Streamflows forecast with Neural Networks

The accurate forecasts and estimations of the amount of sediment transported by rivers are critical concerns in water resource management and soil and water conservation. The identification of appropriate and applicable models or improvements in existing approaches is needed to accurately estimate the suspended sediment concentration (SSC). In recent decades, the utilization of intelligent models has substantially improved SSC estimation. The identification of beneficial and proper input parameters can greatly improve the performance of these smart models. In this regard, we assessed the C-factor of the revised universal soil loss equation (RUSLE) as a new input along with hydrological variables for modeling SSC. Four data-driven models (feed-forward neural network (FFNN); support vector regression (SVR); adaptive neuro-fuzzy inference system (ANFIS); and radial basis function (RBF)) were applied in the Boostan Dam Watershed, Iran. The cross-correlation function (CCF) and partial au...

Establishment of rating curves are often required by the hydrologists for flow estimates in the streams, rivers etc. Measurement of discharge in a river is a time-consuming, expensive, and difficult process, and the conventional approach of regression analysis of stage-discharge relation does not provide encouraging results especially during the floods. Present study is aimed at the application of support vector machines (SVMs) based algorithm for modelling stage-discharge relation including the hysteresis effect. A data set of two discharge-measuring stations located on two Indian rivers has been used for analysis in the present study. A back propagation neural network model was employed in order to compare the performance of the results based on support vector machines based modelling technique. The outcome of the study suggests that the support vector machines works well for both the data sets and produce promising results in comparison to the neural network technique. Finally, the results also suggest the suitability of SVMs algorithm in predicting the looped rating curve having hysteresis effect.

Rainfall is considered as one of the major components of the hydrological process; it takes significant part in evaluating drought and flooding events. Therefore, it is important to have an accurate model for rainfall prediction. Recently, several data-driven modeling approaches have been investigated to perform such forecasting tasks as multilayer perceptron neural networks (MLP-NN). In fact, the rainfall time series modeling (SARIMA) involvesimportant temporal dimensions. In order to evaluate the incomes of both models, statistical parameters were used to make the comparison between the two models. These parameters include the Root Mean Square Error RMSE, Mean Absolute Error MAE, Coefficient Of Correlation CC and BIAS. Two-Third of the data was used for training the model and One-third for testing.

Many river basins of the western U.S. were initially developed for crop irrigation, mostly after the passage of the Reclamation Act of 1906. The diversion for irrigation reduced streamflow and increased salinity downstream. Reduced streamflow, especially reduced bank overflow, caused salinization of floodplains in the presence of a high water table. This was accompanied by the spread of salt-tolerant riparian species, especially saltcedar (Tamarisk sp.). This scenario is repeated in many river systems in the Southwest. There are considerable interests in restoring degradated river systems. Saltcedar control is for example, widely implemented through chemical application or biological control methods. It is viewed as the first step towards restoring the riparian ecosystem. In addition, there has been an expectation that saltcedar control may result in increased flow and reduced salinity. However, this expectation has to be examined carefully. Controlled flooding through reservoir water release has also been attempted in an effort to maintain the original ecosystems. Increasing efforts are also being made to identify and control saline water intrusion in an attempt to lower salinity. The models described in this series were developed mainly to help evaluate river management options on flow and salinity of the stream and the floodplains. The first part deals with water and salt balance in reservoirs. The primary purpose of the model is to predict outflow salinity from the reservoir storage and inflow information in advance of the actual release. A simple two-layer model is used to describe the reservoir processes. The second part addresses water and salt transport through semi-arid river channels. A routing model referred to as ROTO (developed at the Blackland Soil and Water Conservation Laboratory at Temple, Texas) was used to simulate streamflow and salinity on a daily time step. This model also provides the estimate of evaporation and percolation losses of water, the interaction between streamflow and shallow groundwater, and the bank overflow required for estimating salt leaching from the floodplains. Salt transport processes were added to the ROTO model so as to estimate the changes in streamflow salinity. The third part concerns with water and salt storage in the riparian zones. The models are coded in FORTRAN and are available upon request. The models were applied to the middle reach of the Pecos River (Malaga, New Mexico, to Girvin, Texas) and, to a limited extent, to selected sites along the Rio Grande, and were found workable using mostly existing database.

