Day: October 5, 2026

Scientists Map Sky Rivers Over South America — And They Work Just Like Rivers on the Ground

“Flying rivers” (also called aerial rivers) above South America form organized drainage networks that structurally resemble terrestrial river systems, according to a 2026 study led by a National Taiwan University (NTU) interdisciplinary team and published in Nature Communications.

Not the short- lived atmospheric rivers you hear about in weather reports. The team defines aerial rivers as long- term, climatological pathways of moisture transport, the net, statistical flow that integrates all the daily breezes, low- level jets, and transient atmospheric rivers over years.

Aerial view of South America showcasing rivers of clouds, known as 'sky rivers', flowing over the landscape, with mountainous regions and lush greenery visible beneath.
AI generated by Meta

“Flying rivers” (also called aerial rivers) above South America form organized drainage networks that structurally resemble terrestrial river systems, according to a 2026 study led by a National Taiwan University (NTU) interdisciplinary team and published in Nature Communications.

These are long- term preferential pathways of atmospheric moisture transport, not short- lived atmospheric rivers. Moisture enters from the ocean, is often supplemented by forest transpiration (evapotranspiration), and is carried by winds across the continent, eventually falling as rain far downwind. Deforestation or land-use changes in upwind areas can therefore reduce rainfall hundreds or thousands of kilometers away, sometimes across national borders.

Key findings

Researchers analyzed long- term atmospheric moisture transport over South America and developed a mathematical method based on “moisture drainage curves”. These curves track how moisture contributions from upwind areas accumulate.

They identified objective turning points on the curves: beyond these points, additional land area contributes progressively less efficiently to rainfall in a target region. This defines critical upwind basins more rigorously than earlier fixed-percentage approaches. The location of these turning points varies substantially across the continent.

The continental aerial-river system shows four distinct hydrological regimes analogous to land-based river networks:

  • Headwater (G1): Entry points where ocean- enriched moisture first reaches the continent (not always strictly coastal; includes parts of the windward Amazon, Guiana Highlands, and areas of Chile/Patagonia).
  • Drainage (G2): Areas where atmospheric drainage can increase and some water vapor is lost inland (e.g., bands from Colombia across the mid- Amazon and northeastern Brazil).
  • Outfall (G3): Zones where aerial rivers slow abruptly, moisture content drops sharply, and regions rely more on continental evaporation; some lie upstream of major river basins and are particularly relevant for water management.
  • Plain (G4): Farther downwind regions.

Two major continental systems progress from upwind to downwind through the G1- G4 sequence (one tropical system originates from the northeast and extends toward Paraguay and southern Brazil).

Implications

The four types differ in their potential for atmospheric moisture management and susceptibility to ecological tipping points.

This supports more targeted forest conservation and challenges traditional water-resource governance, which often focuses only on surface watersheds or administrative boundaries.

A region’s water security can depend on distant upwind forests. Lead author Prof. Wei Weng (NTU Department of Geography) and co-authors from atmospheric sciences and civil engineering emphasize that the turning- point method makes the concept of aerial rivers more practical for management.

Illustration explaining aerial rivers, showing moisture advection, rainfall, runoff, and the concepts of downwind and upwind regions with farms and animals.
Conceptual diagram of aerial rivers. Credit: Nature Communications (2026). DOI: 10.1038/s41467-026-76303-y

What are “flying rivers”?

Not the short- lived atmospheric rivers you hear about in weather reports. The team defines aerial rivers as long- term, climatological pathways of moisture transport, the net, statistical flow that integrates all the daily breezes, low- level jets, and transient atmospheric rivers over years.

Like a surface river, they are:

  • recharged by evapotranspiration in upwind regions
  • routed by wind patterns
  • discharged by precipitation downwind

Over South America, the Amazon acts like a giant pump, forest transpiration loads the trade winds, they hit the Andes like a wall and get diverted south, raining out over central and southern Brazil, Paraguay, La Plata.

The breakthrough: they have drainage networks like land rivers

Researchers analyzed moisture drainage curves for 724 grid cells across the continent, the relationship between how big an upwind area you include and how much of your rain comes from it.

Using a mathematical method to find the turning point where contribution intensity drops sharply, they found four major drainage types, G1 to G4, that behave like a fluvial system from headwaters to plains: 

G1 – Continental headwater regions
Where aerial rivers first enter the continent from the ocean, e.g. windward Amazon, Guiana Highlands, Chile/Patagonia. The critical upwind basin dominates: 90- 95% of terrestrial- sourced rain comes from a relatively compact ∼930,000 km² area. Only ∼12% of total rain is terrestrial here, so management potential is low.

G2 – Atmospheric drainage expansion
A west- east band from Colombia to mid- Amazon and NE Brazil. Drainage expands, dominance falls. Critical basin ∼410,000 km² supplies 60- 75% of terrestrial rain, about 20- 25% of total rainfall. This is where aerial rivers carry the largest amount of terrestrial moisture across the continent.

G3 – Aerial river outfall
Southern Amazon, Brazilian Highlands, Pampas, parts of Andes. This is where rivers slow abruptly, resembling river outfall areas. Delivery intensity plummets compared to G2, critical basin shrinks to ∼260,000 km² and 40- 55% of terrestrial rain. But these areas are often upstream of major surface river basins, so conserving forests here amplifies both sky and land water.

G4 – Atmospheric plain
Chaco, central-west Brazil, La Plata Basin. Furthest downwind, least efficient delivery, lowest per-area contribution. Turning point at only 30-40%. Yet because G4 depends most heavily on continental moisture, that small critical basin (∼140,000 km²) actually contributes the largest share of total rainfall, 15- 20%, making it the most effective target for intervention. Santa Cruz de la Sierra is the textbook case.

