Artificial Neural Networks (ANNs) are being used increasingly to predict and forecast water resources variables. In this paper, the steps that should be followed in the development of such models are outlined. These include the choice of performance criteria, the division and pre-processing of the available data, the determination of appropriate model inputs and network architecture, optimisation of the connection weights (training) and model validation. The options available to modellers at each of these steps are discussed and the issues that should be considered are highlighted. A review of 43 papers dealing with the use of neural network models for the prediction and forecasting of water resources variables is undertaken in terms of the modelling process adopted. In all but two of the papers reviewed, feedforward networks are used. The vast majority of these networks are trained using the backpropagation algorithm. Issues in relation to the optimal division of the available data, data pre-processing and the choice of appropriate model inputs are seldom considered. In addition, the process of choosing appropriate stopping criteria and optimising network geometry and internal network parameters is generally described poorly or carried out inadequately. All of the above factors can result in non-optimal model performance and an inability to draw meaningful comparisons between different models. Future research efforts should be directed towards the development of guidelines which assist with the development of ANN models and the choice of when ANNs should be used in preference to alternative approaches, the assessment of methods for extracting the knowledge that is contained in the connection weights of trained ANNs and the incorporation of uncertainty into ANN models.

Neural networks for the prediction and forecasting of water resources variables: a review of modelling issues and applications

https://repub.eur.nl/pub/104203/REPUB_104203-OA.pdf

2017 HRS / EHRA / ECAS / APHRS / SOLAECE expert consensus statement on catheter and surgical ablation of atrial fibrillation

In this two-part series, the writers investigate the role of artificial neural networks (ANNs) in hydrology. ANNs are gaining popularity, as is evidenced by the increasing number of papers on this ...

Artificial Neural Networks in Hydrology. I: Preliminary Concepts

This paper forms the second part of the series on application of artificial neural networks (ANNs) in hydrology. The role of ANNs in various branches of hydrology has been examined here. It is foun...

Artificial Neural Networks in Hydrology. II: Hydrologic Applications

This review considers the application of artificial neural networks (ANNs) to rainfall-runoff modelling and flood forecasting. This is an emerging field of research, character- ized by a wide variety of techniques, a diversity of geographical contexts, a general absence of intermodel comparisons, and inconsistent reporting of model skill. This article begins by outlining the basic principles of ANN modelling, common network architectures and training algorithms. The discussion then addresses related themes of the division and preprocessing of data for model calibration/validation; data standardization techniques; and methods of evaluating ANN model performance. A literature survey underlines the need for clear guidance in current modelling practice, as well as the comparison of ANN methods with more conven- tional statistical models. Accordingly, a template is proposed in order to assist the construction of future ANN rainfall-runoff models. Finally, it is suggested that research might focus on the extraction of hydrological 'rules' from ANN weights, and on the development of standard performance measures that penalize unnecessary model complexity.

Hydrological modelling using artificial neural networks

Cardiac Purkinje cells (PCs) are more susceptible to action potential abnormalities as compared to ventricular myocytes (VMs), which could be associated with their distinct intracellular calcium handling. We developed a detailed biophysical model of a mouse cardiac PC, which importantly reproduces the experimentally observed biphasic cytosolic calcium waves. The model includes a stochastic gating formulation for the opening and closing of ryanodine receptor (RyR) channels, simulated with a Monte Carlo method, to accurately reproduce cytosolic calcium wave propagation and the effects of spontaneous calcium release events. Simulations predict that during an action potential, smaller cytosolic calcium wavelets propagated from the sarcolemma towards the center of the cell and initiated larger magnitude cell-wide calcium waves via a calcium-induced-calcium release mechanism. In the presence of RyR mutations, frequent spontaneous calcium leaks from sarcoplasmic reticulum (SR) initiated calcium waves, which upon reaching the cell periphery produced delayed afterdepolarizations (DADs) via sodium-calcium exchanger (NCX) and T-type calcium (ICaT ) channel activation. In the presence of isoproterenol-mediated effects, DADs induced triggered activity by reactivation of fast sodium channels. Based on our model, we found that the activation of either L-type calcium channels (ICaL ), ICaT , sodium-potassium exchanger (INaK ) or NCX is sufficient for occurrence of triggered activity; however, a partial blockade of ICaT or INaK is essential for its successful termination. Our modeling study highlights valuable insights into the mechanisms of DAD-induced triggered activity mediated via cytosolic calcium waves in cardiac PCs and may elucidate the increased arrhythmogeneity in PCs.

