Mukesh K. Tiwari

Wavelet Bootstrap Multiple Linear Regression Based Hybrid Modeling for Daily River Discharge Forecasting

A new hybrid model, the wavelet–bootstrap–multiple linear regression (WBMLR) is proposed to explore the potential of wavelet analysis and bootstrap resampling techniques for daily discharge forecasting. The performance of the developed WBMLR model is also compared with five more models: multiple linear regression (MLR), artificial neural network (ANN), wavelet-based MLR (WMLR), wavelet-based ANN (WANN) and wavelet–bootstrap–ANN (WBANN) models. Seven years of discharge data from seven gauging stations in the middle reaches of Mahanadi river basin in India are applied in this study. Significant input vectors are decomposed into discrete wavelet components (DWCs) using discrete wavelet transformation (DWT) to generate wavelet sub time series that are used as inputs to the MLR and ANN models to develop the WMLR and WANN models, respectively. Effective wavelets are selected by considering several types of wavelets with different vanishing moments. WBMLR and WBANN models are developed as ensemble of different WMLR and WANN models, respectively, developed using different realizations of the training dataset generated using bootstrap resampling technique. The results show that the wavelet bootstrap hybrid models (i.e. WBMLR and WBANN) produce significantly better results than the traditional MLR and ANN models. Hybrid models based on MLR (WMLR, WBMLR) perform better than the ANN based hybrid models (WBANN, WANN). The WBMLR and WMLR models simulate the peak discharges better than the WBANN, WANN, MLR and ANN models, whereas the overall performance of WBMLR model is found to be more accurate and reliable than the remaining five models.

Medium-Term Urban Water Demand Forecasting with Limited Data Using an Ensemble Wavelet–Bootstrap Machine-Learning Approach

AbstractAccurate and reliable weekly and monthly water demand forecasting is important for effective and sustainable planning and use of urban water supply infrastructure. This study explored a hybrid wavelet–bootstrap–artificial neural network (WBANN) modeling approach for weekly (one-week) and monthly (one- and two-month) urban water demand forecasting in situations with limited data availability. The performance of WBANN models was also compared with that of standard artificial neural networks (ANN), bootstrap-based ANN (BANN), and wavelet-based ANN (WANN) models. The proposed WBANN method is aimed at improving the accuracy and reliability of water demand forecasting by incorporating the capability of wavelet transformation and bootstrap analysis using artificial neural networks. Daily and monthly maximum temperature, total precipitation, and water demand data for almost three years obtained from the city of Calgary, Alberta, Canada were used in this study. For weekly and monthly lead-time forecasting,...

Comparison of multi-objective evolutionary neural network, adaptive neuro-fuzzy inference system and bootstrap-based neural network for flood forecasting

Accurate flood forecasting is of utmost importance in mitigating flood disasters. Flood causes severe public and economic loss especially in large river basins. In this study, multi-objective evolutionary neural network (MOENN) model is developed for accurate and reliable hourly water level forecasting at Naraj gauging site in Mahanadi river basin, India. The performance of the developed model is compared with adaptive neuro-fuzzy inference system (ANFIS) and bootstrap-based neural network (BNN) models. The performance of the models is compared in terms of Nash–Sutcliffe efficiency, root mean square error, mean absolute error and percentage deviation in peak (D). The performance of the models in forecasting floods is also evaluated using existing performance evaluation criterion of Central Water Commission, India as well as a multiple linear regression model. A partitioning analysis in conjunction with threshold statistics is carried out to evaluate the performance of the developed models in forecasting floods for low, medium and high water levels. It is found that the performance of MOENN and BNN models is more stable and consistent compared to ANFIS model. For longer lead times, the performance of MOENN model is found to be the best, with its performance in forecasting higher water levels being significantly better compared to ANFIS and BNN models. Overall, it is found that MOENN model has great potential to be applied in flood forecasting.

Evapotranspiration Modeling Using Second-Order Neural Networks

AbstractThis study introduces the utility of the second-order neural network (SONN) method to model the reference evapotranspiration (ET0) in different climatic zones of India. The daily climate data of minimum and maximum air temperatures, minimum and maximum relative humidity, wind speed, and solar radiation from 17 different locations in India were used as the inputs to the SONN models to estimate ET0 corresponding to the FAO-56 Penman-Monteith (FAO-56 PM) method. With the same inputs, for all 17 locations the first-order neural networks such as feed forward back propagation (FFBP-NN) models were also developed and compared with the SONN models. The developed SONN and FFBP-NN models were also compared with the estimates provided by the FAO-56 PM method. The performance criteria adopted for comparing the models were root-mean-squared error (RMSE), mean-absolute error (MAE), coefficient of determination (R2), and the ratio of average output to average target ET0 values (Rratio). Based on the comparisons,...