Two giant systems

The four types of sequence G1 → G2 → G3 → G4 in two super- basins:

  1. Tropical system: from the northeast coast down to Paraguay and southern Brazil
  2. Temperate system: from Patagonia up to La Plata Basin

They align with the austral Hadley and Ferrel cells, with a sharp “fall” between G2 and G3 at 5°S- 13°S where moisture delivery intensity collapses. They ultimately converge in La Plata.

Why it matters for water management

Current water law stops at watershed and political borders. Aerial rivers don’t. The authors show that deforestation in upwind Peruvian Ucayali, for example, can cut downwind rainfall 5-13% and runoff 19- 50% via aerial rivers.

Their framework says: get the drainage curve for your city, identify its type from the map, find the turning point, and you know your critical upwind basin, the upwind forest or land that functions like a headwater to protect. For G3 and G4 regions, targeted conservation or reforestation can secure up to a fifth of total rainfall.

Map of South America illustrating continental aerial river drainage types, including categories such as Headwater, Drainage, Outfall, and Plain, defined by mathematical turning points of moisture drainage curves.
The hydrological structure of aerial rivers (flying rivers) can be classified into four types: headwater, drainage, outfall, and plain regions. Credit: National Taiwan University

Hydrological regimes and drainage systems of aerial rivers across South America

The team used 10 years of data (2006-2015), GPM rainfall, GLEAM evapotranspiration, ERA- Interim winds, in a moisture- tracking model (WAM- 2layers) to trace where the rain over 724 grid cells across South America actually came from.

Then they built moisture drainage curves, size of upwind area vs. % of terrestrial- sourced rain it provides and developed a math method to find the turning point where contribution intensity drops sharply. That’s how they define a critical upwind basin objectively, instead of using arbitrary 40% / 70% / 80% thresholds.

The 4 drainage types = a river system in the sky

They are geographic divisions shaped by long- term planetary transport, corresponding to fluvial positions: headwater, drainage, outfall, and plain. 

G1 – headwater regions: entry points of moisture from ocean. Windward Amazon, Guiana Highlands, much of Chile/Patagonia. Turning point 90- 95%, critical basin avg 930,000 km², but only 12% of rain is terrestrial, so management potential ∼11%.

G2 – atmospheric drainage: west- east band Colombia to mid- Amazon and NE Brazil. Expanding drainage. Turning point 60- 75%, critical basin ∼410,000 km², receives the largest terrestrial moisture load, 20- 25% of total rain manageable. 

G3 – outfall: southern Amazon, Brazilian Highlands, Pampas, Andes. “Occurs where aerial rivers slow abruptly, resembling river outfall areas”. Critical basin ∼260,000 km², 40- 55% contribution, but strategically in upstream areas of major surface basins, so interventions amplify through surface runoff.

G4 – atmospheric plain: Chaco, central- west Brazil, La Plata Basin. Furthest downwind, lowest delivery intensity across continent. Turning point 30- 40%, smallest critical basin ∼140,000 km², but highest dependence on continental moisture, so it actually contributes the largest share of total rainfall (15- 20%) for water- scarce cities like Santa Cruz de la Sierra.

From G1 to G4, basins decrease in both size and dominance, with a pronounced “fall” in delivery intensity at G2- G3 transition between 5°S and 13°S, interpreted as analogous to a river outfall.

Two continental systems

  • Tropical system: NE coast → Paraguay and southern Brazil (much larger)
  • Temperate system: SW origin Patagonia → La Plata Basin

They align with Hadley and Ferrel cells, plus low-level jets and terrain flows. They converge in La Plata.

Why “flying rivers” matter

Unlike surface rivers, aerial rivers are recharged by evapotranspiration. Deforestation in Peruvian Ucayali upwind can cut 5- 13% rainfall and 19- 50% runoff downwind via this link. 

The paper argues water management must move beyond catchment borders. Protecting a critical upwind basin in G3 and G4 is like protecting a headwater, and because in Amazon the downstream region lies upwind of aerial rivers, power relations flip: downstream forest users control upwind rain.

Journal information: Nature Communications  17, Article number: 9341 (2026)

DOI: 10.1038/s41467-026-76303-y

Provided: National Taiwan University

Authors: Wei Weng,
Ping Fu,
Ho Tin Hung,
Kai-Chih Tseng,
Li-Pen Wang,
Yun-Man Hsu,
Min-Hui Lo,
Luís Costa &
Vincent Dubreuil 

Abstract

Aerial rivers, long-term preferential pathways of atmospheric moisture flows, sustain ecosystems and regional hydrology. Maximizing their downwind discharge potential remains elusive without an objective way to identify critical upwind source regions. Here we analyze the drainage patterns of aerial rivers across South America and develop a mathematical method to identify turning points in moisture drainage curves, beyond which the intensity of upwind moisture contribution declines sharply. This provides a criterion for objectively delineating critical upwind basins and reveals pronounced continental variation in their spatial extent. These patterns reflect distinct regimes within the continental aerial river system, enabling its classification into four major drainage types: headwater, drainage, outfall, and plain areas. The four types differ in their potential for atmospheric moisture management and in their susceptibility to ecological tipping points, with implications for refining forest conservation priorities. Together, these findings support more effective land-water planning and moisture governance under a changing climate.

Wrong, Globe and Mail, Market Changes, Not Climate Changes, Are Hurting French Winemakers

A recent article from The Globe and Mail (G&M), “In French wine country, climate change and dwindling demand are a recipe for trouble,” claims that French wine production is suffering because of the effects of climate change and, at a distant second, market changes. This is exactly reversed.