The dependence of agricultural production on water resources is a known fact. Therefore, understanding hydrological processes and events in agricultural production form the basis of effective and reliable management of water resources. Many traditional methods used for the analysis of time-dependent variables in hydrology and meteorology assume linear relationships between these variables. However, the temporal changes of these parameters are quite complex and therefore cannot be easily modeled by conventional estimation methods. Artificial neural networks (ANNs), on the other hand, allow the analysis of nonlinear relationships or processes whose statistical or mathematical calculations cannot be determined in such systems. ANNs have been accepted as a successful model in multidimensional research involving dynamic processes in the field of hydrology. This study, it is aimed at the modeling of stream flows in Bektas Creek. With the aim of modeling, daily meteorological parameters (precipitation, temperature, sunbathing time, relative humidity) measured in Kangal Region and one day delayed flows were used. Streamflow forecasts are simulated with the Generalized Regression Neural Network (GRNN). To reveal the difference of the GRNN model from other ANNs, the same data were also used in the feed-forward neural network (FFNN) model. Model performances were evaluated taking into account the Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and Determination coefficient (R).

In this study, daily rainfall-runoff relationships for Sohu Stream were modelled using an artificial neural network (ANN) method by including the feed-forward back-propagation method. The ANN part was divided into two stages. During the first stage, current flows were estimated by using previously measured flow data. The best network architecture was found to utilise two neurons in the input layer (the delayed flows from the first and second days), two hidden layers, and one output layer (the current flow). The coefficient of determination (R 2) in this architecture was 81.4%. During the second stage, the current flows were estimated by using a combination of previously measured values for precipitation, temperature, and flows. The best architecture consisted of an input layer of 2 days of delayed precipitation, 3 days of delayed flows, and temperature of the current. The R 2 in this architecture was calculated to be 85.5%. The results of the second stage best reflected the real-world situation because they accounted for more input variables. In all models, the variables with the highest R 2 ranked as the previous flow (81.4%), previous precipitation (21.7%), and temperature.

Climate change has influenced several of the water cycle related variables such as rainfall that contribute to increasing natural disasters. To establish new methodologies for rivers level forecasting is necessary for the implementation of early warning systems. In this work, we present results of a multilayer perceptron artificial neural network (ANN) to forecast temporal series of water levels at the outlet of Rio Negro river with 24-hour antecedence. Input data was collected by a set of hydrological monitoring stations composed of water level and rainfall measures acquired with a one-day resolution. Water-level prediction were evaluated by the Nash-Sutcliffe coefficient (NSE) and by the root mean square error (RMSE). The results show consistency between predicted and observed values, especially when combining both water level and rainfall data. In such case, values of NSE reached 0.93 to 0.54 and RMSE between 0.028 and 0.061 for antecedence of 1 to 7 days respectively with implemented topology for the empirical model.

Stream flow (SF) prediction is considered as a very complex due to the hydrological systems of surface water are complex and dynamic. The reliable prediction of stream flow (SF) can be performed by either conceptual or data-driven based models. In the modelling of hydrological processes, the support vector machine (SVM) is a novel, data-driven approach. Hence, six SVM-based models were generated in this study to predict real time hourly SF in the Selangor River Basin from the water level and rainfall of upstream stations. These models composed of six different combinations of input variables and were trained and tested under hourly records of SF, rainfall, and water level over one year (2011). Among the SVM-based models, SVM-M6, which has nine input variables, was the most effective. Under the training and testing data sets, its correlation coefficient and mean absolute error values were 0.992, 0.953, 0.061 and 0.253 respectively.

The prediction of the runoff generated within a watershed is an important input in the design and management of water resources projects. Due to the tremendous spatial and temporal variability in precipitation, rainfall-runoff relationship becomes one of the most complex hydrologic phenomena. Under such circumstances, using soft computing approaches have proven to be an efficient tool in modeling of runoff. These models are capable of predicting river runoff values that can be used for hydrologic and hydraulic engineering design and water management purposes. It has been observed that the artificial neural networks (ANN) model performed well compared to other soft computing techniques such as fuzzy logic and radial basis function investigated in this study. In addition, comparison of scatter plots indicates that the values of runoff predicted by the ANN model are more precise than those found by RBF or Fuzzy Logic model.

