitivity analysis of groundwater level in Jinci Spring Basin (China) based on artificial neural network modeling

Article

Jun 2012

Jinci Spring in Shanxi, north China, is a major local water source. It dried up in April 1994 due to groundwater overexploitation. The groundwater system is complex, involving many nonlinear and uncertain factors. Artificial neural network (ANN) models are statistical techniques to study parameter nonlinear relationships of groundwater systems. However, ANN models offer little explanatory insight into the mechanisms of prediction models. Sensitivity analysis can overcome this shortcoming. In this study, a back-propagation neural network model was built based on the relationship between groundwater level and its sensitivity factors in Jinci Spring Basin; these sensitivity factors included precipitation, river seepage, mining drainage, groundwater withdrawals and lateral discharge to the associated Quaternary aquifer. All the sensitivity factors were analyzed with Garson’s algorithm based on the connection weights of the neural network model. The concept of “sensitivity range” was proposed to describe the value range of the input variables to which the output variables are most sensitive. The sensitivity ranges were analyzed by a local sensitivity approach. The results showed that coal mining drainage is the most sensitive anthropogenic factor, having a large effect on groundwater level of the Jinci Spring Basin.

Modular Neural Networks for Watershed Runoff

Chapter

Jan 2000

Modeling rainfall-runoff at watershed scale is important for water resources management, safe yield computation, and design of flood control structures. Frequent droughts and floods of 1993 in Kansas and other midwestern U.S. are testimonies to the need for good predictive models at the watershed scale. The response of a watershed to precipitation is complicated by various hydrologic components that are distributed within it in a heterogeneous manner. Watershed runoff depends on geomorphologic properties (such as topology, vegetation, soil type) of the watershed and other climatic factors (precipitation, temperature, etc.) of the region. The influence of all these factors is not understood clearly. As a consequence, there exists some skepticism in the use of physically-based models for predicting watershed runoff (Grayson et al., 1992).

Application of Artificial Neural Networks to Forecasting of Surface Water Quality Variables: Issues, Applications and Challenges

Chapter

Jan 2000

Amidst growing concerns for the state of the world’s surface water resources, water quality modeling is assuming increasing importance. Thomann (1998) suggests that we are entering a’Golden Age’ of water quality modeling in which surface water models will make significant contributions towards fuller diagnosis of problems, uncovering surprises, providing a framework for decision-making and lessening future conflicts between environmental interests, managers and those who are regulated. Traditionally, there have been two main philosophical approaches to surface water quality modeling. Process based models consider the underlying physical processes directly, whereas statistical models determine relationships based on historical data sets. Of course, in reality, process based models usually require some degree of statistical calibration to historical data, whereas statistical models should be based, where possible, on relationships that have a physical basis. Recently, artificial neural networks (ANNs) have emerged as alternatives to traditional statistical models in a variety of fields, including water quality modeling.

Effective and Efficient Modeling for Streamflow Forecasting

Chapter

Jan 2000

Artificial Neural Networks are now widely applied in a broad range of fields, including image processing, signal processing, medical studies, financial predictions, power systems, and pattern recognition among others (Kosko, 1992; Refenes et al., 1994; Saund, 1989; Schalkoff, 1992; Suykens et. al. 1996; Vemuri and Rogers 1994). These successes have also inspired applications to water resources and environmental systems (Achela et al., 1998; Chang and Tsang, 1992; Derr and Slutz, 1994; French et al., 1992; Hsu et al, 1997; Hsu et al, 1995; Maier and Danday, 1996; Ranjithan and Eheart, 1993; Roger and Dowla, 1994). Because ANN models have the ability to recursively “learn from the data” they can result in significant savings in time required for model development, and are particularly useful for applications involving complicated, nonlinear processes that are not easily modelled by traditional means. This chapter addresses some issues related to the training of the class of ANNs known as Multi-layer Feedforward Neural Networks (MFNN) which are most commonly used in streamflow forecasting applications. We also present results illustrating the applicability of properly trained MFNNs in prediction of future streamflows from past rainfall and flows, and compare these results to those obtained by other modeling approaches.