An illustration depicting a debate on market changes versus climate changes in the wine industry, featuring graphs and characters discussing wine sales, unsold stock, tariffs, and market challenges amidst barrels of wine.
AI generated by ChatGPT

From ClimateRealism

By Linnea Lueken

A selection of Bordeaux wine bottles displayed in a wooden rack, featuring labels from various châteaux.

A recent article from The Globe and Mail (G&M), “In French wine country, climate change and dwindling demand are a recipe for trouble,” claims that French wine production is suffering because of the effects of climate change and, at a distant second, market changes. This is exactly reversed. While it is true that drought and heatwaves impact grape production and yield, wine itself has persisted through many changes in average global temperature over thousands of years. Market forces are actually the main driver force behind the French wine slump.

This summer’s heat and drought in Europe caused crop loss of grapes in wine regions of France, especially in Bordeaux, where wildfires also destroyed some vineyards. The G&M attributes this weather to climate change, although as Climate Realism contributor and meteorologist Anthony Watts has explained in “Wrong, Newsweek, Europe’s Heat Deaths Are a Policy Failure, Not a Climate Failure,” persistent and severe heatwaves are far from unprecedented in Europe. In addition, as discussed in a second Climate Realism post, recent wildfires are not historically unusual in France, with no data showing a sustained increase in acreage lost over time.

Wine is grown in regions that are hotter than Bordeaux, and colder. Temperature has an effect on the sugars and tannins in the fruit, this is true, but bad weather is something that winemakers have had to deal with as long as wine has been produced. Global climate change is not harming global grape production, according to United Nations data, and data does not indicate that the French climate is spiraling out of control. Only a long-term trend of persistently bad weather would suggest this year’s drought and heat was indicative of climate change, yet there is no such long-term trend. This year’s bad weather was just that, a year of bad weather, like French grape growers have experienced hundreds of times before over the centuries.

Just three years ago, French winemakers were upset because they were producing more wine than they could sell, and they lost hundreds of millions in destroying the excess. Obviously having more wine than patrons wish to imbibe is not a problem caused by climate change, and low production years aren’t either.

As G&M itself admits, France has steadily reduced the production of grapes since peaking in the 1970s. Production peaked in 1979, at a time when many scientists were worried about global cooling with some warning of a possible looming ice age.

Buried several paragraphs into the article, G&M makes the devastating admission that “consumption has fallen by 14 per cent since 2018, largely because of shifting social habits,” and “annual consumption has fallen from more than 100 litres per adult in 1960 to less than 20 litres today.”

These are very significant changes, changes that have been much more damaging to the wine industry overtime than occasional bad weather. Market forces working against them combined with a bad season, for smaller winemakers operating on slim margins, can be devastating.

Another economic issue only briefly mentioned by G&M is that “the cost of fuel and other farm supplies has doubled in the past couple of years.”

That problem would only be exacerbated if climate alarmists have their way regarding fossil fuel use. Food prices are impacted greatly by fuel costs, but especially wine. Wine requires energy for processing and storage. Packaging wine in bottles usually relies on fossil fuel powered machines. Glass and cork prices are impacted heavily by energy costs, so those supplies will become more expensive too.

At the consumer level, the costs add up even more because of shipping. Higher prices for wine that people are already drinking less of only exacerbates the issue, suppressing demand even further.

As one wise interviewee remarked: “You need to adapt to the climate. You need to adapt to the market.”

The wine market has significantly contracted over the past couple of decades. This is obviously bad for winemakers as well as wine drinkers, as lower production usually also means higher prices. Changing drinking habits among Americans and Europeans alike are driving the decline wine sales that G&M bemoans. By contrast, contrary to G&M’s claims, there is no evidence or data to suggest that climate change is to blame for that contraction, even if winemakers experience a poor producing year due to bad weather.

Antarctica’s Last Great Unknown Holds a 4-Metre Climate Threat

East Antarctica’s Wilkes Subglacial Basin (also called Antarctica’s “last great unknown”) holds enough ice to raise global sea levels by an estimated 3- 4 metres and is far less understood than other parts of the continent, prompting a major international call for urgent research.

There is no precise, well- constrained figure for total ice mass lost from the entire Wilkes Subglacial Basin (WSB) over the last ~50 years, because the region remains poorly observed.

Aerial view of a massive ice shelf with cracks, revealing dark water below, illuminated with golden reflections under a cloudy sky.
AI generated by Grok

East Antarctica’s Wilkes Subglacial Basin (also called Antarctica’s “last great unknown”) holds enough ice to raise global sea levels by an estimated 3- 4 metres and is far less understood than other parts of the continent, prompting a major international call for urgent research.

A review paper led by Professor Matt King (University of Tasmania and Australian Centre for Excellence in Antarctic Science) and co- authored by researchers from 14 countries, published in Nature Reviews Earth & Environment (around late September 2026), synthesizes geology, oceanography, climate science, satellite data, and ice- sheet modelling. It highlights major observational gaps that prevent confident predictions of how quickly the region could change or what climate thresholds might trigger large- scale ice retreat.

Key points include:

  • The basin is roughly 1,400 km long by 400 km wide (about half the size of New South Wales, Australia). Much of its ice rests on bedrock deeper than 2,000 m below sea level in places, and the bed deepens inland, a geometry that can enable self- sustaining “runaway” retreat once ocean- driven melting begins at the margins.
  • Access is extremely difficult due to persistent sea ice. No ship has come within ~150 km of the front of the key Cook Glacier; there is little or no in-situ seafloor mapping or ocean- condition data near the ice edge. Few (if any) living scientists have visited core parts of the region.
  • Geological evidence indicates substantial retreat during past warmer periods (e.g., the mid- Pliocene, when CO₂ levels were roughly comparable to recent values). Sparse modern observations show increasing thinning of the Cook Glacier since the 1990s, with current ice discharge around 40 billion tonnes per year.
  • East Antarctica was long viewed as more stable than West Antarctica, but this region shares similar vulnerabilities. Melting here could also disrupt broader ocean circulation and the global climate system, though major collapse is not expected this century under current projections.

