Ashu Jain

Review: Methods used for the development of neural networks for the prediction of water resource variables in river systems: Current status and future directions

Over the past 15 years, artificial neural networks (ANNs) have been used increasingly for prediction and forecasting in water resources and environmental engineering. However, despite this high level of research activity, methods for developing ANN models are not yet well established. In this paper, the steps in the development of ANN models are outlined and taxonomies of approaches are introduced for each of these steps. In order to obtain a snapshot of current practice, ANN development methods are assessed based on these taxonomies for 210 journal papers that were published from 1999 to 2007 and focus on the prediction of water resource variables in river systems. The results obtained indicate that the vast majority of studies focus on flow prediction, with very few applications to water quality. Methods used for determining model inputs, appropriate data subsets and the best model structure are generally obtained in an ad-hoc fashion and require further attention. Although multilayer perceptrons are still the most popular model architecture, other model architectures are also used extensively. In relation to model calibration, gradient based methods are used almost exclusively. In conclusion, despite a significant amount of research activity on the use of ANNs for prediction and forecasting of water resources variables in river systems, little of this is focused on methodological issues. Consequently, there is still a need for the development of robust ANN model development approaches.

Hybrid neural network models for hydrologic time series forecasting

The need for increased accuracies in time series forecasting has motivated the researchers to develop innovative models. In this paper, a new hybrid time series neural network model is proposed that is capable of exploiting the strengths of traditional time series approaches and artificial neural networks (ANNs). The proposed approach consists of an overall modelling framework, which is a combination of the conventional and ANN techniques. The steps involved in the time series analysis, e.g. de-trending and de-seasonalisation, can be carried out before gradually presenting the modified time series data to the ANN. The proposed hybrid approach for time series forecasting is tested using the monthly streamflow data at Colorado River at Lees Ferry, USA. Specifically, results from four time series models of auto-regressive (AR) type and four ANN models are presented. The results obtained in this study suggest that the approach of combining the strengths of the conventional and ANN techniques provides a robust modelling framework capable of capturing the non-linear nature of the complex time series and thus producing more accurate forecasts. Although the proposed hybrid neural network models are applied in hydrology in this study, they have tremendous scope for application in a wide range of areas for achieving increased accuracies in time series forecasting.

Short-Term Water Demand Forecast Modelling at IIT Kanpur Using Artificial Neural Networks

The efficient operation and management of an existing water supply system require short-term water demand forecasts as inputs. Conventionally, regression and time series analysis have been employed in modelling short-term water demand forecasts. The relatively new technique of artificial neural networks has been proposed as an efficient tool for modelling and forecasting in recent years. The primary objective of this study is to investigate the relatively new technique of artificial neural networks for use in forecasting short-term water demand at the Indian Institute of Technology, Kanpur. Other techniques investigated in this study include regression and time series analysis for comparison purposes. The secondary objective of this study is to investigate the validity of the following two hypotheses: 1) the short-term water demand process at the Indian Institute of Technology, Kanpur campus is a dynamic process mainly driven by the maximum air temperature and interrupted by rainfall occurrences, and 2) occurrence of rainfall is a more significant variable than the rainfall amount itself in modelling the short-term water demand forecasts. The data employed in this study consist of weekly water demand at the Indian Institute of Technology, Kanpur campus, and total weekly rainfall and weekly average maximum air temperature from the City of Kanpur, India. Six different artificial neural network models, five regression models, and two time series models have been developed and compared. The artificial neural network models consistently outperformed the regression and time series models developed in this study. An average absolute error in forecasting of 2.41% was achieved from the best artificial neural network model, which also showed the best correlation between the modelled and targeted water demands. It has been found that the water demand at the Indian Institute of Technology, Kanpur campus is better correlated with the rainfall occurrence rather than the amount of rainfall itself.

A comparative analysis of training methods for artificial neural network rainfall-runoff models

This paper compares various training methods available for training multi-layer perceptron (MLP) type of artificial neural networks (ANNs) for modelling the rainfall-runoff process. The training methods investigated include the popular back-propagation algorithm (BPA), real-coded genetic algorithm (RGA), and a self-organizing map (SOM). A SOM was used to first classify the input-output space into different categories and then develop feed-forward MLP models for each category using BPA. The daily average rainfall and streamflow data derived from an existing catchment were employed to develop all ANN models investigated in this study. A wide variety of standard statistical performance evaluation measures were employed to evaluate the performances of various ANN models developed. The results obtained in this study indicate that the approach of first classifying the input-output space into different categories using SOM and then developing separate ANN models for different classes trained using BPA performs better than the approach of developing a single ANN rainfall-runoff model trained using BPA. The ANN rainfall-runoff model trained using RGA was able to provide a better generalization of the complex, dynamic, non-linear, and fragmented rainfall-runoff process in comparison with the other approaches investigated in this study. It has been found that the RGA trained ANN model significantly outperformed the ANN model trained using BPA, and was also able to overcome certain limitations of the ANN rainfall-runoff model trained using BPA reported by many researchers in the past. It is noted that the performances of various ANN models should to be evaluated using a wide variety of statistical performance indices rather than relying on a few global error statistics normally employed that are similar in nature to the global error minimized at the output layer of an ANN.

