STREAMFLOW MODELING AND IMPACT OF CLIMATE CHANGE FOR BEAS BASIN IN HIMALAYA

Abstract

Streamflow modeling is of foremost importance for appropriate planning, designing, development and decision-making activities of water resources. The snow-fed rivers originating from Himalayan basins receive ample amount of runoff generated from snow and glacier melt. Moreover, in a Himalayan basin, like the Beas river basin, snowmelt is a governing factor for runoff production. A systematic modeling of streamflow components in such complex terrain of Himalayas is an important and challenging issue in recent changing environment. A snowmelt model that precisely simulates the three components of streamflow namely snowmelt runoff, rainfall-runoff and baseflow are essential for modeling streamflow in Himalayas. The information on model inputs such as snow cover area, meteorological data, temperature lapse rate and other related parameters are mandatory for snowmelt models. The snowmelt modeling also enables to understand the catchment processes and the major impacts on the streamflow regime due to climate change. A probable change in climatic variables namely temperature and precipitation impact the hydrological processes and control the streamflow. Spatio-temporal variation in magnitude of streamflow due to climate change may affect potential water availability of surface water in future, operation of reservoirs and flood related analysis. Therefore modeling of streamflow and study of climate change impact on streamflow is of paramount importance. Snow is an important key environmental parameter of earth’s climatic system which not only affects the radiation budget of the earth but also influences the streamflow of mountainous rivers. A major source of runoff and groundwater recharge in middle and higher latitudes is the snowmelt runoff from seasonal snow covered areas and Himalayan mountain system forms one of the world’s largest suppliers of freshwater. All the three predominant south Asian rivers namely the Indus, the Ganga and the Brahmaputra river systems originating from the Himalayan mountain system are snow and glacier fed. These rivers receive ample amount of snowmelt water and therefore are considered as the lifeline for the Indian sub-continent. The snow accumulated during winter season turns into a major source of runoff for these rivers during summer period. The spring and summer season runoff of a Himalayan river is mostly a contribution from snowmelt and is a dependable source of water for irrigation, generation of hydroelectric power and drinking water supply. The lack of detailed systematic evaluation of the Himalayan water resources, first of all due to poor and inadequate network of hydro-meteorological observations and secondly, its ii inaccessibility for being rugged, dangerous with harsh climatic conditions are the challenges demanding time for proper assessment of streamflow in these rivers. Even then, the available estimates show that water yield is almost double in the snow-fed higher Himalayan river basins than that of an equivalent peninsular basin mainly because of snowmelt and glacier melt contribution. The perennial nature of these rivers and the appropriate topographic setting of the Himalayan region provide exploitable hydropower potential in these regions. The effect of climatic change on the hydrologic systems, specifically on snow and glacier, have led to the alteration in timing and the amount of runoff in mountainous basins. Streamflow in the Himalayan rivers generated from the melting of snow and glaciers have a direct impact on the water resources of a region under these changing climatic conditions. Therefore, near real time snow cover estimation along with the accurate streamflow simulation and forecast under the changed climatic scenarios is of great importance in managing and planning the water resources of a region and also to minimize risk and loss from devastating floods caused due to rapid melting of snow and glacier under changing climate. Therefore, there exists a pressing need to properly assess the natural water resources in mountainous areas and their judicious utilization during different seasons and years for supporting agricultural production, water supplies, industries, energy generation and for functioning of ecosystem health. Moreover, to plan new upcoming hydropower and river valley projects on the river system of Himalayas accentuated the need of accurate and reliable estimation of runoff from snow and glacier. Therefore the present study envisages the assessment of streamflow estimation including snowmelt runoff and evaluation of the impact of climate change on streamflow. Based on the literature review, the objectives of the present study have