The authors urge a coordinated international campaign (not a single- nation effort) to map the bed and seafloor, improve ocean and ice observations, and refine models. Better understanding would reduce uncertainty for sea- level planning, especially for coastal communities in Australia, the Pacific, and elsewhere. The basin lies largely within the Australian Antarctic Territory.

In short, the “climate threat” is the combination of the basin’s large ice volume, its potentially unstable bed geometry, evidence of past sensitivity to warming, and the current near- total lack of critical local data needed for reliable future projections.

_____________________________________________________________________________________

The runaway (or self- sustaining) ice retreat mechanism is primarily the Marine Ice Sheet Instability (MISI). It occurs in marine-based ice sheets, those grounded on bedrock that lies below sea level, when the bed slopes downward (deepens) toward the interior of the ice sheet. This is called a retrograde or reverse-sloping bed.

This geometry is a key feature of the Wilkes Subglacial Basin (and parts of West Antarctica, such as the Thwaites Glacier region). Once triggered, the process can continue even if the original climate forcing stabilizes or weakens.

Step- by- step mechanism.

Grounding line and ice flux basics
The grounding line is the transition zone where grounded ice (resting on the bed) becomes thin enough to float and form an ice shelf. Ice flux (discharge into the ocean) across the grounding line increases strongly with ice thickness there. Thicker ice flows outward faster.

Initial trigger (usually ocean- driven)
Warmer ocean water reaches the ice- shelf base or grounding zone and increases basal melting. This thins the ice shelf, reducing its “buttressing” effect (the resistance that slows the outflow of grounded ice). The ice upstream accelerates, thins, and the grounding line begins to retreat inland.

Positive feedback on a retrograde bed
On a retrograde slope, as the grounding line moves inland it encounters progressively deeper bedrock. This means the ice at the new grounding-line position is thicker.

  • Thicker ice → higher discharge flux → faster mass loss.
  • Greater mass loss → further thinning and additional grounding-line retreat.
    This is a self- reinforcing positive feedback loop. Theoretical analyses (e.g., Schoof 2007) show that, under these conditions, the grounding line is inherently unstable.

Self- sustaining (runaway) phase
Once the grounding line retreats past certain topographic “pinning points” (localized high points or ridges that temporarily stabilize it), the feedback can drive continued retreat for decades to millennia, even without further ocean warming. Model projections for the Wilkes Subglacial Basin indicate possible retreat rates of up to ~1 km per year in such phases, though exact timings remain highly uncertain due to sparse data.

Additional related processes

  • Ice- shelf weakening or collapse often initiates or accelerates the process by removing buttressing.
  • Buoyancy- driven effects can produce very rapid short- term pulses of grounding-line retreat (hundreds of metres per day in paleo- records) across especially flat beds, when the ice approaches full flotation.
  • A separate (and still debated) process, Marine Ice Cliff Instability (MICI), involves structural collapse of tall, exposed ice cliffs once ice shelves disappear; it can potentially amplify retreat but is not required for classic MISI.

Why it matters for Wilkes Subglacial Basin

Much of the basin’s bedrock lies well below sea level and deepens inland. Past warm periods (e.g., the mid- Pliocene) show evidence of substantial retreat, and current sparse observations already indicate some grounding- line retreat and ice- shelf thinning. Because of the retrograde geometry, crossing a threshold could commit the region to multi- metre sea- level contribution over longer timescales.

In short, the runaway character arises because the bed geometry turns a modest initial retreat into a self-amplifying cycle of thicker ice, faster discharge, and further retreat. Accurate predictions require better mapping of the bed, ocean conditions, and ice properties in these poorly observed regions.

_____________________________________________________________________________________

There is no precise, well- constrained figure for total ice mass lost from the entire Wilkes Subglacial Basin (WSB) over the last ~50 years, because the region remains poorly observed.

Available observations and estimates

  • Cook Glacier (the main outlet draining much of the WSB) currently discharges roughly 40- 41 Gt of ice per year. It has shown increasing thinning rates since the 1990s (one estimate: ~33 ± 12 cm/year over ~25 years) following major ice- shelf loss.
    • Cook West lost most of its ice shelf between ~1973 and 1989 (roughly 1,200 km² of ice shelf), accompanied by a roughly doubling of velocity (from ~700 m/year to ~1,400 m/year). Cook East has accelerated more modestly (~20% since the 1990s).
    • Rignot et al. (2019) report a small dynamic mass loss of ~2- 3 Gt/year averaged for the glaciers draining into the Cook Ice Shelf over 1979- 2017. One related analysis notes a combined increase in discharge of ~3 Gt/year for Cook East and West since 1979.
  • Broader Wilkes Land (which includes the WSB plus neighboring areas such as parts of the Aurora Subglacial Basin) has contributed significantly to East Antarctic mass loss. According to Rignot et al. (2019):
    • Wilkes Land lost mass at rates that helped drive East Antarctica’s overall contribution of 4.4 ± 0.9 mm sea-level equivalent from 1979- 2017.
    • In the most recent interval, they analyzed (2009- 2017), Wilkes Land mass loss averaged 51 ± 13 Gt/year.
  • Recent reviews of the WSB specifically note multidecadal grounding- line retreat, ice- shelf thinning, and some collapse of fringing ice shelves, but without clear large- scale upstream surface lowering of the grounded ice sheet. The basin as a whole still appears relatively close to balance compared with West Antarctica, with changes so far concentrated near the margins.