An evaluation of artificial neural network technique for the determination of infiltration model parameters

Infiltration is a key component in the rainfall runoff models employed for runoff prediction. Conventionally, the hydrologists have relied on classical optimization techniques for obtaining the parameters of various infiltration equations. Recently, artificial neural networks (ANNs) have been proposed as efficient tools for modelling and forecasting. This paper proposes the use of ANNs for calibrating infiltration equations. The ANN consists of rainfall and runoff as the inputs and the infiltration parameters as the outputs. Classical optimization techniques were also employed to determine flow hydrographs for comparison purposes. The performances of both the approaches were evaluated using a variety of standard statistical measures in terms of their ability to predict runoff. The results obtained in this study indicate that the ANN technique can be successfully employed for the purpose of calibration of infiltration equations. The regenerated and predicted storms indicate that the ANN models performed better than the classical techniques. It has been found that the ANNs are capable of performing very well in situations of limited data availability since the differences in the performances of the ANNs trained on partial information and the ANNs trained on the complete information was only marginal and the ANN trained on partial information consisted of a more compact architecture. A wide variety of standard statistical performance evaluation measures are needed to properly evaluate the performances of various ANN models rather than relying on a few global error statistics (such as RMSE and correlation coefficient) normally employed.

Visualisation of Hidden Neuron Behaviour in a Neural Network Rainfall-Runoff Model

This chapter applies graphical and statistical methods to visualise hidden neuron behaviour in a trained neural network rainfall-runoff model developed for the River Ouse catchment in northern England. The methods employed include plotting individual partial network outputs against observed river levels; carrying out correlation analyses to assess relationships among partial network outputs, surface flow and base flow; examining the correlations between the raw hidden neuron outputs, input variables, surface flow and base flow; plotting individual raw hidden neuron outputs ranked by river levels; and regressing raw hidden neuron outputs against river levels. The results show that the hidden neurons do show specialisation. Of the five hidden neurons in the trained neural network model, two appear to be modelling base flow, one appears to be modelling surface flow, while the remaining two may be modelling interflow or quick sub-surface processes. All the methods served to provide confirmation of some or all of these findings. The study shows that a careful examination of a trained neural network can shed some light on the sub-processes captured in its architecture during training.

A DECISION SUPPORT SYSTEM FOR DROUGHT CHARACTERIZATION AND MANAGEMENT

Abstract A decision support system has been developed for drought characterization and management. The purpose of the decision support system is to assist the operators and water managers of the water supply system of the City of Lexington, Kentucky. The motivation of this study was a severe drought that occurred in the state of Kentucky during the summer of 1988. The data derived from the City of Lexington, Kentucky and the Kentucky River Basin were employed in this study. The developed decision support system consists of three components: a water demand forecasting component, a streamflow forecasting component, and an integrated expert system component. The water demand and streamflow forecasting components of the decision support system predict the water consumption for the City of Lexington and the flow in Kentucky River at Lock 10 near Winchester, Kentucky, respectively. The lead time of the forecasting models was taken as five days as they were intended to be employed in developing a short-term drought management policy. Various modeling techniques ranging from regression and time series analysis to the relatively new technique of expert systems and artificial neural networks were explored for forecasting both water demand and streamflow. The integrated expert system component consists of five sub-components. Each sub-component entails developing a knowledge base for a specific purpose. The expert system component integrates all sub-components and characterizes the drought potential in the coming five days and recommends a drought management policy for the week to come. The developed decision support system is capable of running on a persona] computer “and provides a user-friendly platform for decision-makers to explore a wide range of drought management alternatives.

RAINFALL RUNOFF MODELLING USING NEURAL NETWORKS: STATE-OF-THE-ART AND FUTURE RESEARCH NEEDS

ABSTRACT Modeling of rainfall runoff (R-R) processes is useful in many water resources management activities. Traditionally, hydrologists have employed deterministic/conceptual methods for R-R modeling. Recently, Artificial Neural Networks (ANNs) have become popular tools for R-R modeling. This paper reviews the literature on and presents state-of-the-art approaches to ANN R-R modeling. Certain aspects of ANN R-R modeling have been covered in greater detail. These include input selection, data division, ANN training, hybrid modeling, and extrapolation beyond the range of training data. There is a strong need to carry out extensive research on these aspects while developing ANN R-R models.

