My Research Interests

• Climate Hydrology
  • Climate Modelling
  • Climate Variables Downscaling
• Optimization in Water Resource Systems
• Multicriteria Decision Making (MCDM) in Water Resources
• Remote Sensing for Irrigation Management
  
  Earlier while working in IIT, Kharagpur (1994-2002)
  
• Optimal Reservoir Operation Models
• Time Series Analysis in Hydrology
• Rainfall-Runoff Modeling
• Water Allocation Models
• Design Aspects of Stepped Spillway
• Disaggregation Models
• Satellite Remote Sensing in Irrigation Management
• Artificial Neural Networks in Hydrology
• Multicriterion Decision Making (MCDM) in River Basin Development and Management
• Fuzzy Approach for MCDM in River Basin Development and Management
• Genetic Algorithms for Optimal Reservoir Operation

In the recent scenario of climate change, the natural variability and uncertainty associated with the hydrologic variables is gaining importance. Assessment of hydroclimatic teleconnection for Indian subcontinent and its use in basin-scale hydrologic time series analysis and forecasting is investigated. El Nińo-Southern Oscillation (ENSO) is the well established coupled Ocean-atmosphere mode of tropical Pacific Ocean whereas Indian Ocean Dipole (IOD) mode is recently identified. Equatorial Indian Ocean Oscillation (EQUINOO) is the atmospheric component of IOD mode. The potential of ENSO and EQUINOO for predicting Indian summer monsoon rainfall (ISMR) is investigated using Bayesian Dynamic Linear Model (BDLM). A major advantage of this method is that, it is able to capture the dynamic nature of the cause-effect relationship between large-scale circulation information and variability in hydrologic variables. Another new method is developed to capture the dependence between the teleconnected hydroclimatic variables based on the theory of copula. The association of monthly variation of ISMR with the combined information of ENSO and EQUINOO, denoted by monthly composite index (MCI), is also investigated and a relationship is established. The spatial variability of such association is also investigated.

Having established the hydroclimatic teleconnection at a comparatively larger scale, the hydroclimatic teleconnection for basin-scale hydrologic variables is then investigated and established. The association of large-scale atmospheric circulation with inflow during monsoon season into Hirakud reservoir, Orissa, India, has been investigated. The strong predictive potential of the composite index of ENSO and EQUINOO is established including for extreme inflow conditions. Recognizing the basin-scale hydroclimatic association with both ENSO and EQUINOO at seasonal scale, the information of hydroclimatic teleconnection is used for streamflow forecasting for the Mahanadi River basin, Orissa, India, both at seasonal and monthly scale. Information of streamflow from previous month(s) alone, as used in most of the traditional modeling approaches, is shown to be inadequate. It is successfully established that incorporation of large-scale atmospheric circulation information significantly improves the performance of prediction at monthly scale. Adopting the developed approach of using the information of hydroclimatic teleconnection, hydrologic variables can be predicted with better accuracy which will be a very useful input for better management of water resources.

Knowledge of plausible implications of climate change on hydrometeorology of a river basin will prepare us for adapting to the impacts of climate changes on water resources for sustainable management and development. Among the meteorological variables, six "cardinal" variables are identified as the most commonly used in impact studies (IPCC, 2001). These are maximum and minimum temperatures, precipitation, solar radiation, relative humidity, and wind speed. Among the climate scenarios adapted in impact assessments, those given in Intergovernmental Panel on Climate Change's (IPCC's) Special Report on Emissions Scenarios (SRES) have become the standard scenarios. General circulation models (GCMs) are run at coarse spatial resolutions and therefore the climate variables simulated by these models cannot be used directly for impact assessment on a local (river basin) scale. Support vector machine (SVM) is proposed for downscaling monthly sequences of large scale atmospheric variables simulated by third generation coupled Canadian GCM (CGCM3) to monthly sequences of hydrometeorological variables in a river basin. The monthly sequences are subsequently disaggregated to daily sequences using k-nearest neighbor (k-NN) disaggregation technique. The catchment of Malaprabha river (upstream of Malaprabha reservoir) in India is chosen as the case study to demonstrate the effectiveness of the developed models. Implications of climate change on monthly values of each of the six cardinal variables in the region are studied. Results show that precipitation, maximum and minimum temperature, relative humidity and cloud cover are projected to increase in future for A1B, A2 and B1 scenarios. The wind speed is not projected to change in future for all the aforementioned scenarios. On the other hand, the solar radiation is projected to decrease in future for A1B, A2 and B1 scenarios. No trend is discerned with the COMMIT scenario for any of these variables.

