Monitoring, Modeling and Management of Glacial Lakes of Teesta Basin, India.

Natural phenomena like rainfall, coupled with associated activities, have often turned calamities into disasters. In a country like India, endowed with densely populated areas and diverse geographical variations, unfortunate incidents like GLOFs, flash floods, and landslides have frequently proven catastrophically disastrous for the population, causing irreparable loss to life and property. While procedures for disaster management in the aftermath of such incidents exist, there is a pressing need to augment concrete methodologies for the prediction, monitoring, and management of GLOFs, especially concerning hydropower projects.

Interestingly, the Earth's average temperature has risen by 1.1°C since 1850 and is expected to increase further by 1.5°C within a few decades (IPCC, 2021). This rise will intensify the water cycle and accelerate climate change.

The recent flash floods on October 3–4, 2023, have emphasized the necessity for further studies on glacial lakes and their risk assessment. Most of these lakes are located in remote areas at altitudes of around 4,500 to 5,000 meters, making physical assessment a challenging task. To address this, NHPC has initiated a study for monitoring lakes across eight basins in close collaboration with National Remote Sensing Centre, Hyderabad. This study focuses on more than 650 glacial lakes in the Teesta Basin, which are situated within the catchments of NHPC’s four hydropower projects: Rangit, Teesta-V, TLDP-III, and TLDP-IV.

This study aims to integrate Sentinel-1, Sentinel-2, and Landsat 7 and 8 data to measure changes in the areas of glacial lakes over the past 10 years. The 650 lakes will be classified based on risk assessment parameters, including proximity to hydro-projects and settlements, rate of area change, size of the lake, and type of lake. Additionally, subsidence mapping will be incorporated into the classification model for enhanced accuracy.

The Google Earth Engine platform is being utilized to measure changes in lake areas, while Sentinel-1 data is used for time-series analysis of subsidence mapping around the lakes. The output of this study will enable the classification of lakes into five risk categories, which will serve as an input for developing an Early Warning System in later stages.