Sep 06, 2026
7 min read
🏔️ Where the world's glaciers are failing
Good Sunday evening and welcome to your climate data briefing (and happy Father's Day to all the dads and father figures out there 💝 ).
This briefing gives you:
- 1 x Analysis: Tracking global glacier failures
- 2 x Quantish reads: Cereal and El Niño
- 3 x Data updates: Demand, renewables, and fuels
Let's get on the ice. 🧊
Analysis: What 502 glacier failures tell us
With the catastrophe in Nepal sharpening public attention on glaciers, the timing of a new global dataset (published as a pre-print in August) couldn't be more relevant.
Researchers documented 502 glacier failures between 1900 and 2025. It covers 228 glaciers across Europe, Asia, North America, South America, and New Zealand.
The five charts below show what we found when we explored the data.
Two data notes
- Peer review
The paper is currently under review at Earth System Science Data. The underlying dataset is published on Zenodo, an open research repository. While the data is publicly available, it is still being scrutinised by experts and has not yet completed peer review.
- Observation bias
This dataset is not a complete record of every glacier failure since 1900; it's a record of the failures that were documented.
More recent decades show far more events because monitoring and reporting has improved. The data is also strongly biased towards places with better monitoring systems (more on that below).
That muddies any simple claim that glacier failures are becoming more common. But it still gives us many insights.
Chart 1: Total glacier failures
At first glance, the apparent global hotspot for glacier failures is Switzerland.🇨🇭
Nearly one-third of all events in the database (32%) occurred there. The Alps account for more than half.
That clearly says as much about monitoring as it does about glaciers. But it's a useful starting point.
The chart below shows the number of recorded failures by decade. (Note: the 2020s bar covers only the six years to 2025).
Chart 2: Large glacier failures
Europe dominates the full dataset, but not if we filter for large failures. The continent that has recorded the most large glacier failures is Asia.
The researchers define 'large' events as those involving at least one million cubic metres of material.
For scale, if all the steel and concrete in the Sydney Harbour Bridge were packed together, its rough volume would be about 100,000 cubic metres.* So, a large glacier collapse is roughly 10 x Sydney Harbour Bridges' worth of material crashing down. (The August 2026 glacier and bedrock collapse in Nepal is estimated to be tens of millions of cubic metres.)
The chart below shows large failures over time. Just like the overall failures, the recorded number grows, which is undoubtedly thanks to more monitoring. But we can see that large failures make up a greater share of recorded events in Asia and North America.
Chart 3: Glacier detachments
Among the large failures, one type stands out: glacier detachments. Those are the catastrophic events in which a large mass of glacier ice breaks away from its bed.
Of the 28 detachments in the database, 21 occurred in Asia. More than half of those were in Tibet, while another seven were in Tajikistan. The database records none in Europe.
Chart 4: What causes large glacial failures?
A valuable part of this dataset is its record of reported triggers and contributing conditions. The most common factors in large failures were:
- Meltwater input
- Weak bedrock (material beneath the glacier)
- Earthquakes
- Surge-like glacier movement
Many failures had more than one factor and there's a strong seasonal pattern, with failures peaking during the Northern Hemisphere summer.
The chart below shows how often each major factor was cited.
The raw numbers spike in the 2010s, once again reflecting more documented events.
Chart 5: How are contributing factors changing?
But while raw counts are difficult to compare over time, there is a way we can extract meaning from this — by looking at the share of large failures that cite each factor.
Percentage share of mentions removes some of the distortion caused by the growing number of records. It's not perfect because it still can't remove changes in scientific focus or measurement. But it does show patterns:
- Meltwater appears at a higher frequency in recent decades.
- Weak bedrock material beneath glaciers has also become more prominent.
Compare those with earthquake and surge-like activity, both of which jump around from decade to decade, and do not show the same general upward trend over the decades.
Despite the inevitable caveats and limitations of the data, the researchers compiling this data are doing hard yakka for the world's benefit. Already, some patterns are clear: Asia dominates the most destructive end of the glacier failure database. Meltwater is a factor in a growing share of failures.