In this study, daily rainfall-runoff relationships for Sohu Stream were modelled using an artificial neural network (ANN) method by including the feed-forward back-propagation method. The ANN part was divided into two stages. During the first stage, current flows were estimated by using previously measured flow data. The best network architecture was found to utilise two neurons in the input layer (the delayed flows from the first and second days), two hidden layers, and one output layer (the current flow). The coefficient of determination (R 2) in this architecture was 81.4%. During the second stage, the current flows were estimated by using a combination of previously measured values for precipitation, temperature, and flows. The best architecture consisted of an input layer of 2 days of delayed precipitation, 3 days of delayed flows, and temperature of the current. The R 2 in this architecture was calculated to be 85.5%. The results of the second stage best reflected the real-world situation because they accounted for more input variables. In all models, the variables with the highest R 2 ranked as the previous flow (81.4%), previous precipitation (21.7%), and temperature.

Due to the various influencing factors on river suspended sediment transportation, determining an appropriate input combination for developing the suspended sediment load forecasting model is very important for water resources management. The influence of pre-processing of input variables by Gamma Test (GT) was investigated on performance of Support Vector Machine (SVM) with two kernels; Radial Basis Function (RBF) and polynomial in order to forecast daily suspended sediment amount in the period between 1983 and 2014 at Korkorsar basin, northern Iran. The best input combination was identified using GT and correlation coefficient analysis. Then, the SVM model was developed and the suspended sediment amount was forecasted with RBF and polynomial kernels. The obtained results in testing phase showed that GT-SVM (RBF kernel) model can estimate suspended sediment more accurately with the lowest RMSE (14.045 ton/day), highest correlation coefficient (0.88) and highest NSEC coefficient (0.88) than SVM (RBF kernel) model (RMSE = 18.36ton/day, R 2 ¼ 0:79; NSEC ¼ 0:73) and had a better performance than the other models. The results indicated that in forecasting the first nine maximum values of suspended sediment load, GT-SVM (RBF) had a higher capability than the other models and could provide a more accurate estimation from the maximum rate of suspended sediment. The results of this study showed the capability of identifying the priority of the input parameters can change GT to a useful and technical test for input variables pre-processing to forecast the amount of suspended sediments.

The accurate forecasts and estimations of the amount of sediment transported by rivers are critical concerns in water resource management and soil and water conservation. The identification of appropriate and applicable models or improvements in existing approaches is needed to accurately estimate the suspended sediment concentration (SSC). In recent decades, the utilization of intelligent models has substantially improved SSC estimation. The identification of beneficial and proper input parameters can greatly improve the performance of these smart models. In this regard, we assessed the C-factor of the revised universal soil loss equation (RUSLE) as a new input along with hydrological variables for modeling SSC. Four data-driven models (feed-forward neural network (FFNN); support vector regression (SVR); adaptive neuro-fuzzy inference system (ANFIS); and radial basis function (RBF)) were applied in the Boostan Dam Watershed, Iran. The cross-correlation function (CCF) and partial au...

The prediction of the runoff generated within a watershed is an important input in the design and management of water resources projects. Due to the tremendous spatial and temporal variability in precipitation, rainfall-runoff relationship becomes one of the most complex hydrologic phenomena. Under such circumstances, using soft computing approaches have proven to be an efficient tool in modeling of runoff. These models are capable of predicting river runoff values that can be used for hydrologic and hydraulic engineering design and water management purposes. It has been observed that the artificial neural networks (ANN) model performed well compared to other soft computing techniques such as fuzzy logic and radial basis function investigated in this study. In addition, comparison of scatter plots indicates that the values of runoff predicted by the ANN model are more precise than those found by RBF or Fuzzy Logic model.

River streamflow is an essential hydrological parameters for optimal water resource management. This study investigates models used to estimate monthly time-series river streamflow data at two hydrological stations in the USA (Heise and Irwin on Snake River, Idaho). Five diverse types of machine learning (ML) model were tested, support vector machine-radial basis function (SVM-RBF), SVM-Polynomial (SVM-Poly), decision tree (DT), gradient boosting (GB), random forest (RF), and long short-term memory (LSTM). These were trained and tested alongside a conventional multiple linear regression (MLR). To improve the estimation and model performance, hybrid models were designed by coupling the models with wavelet theory (W). The models performance was assessed using root mean square error (RMSE), mean absolute error (MAE), coefficient of determination (R 2), Nash-Sutcliffe efficiency (NSE), and Willmott's index (WI). A sideby-side performance assessment of the stand-alone and hybrid models revealed that the coupled models exhibit better estimates of monthly river streamflow relative to the stand-alone ones. The statistical parameter values for the best model (W-LSTM4) during the test phase was RMSE = 36.533 m 3 /s, MAE = 26.912 m 3 /s, R 2 = 0.947, NSE = 0.946, WI = 0.986 (Heise station), and RMSE = 33.378 m 3 /s, MAE = 24.562 m 3 /s, R 2 = 0.952, NSE = 0.951, WI = 0.987 (Irwin station).

