Prediction of GWL with the help of GRACE TWS for unevenly spaced time series data in India : Analysis of comparative performances of SVR, ANN and LRM

Prediction of Ground Water Level (GWL) is extremely important for sustainable use and management of ground water resource. The motivations for this work is to understand the relationship between Gravity Recovery and Climate Experiment (GRACE) derived terrestrial water change (Delta TWS) data and GWL, so that Delta TWS could be used as a proxy measurement for GWL. In our study, we have selected five observation wells from different geographic regions in India. The datasets are unevenly spaced time series data which restricts us from applying standard time series methodologies and therefore in order to model and predict GWL with the help of Delta TWS, we have built Linear Regression Model (LRM), Support Vector Regression (SVR) and Artificial Neural Network (ANN). Comparative performances of LRM, SVR and ANN have been evaluated with the help of correlation coefficient (rho) and Root Mean Square Error (RMSE) between the actual and fitted (for training dataset) or predicted (for test dataset) values of GWL. It has been observed in our study that Delta TWS is highly significant variable to model GWL and the amount of total variations in GWL that could be explained with the help of Delta TWS varies from 36.48% to 74.28% (0.3648 <= R-2 <= 0.7428). We have found that for the model GWL similar to Delta TWS, for both training and test dataset, performances of SVR and ANN are better than that of LRM in terms of rho and RMSE. It also has been found in our study that with the inclusion of meteorological variables along with Delta TWS as input parameters to model GWL, the performance of SVR improves and it performs better than ANN. These results imply that for modelling irregular time series GWL data, Delta TWS could be very useful. (C) 2018 Elsevier B.V. All rights reserved.