To assess implications of climate change on monthly streamflows in the river basin, daily sequences of the meteorological variables obtained from downscaling and disaggregation models are used as inputs to Soil and Water Assessment Tool (SWAT), besides DEM, land use/land cover and soil data. The SWAT is a physically based, distributed, continuous time hydrological model that operates on a daily time scale. The SWAT model has projected an increase in future streamflows for A1B, A2 and B1 scenarios, whereas no trend is discerned for the COMMIT scenario. Results obtained will be very much useful for effective management of available water resources in the river basin.

Efficient optimization techniques based on swarm intelligence and evolutionary computation principles have been proposed for single and multi-objective optimization in water resources systems. To overcome the inherent limitations of conventional optimization techniques, meta-heuristic techniques such as ant colony optimization (ACO), particle swarm optimization (PSO) and differential evolution (DE) are developed for single and multi-objective optimization. To achieve robust Pareto optimal fronts for multi-objective problems, a novel approach is developed by incorporating Pareto optimality principles into PSO algorithm, called elitist-mutated multi-objective particle swarm optimization (EM-MOPSO). For effectively handling interdependence relationships among decision variables of multi-objective water resource problems, an efficient multi-objective solver, namely multi-objective differential evolution (MODE) is developed. The developed MODE algorithm is evaluated with several test problems and also applied to a case study of Hirakud reservoir to derive operational tradeoffs in the reservoir system optimization. To demonstrate the applicability of the developed optimal operating policies for real time reservoir operation, reservoir inflow forecasting models are developed using soft computing approaches viz., artificial neural networks (ANNs), adaptive network fuzzy inference system (ANFIS) and hybrid particle swarm optimization trained neural network (PSONN). These methods are then applied to a few case studies in planning and operation of reservoir systems in India.

Multicriteria Decision Making (MCDM) has emerged as an effective methodology due to its ability to combine quantitative and qualitative criteria for selection of the best alternative. Several MCDM techniques are adopted for selection or ranking of irrigation planning alternatives. They include (i) fuzzy logic based MCDM methods, namely, similarity analysis (SA) and decision analysis (DA), (ii) Kohonen neural networks (KNN) based classification algorithm (iii) Data Envelopment Analysis (DEA) etc. These techniques are successfully applied to several case studies such as (i) Sri Ram Sagar project, Andhra Pradesh, India, (ii) Jayakwadi irrigation project, Maharashtra, India.

An unsupervised fuzzy classification technique viz., penalized fuzzy c-means algorithm (PFCM) is successfully adopted to classify irrigated area from multi-date multi-spectral remote sensing imageries (IRS LISS I data) into paddy and semi-dry cropped areas in Bhadra command area, Karnataka. Paddy and semi-dry crops were classified with much higher accuracy using PFCM when compared to conventional algorithms. Using this approach, perennial crop (sugarcane) is also discriminated from other crops. These results can be utilized for better irrigation assessment in the command area.

