An artificial neural network model for generating hydrograph from hydro-meteorological parameters

Ahmad, Sajjad; Simonović́, Slobodan P.

doi:10.1016/j.jhydrol.2005.03.032

Cited by 106 publications

(54 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Machine-learning approaches, such as artificial neural networks (ANN) and support vector machines (SVM), are another group of models that have been applied during past decades for simulating various hydrological processes including soil water dynamics (Jiang and Cotton, 2004;Ahmad and Simonovic, 2005;Elshorbagy and Parasuraman, 2008;Zou et al, 2010;Dai et al, 2011;Asefa et al, 2006;Yu and Liong, 2007;Lin et al, 2009;Liu et al, 2010;Deng et al, 2011). These approaches provide a great prediction capacity and do not require the knowledge of soil physical properties.…”

Section: Introductionmentioning

confidence: 99%

A comparison of numerical and machine-learning modeling of soil water content with limited input data

Karandish

Šimůnek

2016

Journal of Hydrology

133

View full text Add to dashboard Cite

Keywords:Machine-learning models ANFIS HYDRUS-2D MLR SVM Soil water content Water saving irrigation a b s t r a c t Soil water content (SWC) is a key factor in optimizing the usage of water resources in agriculture since it provides information to make an accurate estimation of crop water demand. Methods for predicting SWC that have simple data requirements are needed to achieve an optimal irrigation schedule, especially for various water-saving irrigation strategies that are required to resolve both food and water security issues under conditions of water shortages. Thus, a two-year field investigation was carried out to provide a dataset to compare the effectiveness of HYDRUS-2D, a physically-based numerical model, with various machine-learning models, including Multiple Linear Regressions (MLR), Adaptive Neuro-Fuzzy Inference Systems (ANFIS), and Support Vector Machines (SVM), for simulating time series of SWC data under water stress conditions. SWC was monitored using TDRs during the maize growing seasons of 2010 and 2011. Eight combinations of six, simple, independent parameters, including pan evaporation and average air temperature as atmospheric parameters, cumulative growth degree days (cGDD) and crop coefficient (K c ) as crop factors, and water deficit (WD) and irrigation depth (In) as crop stress factors, were adopted for the estimation of SWCs in the machine-learning models. Having Root Mean Square Errors (RMSE) in the range of 0.54-2.07 mm, HYDRUS-2D ranked first for the SWC estimation, while the ANFIS and SVM models with input datasets of cGDD, K c , WD and In ranked next with RMSEs ranging from 1.27 to 1.9 mm and mean bias errors of À0.07 to 0.27 mm, respectively. However, the MLR models did not perform well for SWC forecasting, mainly due to non-linear changes of SWCs under the irrigation process. The results demonstrated that despite requiring only simple input data, the ANFIS and SVM models could be favorably used for SWC predictions under water stress conditions, especially when there is a lack of data. However, process-based numerical models are undoubtedly a better choice for predicting SWCs with lower uncertainties when required data are available, and thus for designing water saving strategies for agriculture and for other environmental applications requiring estimates of SWCs.

show abstract

Section: Introductionmentioning

confidence: 99%

A comparison of numerical and machine-learning modeling of soil water content with limited input data

Karandish

Šimůnek

2016

Journal of Hydrology

133

View full text Add to dashboard Cite

show abstract

“…To overcome these difficulties, artificial neural networks (ANNs) have been proposed [14]. An ANN is a flexible mathematical structure that is capable of identifying complex nonlinear relationships between input and output datasets [2].…”

Section: Resultsmentioning

confidence: 99%

Modeling of Triangular Unit Hydrographs Using an Artificial Neural Network in a Tropical River Basin