Stream flow prediction provides the information of various problems related to the design and effective operation of river balancing system. So it is an essentially important aspect of any watershed modelling. The black box models(ANN) have proven to be an efficient alternative to physical (traditional) methods for simulating streamflow and sediment yield of the catchments. This present study focusses on development of models using ANN for predicting the stream flow for Subarnarekha river basin. By reviewing the earlier research works, it is observed that the procedure addresses the selection of input variables, the definition of model architecture and the strategy of the learning process. The input vectors used for the models were daily rainfall, mean daily evaporation, mean daily temperature and lag streamflow. Further, it is observed that the model performance was evaluated by statistical parameters like root-mean square error measures (RMSE), Nash-Sutcliffe efficiency (N-S) and ...

Runoff prediction, as a nonlinear and complex process, is essential for designing canals, water management and planning, flood control and predicting soil erosion. There are a number of techniques for runoff prediction based on the hydro-meteorological and geomorphological variables. In recent years, several soft computing techniques have been developed to predict runoff. There are some challenging issues in runoff modeling including the selection of appropriate inputs and determination of the optimum length of training and testing data sets. In this study, the gamma test (GT), forward selection and factor analysis were used to determine the best input combination. In addition, GT was applied to determine the optimum length of training and testing data sets. Results showed the input combination based on the GT method with five variables has better performance than other combinations. For modeling, among four techniques: artificial neural networks, local linear regression, an adaptiv...

Rainfall forcasting is a non-linear forecasting process that varies according to area and strongly influenced by climate change. It is a difficult process due to complexity of rainfall trend in the previous event and the popularity of Adaptive Neuro Fuzzy Inference System (ANFIS) with hybrid learning method give high prediction for rainfall as a forecasting model. Thus, in this study we investigate the efficient membership function of ANFIS for predicting rainfall in Banyuwangi, Indonesia. The number of different membership functions that use hybrid learning method is compared. The validation process shows that 3 or 4 membership function gives minimum RMSE results that use temperature, wind speed and relative humidity as parameters.

Abstract: Engineering project design, environmental impact analysis and water resources problems, particularly hydrological simulation and forecasting often require the estimation of streamflow. Simulation and forecasting techniques of streamflow time series have been applied widely for determining the dimensions of hydraulic works for risk assessment in urban water supply systems and irrigation, optimal operation of reservoir systems, optimization hydroelectric production, and so on.

The dependence of agricultural production on water resources is a known fact. Therefore, understanding hydrological processes and events in agricultural production form the basis of effective and reliable management of water resources. Many traditional methods used for the analysis of time-dependent variables in hydrology and meteorology assume linear relationships between these variables. However, the temporal changes of these parameters are quite complex and therefore cannot be easily modeled by conventional estimation methods. Artificial neural networks (ANNs), on the other hand, allow the analysis of nonlinear relationships or processes whose statistical or mathematical calculations cannot be determined in such systems. ANNs have been accepted as a successful model in multidimensional research involving dynamic processes in the field of hydrology.

This study, it is aimed at the modeling of stream flows in Bektas Creek. With the aim of modeling, daily meteorological parameters (precipitation, temperature, sunbathing time, relative humidity) measured in Kangal Region and one day delayed flows were used. Streamflow forecasts are simulated with the Generalized Regression Neural Network (GRNN). To reveal the difference of the GRNN model from other ANNs, the same data were also used in the feed-forward neural network (FFNN) model. Model performances were evaluated taking into account the Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and Determination coefficient (R).

Keywords: Meteorological Parameters, Streamflow Prediction, Generalized Regression Neural Network, Feed-Forward Neural Network.