• An integrated model for optimal reservoir operation for irrigation of multiple crops was developed considering reservoir water balance, storage related evaporation losses, conveyance losses, soil moisture balance in the crop root zone, soil heterogeneity, crop root growth along with the stochastic nature of inflows into the reservoir and rainfall in the command area. A combination of Linear program and Stochastic dynamic program were used to maximise the expected sum of relative yields from irrigated crops. This model was applied to a single purpose reservoir, Malaprabha reservoir in Karnataka state. This work is published in Water Resources Research Journal of American Geophysical Union (AGU). This work is well appreciated and reprinted in Water Resources Journal of UN ESCAP
• Mulitobjective Fuzzy linear programming (MOFLP) was used to develop optimal operating policy for reservoir and this work was presented in International Conference on Civil Engineering for Sustainable Development, February 1997, Roorkee, India.
• Stochastic linear programming was used to develop optimal operating policy for Hirakud reservoir in Mahanadi basin, Orissa state and this work was presented in International conference on Large scale water resources development in developing countries, October 1997, Kathmandu, Nepal.
• A multilevel optimisation model was developed to decide the cropping pattern and optimal irrigation allocations at the seasonal level (using dynamic programming) and optimal irrigation allocations within a season (weekly) for a single crop for the given level of seasonal allocation (using stochastic dynamic programming) at the intra-seasonal level. This work was published in ASCE Journal of Irrigation and Drainage Engineering.

• Auto Regressive Moving Average (ARMA) models were fitted to consider the stochastic nature of the streamflows. The program developed for the purpose was applied to different streamflow data followed by validation. The results were published in Hydrological Sciences Journal of IAHS, U.K.
• A Markov mixture model was developed to forecast the streamflow and the results were compared with the forecasts from the family of ARMA models. The results were published in the proceedings of Conference on Hydromechanics and Water Resources, May 1990, Bangalore.
• Spectral analysis of streamflow data was carried out to identify different periodicities. The results were published in Journal of Indian Water Resources Society.
• Hydrological drought forecasting is made with the help of Markov mixture modelling and was presented in All India seminar on Natural Disaster: Causes and Management, December 1995, Bangalore.

• Applied Multiple Regression analysis to study the Rainfall-Runoff relationships of small watersheds in Tamilnadu.
• Least Absolute Deviations approach has been adopted for Linear multiple regression analysis of Rainfall-Runoff data and a code was developed for this purpose using LP. This work was presented in National seminar on Water and Environment, December 1994, Trivandrum.
• Watershed Bounded Network Model (WBNM) was used to model the runoff from a watershed with the help of watershed topology and drainage map and was presented in International symposium on Emerging Trends in Hydrology, September 1997, Roorkee.

• A Linear Programming model was developed to determine the seasonal irrigation allocations corresponding to soil moisture balance for different evaporation rates and rainfall at different levels of accedence probability levels. This work was published in the proceedings of VIII Congress of International Association for Hydraulic Research (IAHR), Asia & Pacific Regional Division, October 1992, Pune.
• A model to allocate water for Drinking, Irrigation and Industrial purposes to different demand nodes for the given availability of water from different input nodes (Ground/Surface) was developed. A computer program was developed for this purpose and was applied to some case studies.
• Linear Programming Gradient (LPG) method was used for the optimal design of water distribution system for a specified layout. A computer program was developed in LINGO for this purpose. This work was published in the Journal of Indian Water Works Association which received best paper award for that year.

• Design aspects of Stepped Spillway over an embankment dam (Jambhira) in Orissa was studied and modifications were suggested for the proposed structure as a part of consultancy work to Orissa Govt.
• Experimental studies were conducted on a monolithic stepped spillway structure to study the energy dissipation characteristics of the spillway. The results were published in the proceedings of International conference on Dam Safety evaluation, November 1996, Trivandrum, India.

• A new approach is developed for space-time disaggregation of streamflow for obtaining the data at number of upstream stations for different months (Shorter time period) given the seasonal data at a down stream station. The methodology uses k-nearest neighbors coupled with liner optimization. The capability of this model is demonstrated through a case study of a river basin in USA. The results obtained were published in Water Resources Research Journal of AGU.

