Evaluation and forecasting of daily groundwater outflow in a small chalky watershed

This paper deals with prediction of the response of karstic springs by means of artificial neural networks (ANNs). A feed-forward back propagation ANN with three layers has been developed, to predict flow rates of two karstic springs, located at Rouvas area, Crete, Greece, using rainfall data as input. While the number of neurons of the input and output layers was determined by choice of data and desired output respectively, the number of neurons of the hidden layer was decided by means of numerous tests. Data used in ANN training and testing include daily and monthly precipitation depths (from September, 2006 to December, 2010) and measured flow rates of the two springs (from April, 2007 to December, 2010. Results show that the trained artificial neural network performed well, although flow rate measurements were not very regular. Moreover, the possibility of estimating the flow rate of one spring, based on measurements of the other has been investigated. Again the ANN gave satisfactory results. All spring flow rate and rainfall measurements are presented as an appendix, to facilitate further scientific research in the area of ANN application to water resources management.

show abstract

“…Combinations of ANNs with deterministic models have also been used (e.g. Lallahem and Mania, 2003;Chen and Adams, 2006).…”

Section: Artificial Neural Network and Their Usefulnessmentioning

confidence: 99%

Measurement and prediction of karstic spring flow rates

Darivianakis

Katsifarakis²,

Vafeiadis³

2015

Issue 2

View full text Add to dashboard Cite

show abstract

“…After the numbers of neurons in the first, hidden and last layers were fixed, and in order to make the results statistically more plausible, three different partitions of the database were used for the simulation. The database was divided into training, validation and test groups: 50% of the data was selected for the training phase, 25% for that of validation and the remaining 25% was used to test the network's prediction ability (Lallahem and Mania, 2003).…”

Section: E D I T E R R a N E A N S E Amentioning

confidence: 99%

On the use of finite‐difference and neural‐network models to evaluate the impact of underground water overexploitation

Hani

Lallahem

Mania

et al. 2006

Hydrological Processes

Self Cite

View full text Add to dashboard Cite

Abstract:The purpose of this study is to include expert knowledge as one part of the modelling system and thereby offer the chance to create a productive interactive system between expert, mathematical model, ASM, and artificial neural networks (ANNs). An attempt to determine outflow-influencing parameters in order to simulate spring flow is presented. The Bouteldja dune aquifer is fed by rains and streaming water on the sandy argillaceous relieves in the Est. The lateral passage to the gravel of the Bouteldja Plain is marked by numerous bogs that correspond to the piezometric level. These bogs have long been an environment for migratory birds and a natural reserve for many species. However, the continued exploitation of about 30 wells has negatively influenced the hydrodynamic equilibrium of the aquifer and has brought a diminution of the sources' capacity.In this study, we tried by using a hydrodynamic model and the neural network to ascertain the state of the resources and to identify the factors responsible for the decreasing flows of the three principal springs of the area (Bougles, Bourdim and Titteri) by using neural networks. The results obtained show a continued exhaustion of the reserve since 1986 with a large cone of depression. The ANNs show that the decrease in flows of the springs is not only due to the unfavourable climatic conditions, but also to the intensive exploitation of the aquifer. These results show that the groundwater reserves are decreasing over time, thus highlighting the need to take some urgent measures to stop this phenomenon.

show abstract

“…Their applications range from the forecasting of hourly and daily river stages or discharges (Thirumalaiah & Deo, 1998;Campolo et al, 1999Campolo et al, , 2003Imrie et al, 2000), groundwater modelling (Yang et al, 1997;Lallahem & Mania, 2003), and reservoir operation (Jain et al, 1999;Hasebe & Nagayama, 2002) to rainfall-runoff modelling (Shamseldin, 1997;Tokar & Markus, 2000).…”

Section: Introductionmentioning

confidence: 99%

Comparison of fuzzy inference systems for streamflow prediction

Özger

2009

Hydrological Sciences Journal

View full text Add to dashboard Cite

A comparison of the Mamdani and the Takagi-Sugeno (TS) fuzzy inference systems is presented for predicting streamflow values. The TS fuzzy rule base uses linear functions of inputs to predict the output, whereas the Mamdani version of inference determines outputs through fuzzy sub-sets. A genetic algorithmtrained Mamdani system is applied for streamflow forecasting. All the uncertainties and model complications are treated in linguistic expressions in the form of IF-THEN statements. Fuzzy membership functions, rules and the type of defuzzification method are adjusted until the best correlation between measured and predicted values is reached. The two methods are then applied to flow predictions on the Euphrates River in Turkey, without employing exogenous variables such as rainfall. The advantages and disadvantages of the models are discussed using the case study. It is shown that the Mamdani type of fuzzy inference modelling outperforms the Takagi-Sugeno approach in terms of error criteria comparisons, but neither of the two outperforms a standard ARMA(2, 2) model.Key words fuzzy logic; genetic algorithms; Mamdani; Takagi-Sugeno; prediction; streamflow; Euphrates RiverComparaison de systèmes d'inférence floue pour la prévision de débit en rivière Résumé Une comparaison des systèmes d'inférence floue de Mamdani et de Takagi-Sugeno (TS) est présentée vis-à-vis de la prévision de débit en rivière. La règle floue de base de TS utilise des fonctions linéaires des entrées pour prévoir la sortie, tandis que la version de Mamdani de l'inférence détermine les sorties à partir de sous-ensembles flous. Un système de Mamadani entraîné par algorithme génétique est appliqué pour la prévision de débit en rivière. Toutes les incertitudes et les complications de modélisation sont traitées sous la forme d'expressions linguistiques du type SI-ALORS. Les fonctions d'appartenance et les règles floues, ainsi que la méthode de défuzzification, sont ajustées jusqu'à l'obtention de la meilleure corrélation entre les données mesurées et prévues. Les deux méthodes sont ensuite appliquées à la prévision des débits de l'Euphrate en Turquie, sans utiliser de variables exogènes telles que la pluie. Les avantages et les désavantages des modèles sont discutés pour l'étude de cas. Il apparaît que la modélisation par inférence floue de type Mamdani donne de meilleurs résultats que l'approche TS en termes de comparaison des critères d'erreur, mais qu'aucune des deux ne surpasse un modèle standard ARMA(2, 2).

show abstract

Evaluation and forecasting of daily groundwater outflow in a small chalky watershed

Cited by 30 publications

References 11 publications

Measurement and prediction of karstic spring flow rates

Measurement and prediction of karstic spring flow rates

On the use of finite‐difference and neural‐network models to evaluate the impact of underground water overexploitation

Comparison of fuzzy inference systems for streamflow prediction

Contact Info

Product

Resources

About