been set such as to interrelate snowmelt modeling and impact of climate change on streamflow, as follows: • Snow cover assessment using multiple satellite images and assessment through a relationship developed between SCA and temperature. • Determination of seasonally and topographically varying TLR using LST data estimated from thermal satellite images. • Snowmelt runoff simulation for the Beas river using snowmelt runoff model. • Trend analysis of temperature, precipitation and discharge data of Beas basin. • Application of future scenarios on the basis of data analysis to investigate the impact of climate change on runoff. The above objectives have been accomplished for the Beas river basin up to Pandoh dam. The Beas river originates from Beas Kund, a small spring near Rohtang Pass at an elevation of 4085 m and is an important river of the Indus river system. The length of the Beas iii river up to Pandoh dam is 116 km. Among its major tributaries namely Parvati, Thirthan, Sainj Khad, Bakhli Khad, Parvati and Sainj Khad are glacier-fed. The catchment area of Beas river upstream of Pandoh dam is 5384.9 sq. km. In the present study, snowmelt runoff has been modeled using SNOWMOD model. The important input of this model is snow cover area (SCA) and temperature lapse rate beside meteorological data and other parameters. Therefore, the main emphasis has been on estimation of SCA using different approaches and seasonally/topographically varying TLR. Further to evaluate the impact of climate change, trend analyses as well as generation of future scenarios have been the major areas of concern in this study. Finally, impact of climate change on streamflow has been studied on the basis of future scenarios. Snow cover assessment Vastness of the river basin area, time, and manpower constraints in data collection and seasonal changes in the information require fast inventory of the snow cover areas and its mapping. The scarce availability of field based information on the snow cover area (SCA) is a big constraint, rather a challenge, for carrying out snowmelt runoff studies in such a complex rugged terrain of Himalaya. Innovative space based sensor technology and digital image processing tools provide researchers and scientists to envisage snow cover monitoring and mapping for hydrology and water resources management and its impacts due to climate change. Remote sensing has emerged as a powerful active tool with its proficiency in mapping the snow cover area. Earlier NOAAA-AVHRR and IRS-WiFS data products were used in Himalayan basins. However, getting cloud-free data for most part of the year, particularly in the Himalayan region, is again a big deal. Consequently, Moderate Resolution Imaging Spectro radiometer (MODIS), a NASA space satellite, has been used for determination of SCA for the study area as it provides cloud-free scenes at a regular interval.The MODIS 8-day composite maps for a period of six years (2000-2005) were downloaded and processed to obtain snow cover depletion curves for the Beas basin. To check the efficacy of MODIS SCA, IRS-1C/1D WiFs and AWiFS satellite data have also been used for some of the dates. As per the analysis it was found that SCA estimated using MODIS is quite satisfactory compared to SCA obtained using WiFS and AWiFS satellite data. SCA was estimated mainly using MODIS satellite data available from 2000 onwards. A method using mean air temperature has been proposed to obtain SCA information when the satellite data was not available. An exponential relationship between SCA and CMAT has been generated for preparation of depletion of SCA. This iv technique also supports in providing SCA for the previous years when the satellite data is not available or is cloud covered and hence reducing the quantity of satellite imageries. Determination of Temperature Lapse Rate Temperature Lapse Rate (TLR) is an important parameter necessary for temperature-based conceptual snowmelt runoff simulation. Earlier TLR was estimated from the ground based air temperature data recorded at meteorological stations. However, it is very difficult to establish meteorological stations in the complex Himalayan terrains with rugged and undulated topography and the available sparsely located network represents only local temperature. Thus, estimation of representative TLR values is extremely difficult and hence a fixed TLR value (=0.65 o Spatio-temporal TLR was estimated for different dates (2001 to 2009) employing satellite based MODIS-LST maps and USGS-DEM for the Beas basin. In case of seasonal variation, TLR was the lowest during monsoon season whereas its variation with respect to topography showed that the lapse rate was low in lower altitudes than in higher altitudes. C/100m) calculated