Forecasting future response behaviour of a semi-arid catchment in terms of runoff coefficient being trivial, an attempt has been made to apply an artificial neural network (ANN) model to forecast the runoff coefficients (ROC) for the rapidly urbanizing Notwane catchment system in Botswana. Runoff coefficients computed from 1978 to 2000, by the water balance technique have been used to develop the optimal network architecture with appropriate choice of the size of input vectors, number of hidden layers and number of neurons in the hidden layers, training algorithms and transfer functions for the network. The developed network has then been used to simulate the runoff coefficients for the above period. Based on its performance in terms of reproducibility of the water balance runoff coefficients, the network was used to forecast the runoff coefficients up to 2020 and it was found that the increase in runoff coefficients is about 1% per year. Based on the weights given by the network to the input variables, it was found that while about 48% contribution came from climatic factors, 52% came from the land use/land cover.

This research was conducted to present an integrated rainfall-runoff model based on the physical characteristics of the watershed, and to predict discharge not only in the outlet, but also at any desired point within the basin. To achieve this goal, a matrix of hydro-climatic variables (i.e. daily rainfall and daily discharge) and geomorphologic characteristics such as upstream drainage area (A), mean slope of watershed (S) and curve number (CN) was designed and simulated using artificial intelligence techniques. Integrated Geomorphology-based Artificial Neural Network (IGANN) model with Root Mean Squared Error (RMSE) of 0.02786 m3 s-1 and Nash-Sutcliffe Efficiency (NSE) of 0.9403 and Integrated Geomorphology-based Adaptive Neuro-Fuzzy Inference System (IGANFIS) model with RMSE of 0.02795 m3 s-1 and NSE of 0.94467 were able to predict the discharge values of all hydrometric stations of the Chalus River watershed with a very low error and high accuracy. The results of cross validation stage confirmed the efficiency of models. Hydro-climatic variables and geomorphologic parameters selected in the study were: discharge of one day ago, discharge of two days ago, rainfall of current day and rainfall of one day ago and S, CN and A, respectively. In addition, the IGANN model shows superiority compared with the IGANFIS model.

https://www.ijert.org/application-of-artificial-neural-network-in-hydrology-a-review

https://www.ijert.org/research/application-of-artificial-neural-network-in-hydrology-a-review-IJERTV4IS060247.pdf

In this research paper exhaustive review is conducted on artificial neural network (ANN) which is employed in the field of hydrological related problems, whereas the conventional approaches are cumbersome and complex in view of computational analysis. Artificial intelligence operation can be well exemplified through application in rainfall-runoff modelling, modelling streamflow, water quality modelling, and application in ground water. A good physical understanding of the hydrologic process being modelled can help in selecting the input vector and designing a more efficient network. This review provides examples for ANN model that provides reasonable accuracy for hydrological problems, and a more effective tool for engineering applications.

For highly nonlinear and complex problems like rainfall runoff models, artificial neural networks (ANNS) are highly developed empirical models available and are increasingly common in the analysis of hydrology and water resource problems. The aim of this study was to develop rainfall runoff models for Barak basin. In this study, feed-forward back propagation neural network (FFBPNN) and layer recurrent neural network (LRNN) are used to predict runoff during monsoon period. The performance of both the models (FFBPNN, LRNN) for runoff over 33 years of data (1979-2011) has been evaluated for two watersheds namely Katakhal and Sonai of Barak basin. Two neural networks with the evaluation criteria mean square error (MSE), root mean square error (RMSE) and coefficient of regression (R 2) FFBPNN perform best with architecture 3-9-1 following Logsig function. This study confirms that FFBPNN and LRNN are useful techniques for predicting runoff in both of the watersheds of Barak Basin.