Quantish reads
How the European summer's heat and drought impacted crops. 'Shrunken potatoes in the Netherlands, reduced carrot harvests in France, dried-up rice fields in Italy and scorched olive groves in parts of the Mediterranean region'. France will suffer heavy losses in its cereal production, such as wheat, barley and oats. The record heatwave in June contributed to more than €2 billion in grain production losses. (CarbonBrief)
The intensity of an El Niño is measured by how far temperatures in the tropical Pacific rise above normal. At two degrees Celsius above average, it becomes a Super El Niño. Scientists expect this year's event to be the strongest in at least 80 years of precise record keeping — far exceeding the previous extreme El Niños of 1997, 2015, and 2023. It will 'alter weather patterns across the planet'. (The New York Times)
Data updates
All updates show data for the most recent 28 days (4 weeks) available. Brackets show comparison with the same period last year.
⚡ Electricity demand
Underlying electricity demand in the NEM. Period: 28 days to 1 September.
• Total demand: 17,185.61 GWh (+2.2%)
• Average daily demand: 613.77 GWh (+2.2%)
• Maximum daily demand: 664.22 GWh (+1.3%)
• Minimum daily demand: 560.04 GWh (+1.4%)
💡 Demand was higher across the board. Even the lowest day used 1.4% more electricity than its equivalent in 2025.
🧮 Underlying demand = operational demand + rooftop solar. This is a new OnlyFacts dataset.
☀️ Renewables share
Proportion of electricity generated from renewables. Period: 28 days to 5 September, except NTESMO/NT to 30 August.
Main electricity grids
🥇 NEM: 45.0% (+3.8 percentage points)
🥈 SWIS: 33.9% (-3.3 points)
🥉 NTESMO: 28.5% (+3.6 points)
States and territories
🥇 Tas: 92.5% (+1.4 points)
🥈 SA: 79.5% (+3.8 points)
🥉 Vic: 41.7% (no change)
NSW: 39.7% (+4.5 points)
Qld: 38.3% (+7.4 points)
WA: 33.9% (-3.3 points)
NT: 28.5% (+3.6 points)
💡 Queensland recorded the largest rise in renewable share. Western Australia was the only jurisdiction to go backwards.
🔌 Fuels powering the grids
Contribution to electricity supply. Period: 28 days to 5 September, except NTESMO to 30 August.
NEM
☀️ Solar: 20.5% (+2.4 percentage points)
🌬️ Wind: 17.3% (+0.7 points)
🔋 Battery discharging: 2.5% (+1.7 points)
⚫ Black coal: 36.4% (-2.1 points)
🟤 Brown coal: 14.6% (+0.9 points)
🔥 Gas: 2.6% (-3.5 points)
💡 Solar supplied one-fifth of NEM electricity. Batteries nearly tripled their output. Gas and black goal generation dropped.
SWIS
🔥 Gas: 42.6% (+9.2 percentage points)
⚫ Coal: 19.3% (-7.9 points)
☀️ Solar: 19.1% (+0.2 points)
🌬️ Wind: 11.4% (-4.5 points)
🔋 Battery discharging: 6.2% (+2.9 points)
💡 Western Australia burned less coal but much more gas. Wind's share dropped. Batteries nearly doubled their output.
NTESMO
🔥 Gas: 71.5% (-3.6 percentage points)
☀️ Solar: 28.0% (+3.6 points)
💡 Solar supplied more than a quarter of NTESMO electricity.
Final Thought
'You have to defy some odds.'
— Jai Arrow
That's your climate data briefing for this week. Wishing you a great week ahead.
💛 Juliette and the OnlyFacts team
Need to make data easier to understand and harder to ignore?
Data sources:
Paper: Bashkova, E., Rupper, S. B., Shugar, D. H., and Forster, R. R.: Global database of glacier failures (1900–2025), Earth Syst. Sci. Data Discuss. [preprint], https://doi.org/10.5194/essd-2026-481, in review, 2026.
Dataset: Bashkova, Ekaterina, and Rupper, Summer, 2026. Global database of glacier break-off events (1900-2025). Zenodo. https://doi.org/10.5281/zenodo.19477908.
*Data notes:
- The researchers compiled the database from scientific papers, technical reports, hazard inventories and verified media reports.
- The data included one New Zealand entry, which had no event date. It was excluded from the charts.
- Sydney Harbour Bridge calculations:
