In late May 2021, 172 runners set out to tackle a 100km ultramarathon in north-west China. By the next day 21 of the runners had been killed by hypothermia after an unexpectedly intense storm brought freezing temperatures, strong winds and hail to an upland section of the course. Weather forecasts had predicted a cold front, but had not captured how extreme the conditions would be.
There are no meteorological stations in the area and survivor reports are subjective, but now a new hyper-local weather model – using topographic data at tens of metre resolution rather than kilometres – indicates that the intense wind and rain caused temperatures to drop by 6.7°C. Taking the blizzard-like conditions into account and the effect of wet clothes on body temperature, the study – which is published in the Journal of Geophysical Research: Atmospheres – estimates that runners would have experienced an apparent temperature of -10°C.
Storms like this are common at high altitudes on mountains like Everest, and while rarer at lower elevations, their sudden arrival makes them particularly dangerous. The researchers suggest hyper-local weather models can improve forecast accuracy for mountain events, where steep mountain slopes generate highly localised effects on wind, rain and temperature at a scale too small to be picked up by conventional weather forecasts.