Saefulloh¹

2018

GEOMATE

View full text Add to dashboard Cite

Rainfall-runoff models are crucial for estimating floods in a river basin. Most watersheds in Indonesia have a data deficiency problem, especially in natural watersheds (ungauged river basins), which may affect the accuracy of design and planning of water resources. Most synthetic unit hydrograph methods are not in accordance with the characteristics of Indonesian watersheds, and adjustments should be made to obtain accurate results. This study aimed to develop a simple triangular unit hydrograph generated by using a neural network for different watersheds in Indonesia. The triangular unit hydrograph consists of the peak discharge, time to peak, and time base developed using a neural network with a learning process from the observed unit hydrograph, and the result will be compared to the Snyder-Alexeyev synthetic unit hydrograph after being adjusted to obtain accurate results in comparison to observed data. An artificial neural network (ANN) model was developed by inputting basin characteristics such as catchment area (A), river length (L), basin slope (S), shape factor (F), and runoff coefficient (C). The model will generate the output of a triangular synthetic unit hydrograph consisting of peak discharge (Qp), time to peak (Tp), and time base (Tb). A case study is discussed in tropical river basins mostly on Java Island, where flood events are frequent. The simulation result from applying an ANN using generalized reduced gradient neural network (GRGNN) methods is significantly in line with historical data. The ANN simulation shows more accurate results than the adjusted Snyder-Alexeyev unit hydrograph. The results indicated that the synthetic unit hydrograph generated by an ANN can be applied to an ungauged river basin.

show abstract

“…ANNs revealed to be a promising alternative for rainfall-runoff modeling (Ahmad & Simonovic, 2005;Rajukar et al, 2004), streamflow prediction (Muttiah et al, 1997;Maier & Dandy, 2000;Dolling & Varas, 2002;Sivakumar et al, 2002;Kisi, 2004;Cigizoglu & Kisi, 2005;Cigizoglu, 2008) and reservoir inflow forecasting (Saad et al, 1996;Jain et al, 1999). Recently, Coulibaly et al (2001b) and Kisi & Cigizoglu (2007) reviewed ANN-based models developed over the last years in hydrology, showing the extensive use of multi-layer feed-forward neural networks (FFNN), trained by standard back propagation (BP) algorithm (Magoulas et al, 1999).…”

Section: Introductionmentioning

confidence: 99%

Prediction of water flows in Colorado River, Argentina

Pierini

Gómez

Telesca

2012

lajar

View full text Add to dashboard Cite

ABSTRACT. 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. Keywords: prediction, time series, neural networks, autoregressive models, flows, Colorado River, Argentina. Predicción de caudales en río Colorado, ArgentinaRESUMEN. La identificación de modelos adecuados para predecir caudales diarios es importante para la planificación y la gestión de almacenamiento de agua en los embalses de la Argentina. La predicción a largo plazo del caudal es crucial para la regulación de los embalses y centrales hidroeléctricas, evaluar la protección del medio ambiente y el desarrollo sostenible, garantizar el correcto funcionamiento del abastecimiento público de agua en ciudades como Catriel, 25 de Mayo, río Colorado y también, eventualmente, en Bahía Blanca. En este trabajo, se analizan series de tiempo de caudales de agua, arriba del embalse y de la planta hidroeléctrica en Buta Ranquil, para modelar y predecir las fluctuaciones diarias. Se comparan los resultados obtenidos mediante el uso de una red neuronal artificial (ANN) de tres capas y un modelo autoregresivo (AR), con 18 años de datos, cuyos últimos 3 años se utilizan para la validación del modelo por medio de la raíz del error cuadrado medio (RMSE) y medida de certeza (Skill). Para predecir o rellenar el caudal diario, los resultados indican que el mejor desempeño es del ANN con respecto al modelo AR.

show abstract

An artificial neural network model for generating hydrograph from hydro-meteorological parameters

Cited by 106 publications

References 21 publications

A comparison of numerical and machine-learning modeling of soil water content with limited input data

A comparison of numerical and machine-learning modeling of soil water content with limited input data

Modeling of Triangular Unit Hydrographs Using an Artificial Neural Network in a Tropical River Basin

Prediction of water flows in Colorado River, Argentina

Contact Info

Product

Resources

About