Snowmelt Runoff Modelling of Upper Beas Catchment Using RS and GIS Techniques

Abstract

The mountain region covers a substantial portion of the land and serves as a crucial source of fresh water in the form of perennial rivers. These rivers are mostly snow-fed, which imposes high discharge during the monsoon and summer, leading to peak discharge. Avalanches from the emergence of glacial lakes during the spring season are a yearly phenomenon that causes flooding, and underestimating seasonal snowmelt is a yearly catastrophe. Drought poses a frequent disaster during the dry season when water entirely disappears from tributaries and even in the main rivers. The upper Beas River basin, extending up to the Pandoh Dam as the outlet point, witnesses substantial snowfall between October and February, with the water turning muddy from March to September. To ensure effective water resource management, disaster preparedness, and drought mitigation, it is essential to implement a comprehensive approach involving long-term monitoring and forecasting of glaciers and snowmelt. This can be achieved through accurate and efficient hydrological modeling. The first objective of the current research was to apply advanced Remote Sensing and GIS techniques to find the specific areas contributing to high surface runoff and sediment production rates in the mainstream during the monsoon and summer seasons using geo-morphometric analysis and Principal Component Analysis (PCA). The second objective was to establish a suitable hydrological model for precise measurement of snow accumulation and depletion using statistical techniques such as Seasonal Auto-Regressive Moving Average (SARIMA) models. Three hydrological models viz. SRM, MIKE HYDRO RIVER, and SWAT were then employed for the simulation of discharge based on evaluation criteria like R2, PBIAS, NSE, and RSR. Finally, the simulated discharge computed using the best-fit model was utilized for the identification of potential hydropower sites along the flow channels. The outcomes of the geo-morphometric analysis indicated that surface runoff and sediment production rates (SPR) ranged from 3.576 to 5.240 sq. km/sq. km and 0.101 to 0.234 ha-m/100sq.km/year, respectively. These rates were consistently observed annually, particularly during the monsoon season. The primary contributors to these rates were the upper glaciated regions. Consequently, the sub-catchments were categorized into three groups: medium, high, and very high. This classification was based on the observation that sub watersheds situated in the upper regions exhibited elevated levels of runoff and sediment load. The empirical SARIMA models proposed independently for snow accumulation as (1,1,1) (0,1,3)19 and for snow depletion as (1,1,1) (1,1,2)27, demonstrated a favorable agreement with the observed data. The R2 values were 0.83 for snow accumulation and 0.89 for snow depletion, indicating a strong correspondence between the model predictions and the actual observations. Furthermore, the attained results give confidence in utilizing the proposed models for forecasting weekly snow accumulation and depletion. The comparison results of three hydrological models in terms of parameters and evaluation criteria showed that all three models successfully simulated runoff discharge within acceptable limits. However, the SRM model outperformed the other two in terms of evaluation criteria, but the model does not consider the contribution of base flow in the stream as one of the major components of making the river perennial. The second best fit model (MIKE HYDRO RIVER), as the model accounts for both stream flow and base flow, was used for the identification of potential hydropower sites along the flow channels having orders 5 or more. A total of 131 hydropower potential sites were identified along the main 13 streams for the run-of-river (ROR) plants that do not require any upstream storage construction.