Context and limitations

Satellite records (altimetry, gravity, velocity) become denser after the 1990s- 2000s, so pre- 1990 estimates rely more heavily on sparse imagery and have larger uncertainties. Cumulative losses over a full 50- year period (roughly 1976- 2026) for the exact WSB catchment are therefore not reliably quantified in the literature and remain much smaller than the multi- metre sea- level potential stored in the basin.

In summary, measurable ice loss has occurred mainly through coastal outlets (especially Cook Glacier) via acceleration, thinning, and ice-shelf reduction since the 1970s- 1990s, on the order of a few Gt per year for the core WSB outlets, but the interior of the basin has not yet shown major widespread drawdown. This limited observational record is one reason scientists are calling for urgent targeted research.

_____________________________________________________________________________________

Wilkes Subglacial Basin (WSB, East Antarctica) vs. Thwaites Glacier (West Antarctica)

Both are marine- based systems grounded largely below sea level on beds that deepen inland (retrograde slopes). This geometry makes them susceptible to Marine Ice Sheet Instability (MISI), positive feedback in which grounding- line retreat can become self- sustaining. Beyond that shared vulnerability, they differ sharply in scale, current activity, data coverage, and near- term risk.

Size and Sea- Level Potential

FeatureWilkes Subglacial BasinThwaites Glacier
Approximate area~400 km wide × 1,400 km long inland~192,000 km² (roughly Florida-sized)
Sea- level equivalent3- 4 metres if fully lost~65 cm if fully lost
Role in larger systemMajor East Antarctic marine basinKeystone of the Amundsen Sea sector; its collapse could unlock additional West Antarctic ice (>3 m total potential in some scenarios)

WSB holds far more ice. Thwaites is smaller but strategically important because it props up neighboring ice.

Current Observed Changes and Mass Loss

  • Thwaites: Highly active and rapidly changing. It currently loses roughly 50 Gt of ice per year more than it gains from snowfall (net imbalance). Ice discharge has increased substantially since the 1990s; grounding- line retreat has accelerated; the glacier contributes ~4% of current global sea-level rise. Net loss since ~2000 exceeds 1,000 Gt in the broader basin. Ocean- driven basal melting is intense.
  • WSB (mainly via Cook Glacier): Much quieter so far. Cook Glacier discharges ~40 Gt/year. There has been multidecadal grounding-line retreat, ice- shelf thinning and or collapse (notably Cook West ice shelf largely lost in the 1970s- 1980s), and increasing thinning rates since the 1990s, but no clear large- scale upstream surface lowering of the grounded ice sheet. Dynamic losses remain small (a few Gt/year for the main outlets). The basin as a whole still appears relatively close to balance compared with Thwaites.

Thwaites is already in a phase of rapid mass loss; WSB shows early warning signs but has not yet entered a comparable acceleration.

Vulnerability and Timescales

  • Both can undergo runaway retreat once the grounding line passes key pinning points.
  • Thwaites is already responding strongly to warm Circumpolar Deep Water intrusion. Recent modelling suggests it is unlikely to fully collapse this century but could still contribute several centimetres of sea-level rise by 2100; longer- term (centuries) risks remain high, and some loss may continue even if ocean melting is reduced because of past forcing.
  • WSB has the same geometric vulnerability. Paleorecords show substantial retreat during past warmer periods (e.g., Pliocene, some Pleistocene interglacials). Models indicate that once past thresholds, retreat rates could reach ~1 km/year, with possible multi- millennial commitment from decisions made this century. However, the timing of any large-scale response remains highly uncertain due to data gaps.

Exploration and Knowledge Gaps

  • Thwaites: One of the most intensively studied glaciers on Earth. The multi- year International Thwaites Glacier Collaboration has produced extensive fieldwork, oceanography, radar, modelling, and satellite analysis.
  • WSB: Described as Antarctica’s “last great unknown”. No ship has approached within ~150 km of the Cook Glacier front; seafloor mapping and in- situ ocean data near the ice edge are essentially absent. Few living scientists have visited core areas. This is why a recent international review called for urgent coordinated research.

Summary

Thwaites is the active, well- observed “Doomsday Glacier” already contributing significantly to sea- level rise and serving as a potential trigger for wider West Antarctic losses.

The Wilkes Subglacial Basin is a much larger reservoir of ice with the same fundamental instability mechanism, but it remains far less dynamic and far less studied. Its potential impact is several times greater than Thwaites alone, yet the lack of basic observational data means scientists cannot yet confidently predict when (or how rapidly) it might cross a tipping point.

In short: Thwaites is the present- day concern; WSB is a high- stakes future risk that is still poorly constrained.

_____________________________________________________________________________________

Dynamic instability of the Wilkes Subglacial Basin Ice Sheet, East Antarctica

The Wilkes Subglacial Basin (WSB) is a marine- based sector of the East Antarctic Ice Sheet that holds enough ice to raise global mean sea level by 3- 4 metres. Its bed geometry (deeply below sea level and deepening inland) makes it potentially vulnerable to unstable, self- sustaining retreat (Marine Ice Sheet Instability).

The review synthesises:

  • Paleo evidence: Substantial retreat occurred during the Pliocene and other past warm periods, although exact temperature thresholds remain uncertain.
  • Modern observations: Multidecadal grounding- line retreat, plus thinning and collapse of key fringing ice shelves, but without clear upstream surface lowering of the grounded ice sheet.
  • Future projections: Models show the potential for rapid retreat (up to ~1 km per year) once the grounding line passes bedrock pinning points. Timings are highly uncertain, and decisions this century could commit the basin to multi-millennial change.

The authors conclude that the potential consequences (sea- level rise, disruption of ocean circulation, carbon cycling, and marine ecosystems) are severe enough to warrant an urgent, coordinated, multi-national and multi- year programme of fieldwork and modelling to close the large observational gaps.