Since fresh water is limited while agricultural and human water demands are continuously increasing, optimal prediction and management of streamflows as a source of fresh water is crucially important. This study investigates and demonstrates how data preprocessing and data mining techniques would improve the accuracy of streamflow predictive models. Based on easily accessible Snow Telemetry data (SNOTEL), four streamflow prediction models – autoregressive integrated moving average (ARIMA), artificial neural networks (ANNs), a hybrid-model of ANN and ARIMA (ANN-ARIMA), and an adaptive neuro fuzzy inference system (ANFIS) – were developed and utilized in a streamflow prediction process on Elephant Butte Reservoir. Utilizing the statistical correlation analysis and the extracting importance degrees of predic-tors led to efficiently select the most effective predictors for daily and monthly streamflow to Elephant Butte Reservoir. For the daily prediction time step, by preprocessing the historical data and extracting and utilizing the extracted climate variability indices through data mining techniques, the ANFIS model achieved a superior streamflow prediction performance for Elephant Butte Reservoir compared to the other three evaluated prediction models. Additionally, for predicting monthly streamflow to the Elephant Butte Reservoir, ANFIS showed significantly higher accuracy than the ANNs. As an optimal application of the developed predictive expert systems, successful integrating the prediction models in integrated reservoir operations balanced the need for a reliable supply of irrigation water against losses through evaporation. The optimal operation plan significantly minimizes the total evaporation loss from both reservoirs by providing the optimal storage levels in both reservoirs. This study provides the conceptual procedures of non-seasonal (ARIMA) model, and since the model is univariate, it demonstrates a strongly-reliable inflow prediction when existing information is limited to streamflow data as a predictor.

Rainfall runoff is a very complicated process due to its nonlinear and multidimensional dynamics, and hence difficult to model. There are several options for a modeller to consider, for example: the type of input data to be used, the length of model calibration (training) data and whether or not the input data be treated as signals with different frequency bands so that they can be modelled separately. This paper
describes a new hybrid modelling scheme to answer the above mentioned questions. The proposed methodology
is based on a hybrid model integrating wavelet transformation, a modelling engine (Artificial
Neural Network) and the Gamma Test. First, the Gamma Test is used to decide the required input data dimensions and its length. Second, the wavelet transformation decomposes the input signals into different frequency bands. Finally, a modelling engine (ANN in this study) is used to model the decomposed signals separately. The proposed scheme was tested using the Brue catchment, Southwest England, as
a case study and has produced very positive results. The hybrid model outperforms all other models tested. This study has a wider implication in the hydrological modelling field since its general framework could be applied to other model combinations (e.g., model engine could be Support Vector Machines, neuro-fuzzy systems, or even a conceptual model. The signal decomposition could be carried out by Fourier
transformation).

The relationship between rainfall and runoff is highly complex and nonlinear, and depends on a large number of parameters and characteristics of a watershed. The purpose of this paper is to investigate the relationship between rainfall and runoff, applying Artificial Neural Networks (ANN). The Hydrological watershed of Pineios river upstream of Sarakina's bridge has been used as the case study. In this study, daily river flow data were obtained from Sarakina's gauging station which is located at the exit of the watershed. Similarly, daily rainfall data were obtained from three rain gauging stations located in the watershed. An Artificial Neural Network featuring three layers (input, hidden, output) was chosen for this application, while the simple Backpropagation was used as the training method. The network is able to predict the river flow after a certain rainfall taking under consideration the river flow data of the day before.

Stream flow forecasting can be an appropriate indicator in estimating future conditions for water resources management. The present study aimed to compare the efficiency of Support Vector Machine (SVM), Adaptive Neural Fuzzy Inference Systems (ANFIS) and conceptual hydrological model of MIKE11/NAM in simulating the daily stream flow. The studied area is Eskandari basin located in Iran. For this purpose, a ten-year period (1999-2009) of daily data including rainfall, runoff, temperature and evaporation were used. Furthermore, the performances of the models in flow simulation were investigated using statistical indicators of correlation coefficient (R 2 ), Root Mean Square Error (RMSE) and the Nash-Sutcliffe (NS) coefficient. The results showed that every three models possess an appropriate performance and efficiency in the studied area. During testing (verification) period, SVM with the highest correlation coefficient (R 2 =0.99) and lowest RMSE equal to (RMSE=2.13 s m / 3 ), had a better performance than ANFIS model (R 2 =0.82, RMSE=3.21 s m / 3 ) and NAM model (R 2 =0.75, RMSE=3.48 s m / 3 ). In addition, Nash-Sutcliffe coefficient for SVM, ANFIS and NAM models were 0.99, 0.79 and 0.70, respectively.

This study provides a unique opportunity to analyze the issue of flow forecast based on the soil and water assessment tool (SWAT) and artificial neural network (ANN) models. In last two decades, the ANNs have been extensively applied to various water resources system problems. In this study, the ANNs were applied to the daily flow of the Pracana basin in Portugal. The comparison of ANN models and a process-based model SWAT was established based on their prediction accuracy. The ANN model was found to be more successful than the SWAT in relation to better forecast of peak flow. Nevertheless the SWAT model results revealed a better value of mean squared error. The results of this study, in general, showed that ANNs can be powerful tools in daily flow forecasts.

