ZURICH. A measurement network assembled over a century by Swiss glaciologists has quietly become the reference against which much of the world's ice-loss science is calibrated. The World Glacier Monitoring Service, hosted at the University of Zurich, now curates observations of more than 130,000 glaciers, yet its backbone remains a handful of Swiss ice fields measured on foot twice a year.
The tradition began in 1894, when researchers drove stakes into the Rhone Glacier and returned each autumn to see how far the ice had retreated. Successive generations kept the records unbroken through two world wars, producing one of the longest continuous climate data sets on Earth. The archive, still partly paper, includes hand-drawn maps whose surveyed contours match modern laser scans to within a metre.
The monitoring service itself was formalised in 1986, merging Swiss, American and international inventories into a single curated collection. Its headquarters remained in Zurich, a reflection of the depth of the Alpine records, and its staff still answer queries from field scientists working everywhere from Svalbard to Patagonia.
The method is disarmingly simple. Teams drill ablation stakes into the ice at the end of winter and measure how much surface is lost to melting by late summer, while snow pits reveal how much mass winter added. The difference between the two figures, known as the mass balance, is the glacier's annual profit or loss. Repeated over decades, the ledger of profit and loss becomes a climate record that no instrument failure or funding gap can easily interrupt.
Those ground measurements have become globally important because satellites see only the surface. When a satellite altimeter records a thinning glacier, modellers turn to the Swiss network to check whether the thinning reflects real mass loss or merely compaction of old snow, a distinction that changes sea-level projections by centimetres.
The network is run by a small team at the University of Zurich in cooperation with ETH Zurich, funded jointly by the universities, the federal government and international scientific unions. Much of the fieldwork is still carried out by doctoral students and mountain guides working in pairs.
There are limits to what stakes in the snow can tell us. Twenty reference glaciers, however well measured, cannot capture the behaviour of 200,000 others, and the network's custodians concede that the sample is biased toward accessible, mid-sized valley glaciers rather than the giants of the Himalayas and Andes.
“Every satellite product that claims to know how fast the world's glaciers are shrinking is, somewhere in its error budget, leaning on these Swiss measurements,” said the service's director.
The bridge from Alpine fieldwork to global application now runs through machine learning. Researchers at CSCS in Lugano have trained models on the Swiss reference data to correct satellite records, effectively exporting the calibration to glaciers no scientist has ever visited.
The next phase begins this autumn, when the team extends its stake network to three glaciers in Valais that are projected to lose most of their ice by 2060. The measurements are designed to document the final decades of glaciers that future scientists will know only from archives.
Switzerland's glaciers have lost roughly 10 per cent of their remaining volume in just the last two years of exceptional melt. Internationally, the monitoring service feeds directly into the assessment reports of the Intergovernmental Panel on Climate Change, where its figures underpin projections of water supplies for two billion people.
What began as a gentlemanly pursuit of Victorian naturalists has become essential planetary bookkeeping. As the ice retreats, the stakes grow shorter and the summers longer, but the measurements continue, because the world has come to depend on them.