A redesign of Switzerland’s public glacier-monitoring portal concealed small but real retreats at 37 of 108 measurement sites, according to an audit published on Tuesday. The review found that rounded figures and hidden uncertainty ranges made the sites appear unchanged on the portal’s default map, even though raw measurements showed ice-front retreats of between 1.2 and 4.8 metres.
The Swiss Alpine Change Observatory, which runs the service from Sion, said the redesign was intended to make dense scientific data easier to read on phones. But the audit by the fictional-sounding Office for Public Research Audits concluded that the new compact view gave readers “an incomplete and potentially misleading impression” of glacier movement. The portal receives about 420,000 visits a month.
What the audit found
Auditors examined 312 map panels and compared the default display with the underlying measurements and full research charts. In the redesign, figures were rounded to the nearest five metres, while uncertainty ranges — estimates of possible measurement error — appeared only after users opened a separate detail view. Thirty-seven sites were therefore labelled “no shift” despite recorded retreats below five metres.
A zero on the screen is not the same thing as a glacier standing still.
Roger Gautier-Clerc, audit lead
The audit did not find that the observatory changed its measurements or scientific conclusions. It did find that the display buried the 1990 reference line on mobile screens and gave the “no shift” label the same visual weight as clear retreat. Auditors tested the interface with 24 users; 15 initially described the affected glaciers as stable.
Glaciologists are divided over whether the small movements should count as meaningful trends. Some say annual changes of a few metres can sit within measurement uncertainty, particularly where snow and seasonal melt complicate the glacier’s edge. Others say hiding both the uncertainty and the direction of change invites the public to read a cautious measurement as reassurance.
A fight over the default
Gabriela Leite, a glaciologist at the observatory, said the team had approved rounding to reduce clutter but had not expected the label to be read as a verdict. “We should have kept the direction of travel visible, even where the exact number is uncertain,” she said. She called for the portal to restore the reference line and show the full range beside each figure.
Observatory director Uli Saefullah rejected accusations that the redesign softened the picture of Alpine ice loss. The agency spent CHF 1.9 million (about €2 million) on the portal overhaul, he said, and will publish a corrected version by 15 October. The next update will keep compact panels but replace “no shift” with a neutral range where results fall within measurement uncertainty.
The audit was commissioned in July after researchers at three Swiss universities complained that figures on the phone view did not match the charts they used in papers. The observatory said it had not received a formal complaint before then. Auditors also found that the design team tested screen readability but did not ask researchers to review how the labels could be interpreted.
The findings have angered scientists who fear the simplified map will travel farther than its caveats. Leandro Castro-Andrade, a researcher who uses the portal in public briefings, said people often see the map before they see a technical paper. “If we want trust, the first screen must show what the data actually says,” he said.
The observatory will keep the existing portal online while it makes the changes, and the audit office has asked it to publish a list of corrected panels. Switzerland’s glacier network tracks ice-front positions across the Alps, where retreat affects water planning and mountain hazards. For now, readers will need to open each detail view to see the full measurements.
