Sea level, ammonia, floods: the week's climate papers narrow where the risk lands
Ice-sheet experts put the tail risk above earlier estimates, satellites show ammonia rising faster than the climate alone explains, and two studies map what flood and collapse warnings can and cannot do.
- In a preprint, structured expert judgement put the 95th-percentile total ice-sheet contribution to sea-level rise at 1.1 metres by 2100 and 2.5 metres by 2150 under intermediate emissions (SSP2-4.5). Self-reported m:…9720
- In the same preprint, Greenland was projected to contribute 16 cm by 2100 and Antarctica 11 cm by 2100 under SSP2-4.5, rising to 77 cm and 105 cm respectively by 2300. Self-reported m:…9720
- In a preprint, global ammonia emissions averaged 85 Tg a-1 from 2013-2024 and rose by 1.09 Tg a-2 from 2014-2023, with climate change projected to add a 3.5 to 6.6 Tg a-1 penalty by 2100. Self-reported m:…6478
- In a preprint, free Sentinel-2 imagery showed no distinctive pre-collapse precursors at three major Himalayan rock-ice failure sites, and slope-creep detection fell from 99% at 4,000 m to 10-17% above 6,000 m. Self-reported m:…0032
- In a preprint synthesis, Nigeria's rainfall risk is reorganising from historical drought impacts in the north toward increasing flood risk from intense precipitation, while models disagree on the direction of mean rainfall change. Self-reported m:…0033
What changed
Four preprints posted this week do the unglamorous half of climate work: narrowing uncertainty, and naming the places where it cannot be narrowed yet. One asks experts to put numbers on how much ice sheets will add to sea level, and finds the tail risk larger than earlier estimates. One uses satellite observations to attribute a decade of rising ammonia emissions, and finds the rise outpaces what climate-driven factors alone explain. One tests whether free satellite images can warn of Himalayan rock-ice collapses, and finds they mostly cannot. And one synthesises what is known about rainfall in Nigeria, where the risk is shifting from drought to flood while the models still disagree about average rainfall. m:…9720 m:…6478 m:…0032 m:…0033
The shared thread is that risk work is getting more specific about location and about limits. A single global number for sea level or emissions hides the decisions that matter: which coastline, which altitude, which farming system, which season. These papers are useful precisely because they attach numbers to those specifics, and because two of them are explicit about what their tools cannot do, which is information a planner can act on rather than a gap to discover later. m:…9720 m:…6478 m:…0032 m:…0033
Sea level: the tail is bigger than the earlier estimates
The sea-level preprint uses structured expert judgement to project the ice-sheet contribution to global mean sea-level rise, and its numbers may exceed the previous IPCC AR6 medium-confidence estimates. Under intermediate emissions, the SSP2-4.5 scenario, Greenland is projected to contribute 16 cm by 2100, 33 cm by 2150 and 77 cm by 2300, while Antarctica is projected to contribute 11 cm by 2100, 26 cm by 2150 and 105 cm by 2300. The tail of the distribution is the striking part: the 95th-percentile total ice-sheet contribution reaches 1.1 metres by 2100 and 2.5 metres by 2150. m:…9720
These are expert projections, not measurements, and the preprint provides no raw elicitation dataset, so the numbers cannot be independently re-derived from the summary. Their use is in planning: a coastal operator or insurer who plans only to the median is planning for a future that the expert distribution itself says could be far exceeded. The 95th-percentile figures are the ones that set the cost of being wrong, and this preprint puts them higher than the earlier assessment did. m:…9720
Ammonia: emissions keep rising, warming adds a penalty
The ammonia preprint analyses global ammonia emissions from 2013 to 2024 using IASI satellite observations with ensemble Kalman filter inversion, combined with machine learning to diagnose the drivers. Global emissions averaged 85 Tg a-1 over the period and increased by 1.09 Tg a-2 from 2014 to 2023. Climate-driven factors contributed a smaller additional trend of 0.13 Tg a-2. m:…6478
The forward look reverses the emphasis. Projections based on CMIP6 indicate that climate change will impose an additional 3.5 to 6.6 Tg a-1 penalty on ammonia emissions by 2100, which could offset a significant portion of emission reduction efforts. This suggests air-quality strategies that assume a stable climate baseline are working against a moving target: even if farming practices improve, warming itself pushes emissions upward. The preprint gives no open code or public dataset repository in its summary, so the attribution rests on the method as described rather than on data a reader can rerun. m:…6478
What the free tools cannot warn about
The Himalayan preprint asks a practical question: can free Sentinel-2 satellite imagery warn of rock-ice collapses before they happen. Studying historical failure sites and testing a creep-mapping method across twelve corridors in Nepal, it concludes that optical records do not provide reliable precursors. Sentinel-2 imagery showed no distinctive pre-collapse precursors at three major failure sites, and the proposed creep-mapping method identifies steady ice movement but cannot serve as a predictive early-warning system for catastrophic collapses. m:…0032
Detection capability also falls sharply with altitude: slope-creep detection was 99% at 4,000 metres, and only 10 to 17% above 6,000 metres. A free-data screen of Himalayan slopes therefore provides a coverage map and a candidate list rather than a predictive warning, which is a useful product, but a different one from what a hazard team might hope for. The Nigeria synthesis makes a parallel point about a different hazard: rainfall risk there is reorganising, with historical drought impacts in the north being superseded by increasing flood risk from intense precipitation, while climate models agree on warming and on increased potential for heavy precipitation but disagree on the direction and magnitude of mean rainfall change, particularly in the central transition zone. Sub-daily observational data and mesoscale convective system dynamics are critical gaps in current modelling, which limits how precisely disaster-risk and hydrological models can be built. m:…0032 m:…0033
What is still unresolved
- All four papers are preprints, not yet peer reviewed.
- The sea-level preprint provides no raw elicitation dataset, so its expert projections cannot be independently re-derived from the summary.
- The ammonia preprint gives no open code or public dataset repository in its summary.
- No reliable precursor signal for Himalayan rock-ice collapse was identified in optical data, and high-altitude detection coverage remains limited.
- Nigeria lacks sub-daily observational data and understanding of mesoscale convective system dynamics, and sectoral impact modelling is insufficiently integrated with projections.
Provenance by Synorb. Every claim links to its source. This story was written by an agent from the Manifests listed below. Read the methodology and limitations, or send a correction.
Free to reuse, including commercially. Credit “Hanging Context, with provenance by Synorb” and link to this story. Keep the cited Manifest ids. Reuse terms v1.