This study provides a unique opportunity to analyze the issue of flow forecast based on the soil and water assessment tool (SWAT) and artificial neural network (ANN) models. In last two decades, the ANNs have been extensively applied to various water resources system problems. In this study, the ANNs were applied to the daily flow of the Pracana basin in Portugal. The comparison of ANN models and a process-based model SWAT was established based on their prediction accuracy. The ANN model was found to be more successful than the SWAT in relation to better forecast of peak flow. Nevertheless the SWAT model results revealed a better value of mean squared error. The results of this study, in general, showed that ANNs can be powerful tools in daily flow forecasts.

This study provides a unique opportunity to analyze the issue of flow forecast based on the soil and water assessment tool (SWAT) and artificial neural network (ANN) models. In last two decades, the ANNs have been extensively applied to various water resources system problems. In this study, the ANNs were applied to the daily flow of the Pracana basin in Portugal. The comparison of ANN models and a process-based model SWAT was established based on their prediction accuracy. The ANN model was found to be more successful than the SWAT in relation to better forecast of peak flow. Nevertheless the SWAT model results revealed a better value of mean squared error. The results of this study, in general, showed that ANNs can be powerful tools in daily flow forecasts.

This study provides a unique opportunity to analyze the issue of flow forecast based on the soil and water assessment tool (SWAT) and artificial neural network (ANN) models. In last two decades, the ANNs have been extensively applied to various water resources system problems. In this study, the ANNs were applied to the daily flow of the Pracana basin in Portugal. The comparison of ANN models and a process-based model SWAT was established based on their prediction accuracy. The ANN model was found to be more successful than the SWAT in relation to better forecast of peak flow. Nevertheless the SWAT model results revealed a better value of mean squared error. The results of this study, in general, showed that ANNs can be powerful tools in daily flow forecasts.

The objective of this study is to validate a flow prediction model for a hydrometric station using a multistation criterion in addition to standard single-station performance criteria. In this contribution we used cluster analysis to identify the regional flow height, i.e., waterlevel patterns and validate the output of an artificial neural network (ANN) model of the Alportel River in Portugal. Measurements of precipitation, temperature, and flow height were used as input variables to the ANN model with a lead time of 12 h. The lead time of 12 h is assumed to be appropriate for a short-term hydrological prediction since it is meaningful for physical processes. The ANN model with three inputs, four hidden neurons, and ten epochs was tested using the new model-validation criterion. The high performance of the model (i.e., Nash-Sutcliffe coefficient is equal to 0.922) was confirmed by the cluster-analysis criterion. It can be concluded that a multistation-based approach can be used as an additional validation criterion and might result in a rejection of a model which initially passed a single-station validation criterion.

In coastal and open ocean human activities, there is an increasing demand for accurate estimates of future sea state. In these activities, predictions of wave heights and periods are of particular importance. In this study, two different neural network strategies were employed to forecast significant wave heights and zero-up-crossing wave periods 3, 6, 12 and 24 h in advance. In the first approach, eight simple separate neural nets were implemented to simulate every wave parameter over each prediction interval. In the second approach, only two networks provided simultaneous forecasts of these wave parameters for the four prediction intervals. Two independent sets of measurements from a directional wave buoy moored off the Portuguese west coast were used to train and to validate the artificial neural nets. Saliency analysis of the results permitted an optimization of the networks' architectures. The optimal learning algorithm for each case was also determined. The short-term forecasts of the wave parameters verified by actual observations demonstrate the suitability of the artificial neural technique. r

This study provides a unique opportunity to analyze the issue of flow forecast based on the soil and water assessment tool (SWAT) and artificial neural network (ANN) models. In last two decades, the ANNs have been extensively applied to various water resources system problems. In this study, the ANNs were applied to the daily flow of the Pracana basin in Portugal. The comparison of ANN models and a pro- cess-based model SWAT was established based on their prediction accuracy. The ANN model was found to be more successful than the SWAT in relation to better forecast of peak flow. Nevertheless the SWAT model results revealed a better value of mean squared error. The results of this study, in general, showed that ANNs can be powerful tools in daily flow forecasts.

Keywords: Flow, Forecast, Artificial neural networks, SWAT, Portugal