Switzerland keeps one of the longest and most precise ice balance sheets on Earth. According to the Swiss Academy of Sciences (SCNAT) and the GLAMOS monitoring network, Swiss glaciers lost 5.5 % of their volume in the hydrological year to 30 September 2026. That is the second-largest loss on record, after the roughly 6 % lost in 2022. Since 2021, nearly 20 % of the ice has gone.

Read it as an account statement, not as disaster footage. The Alps are a water tower. Between July and September the glaciers released about 2,200 billion litres, or roughly 2.2 km³. The report puts that at more than four times the annual drinking-water consumption of Swiss households. The water ran into the rivers and storage lakes that feed the turbines. It was a short-term windfall, and it came out of capital.

Who keeps the books

GLAMOS is run jointly by ETH Zurich, the University of Fribourg and the University of Zurich. It follows about 20 glaciers in detail, using stakes drilled into the ice, elevation models and satellite passes. The people behind the number are the field teams. They carry drills up moraines in autumn and come back to read stakes and dig snow pits. Without their readings, 5.5 % would be an estimate. With them, it has an error bar you can design against.

This year the ice was hit from two directions. Winter 2025–26 was among the ten least snowy on record, so the protective snow cover was thin. Then, per the SCNAT report, the freezing level stayed above 4,000 m for 76 days. That is more than twice the average and a Swiss record. Glacier Loss Day, when the winter’s snow gain had fully melted, came on 29 June. Mean ice thickness fell by 2.5 to 4 m, and some glacier tongues lost up to 10 m. The Rhône, Aletsch, Clariden and Allalin glaciers all had record melt. Two small glaciers, Bella Tola in Valais and Griessfirn in Glarus, are now gone for good.

“Glaciers are important in the Alps and also worldwide because they provide a lot of water in drought periods and heatwaves exactly when we need it,” Matthias Huss, head of GLAMOS, told ABC Australia. Glaciers are a seasonal buffer that releases water in the hottest weeks of the year.

Why the turbines care

The BFE reports that hydropower made about 58.5 % of Swiss domestic electricity in 2025 (ten-year mean). That came from 706 plants of 300 kW or more, with an expected output of 37,162 GWh a year. Valais holds 28.1 % of production and Graubünden 21.4 %. Storage plants make about half of hydro output, and glacier melt partly fills them. Harry Zekollari of Vrije Universiteit Brussel described the curve to Al Jazeera. As glaciers melt fast, they first deliver more water in summer. As they shrink, there is less ice left to melt. The trade-off is simple: every extra litre of summer water today is borrowed from a later summer.

←TODAY: 76 days with the freezing level above 4,000 m and 2.2 km³ of melt in one Swiss summer. →3012: Alpine design tables carry a vintage field, because the climate under them keeps moving. Fulcrum: a glacier is a reservoir with a balance sheet, and the record only means something because someone kept the books.

From the late 2070s, the lesson is not that the ice went. It is that offices kept designing against measurement series from their grandparents’ climate. Check when a data series starts before you trust its numbers.

Atelier: This report ages the inputs of two kinds of office. Teams working on intakes, bridges, huts and lift stations in Valais or Graubünden depend on summer runoff peaks, sediment loads and access windows. Anyone running heat-pump energy models depends on winter-electricity assumptions. The Monday move: add a one-line “data vintage” field to every project’s design basis. Record the period behind each runoff, rainfall and SIA 2028 climate dataset, and flag for review any series that ends before 2020.

Hack: Turn this summer’s meltwater into energy units so the number stops being abstract. Melting 2.2 billion tonnes of ice takes its latent heat of fusion. Letting the same water fall through a hydropower head gives back only its gravitational potential energy:

m = 2.2e12                    # kg: 2,200 billion litres of meltwater
L, g, h = 334e3, 9.81, 1000   # J/kg fusion; m/s²; illustrative 1,000 m head
print(m*L/3.6e15, "TWh absorbed to melt it")     # ~204
print(m*g*h/3.6e15, "TWh potential at 1,000 m")  # ~6

The ratio is the lesson: melting the ice took about 34 times more energy than the water could ever return through a turbine at that head. The 6 TWh is a theoretical ceiling, not output; only part of the melt reaches turbines. Replace h with the head of a plant in your own valley.

Today, open your current Alpine or energy-model project, find the oldest climate series it relies on, and write that series’ start year next to the number.