The future of forecasting: Predicting avalanches to save lives

In the five months leading up to March 2022, there were nearly 450 recorded avalanches in three of the countries in which AKDN works. In the mountainous regions of Afghanistan, Pakistan and Tajikistan, these avalanches not only threaten lives, but also destroy homes and livelihoods, blocking access to remote villages and cutting off supplies.

Here the Aga Khan Agency for Habitat (AKAH) has established a community avalanche education programme, and comprehensive hazard assessment and monitoring systems. While 98 percent of the avalanches occurred in the areas where it works, the efforts of AKAH and the communities with which it works meant that these were only responsible for 13 percent of the fatalities. But avalanches are difficult to predict, especially in areas without Internet connections or financial resources to use new technologies. As rising temperatures and increased rain intensify the problem, better forecasting tools are needed.

Working within an EU-funded project, E3C, the Mountain Societies Research Institute (MRSI) at the University of Central Asia was tasked to find a way to map avalanches in Afghanistan. MSRI designed and created a Snow Avalanche Frequency Estimation (SAFE) model, which AKAH helped validate, using incident data and analytics from its avalanche preparedness programme in Afghanistan. SAFE uses the NASA/US Geological Survey series of Landsat satellites to show where nearly 810,000 avalanches have occurred over the last 32 years. An automated process detects avalanche deposits to indicate what valleys have historically been most impacted, providing more comprehensive longitudinal data than new technologies, at low cost. The most vulnerable villages and roads have been mapped, enabling stakeholders to prioritise mitigation measures such as trenches or snow-supporting structures called galleries.

Next, the team plans to run the model on the main roads of Gorno-Badakhshan Autonomous Oblast, Tajikistan. AKAH is also testing an automated weather monitoring system in Tajikistan that will provide real-time data on critical weather parameters in avalanche-prone terrain to further improve forecasting.

SAFE is available online on the Google Engine and has been designed to be simple and universal to use. The research paper, Snow Avalanche Frequency Estimation (SAFE): 32 years of monitoring remote avalanche depositional zones in high mountains of Afghanistan, has been published in The Cryosphere, a not-for-profit international scientific journal dedicated to the publication and discussion of research articles, short communications and review papers on all aspects of frozen water and ground on Earth and on other planetary bodies.