Rainfall-Runoff Modeling Using Artificial Neural Networks

Article

Jul 1999

An Artificial Neural Network (ANN) methodology was employed to forecast daily runoff as a function of daily precipitation, temperature, and snowmelt for the Little Patuxent River watershed in Maryland. The sensitivity of the prediction accuracy to the content and length of training data was investigated. The ANN rainfall-runoff model compared favorably with results obtained using existing techniques including statistical regression and a simple conceptual model. The ANN model provides a more systematic approach, reduces the length of calibration data, and shortens the time spent in calibration of the models. At the same time, it represents an improvement upon the prediction accuracy and flexibility of current methods.

River Stage Forecasting Using Artificial Neural Networks

Article

Jan 1998

Methods to continuously forecast water levels at a site along a river are generally model based. Physical processes influencing occurrence of a river stage are, however, highly complex and uncertain, which makes it difficult to capture them in some form of deterministic or statistical model. Neural networks provide model-free solutions and hence can be expected to be appropriate in these conditions. Built-in dynamism in forecasting, data-error tolerance, and lack of requirements of any exogenous input are additional attractive features of neural networks. This paper highlights their use in real-time forecasting of water levels at a given site continuously throughout the year based on the same levels at some upstream gauging station and/or using the stage time history recorded at the same site. The network is trained by using three algorithms, namely, error back propagation, cascade correlation, and conjugate gradient. The training results are compared with each other. The network is verified with untrained data.

A recurrent neural network approach using indices of low-frequency climatic variability to forecast regional annual runoff

Article

Oct 2000

This paper evaluates the potential of using low-frequency climatic mode indices to forecast regional annual runoÄ in northern Quebec and the Labrador region. The impact of climatic trends in the forecast accuracy is investigated using a recurrent neural networks (RNN) approach, time-series of inflow to eight large hydropower systems in Quebec and Labrador, and indices of selected modes of climatic variability: El Nino-Southern Oscillation (ENSO), North Atlantic Oscillation (NAO), Pacific-North American (PNA), BaÅn Island-West Atlantic (BWA) and sea-level pressure (SLP) at Iceland. A wavelet analysis is used to show that the selected climatic patterns are related to annual runoÄ from 1950 to 1996 in northern Quebec. The forecast results indicate that the use of BWA, PNA and ENSO indices results in better forecast skill than the use of SLP or NAO. Overall, the use of the BWA index is found to provide the best forecast improvement (38% on average), whereas the use of PNA provides 28% of improvement on average. Using the SLP index improves the forecast accuracy by 4%, and the use of an ENSO indicator leads to an improvement of 6%. The NAO index used here is found to provide only a modest improvement. Copyright # 2000 John Wiley & Sons, Ltd.

River flow forecasting through conceptual models

Article

A Non-Linear Rainfall-Runoff Model Using an Artificial Neural Network

Article

Mar 1999
J HYDROL

A rainfall–runoff model that can be successfully calibrated (i.e., yielding sufficiently accurate results) using relatively short lengths of data, is desirable for any basin in general, and the basins of developing countries like India, in particular, for which scarcity of data is a major problem. An artificial neural network paradigm, known as the temporal back propagation neural network (TBP-NN), is successfully demonstrated as a monthly rainfall–runoff model. The performance of this model in a “scarce data” scenario (i.e., the effects of using reduced calibration periods on the performance) is compared with Volterra-type Functional Series Models (FSM), utilising the data of the River Lee (in the UK) and the Thuthapuzha River (in Kerala, India). The results confirm the TBP-NN model as being the most efficient of the black-box models tested for calibration periods as short as six years.

River flow prediction using artificial neural networks: Generalisation beyond the calibration range

Article

Jun 2000
J HYDROL

Artificial neural networks (ANNs) provide a quick and flexible means of creating models for river flow prediction, and have been shown to perform well in comparison with conventional methods. However, if the models are trained using a dataset that contains a limited range of values, they may perform poorly when encountering events containing previously unobserved values. This failure to generalise limits their use as a tool in applications where the data available for calibration is unlikely to cover all possible scenarios.This paper presents a method for improved generalisation during training by adding a guidance system to the cascade-correlation learning architecture. Two case studies from catchments in the UK are prepared so that the validation data contains values that are greater or less than any included in the calibration data. The ability of the developed algorithm to generalise on new data is compared with that of the standard error backpropagation algorithm. The ability of ANNs trained with different output activation functions to extrapolate beyond the calibration data is assessed.