River-Flow Forecasting Using Higher-Order Neural Networks

In this paper, we propose a novel neural modeling methodology for forecasting daily river discharge that makes use of neural units with higher-order synaptic operations (NU-HSOs). For hydrologic forecasting, conventional rainfall-runoff models based on mechanistic approaches in the literature have shown limitations attributable to their overparameterization and complexity. With the use of neural units with quadratic synaptic operation (NU-QSO) and cubic synaptic operation (NU-CSO), as suggested in this paper, the refined neural modeling methodology can overcome the intricacy and inefficiency of conventional models. In this paper, neural network (NN) models with NU-HSO are compared with conventional NNs with neural units with linear synaptic operation (NU-LSO) for forecasting river discharge. This study was conducted using 1- to 5-day lead time forecasting in the Mahanadi River basin at the Naraj gauging site to evaluate the effectiveness of the higher-order neural networks (HO-NNs). Performance indices for the prediction of daily discharge forecasting indicated that NNs with NU-CSO and NNs with NU-QSO achieved better performance than NNs with NU-LSO even with a lower number of hidden neurons. Thus, this study shows that HO-NNs can be effective in hydrologic forecasting. DOI: 10.1061/(ASCE)HE.1943-5584.0000486.

Prioritization of agricultural sub-watersheds in semi arid middle region of Gujarat using remote sensing and GIS

In this study, proven capability of remote sensing and GIS are used for watershed prioritization. 19 different sub-watersheds are prioritised through geomorphological analysis and suitable structures are proposed for soil and water conservation in a Limkheda agricultural watershed situated in semi arid middle region of Gujarat, India. Remote sensing images such as SRTM are used to delineate the watershed and to generate slope thematic maps, soil maps are applied to generate soil type, whereas LISS III remote sensing image is used for generating land use maps. Prioritization of sub-watersheds using geomorphological analysis is carried out by seven different linear and shape parameters. Then different sub-watersheds are prioritised by assigning ranks using compound parameter. After prioritization, land use, soil type and land slope categories of sub-watersheds are integrated to propose suitable soil and water conservation structures. In this study, it is proposed that soil conservation measures should be adopted as per the priority assigned to reduce the adverse effect on the land and environment. Overall, it is concluded in this study that delineation of watersheds into sub-watersheds and prioritization of these sub-watersheds are very relevant, helpful and important in semi-arid regions of middle Gujarat, where there is high diversity in agricultural practices and size of land holdings. Adaptation of soil conservation measures priority-wise will not only reduce the soil erosion but also increase the water availability in the surface and as groundwater and will further reduces the possibility of droughts as well as floods and finally environmental hazards.

An ensemble wavelet bootstrap machine learning approach to water demand forecasting: a case study in the city of Calgary, Canada

Abstract This paper explores a hybrid wavelet, bootstrap and neural network (WBNN) modeling approach for daily (1, 3 and 5 day) urban water demand forecasting in situations with limited data availability. This method was tested using 3 years of daily water demand and meteorological data for the city of Calgary, Alberta, Canada. The performance of the WBNN method was compared to that of three other methods: traditional neural networks (NN), wavelet NNs (WNN), and bootstrap-based NN (BNN) models. While the hybrid WBNN and WNN models equally provided 1-day lead-time forecasts of greater accuracy than those obtained with other methods, for longer lead-time (3- or 5-day) forecasts the WBNN model alone outperformed the other models. The confidence bands generated using the WBNN model displayed the uncertainty associated with the forecasts.

Water demand forecasting using extreme learning machines

Abstract The capacity of recently-developed extreme learning machine (ELM) modelling approaches in forecasting daily urban water demand from limited data, alone or in concert with wavelet analysis (W) or bootstrap (B) methods (i.e., ELM, ELMW, ELMB), was assessed, and compared to that of equivalent traditional artificial neural network-based models (i.e., ANN, ANNW, ANNB). The urban water demand forecasting models were developed using 3-year water demand and climate datasets for the city of Calgary, Alberta, Canada. While the hybrid ELMB and ANNB models provided satisfactory 1-day lead-time forecasts of similar accuracy, the ANNW and ELMW models provided greater accuracy, with the ELMW model outperforming the ANNW model. Significant improvement in peak urban water demand prediction was only achieved with the ELMW model. The superiority of the ELMW model over both the ANNW or ANNB models demonstrated the significant role of wavelet transformation in improving the overall performance of the urban water demand model.

Reservoir Inflow Forecasting Using Extreme Learning Machines

Accurate and reliable forecasting of reservoir inflow is necessary for efficient and effective water resources’ planning and management. In the present study, the capacity of recently developed extreme learning machines (ELMs) modeling approach in forecasting reservoir inflows is assessed and compared to that of equivalent traditional artificial neural network-based models. Performance of wavelet analysis technique is also explored by developing wavelet-based ELMs (WELMs) and wavelet-based ANNs (WANNs) models. Seven years of reservoir inflow data along with outflow data from two upstream reservoirs in the Damodar catchment along with rainfall data of 5 upstream rain gauge stations are considered in this study. Out of 7 years’ daily data, 5 years’ data are used for training the model, one-year data are used for cross-validation, and remaining one-year data are used to evaluate the performance of the developed models. Different performance indices indicated better performance of ELM and WELM models in comparison with MLR, ANN, WMLR, and WANN models. This study demonstrated the effectiveness of proper selection of wavelet functions and appropriate methodology for wavelet-based model development. ELM models were also computationally efficient as demonstrated by faster running time, and consequently, this study advocates the superiority of the WELM model and the significant role of wavelet transformation in order to improve the model’s overall performance for reservoir inflow forecasting modeling.