Investigation of sensitivity of popular training methods to initial weights in ANN rainfall-runoff modeling

A attempt has been made to understand the ground water quality by using the water quality index (WQI) in the mining region of Goa. WQI, a technique of rating water quality, is an effective tool to assess spatial and temporal changes in ground water quality. Forty five groundwater samples were collected from open and tube wells during summer, monsoon, post-monsoon and winter seasons. The groundwater samples were subjected to comprehensive physio-chemical analysis involving major cations (Ca2+, Mg2+, Na+, K+, Fe2++ ) anions ( HClO3-, Cl, SO4 2-, NO3-, F-, PO4 3-) besides general parameters (pH, EC, TDS, alkalinity, total hardness, color, turbidity). The water quality index rating was calculated to quantify overall water quality for human consumption. For calculating WQI 10 parameters, namely pH, TDS, total hardness, chloride, nitrate, turbidity, fluoride, iron, calcium hardness, magnesium hardness were considered. The values of WQI have been affected mainly by the concentration of dissolved ions (F, NO, Ca and Mg) in ground water. Concentration of dissolved solids found to be more during monsoon season. It may be due more seepage and movement of ground water due to excessive rainfall there. The values of WQI of the samples were found in the range of 8-12 for all the seasons and considered to be in the very good category.U Heat Island is a more threatening meteorological phenomenon in the longer summer tropical cities with already critical thermal regime than in the longer winter temperate cities. Further, the population growth in the tropical megacities is far greater than their counterpart temperate cities. The present paper enquires the changing trends in the urban climates. Cities are recognized as urban heat islands because their temperatures are higher by 3°C to 9°C in comparison to the temperatures of the surrounding rural areas. This thermal differential is generally proportional to the city size and morphology. Larger is the city size and more diverse are its residential, industrial and commercial functions, larger and higher is its heat dome. Urban Heat Island has become a major environmental issue particularly in view of rapid urban sprawl in the developing countries. Urban Heat Islands are the nuclei of global warming and climate change. Urban Heat Island is not a cognizable problem of towns and small cities. It is a matter of great concern to the urban planners, environmentalists and citizens of the large cities, particularly the million and megacities. However, this problem is assuming a serious dimension in a fast emerging megacity of Hyderabad. The present paper estimates heat island intensity of Hyderabad. In the wake of its rapidly growing economy, size and population, the city has experienced a population growth from 1.79 million in 1971 to 7.74 million in 2011. The corresponding built-up area has recorded a growth from 298.5 sq.km. to a sprawling 851 sq.km. The paper enquires the adverse impact of this built-up growth on the daytime as well as nocturnal temperature rise. This has rendered the urban climate increasingly taxing to the human health and comfort.M seasonal, and annual variations of salinity profiles over different sectors of the Bay of Bengal (BoB) are investigated using seven years of Argo data. The salinity profile analysis together with the analysis of variability in surface circulation and precipitation utilized to understand interannual and seasonal variability in salinity profiles over three sectors of BoB i.e., northern (NBoB), central (CBoB), and southern (SBoB). The influence of massive river outflow close to river mouths in producing the observed sea surface salinity minima in the coastal northern BoB during November-December is highlighted. Seasonal changes in salinity profiles are primarily caused by freshwater flux, mixing processes and advection. In general, NBoB remains fresher as compare to CBoB and SBoB throughout the year. Interannual variability of salinity structure was found to be maximum in NBoB, particularly in post-monsoon (ON) and winter (DJF) seasons, where the differences in surface salinity between the years were found to be up to 2 psu. CBoB shows minimum interannual variations in salinity profiles, except unusual decrease in surface salinity on two occasions. Analysis suggest crucial role of coastal currents, gyres and surface circulation in controlling seasonal and interannual variability in salinity profiles. Some unusual features observed in salinity profiles during pre-monsoon season of year 2009 in the SBoB, which is analyzed further with other data sets and discussed in detail.E fluoride concentrations have been reported in ground waters of more than 20 developed and developing countries around the world including India where 20 states are facing acute fluorosis problems. In this paper, the teeth of 152 school children (06 to 11yrs) of Govt. Primary school, Gandhoniya, Barkagaon, Hazaribag, and Jharkhand, India were surveyed and collected data were compared with Dean’s Index. The results clearly indicate that the most of them are suffering from dental fluorosis consequently creating a lot of problems such as poor health, endurance of dental abscesses, inability to chew food well, embarrassment about discoloured & damaged teeth and distraction from play & learning in the area. School based oral health check-up programme be organised regularly for awareness of general mass. It should include screening, referral and case management to ensure the timely receipt of oral health care from health professionals in the community.

