Groundwater modelling in Ganga-Ramganga interbasin using fuzzy logic and ANFIS
Loading...
Date
2006-06
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
G.B. Pant University of Agriculture and Technology, Pantnagar - 263145 (Uttarakhand)
Abstract
The realisation of the concept of natural resources and its conservancy is presently looked upon as one of the main interests of our civilisation. Water has a unique position among the natural resources and is one of the major components of our economic advancement. Keeping this in view, the present study was conducted in the selected area of Ganga-Ramganga interbasin of Uttar Pradesh with specific objective to develop groundwater models using Fuzzy Logic Rule Based Algorithm, and Adaptive Neural Fuzzy Inference System (ANFIS), and to evaluate the models’ performance on the basis of performance indicators. Various components of groundwater recharge and discharge were estimated for preparation of input data set of the study area. The Fuzzy logic rule based algorithm technique was adopted to develop the groundwater model. Groundwater recharge, groundwater discharge and previous water table elevation above mean sea level were considered as input and the water table elevation as consequence variable for the model development. All input and output variables were separately divided into seven subsets. The Fuzzy rule base was formed based on the basis of historical data and intuition. The centroid defuzzification method was adopted to obtain crisp value. For developing such models the area was divided in to 43 polygonal nodes to account for large spatial variation in the region. Thus as many as 86 models were developed in order to predict pre- and post-monsoon water table elevations for 43 nodes. All the models performed well when evaluated using statistical and hydrological performance indicators. Keeping the potential of Adaptive Neuro-Fuzzy Inference System (ANFIS) technique in view, it was applied for developing a single model separately for, pre- and post-monsoon conditions for all 43 nodes. The input and output variables were same as in Fuzzy logic rule based models. The water table elevation predicted by ANFIS model was compared with the observed values and performance of the model was tested using various performance criteria. The results revealed that both models performed well for the prediction of the water table elevation. When compared, the Fuzzy logic rule based models performed better than ANFIS models on the basis of performance indicators. However, the ANFIS technique had advantage of having single model for whole study area for each season.
Description
Keywords
null