This is the scientific paper that prompted the recent press coverage describing the WSB as “Antarctica’s last great unknown.”

The full text is behind a Nature paywall and institutional access, but the abstract and the associated University of Tasmania / ACEAS press materials provide the core findings that have been discussed in our prior exchanges.

Journal information: Nature Reviews Earth & Environment (published 30 September 2026)

DOI: 10.1038/s43017-026-00827-6

Provided: University of Tasmania

Authors: Matt A. King,
Nancy A. N. Bertler,
Sridhar Anandakrishnan,
Guilhem Barruol,
Michael J. Bentley,
Xavier Crosta,
Poul Christoffersen,
Florence Colleoni,
Laura De Santis,
Carlota Escutia,
Johan Etourneau,
Dimitris Evangelinos,
Fausto Ferraccioli,
Nicholas R. Golledge,
Matt Harris,
Petra Heil,
Nicole A. Hill,
Indi Hodgson-Johnston,
Coen Hofstede,
Xiaoxia Huang,
Richard Jones,
Serena Lagorio,
Won Sang Lee  (이원상),
Andrew Mackintosh,
Felicity S. McCormack,
Robert M. McKay,
Maxence Menthon,
Taryn Noble,
Marilyn N. Raphael,
Stephen Rintoul,
Martin Siegert,
Craig Stevens,
Craig Stewart,
Chris Stokes,
Yusuke Suganuma,
Fiorenza Torricella,
Yu Wang,
Trevor Williams &
Chen Zhao  (赵晨) 

Abstract

The Wilkes Subglacial Basin (WSB), a marine-based sector of the East Antarctic Ice Sheet, contains enough ice to increase the global-mean sea level by 3–4 m. The geometry of the WSB makes its ice sheet potentially vulnerable to unstable retreat, which could lead to rapidly increasing sea levels as well as meltwater production that could potentially disrupt ocean circulation, carbon cycling and marine ecosystems. This Review explores the climate vulnerability of the WSB in the context of future global warming. Paleo marine-sedimentary and model evidence suggests that substantial WSB ice-sheet retreat occurred during Pliocene and other past warm periods, albeit with uncertain temperature thresholds. Observations show unattributed multidecadal retreat of the grounding line and thinning and collapse of key fringing ice shelves but without clear upstream surface lowering. Projections of ice-sheet extent under future emissions scenarios indicate rapid retreat (up to ~1 km yr−1) beyond bedrock pinning points, but with uncertain timings and the possibility of commitment to multimillennial change during this century. Given the potentially severe consequences of WSB ice-sheet change, a multinational and multiyear fieldwork and modelling programme is essential to reduce projection uncertainty.

Rediscovering Thomas Jefferson’s revolutionary charge against abusive taxation

Thomas Jefferson would not be a fan of the energy tax plan his gubernatorial successor has coerced into place this year at the expense of ratepayers and without public input.

By contrast, Gov. Abigail Spanberger (D), who assumed office this past January, has implemented what some policy analysts aptly describe as a “backdoor carbon tax” that could add $1 billion in electricity costs over the next year. That process began this past summer when Spanberger had Virginia rejoin the Regional Greenhouse Gas Initiative, widely known as RGGI, a multistate climate change plan advertised as a “cap and trade” regulatory scheme.

A historical depiction of Thomas Jefferson looking shocked at a tax bill titled 'Virginia Electricity Bill', featuring details about energy charges and a significant increase in carbon allowance pricing.
AI generated by ChatGPT

From CFACT

By Kevin Mooney

Virginia ratepayers smacked with 153 percent increase in auction prices used to determine carbon tax.

Thomas Jefferson would not be a fan of the energy tax plan his gubernatorial successor has coerced into place this year at the expense of ratepayers and without public input.

The second governor of Virginia and third president of the United States argued that small farmers who lived off their own land should pay little or no federal taxes. Jefferson also opposed the Excise Whiskey Tax of 1791 because, in his view, the tax provided the federal government with too much power. Moreover, this founding father, along with his fellow Virginians, took on the Stamp Act of 1765 that led to the American Revolution.

By contrast, Gov. Abigail Spanberger (D), who assumed office this past January, has implemented what some policy analysts aptly describe as a “backdoor carbon tax” that could add $1 billion in electricity costs over the next year. That process began this past summer when Spanberger had Virginia rejoin the Regional Greenhouse Gas Initiative, widely known as RGGI, a multistate climate change plan advertised as a “cap and trade” regulatory scheme. Spanberger bypassed the traditional public comment period along the way toward cementing the climate initiative back into place. She also did not allow for a vote from the state’s Air Pollution Board, which would be the normal procedure.

The 153 Percent Increase in Auction Prices

Power generators in RGGI states must purchase an allowance for every ton of carbon dioxide their plants emit from using coal, oil, or natural gas. They then pass the cost of that allowance on to consumers in their electricity bills. That’s where the carbon tax comes into play.

RGGI holds quarterly auctions to determine what the allowance costs will be. The latest round came in with an enormous price hike, meaning Virginia ratepayers will need to hold onto their wallets. The September 9 auction came in at $37.65 per allowance in comparison to $14.88 in December 2023, which was the last auction Virginia participated in before Republican Gov. Glenn Youngkin exited RGGI. That’s a 153 percent increase.

A carbon tax is a tax on every human activity imaginable since it ties a fee to greenhouse gas emissions that flow out of fuel production and consumption. Even the most mundane activities would involve carbon taxes. Anyone going out to the store to buy groceries, deliver mail, or pick up the kids at school will be paying carbon taxes. This is what Spanberger has unleashed in a state where Jeffersonian antipathy toward intrusive government policies once held sway, and she’s not done yet.

