HYDROLOGICAL MODELING TO STUDY THE INTERACTIONS OF LAND USE−CLIMATE−HYDROLOGY FOR SUSTAINABLE RIVER BASIN MANAGEMENT

Abstract

Sustainable natural resources management at local and regional scale is a prime concern for growth, development, conservation, and protection of the environment, and prosperity of the nation. To mitigate the present and future anthropogenic activities and climatic change, field-based investigations of land and water resources are time-consuming and challenging to manage and sustain the agriculture food production in developing countries like India. Indian River basins are one of the most influencing natural systems owing to land dynamics and the uncertain climatic events, such as cloud bursting and heavy rainfall occurred in Uttarakhand during the year 2013, then in Chennai (2015), and the recently in Kerala (2018). The present study has been focused to model the complex hydrological process of the Betwa River basin, part of the lower Yamuna River basin located in central India, for sustainable management of land and water resources considering future climate change and land use change. In this context, hydrological modelling can be considered as a valuable technique for the simulation of basin-wide various hydrologic components i.e. stream flow (FLOW), sediment yield (SYLD), evapotranspiration (ET) and water yield (WYLD) etc. In this study, various satellite imageries have been utilized to prepare the historical land use/land cover (LU/LC) maps for the years 1972, 1976, 1991, 2001, 2007, 2010 and 2013 using a maximum likelihood supervised classification method. Further, an integrated Cellular Automata-Markov Chain (CA-MC) model based on Geographical Information System (GIS)-based Multi-Criteria Evaluation (MCE) and the Multi-Objective Land Allocation (MOLA) methods has been employed to predict the future LU/LC maps for the years 2020, 2040, 2060, 2080 and 2100. Future problems such as food security and surface water resources availability are successfully discovered through CA-MC model. To study the relationships between land cover dynamics and hydro-climatic variables, the MODIS NDVI (MOD13Q1) and land cover (MCD12Q1) time-series datasets have been used for correlation analysis, and then Multiple Linear Regression (MLR) models were prepared at monthly, seasonal and annual time-scale over the period of 2001-2013. The Savitzky-Golay filtering method was employed to de-noise and smoothing of the NDVI time-series data using TIMESAT software. The land greening and land degradation under dry spell, wet spell and combined dry and wet spells were analyzed employing a conceptual framework, representing four concepts of climatic greening, climatic degradation, non-climatic greening and non-climatic degradation etc. The developed conceptual framework approach can be applied effectively in other river basins having different land cover and hydro-climatic conditions. Further, hydrological modelling considering numerous medium to large sized water storages (7 reservoirs and 2 weirs) located on main channel as well as tributary channel of the Betwa river has been carried out using the Soil and Water Assessment Tool (SWAT). With the required spatial, storage and outflow information, these water storages were successfully implemented and managed for reliable hydrological simulation using the SWAT model. Monthly calibration, validation, sensitivity and uncertainty analyses have been carried out using the SWAT- Calibration and Uncertainty Programs (CUP) Sequential Uncertainty Fitting version-2 (SUFI-2) algorithm for the years 2003-2013. The observed and simulated hydrographs for both the streamflow and sediment indicates a good performance of the SWAT model. The model ii performance was high for the Garrauli gauging site without any upstream water storage structure, as compared to the gauging sites with upstream water storage structures. This analysis shows that better information of the water storage structures promises a significantly improved hydrological simulation using the SWAT model. The India Meteorological Department (IMD) data benchmarked in calibrated and validated SWAT model was replaced by the downscaled and bias-corrected (quantile mapping method) Global Climate Model (GCM) data of the Max-Planck-Institute-Earth System Model-Medium Resolution (MPI-ESM-MR) model. In this study, the MPI-ESM-MR model data of RCP 8.5, a worst-case climate scenario, has been considered for hydrologic simulation at the severe climate condition in future. Land use data of the years 2013 and 2040, and the GCM-derived climate variables were categorized into five periods i.e. baseline 1986 (1986-2005), horizon 2020 (2020-2039), horizon 2040 (2040-2059), horizon 2060 (2060-2079), and horizon 2080 (2080-2099) and used to assess the land use and climate change impact on hydrological simulation of streamflow (FLOW), sediment yield (SYLD), Evapotranspiration (ET) and water yield (WYLD). It was found that climate change impact is dominant over the impact of land use change in future. Further, a conceptual framework has been developed to assess the individual as well as combined impacts of land use and climate change. The proposed conceptual framework can be used effectively for watershed analysis with given limitations. Furthermore, based on the future simulations, critical sub-watersheds of the study area were identified and then prioritized for effective implementation of Best Management Practices (BMPs). In this study, the over-land as well as in-stream BMPs has been implemented to reduce the streamflow and sediment yield in future. Four over-land BMPs namely tillage management, contour farming, residue management and strip cropping for agriculture land, and five in-stream BMPs namely grassed waterways, streambank stabilization, grade stabilization structures, porous gully plugs and recharge structures for main and tributary river channels have been considered in this study. Sensitivity and uncertainty analysis of BMPs parameters were also carried out for an effective management and implementation of BMPs in the river basin. The effectiveness of BMPs implementation was estimated by percent reduction and sensitivity index of the model parameters. It was found that strip cropping is the most effective agriculture land operation which reduces streamflow in the range of 11.07% to 13.97% and sediment yield in the range of 21.04% to 37.28% for soil and water conservation of the river basin in future. Furthermore, the in-stream BMPs namely grassed waterways and streambank stabilization can be an effective intervention for sediment yield reduction (about 20% to 60%), and grade stabilization structures for streamflow reduction (about 6% to 10%) within the main river channel. Overall, this study provides connectivity of land use change, climate change, and hydrological modelling for the research communities focusing sustainable river basin management, and may also provide valuable guidelines to the users interested in water resources development, planning and management in agriculture dominant large river basin.