HYDROLOGIC CHARACTERISTIC ANALYSIS IN GLACIERIZED BASIN CONSIDERING CLIMATE IMPACT, LAND USE CHANGES AND ELEVATION DEPENDENT BEHAVIOUR OF METEOROLOGICAL DATA

Abstract

The rainfall and meltwater from upstream mountains of Himalaya affect domestic and irrigational water demand, hydropower generation and sustainability of Himalayan bio-diversity and environment. Increase in temperature and uncertain rainfall patterns with improper management of land resources cause frequent floods, loss of snowpack, rapid glaciermelt, changes in timing of streamflow and increased sediment load. Due to global warming this phenomenon is projected to continue doing so in the coming decades. Therefore, accurate estimation of the quantities of various runoff components, streamflow and sediment yield is extremely essential to overcome the challenges of water resources management in mountain catchments. It is equally important to use reliable tools and approaches for estimating past and future changes in melt components, sediment load, rainfall runoff and streamflow dynamics to overcome the climate change impacts. The scarcity of ground-based observations in this region affects the reliability of hydrological modeling and water resource management. Therefore, it is essential to estimate further data sources than usually applied data. A trend attribution study and simulation experiments were carried out to examine the effect of temperature and precipitation variation on the changes in streamflow and runoff components in the three altitude regions of the Beas river basin. Initially VIC-Glacier hydrology model was used to (i) Analyze the contribution of runoff components to discharge and (ii) Quantify climate impact in regulating the trend of streamflow and its components. Mann-Kendall test and Sen’s slope estimator test was used to identify the trend and it’s magnitude of meteorological and hydrological components. Results indicated that the VIC-Glacier hydrology model more precisely analyzed the sensitivity of hydrological processes. Rainfall runoff (47.12%) and snowmelt runoff (49.21%) were dominant contributors to the total runoff of low (Beas up to Pong dam) and high-altitude (Beas up to Manali) basins. Whereas, both of these runoff components significantly contributed to the total runoff (20.89% and 26.00%) of the middle altitude basin (Beas up to Pandoh dam). Seasonal (56-65%, 65-73% and 70-77%) and annual (34.00%, 36.00% and 37.96%) baseflow contribution was also notable for these three altitude regions. At the seasonal (except winter) and annual scale in higher, middle and lower altitudes, streamflow showed a significant downward trend. Instant melting of accumulated snowpack during winter (due to an increase of winter temperature) resulted in rise in streamflow during this season. Decreasing streamflow for Pong dam and Manali was attributed predominantly to rainfall runoff and snowmelt runoff, which contributed to 70% and 75% to overall decrease in basin runoff. The decline in both of these runoff components accounted for up to 61% and 54% in controlling the total runoff trend (negative) of Manali. The results indicated that temperature and precipitation variations were translated into the hydrologic regime’s change of the Beas river basin by way of changes in runoff components. This study also helps us to improve our understanding of the variation of streamflow and its components under climate change. Temperature and precipitation products from four reanalyses (MERRA, ERA-Interim, JRA-55 and CFSR), one global meteorological forcing data (WFDEI), three satellites (IMERG, MSWEP, and TRMM-RT) and one global gridded rain gauge (APHRODITE) data were compared with observation to understand uniformity/disparity between them. The study also provided a comprehensive evaluation of streamflow simulated by above-mentioned different sets of precipitation products, which are based on different retrieval algorithms with varying native temporal resolution. VIC-Glacier hydrology model was used to evaluate the hydrologic utility of gridded rainfall products. The assessment of the performance of hydrometeorological variables involved several statistical techniques, such as PBIAS, RMSE, RE, R2, NSE, CC, categorical index and extreme frequency index. Moreover, climate data from ERA-Interim reanalysis was used to (i) Simulate streamflow with limited climate stations in data sparse Beas river basin and (ii) Reproduce elevation dependency that can help to provide accurate lapse rate and altitude gradient of temperature and precipitation for snowmelt runoff modeling. Interpolated station temperature and rainfall data IMD was used as a reference point. Linear regression method and hydrological modeling approach (using snow module of ARCSWAT) were used to understand elevation dependent behavior of them. Results conveyed that there was a good match between ERA-Interim and observed temperature. Gridded rainfall products correlated in-situ observational data with over/underestimation both seasonally and annually. Higher RMSE proved disparity between different precipitation datasets. JRA-55, CRSR, WFDEI, IMERG and MSWEP rainfall produced errors in detecting extreme rainfall events in the monsoon season. ERA-Interim temperature and rainfall possessed a satisfactory trend and variation with elevation. Streamflow and sediment yield simulated using this reanalysis data better represented the observed hydrograph pattern. In contrast, other gridded rainfall datasets did not agree well with observed data-simulated hydrograph even at capturing the peak flow. This can be attributed to their inability to accurately measure stratiform precipitation between intense periods of precipitation during extreme events. Thus, the vital demand of reliable climatic and hydrologic data of fine spatial and temporal resolution triggered the employment of ERA-Interim as a surrogate in most of the hydrological modeling exercises of mountain region. The study suggests that uncertainties of station observation and gridded rainfall estimate to produce orographic rainfall pattern of the study basin is attributed to errors in measurement, retrieval algorithm and post processing techniques. Application and validation of a coupled modeling approach with an ensemble of six RCMs from CORDEX (CSIRO-ACCESS1, CSIRO-CCSM4, IITM-RegCM4, MPI-ESM-LR, NORESM1 and SHMI-RCA4-V2-ICHEC-EC-EARTH) and VIC-Glacier hydrology model were carried out for Beas river basin up to Pong dam to estimate sensitivity of hydrologic components to climate change. CA-Markov model (combined model of Markov and cellular Automata) was used to predict future LULC for 2030 and 2040 using the LULC map of 1985, 1995, 2005, 2030 and 2040. Water and energy balance components were estimated under combined and isolated effects of LULC and climate change. According to results, climate models project an increment of future temperature and rainfall on both seasonal and annual basis. Increase/decrease in deciduous broadleaf forest/grassland affects an increase in the future surface runoff and ET (under constant climate conditions and land use and land cover change only). Rainfall runoff (25-45% under RCP 4.5 and 45-55% under RCP 8.5), glaciermelt runoff (52-85% under RCP 4.5 and 64-88% under RCP 8.5) and evapotranspiration (25-30% under RCP 4.5 and 28-34% under RCP 8.5) are expected to increase both seasonally and annually in the near (2006-2040) and far future (2041-2070). Meanwhile snowmelt runoff is likely to be decreased (44-60% under RCP 4.5 and 87-89% under RCP 8.5) during 2006-2040 and 2041-2070 due to increasing temperature. Under the influence of increased rainfall, future discharge is envisaged to rise from 12-16% and 16-20% during 2006-2040 and 2041-2070. A significant decrease in snow depth and snow cover in the future is attributable to more rain on snow events. Results suggest that increased discharge has the potential to cause major floods in this basin. Over the long-term snow and glaciers will become much reduced in extent and their contribution to river flow will decrease. By replicating snow/glacier studies in reference catchments, incorporating high-resolution data and considering the potential sources of uncertainty in a hydrologic variable the result of this thesis can be taken as indicative of how hydrological regimes changes in the present and future time periods under changing climate. This analysis is more reliable and used to develop strategies for adapting to the rapidly changing environmental conditions.