Development of stage-discharge relationship using simple rating curve, ANN and data transformation techniques

Andreas Schmid has a Ph.D. in Biochemical Engineering-Environmental Engineering and now is a full Professor in water and wastewater treatment at the Faculty of Engineering at the University of Applied Sciences Hof, Germany. More than 20 years of working experiences in industry built the fundament to derive applied sciences at his current position. He received several awards for his research in the water and wastewater sector and holds a couple of national and international patents in environmental techniques. At present his research focus concentrates on cavitation technologies and relating applications in environmental engineering especially elimination of industrialand micro-pollutants. Improving mass transfer in gas-liquid systems by “supercavitation”Zhaodong Feng obtained his Ph.D. from University of Kansas (1992), M.S. from University of Washington (1987) and M.A. from Lanzhou University (1982). He did two-year postdoctoral study at Columbia University (1992-1994) and was a Professor at Montclair State University (1996-2008). He was a Yangtze-Scholar Professor at Lanzhou University (2000-2010) and Tianshan-Scholar Professor at Xinjiang University (2011-2013). Surface runoff responses to climate change and LUCC in the Asian arid zoneM scientists throughout the world have been analyzing the climate and environmental factors that can affect our health or ecology, and the level of risk. Every year Nepal experiences natural calamities such as draughts, and Glacier Lakes Outburst Floods (GLOF) etc. Likewise the impacts of global warning are felt in agriculture, energy production, and pollution also, which ultimately affects the economy of the nation. Research recommends a sustainable policy framework on how the imbalance of climate change could significantly be reduced by using different appropriate measures.The goal of this study is to develop different scenarios of water resource availability in near upstream of Kaligandaki River basin, under climate change-induced parameters such as precipitation and temperatures variability, rainfall extreme floods, droughts etc. Climate models suggest that global warming could bring warmer, drier conditions to Nepalese high Mountains due to the large topographical differences of the climate parameters. A detailed knowledge of mass-balance observation and discharge measurement are considered and the combination of both will be analyzed by means of either Water Balance Model (WatBal) or General Climate Model (GCM) with multicriteria model performance evaluation. For this discharge measurements should be taken during the melting season which demonstrates that timing of runoff. Mostly, the Water Balance Model CLIRUN3 was combined with years of basic climate information records (precipitation, potential evapotranspiration and water flow) to simulate monthly river runoff in the river basin. If both temperature and precipitation increase, the mean runoff value in the region will be reduced by considerable amount from monsoon to non-monsoon season., this will help to formulate numerical flow line glacier model on high Mountains of Nepal at upstream of Kaligandaki river and forcing mechanisms for flows in the next few decades. This is an indication that with extreme events, hydro hazards, depleting permafrost areas and glacier melts have close links with river flows and sediment. At the end, in next few decades in climate continue changing more than at present rate. Lekha Nath Bagale, Hydrol Current Res 2013, 4:2 http://dx.doi.org/10.4172/2157-7587.S1.008A daily water balance model was developed using daily rainfall data, contributing roof area, leakage/evaporation loss factor, available storage volume, tank overflow and rainwater demand. In order to assess reliability of domestic rainwater tanks in augmenting partial household water demand in Sydney area, the developed water balance model was used for three different climatic conditions (i.e. dry, average and wet years). The traditional practice of rainwater harvesting volume/size design is based on historic annual average rainfall data. However, design of rainwater harvesting volume based on annual average rainfall data is not realistic. As a stormwater harvesting system designed considering average rainfall will not provide much benefit for a critical dry period. In several earlier studies, a single representative year was selected for each of the dry, average and wet years. Dry, average and wet years were defined for the years having an annual rainfall of 10 percentile, 50 percentile and 90 percentile values respectively. However, as a particular year may have an unusal rainfall pattern, this study considered five respective years for each of the dry, average and wet years. Model was used for selected five years and average outcomes were calculated. To assess the spatial variablity, the model was used for the performance analysis at four different regions of Sydney (Australia); North, Central, South-East and South-West. These four different regions of Sydney are characterised by notable different topography and rainfall characteristics. Reliability is defined as percentage of days in a year when rainwater tank is able to supply the intended partial demand for a particular condition. For the three climatic conditions, a number of reliability charts were produced for domestic rainwater tanks in relations to tank volume, roof area and number of people in a house (i.e. water demand). It is found that for a relatively small roof size (100m 2 ), 100% reliability cannot be achieved even with a very large tank (10,000L). Reliability becomes independent of tank size for tank sizes larger than 4,000~7,000L depending on the location. This is defined as threshold tank size, relationships with threshold tank sizes and annual rainfall amounts are then established for all the locations.