IMPACT OF CLIMATE AND LANDUSE CHANGES ON GROUNDWATER IN A CANAL COMMAND

Abstract

Climate change has been emerging as one of the challenges in the global environment. It is responsible for various substantial impacts on hydrological cycle, which influences the climatic parameters likely temperature and rainfall. Hence, it is necessary to predict climate variable at regional and local scale, to know the future climatic condition, which ultimately becomes helpful to perform planning and management of the water resources availability. It also affect the agricultural production based economy leading to food deficits. However, appropriate measures, thoroughly evaluation of the situation and cognizance among the people may help in productive management. In the present study, historical data have been used for the climate trend analysis and future prediction of climate variable (viz., temperature and precipitation) is carried out to understand the climatic condition of the study area, Indira Sagar Canal command area, Madhya Pradesh (MP), India. Landuse change detection and future prediction has been carried out to estimate the effect of future landuse and landcover changes on groundwater recharge. Groundwater model have been developed for study area to estimate groundwater level. The relationship between rainfall and recharge has been develop to evaluate the impact of climate parameter on groundwater recharge at present and in future scenrio. The trend analysis is carried out at regional and local scale (state and canal command level) for rainfall and temperature (Tmax, Tmin and Tmean) with more than 100 years data for the entire MP,and for the study area, which is a part of the Narmada river basin. The rainfall trend for entire MP shows a decreasing trend annually and particularly in the monsoon time. Monsoon season is the major contributor of rainfall in a year and decrease in monsoon rainfall is reflected in the reduced total annual rainfall of MP. The magnitude of the decreasing trends in the annual rainfall varies from 0.01 mm/ year (Gwalior) to 1.58 mm/ year (Balaghat) for MP is observed from 1901 to 2011. For Tmax, Tmin and Tmeanof MP have shown a significant increase in almost all the seasons and in the annual series. Trend analysis for rainfall and temperatures are carried out for 111 years (1901-2011) and 105 years (1901-2005) respectively at 95% and 99% confidence level. The net increase or decrease in rainfall and temperatures over the study area is performed to assess the effect of climate.At local level, detection in longterm changes in precipitation and temperature characteristics in annual, monthly and seasonal basis have been determined by both parametric and non-parametric tests. In the case of rainfall, winter season indicates the downward trend and only the month of August shows the statistically significant trend at 95% of confidence level, for a study area. In the case of ABSTRACT ii temperature, the monotonic trends in the annual temperature time series were positively significant at 0.1 level of significance. Change point year has been 1964 for the Barwani and West Nimar stations, wheareas 1950 is noticed for East Nimar station. The spatial analysis of the monotonic trends indicated that the stronger increasing trend of temperature series at Indira Sagar canal command area. Overall results show a net deficit in the rainfall amount and the net gain in the temperature amount. Projection of future rainfall has been generated by LS-SVM (Least-Square Support Vector Machine) and SDSM (Statistical Downscaling Model) methods using GCM (General Circulation Model) data. Future projection of climate data is done for three stations of the study area -viz., Barwani, East Nimar (Khandwa) and West Nimar (Khargone). The rainfall, minimum and maximum temperature show increase in future in both the models. This study concludes coefficient of correlation (R) gives the better performance in LS-SVM as compare to SDSM. Mean absolute error (MAE) in case of rainfall has been found in range of 0.52-0.54 for SDSM and for LS-SVM within a range of 0.58-60. RMS error in case of maximum temperature has been found in range of 1.29-1.91 for SDSM and for LS-SVM within a range of 1.11-1.22. The Maximum Likelihood Classification (MLC) is used for the landuse classification of 1990, 2000 and 2010. Markov Chain Model has been used to project future landuse change of 2020 and 2030. The prediction of the follow year 2020 and 2030 was done by Markov Chain model using explanatory variables viz. distance form streams, urban, roads and slope, elevation and evidence likelihood maps with the spatial accuracy of 80.12% and 76.58% that shows matched area between the classified 2010 image and predicted 2010 image. The transition probability shows two classes of water body and settlement are consistent with the highest probability of self-retention and low probability of change in the future in 2020 and 2030 landuse. Barren land and vegetation indicate highest probability of change to other classes from 2010 to 2020 and 2020 to 2030. Agricultural lands show increases from 1990 to 2010, whereas in 2020 and 2030 there is a slight decrease in the agricultural land due to the encroachment of the continuously increasing settlement in these areas. In 2030, the transfer of vegetated area to other classes has reduced, while transfer of barren land to agriculture and barren land and agriculture to settlement is more. Moreover, three-dimensional conceptual groundwater model is developed to calibrate the model with known limited available information. Calibratation and Validation of the model is done for using six and four years with observed groundwater levels. The observed and estimated groundwater levels for the period 2001 to 2010 are plotted to check the trend in iii groundwater prediction through developed GWM in Simulation stage. The simulatation statistics for (2001 to 2010) time steps shows that the normalized RMS from 2.17% to 5.24 %, which should be less than 10 to 15% for a good simulation and RMS varies from 1.114 m to 3.937 m. Developed an integrated framework of climate parameter(rainfall) with groundwater recharge to predict the future groundwater recharge at 2020s and 2030s scenario. The linear regression relationship has been formulated between observed recharge and rainfall. Therefore, the predicted rainfall is used in the developed equation to obtain the future groundwater recharge. For monsoon season, the average predicted recharge for the entire Indira Sagar command area are expected to vary in range of 81.4 to 61 mm/year with respect to their corresponding average rainfall 845 to 633 mm/year for the period of 2011-2020. Similary, during the period of 2021-2030, average values of rainfall and corresponding recharge are expected to vary within the range of 868.6 to 623.62 mm/year and 83.21 and 59.29 mm/year respectively. However, the existence of uncertainty in the model outputs could not be ignored and there are also complex relations within the interactions of climate, landuse, hydrological cycle and to come to any implication is quite difficult. In order to evaluate the impact of change in climate and landuse, affecting the environment and natural resources need to assess. Therefore, this study has highlighted the importance of climate analysis at the local and regional level. Along with the impact of human intervention that may be considered in the decision making process, which is influencing the groundwater recharge of the area, and proper management processes of groundwater might be helpful in the future.