In September, the governor signed an executive order that could pull hundreds of artificial intelligence data centers into RGGI. If Spanberger succeeds, these data centers could soon be competing with Dominion Energy Virginia and other retail electricity producers in RGGI’s quarterly auctions. This added layer of participation would only serve to put upward pressure on consumer electricity bills.

But what should most unsettle Virginia ratepayers is the fact that the specter of higher costs associated with RGGI was evident even before the September auction. Back in June, Federal Energy Regulatory Commission (FERC) Chairman Mark Cristie noted that a filing from Dominion with the Virginia State Corporation Commission for the purpose of recovering RGGI costs tells a sobering story. The typical residential consumer is projected to pay $13 more per month, which adds up to $156 per year. Ouch!! That’s bad enough, but keep in mind this filing was made before the higher prices kicked in with the September auction.

The new clearing price of $37.65 per ton of carbon emissions RGGI announced on September 11 is 8 percent higher than the price set in June and 69 percent higher than it was a year ago in September 2025.

Where is All This Money Going? 

Stephen Haner, an energy and environmental policy analyst for the Jefferson Forum, has some answers. Virginia’s new sale of carbon allowances is expected to pull in $259 million from electricity producers along with allowance market speculators. As a matter of law, more than half of this amount is to be used for energy conservation projects benefiting low-income households. Some of the money may go into flood mitigation efforts.

In June, the Virginia General Assembly voted to create a refund for residential consumers, small businesses, and churches paying the carbon tax. Only about 45 percent of the tax is expected to go back into rebates.

Dominion’s largest customers (think in terms of large industrial centers, retailers, and possibly data centers) will not receive any rebates. So, who are the real winners in the RGGI scheme? It would seem government officials looking to pull in more tax dollars, alongside a few well-connected special interests, come out on top.

Haner told Restoration News that a genuine “cap and trade” system is not at work in Virginia and utilities can have incentives to hold onto their allowances rather than selling them. He said,

RGGI is not about carbon reduction. It is about revenue enhancement. The allowances are granted to the states, not to the utilities. The states sell the allowances either to a utility or a secondary investor. There is a cap, and that cap does go down with every contract year. There is trading, and, in theory, if a utility had a surplus of allowances, it could sell them on the secondary market, but it makes more sense to hold them and buy fewer next time. If you bought it at $25, the one you don’t have to buy might be $35, so better to hold it. But in the true cap-and-trade model, the allowance would have been issued to the utility based on its initial need, and then the shrinking cap would be giving it fewer and fewer allowances to use, forcing it to go bid to buy from someone else.

Virginia ratepayers appear to be stuck with carbon taxes for the time being. But it’s possible that demand from data centers could push RGGI to the breaking point. Spanberger’s ploy to avoid public comments could also lead to legal challenges. Other RGGI states might also balk at the rising prices coming out of Virginia. There’s also the possibility that these other states could exit the program. In Delaware, there’s a bill to do exactly that. The other problem Spanberger and crew have is that the one large energy-producing state that had briefly been in RGGI has already pulled out. That would be Pennsylvania, which never participated in any auctions to begin with because of legal complications. Gov. Josh Shapiro (D) pulled Pennsylvania out of RGGI last November. Apparently, it had become clear to the governor that the coalition of industry and labor groups opposed to the plan had the upper hand in court.

Another cautionary tale for Virginia comes from the Institute for Energy Research (IER), which has released an updated version of its report “Blue States, High Rates.” IER points to RGGI participation as one of the key factors behind rising energy prices in certain states.

Virginia had a reprieve from carbon taxes under Youngkin, whose resistance to RGGI offers a sharp Jeffersonian contrast to Spanberger. The good news is that it may not be too late for the state to turn back. The Virginia governor cannot run for consecutive terms, and the carbon tax plan could become very unpopular very fast.

As president, Jefferson eliminated the excise tax on whiskey in 1802. By then, the country had ripened toward his message. Jefferson believed the phrase “to provide for the general welfare” was meant as a strict limit, so taxation was constrained to funding enumerated constitutional powers.

That’s not what the carbon tax is about since it provides government officials with wide latitude to put a price on productive activity. Virginia residents who find they can’t move more than two inches without paying the carbon tax will hopefully come to understand the unfortunate reality they now face and vote to restore their earlier, prouder Jeffersonian traditions.

This article originally appeared at Restoration News

Moon & Sun’s Tiny Tugs Can Trigger Slow Earthquakes Through Fault Resonance

Tidal stresses from the Moon and Sun, tiny gravitational forces that flex Earth’s crust, can trigger slow earthquakes through a resonance effect, according to recent modeling published in JGR Solid Earth (Zhou et al., 2026).

The spring- block model (also called the slider- block or spring- slider model) is a simple but powerful mechanical idealization used in earthquake science, especially for studying rate- and- state friction and the conditions under which faults stick, slip, or produce slow earthquakes.

Diagram illustrating how the gravitational pulls of the Moon and Sun can trigger slow earthquakes through fault resonance, featuring Earth and its crust with visual representations of tidal forces and fault mechanics.
AI generated by Grok

Tidal stresses from the Moon and Sun, tiny gravitational forces that flex Earth’s crust, can trigger slow earthquakes through a resonance effect, according to recent modeling published in JGR Solid Earth (Zhou et al., 2026).

Slow earthquakes (or slow- slip events) release energy over hours to months rather than seconds. They are usually imperceptible at the surface but occur on some plate-boundary faults and can precede or relate to larger conventional quakes.

Solid- Earth tides (distinct from ocean tides) produce stresses of only a few kilopascals, comparable to a gentle hand press, yet observations show they can modulate or trigger these slow events on certain faults (for example, in southwest Japan and the Cascadia subduction zone, where activity often peaks at ~12- and 24- hour tidal intervals).