from air temperature is most commonly and often used in snowmelt studies. Moreover, since TLR can vary both seasonally and regionally within different elevation zones, the use of fixed values of TLR in the snowmelt model may not be appropriate. Nowadays, remote sensing satellites provide a straightforward and consistent way to observe Land Surface Temperature (LST) over large scales with more spatially detailed information and these data well represent the terrain. Snowmelt runoff simulation Streamflow was modeled for the Beas river basin using SNOWMOD, which is a temperature index based snowmelt runoff model. The model simulates the three components of streamflow i.e. snowmelt runoff, rainfall-runoff and baseflow independently and their sum provides the total streamflow. Employing the above variable TLR values the model was calibrated using daily data of 3 years (2002-2005) and these parameters obtained during calibration were used for validation of streamflow for a period of twelve years i.e. 1990-1993, 1993-1996, 1996-1999 and 1999-2002. The model efficiency (R2) for snowmelt season of 1990-2002 varied from 0.68 to 0.80 and the difference in volume varied from -9.63% to 9.15%. During this period the snowmelt contribution varied from 24.07% to 34.62%, and rainfall from v 30.48% to 41.85%. The average runoff generated from the snowmelt during the ablation period was 38.2% whereas remaining 61.8% was contributed from rainfall. Trend analysis of temperature, precipitation and discharge data of Beas basin For better understanding the trends in climatic variables namely temperature, rainfall and discharge of the Beas basin trend analysis was carried out using parametric (linear regression) and non-parametric (Mann-Kendall and Sen’s slope Estimator) approaches both at annual and seasonal scales. The hydro-meteorological data of Beas river were collected from BBMB, Pandoh, Himachal Pradesh. Majority of stations showed increasing trends in mean annual temperature whereas one station (Manali) showed a decreasing trend over the last three decades. The annual rainfall indicated increasing trend at Banjar station, and decreasing trend at all other four stations with maximum decrease of -8.07mm/year at Sainj. Seasonal analysis of rainfall trends showed that all stations during pre-monsoon, post-monsoon, and winter season experienced decreasing trend whereas all stations experienced increasing trend in monsoon seasons. A decreasing trend in discharge was observed at three stations Bakhli, Pandoh and Thalout while the other two stations Sainj and Tirthan indicated increasing trend at annual time scale. Impact of climate change on runoff For climate change impact assessment, in two ways scenarios were generated and applied. First of all, ten hypothetical scenarios for (T + 1°C, P + 0%), (T + 2°C, P + 0%), (T + 1°C, P + 5%), (T + 2°C, P + 5%), (T + 1°C, P + 10%), (T + 2°C, P + 10%), (T + 1°C, P - 5%), (T + 2°C, P - 5%), (T + 1°C, P - 10%), (T + 2°C, P - 10%) with respect to baseline scenario (T + 0, P + 0%) were generated to study the impact of climate change on streamflow. An increase of 1°C and 2°C in temperature increased the mean annual streamflow by about 9% and 8.69%, respectively. Further, to assess the impact of temperature, projected temperature scenarios based on IPCC SRES A1B were generated and applied. To account for the change of SCA, modified snow cover depletion curves were prepared using exponential relationship between SCA and air temperature. The model was simulated with the change in temperature and modified depletion curves for the years 2040-43, 43-46, 46-49, 49-52 and 2093-96, 2096-99. The total vi flow reduced for the years 2040-43, 2043-46, 2046-49, 2049-52 while for the years 2093-96 and 2096-99, it increased marginally. The snowmelt runoff decreased in all future scenarios. The stream/snowmelt runoff modeling and impact of climate change studies for a Himalayan basin incorporating the unique technique of estimating SCA using mean air temperature and seasonally and topographically varying TLR have been used for the Beas basin for the first time, and promising results obtained. Overall, the significantly improved TLR and SCA estimates enhanced the model efficiency in simulating the streamflow. The assessment of hydrological impacts of climate change at basin scale for any Himalayan region is unique and crucial. The hypothetical and future projected scenarios furnished with the plausible future streamflow under the changing climatic conditions for the basin. Therefore, such studies are of great significance for accurate streamflow simulation and forecast of Himalayan rivers under contemporary and changing climate for proper planning and management of precious water resources.