Assessment of impacts of land use/land cover changes upstream of a dam in a semi‑arid watershed using QSWAT

Abstract

The study of hydrodynamic responses of the watershed influenced by the changes occurring in the climate and pattern of land use is vital in the management of sustainable water resources. The open-source soil and water assessment tool QSWAT has been adopted in this study to link the meteorological factors with land-surface hydrology and present a complete response of an ungauged watershed in the Krishna basin. The main objective of the present work is to assess the adaptability of the QSWAT model for the selected semi-arid watershed, which has undergone drastic land use/land cover (LULC) changes due to the construction of a dam. The impact of the LULC changes on the hydrodynamic response of the watershed is analysed. For automatic calibration and uncertainty analysis, SUFI-2 algorithm is used. Initially, the model adaptability for the water- shed is assessed by simulating for 32 years, of which 27 years (from 1982 to 2008) are used for calibration and 5 years (from 2008 to 2013) for validation. Further, Landsat satellite images along with 14 LULC classifications for the year 1998 and 2009, indicating scenarios of pre- and post-construction of the dam, respectively, are used as input to QSWAT to analyse the influence of LULC changes on the water balance components. The comparison indicates decrease in the agricultural area, barren land and urban built-up area. The annual water yield and surface runoff of the watershed have been reduced by 28.97% and 29.91%, respectively. An increment of 6% evapotranspiration loss with a decreasing trend in rainfall is noted which is alarming. The simulated results indicate that the hydrological responses are influenced by the LULC changes. The basin LULC and hydrological components have been affected by the storage reservoir created due to the dam. The QSWAT seems to be reliable tool as there is a good agreement between the simulated and observed flows. The obtained model performance indices: the NSE and R2 calibration values are 0.89 and 0.96, respectively, and the values for validation are 0.79 and 0.83 respectively, hence indicating a strong and predictive capability of the model to the ungauged watersheds with drastic LULC changes. Apart from establishing sustainable water resources management techniques in the watershed, there is a stressing need for such analysis before any human intervention into the natural system like construction of dams.