Delayed afterdepolarization-induced triggered activity in cardiac purkinje cells mediated through cytosolic calcium diffusion waves.

Information on the heights of ocean waves at a site can be collected by a variety of instruments-each involving different methods of data retrieval and synthesis. Typically, sensing of wave heights by satellite requires the wave information to be presented in the form of values that are averaged over space and time intervals. In order to use such data for operational applications, it then becomes necessary to derive the wave heights over shorter intervals from their values available over long durations. This paper attempts to do this by employing the technique of neural networks. A simple three-layered feedforward network trained with the supervised backpropagation technique was used. The values of monthly mean significant wave heights available at different grid locations around the Indian coastline were given as input to obtain the output of weekly mean wave heights at the same locations. Analysis of the results indicated usefulness of the neural network technique in wave height interpolation problems.

Neural Networks for Wave Height Interpolation

The significant wave height and average wave period form an essential input for operational activities in ocean and coastal areas. Such information is important in issuing appropriate warnings to people planning any construction or instillation works in the oceanic environment. Many countries over the world routinely collect wave and wind data through a network of wave rider buoys. The data collecting agencies transmit the resulting information online to their registered users through an internet or a web-based system. Operational wave forecasts in addition to the measured data are also made and supplied online to the users. This paper discusses operational wave forecasting in real time mode at locations where wind rather than wave data are continuously recorded. It is based on the time series modeling and incorporates an artificial intelligence technique of genetic programming. The significant wave height and average wave period values are forecasted over a period of 96 hr in future from the observations...

https://journals.sagepub.com/doi/pdf/10.1260/1759-3131.5.4.249

Real Time Wave Forecasting Using Wind Time History and Genetic Programming

Structures built in the sea are traditionally designed according to historical climate observations or hindcasts. In geographical locations typical of India, such designs do not take the effect of future climate change into account. For structural safety, considerations of such effects are highly desirable. Many recent studies have demonstrated likely changes in the extreme wave conditions at different offshore locations. This paper reports similar results in the Mumbai High region, where the majority of India's offshore oil platforms are located. Using historical wind data provided by the National Centre for Environmental Prediction/National Centre for Atmospheric Research and a Canadian general circulation model for future data (CMIP5-RCP8.5), a numerical wave model of the past and future wave conditions was simulated over a 40-year period. A statistical analysis of both past and projected datasets obtained significant wave heights with a 100-year return. The comparison of wave heights derived from past...

Projected impact of climate change on waves at Mumbai High

Bhat, S.; Jain, P., and Deo, M.C., 2019. Application of regional climate models for coastal design parameters along India. Journal of Coastal Research, 35(1), 110–121. Coconut Creek (Florida), ISSN 0749-0208.Coastal structures are generally designed on the basis of significant wave height, Hs2, corresponding to a return period of 100 years. This design wave height is derived by fitting an appropriate statistical distribution to a set of long-duration Hs data, either observed or simulated, using historical wind conditions. This study investigates what might happen if wind conditions projected on the basis of regional climate models (RCMs), reflecting the effect of climate change, are instead used. At a series of 91 coastal stations spread over the entire 7000-km-long coastline of India, waves were simulated for two time slices of 27 years each in the past and future using a numerical wave model driven by an ensemble of RCMs. The RCMs were earlier run for a moderate global warming scenario. The long-term statistical analysis of resulting wave data was carried out, and Hs with a 100-year return period were derived at each location. It was noticed that if projected wind were used in the place of past wind, an increase in the magnitude of design Hs is likely to be encountered at most stations, although there might be some locations where a decrease instead of an increase might happen. Because the west coast is subjected to different wind conditions than the east coast, it also showed different behavior in terms of the change. These observations were supported by comparative analysis of trends, probability distributions, and wave rose diagrams of past and projected wave data.

Makarand Deo

Papers

Delayed afterdepolarization-induced triggered activity in cardiac purkinje cells mediated through cytosolic calcium diffusion waves.

Neural Networks for Wave Height Interpolation

Real Time Wave Forecasting Using Wind Time History and Genetic Programming

Projected impact of climate change on waves at Mumbai High

Application of Regional Climate Models for Coastal Design Parameters along India