• Digital and Visual interpretation techniques were used to monitor the deviations in actual cropping pattern from that recommended in Rajolibanda Diversion Scheme (RDS), A.P. A technical report was sent to the concerned authorities including Ministry of Water Resources, Govt. of India.
• A Diagnostic analysis of the canal system performance in RDS, A.P. was carried out with the aid of SRS techniques. Temporal and spatial distribution of irrigation supply was evaluated in Electronic Spreadsheet environment. A technical report was sent to the concerned authorities in Govt. of AP and Ministry of Water resources.
• Satellite evaluation of current status of irrigation management in RDS, AP was published in the Bulletin of Natural Resources Management System (NNRMS), Bangalore.
• Evaluation of irrigation management in Bhadra project command area in Karnataka was made with Satellite remote sensing techniques. This work is well appreciated by International Water Management Institute, Sri Lanka.
• Texture labelling and texture spectrum analysis of microwave data (C-band) of ERS-1 SAR was carried out to study the land use/ land cover of Bhadra project command area in Karnataka. This work was presented in National symposium on microwave remote sensing and users meet, January 1994, Ahmedabad.
• The following three works were presented in Workshop on Remote sensing and GIS applications in water resources engineering, September 1997, Bangalore: Satellite Remote Sensing in Irrigation Management ; Energy Interactions on Earth Surface ; Digital Image Processing Techniques for Image Enhancement and Information Extraction.

• Artificial Neural Networks were utilised to develop Rainfall-Runoff model for some watersheds in Tamilnadu. For this purpose a general purpose interactive program is developed in C Language. This work was presented in 24th National conference on Fluid Mechanics and Fluid Power, December 1997, Calcutta.
• Dept. of Science and Technology, Govt. of India, has awarded a project under Young Scientist Scheme to carry out research on this topic.
• Recurrent Neural Networks were used to carry out the Hydrologic time series analysis and Forecasting. These models were successfully applied to forecast monthly river flow data of Hemavathi river in South India. This work is communicated to ASCE J. Hydrology for possible publication.
• Artificial Neural Networks were employed to temporally disaggregate the seasonal rainfall data at a given station to monthly rainfall data. This approach can be used to obtain short-term (say monthly) data from a long-term seasonal forecast (say monsoon season). The results will be presented in ‘ICIWRM-2000’ held in December, 2000 in Roorkee, India.

• Ranking of river basin planning and development alternatives under multi criterion environment, including both qualitative and quantitative aspects is carriedout using ELECTRE (ELimination Et Choice Translating REality) for Krishna river basin. A total of 7 reservoirs and a diversion network are considered for formulation of 24 alternative systems with 18 criteria of which 9 are qualitative in nature. This work was Published in Hydrological Sciences Journal of IAHS, U.K.
• Selection of the best alternative plan in irrigation development strategies is examined in the multiobjective context. The study deals with three conflicting objectives, net benefits, agricultural production and labour employment. Five stage procedure is adopted combining Individual Optimization, Multiobjective Optimization, Cluster Analysis, Multicriterion Decision Making (MCDM) methods and Correlation Analysis. Two MCDM methods, PROMETHEE-2 and newly developed EXPROM-2 are employed in the evaluation. The methodology is applied to a case study of Sri Ram Sagar Project, Andhra Pradesh, India. This work is published in Journal of Agricultural Systems.

• A methodology, RANFUW, (Ranking FUzzy Weights) for Ranking alternatives in multi criterion environment employing multiple experts opinion using Fuzzy Weights with Maximising set and Minimising set was developed. RANFUW was applied to Krishna river basin planning and development. This work was published in the form of two papers in Fuzzy sets and Systems journal, UK.
• Performance of ELECTRE and RANFUW were evaluated through a case study and was published in the proceedings of International conference on Aspects of conflicts in reservoir development and management, September 1996, London, U.K.
• Irrigation planning model was developed for an existing system with Mutli Objective Fuzzy Linear Programming and the results are published in ISH Journal of Hydraulic Engineering.

• Genetic Algorithms were employed to optimize the hydro power generation from a reservoir and the results were presented in 'Xth World Water Congress', Melbourne, Australia