In a bid to comprehend the potential dangers of glacial lakes, the Glacial Lake Outburst Floods in Central Asia (GLOFCA) project has emerged as a significant contributor to regional best practices. By leveraging cutting-edge technologies and international expertise, the project aims to create a safer future for Central Asian communities.
One of GLOFCA’s main contributions lies in creating a comprehensive and up-to-date inventory of glacial lakes across the region. This step is crucial as it allows experts to monitor changes over time and identify lakes that may pose potential hazards. Central Asia’s unique terrain has shown that new lakes can form rapidly, making regular mapping essential, even during adverse weather conditions or cloud cover.
To achieve this, the project is developing the Glacial Lakes Inventory (GLI) toolbox, utilizing the python Tkinter library. This toolbox will monitor lake dynamics and provide statistics on surface area changes, the appearance of new lakes, and the disappearance of existing ones. The toolbox relies on Sentinel-2 Normalised Difference Water Index (NDWI) to detect glacial lakes, and an exciting deep learning-based methodology that fuses information from Sentinel-1 Synthetic Aperture Radar (SAR) and Sentinel-2 optical satellite data, allowing high-resolution monitoring year-round.
Mapping and monitoring glacial lakes using satellite sensors come with challenges. Many of these lakes are small and frozen for a significant part of the year, making detection difficult. Furthermore, clouds, cast shadows, lake turbidity, and atmospheric conditions pose additional hurdles. However, the GLOFCA project’s state-of-the-art algorithms and techniques show promising results, even detecting lakes as small as 0.01 km² in area.
The GLOFCA project employs an advanced deep learning algorithm called the Glacial Lakes Monitoring (GLM) network. This network uses a combination of satellite data from both Sentinel-1 SAR and Sentinel-2 sensors to create a comprehensive and data-driven approach to glacial lake monitoring. The deep learning process involves semantic segmentation, classifying each pixel as either lake or background, ultimately providing experts with valuable data on glacial lake extents.
The GLOFCA project not only focuses on lake mapping but also advances susceptibility assessment. By combining regional expertise and international best practices, the project presents a comprehensive checklist for experts to evaluate factors that contribute to GLOF susceptibility. This guidance aids in prioritizing glacial lakes for further monitoring and action.
As GLOFCA continues its mission, Central Asian communities can rest assured that innovative technologies and international collaborations are making great strides in understanding glacial lake hazards. By harnessing the power of satellite data and deep learning, the project is laying the groundwork for safer, more resilient communities in the face of potential glacial lake outburst floods.
For detailed information, please refer to Chapter 4 GLOFCA contribution to regional best practices of the comprehensive document titled: Glacial Lakes Outburst Flood: Best Practice Guidance