Daily Reservoir Inflow Forecasting Using Artificial Neural Networks With Stopped Training Approach

Article

May 2000
J HYDROL

In this paper, an early stopped training approach (STA) is introduced to train multi-layer feed-forward neural networks (FNN) for real-time reservoir inflow forecasting. The proposed method takes advantage of both Levenberg–Marquardt Backpropagation (LMBP) and cross-validation technique to avoid underfitting or overfitting on FNN training and enhances generalization performance. The methodology is assessed using multivariate hydrological time series from Chute-du-Diable hydrosystem in northern Quebec (Canada). The performance of the model is compared to benchmarks from a statistical model and an operational conceptual model. Since the ultimate goal concerns the real-time forecast accuracy, overall the results show that the proposed method is effective for improving prediction accuracy. Moreover it offers an alternative when dynamic adaptive forecasting is desired.

Short Term Streamflow Forecasting Using Artificial Neural Networks

Article

Jan 1999
J HYDROL

The research described in this article investigates the utility of Artificial Neural Networks (ANNs) for short term forecasting of streamflow. The work explores the capabilities of ANNs and compares the performance of this tool to conventional approaches used to forecast streamflow. Several issues associated with the use of an ANN are examined including the type of input data and the number, and the size of hidden layer(s) to be included in the network. Perceived strengths of ANNs are the capability for representing complex, non-linear relationships as well as being able to model interaction effects. The application of the ANN approach is to a portion of the Winnipeg River system in Northwest Ontario, Canada. Forecasting was conducted on a catchment area of approximately 20 000 km2. using quarter monthly time intervals. The results were most promising. A very close fit was obtained during the calibration (training) phase and the ANNs developed consistently outperformed a conventional model during the verification (testing) phase for all of the four forecast lead-times. The average improvement in the root mean squared error (RMSE) for the 8 years of test data varied from 5 cms in the four time step ahead forecasts to 12.1 cms in the two time step ahead forecasts.

River Flow Forecasting Through Conceptual Models: Part 1. — A Discussion of Principles

Article

Apr 1970
J HYDROL

The principles governing the application of the conceptual model technique to river flow forecasting are discussed. The necessity for a systematic approach to the development and testing of the model is explained and some preliminary ideas suggested.

Bayesian Interpolation

Article

Jul 1991

David J. C. Mackay

Although Bayesian analysis has been in use since Laplace, the Bayesian method of model--comparison has only recently been developed in depth. In this paper, the Bayesian approach to regularisation and model--comparison is demonstrated by studying the inference problem of interpolating noisy data. The concepts and methods described are quite general and can be applied to many other problems. Regularising constants are set by examining their posterior probability distribution. Alternative regularisers (priors) and alternative basis sets are objectively compared by evaluating the evidence for them. `Occam's razor' is automatically embodied by this framework. The way in which Bayes infers the values of regularising constants and noise levels has an elegant interpretation in terms of the effective number of parameters determined by the data set. This framework is due to Gull and Skilling. 1 Data modelling and Occam's razor In science, a central task is to develop and compare models to a...

A recurrent neural networks approach using indices of low-frequency climatic variability to forecast regional annual runoff

2755

Coulibaly

Improving extreme hydrologic events forecasting using a new criterion for ANN selection

Abstract

No full-text available

Recommended publications

McMaster Engineering: Canada Excellence Research Chairs Faculty Position Available...

Black Academic Excellence Recruitment Opportunities at McMaster Engineering

McMaster Engineering is hiring widely for faculty positions

Empowering older adults to live and age on

Efficiently solving the dynamics of many-body localized systems at strong disorder

Time discontinuous Galerkin methods with energy decaying correction for non‐linear elastodynamics

Experimental determination of free surface levels at open-channel junctions: By MARCIA MARIA LARA PI...

On the meaning and the accuracy of the pressure-flow technique to determine constriction areas withi...