WATER BALANCE STUDY OF MARSYANGDI RIVER BASIN IN NEPAL USING SWAT MODEL

Abstract

The Marsyangdi River is originated from the northeast of the Annapurna Himalayan range and flowing southward, ultimately joins to the holy Ganga River. The total watershed area of the Marsyangdi River is 4037 Km² at the Bimalnagar outlet which is considered in this study as study area. The climate of the Marsyangdi basin has basically governed by summer monsoon (June to September) which received ¾ of annual precipitation. This landscape is extending from higher Himalayan to lower tropical region. This results in significant weather variation in the region. Generally, the basin climate varying with elevation with cold high alpine to humid and hot tropical type class. The majority of the precipitation for this region is concentrated during the monsoon month (June to September) and the upper part of basin is all time snow covered which maintains the river’s perennial characteristics. The Marsyangdi river basin has great potential for hydropower generation with tremendous tourism scope such as rafting, kayaking, etc. The Himalayan River’s hydrology is supposed to be most vulnerable, fragile hydrology and more sensitive to climate change due to the effects of the seasonal, latitudinal and altitudinal shift in freezing line. Rising air temperature, snow melting, erratic rainfall, desertification are key indicators for climate change within the basin. In this study, hydrological modelling of the Marsyangdi river basin was carried out using the ‘Soil and Water Assessment Tool’ (SWAT). The major objectives of the study were (i) assessment of water availability, (ii) estimation of different water balance components and (iii) to assess and simulate the impact of climate change on streamflow. Hydrological and meteorological data were obtained from DHM Nepal. Part of the meteorological data, solar radiation and wind speed data were auto simulated by ArcSWAT. Other inputs such as SRTM-DEM, Soil map and LULC map were obtained from USGS, FAO (SOTAR) and ICIMOD Nepal respectively. The SWAT model was calibrated for daily discharge using SWAT CUP/SUFI-2 algorithm for the periods from 1994 to 2005. The model was validated for periods from 2006 to 2009 where warm up period of 1992 to 1993 was taken. Model performance was examined using various model performance criteria reported in the literature such NSE, PBIAS, R² and RSR. For daily flow calibration period (1994-2005), the value of R², NSE, RSR and PBIAS were 0.87, 0.85, 0.39 and 13.2, and for validation period (2006-2009) the values were 0.76, 0.75, 0.49 and 14.1 respectively. On the basis of these statistical results and graphical comparisons it was found that the performance of the SWAT model was “very good” for R², NSE and RSR objective functions and “good” for PBIAS objective function in both calibration and validation stage. Climate parameters from two finer resolutions Regional climate models (RCMs) viz.