How the mechanism works

Researchers used a simplified spring- block model of a single fault patch combined with rate- and- state friction laws (friction that depends on slip speed and the evolving “state” of the contact surface).

Key findings include:

  • Faults that are already sliding stably can be nudged out of that state by weak periodic tidal forcing.
  • When the period of the tidal stress matches the fault’s natural response timescale (the time needed for friction and other interface properties to adjust), resonance amplifies the effect, analogous to pushing a swing at the right rhythm or rubbing a wet finger around a wine-glass rim. Even small amplitudes can then produce accelerated slip.
  • Outcomes depend on both amplitude and period:
    • Low- amplitude forcing may leave the fault sliding quietly.
    • Above a threshold amplitude, and with a matching period, the fault can produce slow slip events (or, under some conditions, faster events).
  • Triggering can align with peak tidal stress, the peak rate of stress change, or more complex/chaotic patterns, depending on the specific fault properties and the phase of the tidal cycle. sciencealert.com

The model reproduces observed tidal modulation of slow earthquakes and low- frequency earthquakes. Because tidal forcing is predictable, matching observed seismic patterns to tidal cycles can help constrain fault properties (such as frictional strength and critical slip distance).

Caveats and context

The simulation treats an isolated fault patch and is therefore best suited to repeated local low- frequency events rather than large- scale tremor involving many patches.

Real sensitivity also depends on local geology (fluid pressure, rock type, etc.). Tidal stresses are far too weak to trigger large “fast” earthquakes on their own in most cases; any influence on major quakes remains subtle and is not useful for deterministic prediction.

Overall, the work supplies a physical framework explaining why such minute, continuous tidal stresses can still matter for slow slip, and it may ultimately improve understanding of stress buildup and release on subduction zones that host megathrust earthquakes. The Phys.org piece by Paul Arnold reports on this same body of research.

_____________________________________________________________________________________

The spring- block model (also called the slider- block or spring- slider model) is a simple but powerful mechanical idealization used in earthquake science, especially for studying rate- and- state friction and the conditions under which faults stick, slip, or produce slow earthquakes.

Basic setup

Imagine a block (representing a patch of a fault) sitting on a rough surface (the opposing side of the fault).

  • The block is attached to a spring.
  • The other end of the spring is pulled at a constant slow speed (this represents the long- term tectonic loading or plate motion).
  • Friction acts between the block and the surface, resisting motion.

As the far end of the spring is pulled, the spring stretches and stores elastic energy. When the force in the spring exceeds the frictional resistance, the block slips forward. After slipping, the spring relaxes, friction may heal, and the cycle can repeat.This captures the essential competition between:

  • Elastic loading (the spring), and
  • Frictional resistance on the fault surface.

Rate- and- state friction (the key ingredient in the paper)

In modern versions of the model (including the one used by Zhou et al. 2026), friction is not a fixed number. It depends on:

  • Slip velocity (the “rate” part): how fast the surfaces are sliding relative to each other.
  • State variable θ\theta: a measure of the “quality” of the contact (how well the microscopic asperities are interlocking or how much time has passed since the last slip). The state evolves with time and slip according to an evolution law (commonly the aging or slip law).

Two important frictional parameters appear:

  • (a): controls the direct (instantaneous) velocity dependence of friction.
  • (b): controls how much friction evolves (weakens or strengthens) with the state variable.

The difference a−ba – b (or the ratio a/ba/b) determines the overall behavior:

  • If a−b<0a – b < 0 (velocity- weakening), the fault can become unstable.
  • If a−b>0a – b > 0 (velocity- strengthening), sliding tends to be stable.

Critical stiffness kc

The spring has a stiffness (k) (force per unit stretch). There is a critical value kck_c that depends on the frictional parameters, normal stress, and the characteristic slip distance dcd_c:

kc∝(b−a)σdck_c \propto \frac{(b – a)\sigma}{d_c}

  • When k<kck < k_c the system is unstable → classic stick-slip (regular earthquakes).
  • When k>kck > k_c the system is stable → continuous stable sliding (creep).
  • When (k) is only slightly larger than kck_ck_c (the regime studied in the paper), the fault is stable but marginally so. Small external perturbations can push it into large slip-velocity amplifications via resonance.

How the paper uses the model

Zhou et al. (2026) focus exactly on this near-critical, velocity- weakening, stably sliding regime. They add periodic (tidal- like) stress perturbations, either changes in normal stress or shear stress, and show that:

  • When the period of the tidal forcing is close to the natural frictional response time of the fault, resonance occurs.
  • Even tiny perturbations (a few kPa) can then produce bursts of accelerated slip (slow earthquakes or even faster events).

The model is deliberately simplified (a single block = a single fault patch), which makes the mathematics tractable and allows clear identification of the controlling dimensionless parameters (normalized period and amplitude).

Real faults are more complex (many interacting patches, fluids, etc.), but the spring- block model isolates the essential frictional dynamics that can explain why some faults are so sensitive to the tiny stresses from the Moon and Sun.

In short: the spring- block model turns the complicated physics of a tectonic fault into a manageable mechanical system whose behavior is governed by the competition between spring stiffness and rate- and-state friction, exactly the tool needed to understand tidal triggering of slow earthquakes.

Earth’s crust really does experience continuous, tiny stress changes caused by the gravitational pull of the Moon and the Sun. These are called tidal stresses (or solid- Earth tides).

Why they exist

Just as the Moon and Sun raise ocean tides by pulling on the water, they also slightly deform the solid rock of the Earth itself. The solid Earth is elastic, so it bulges by a few tens of centimeters. That deformation produces stresses throughout the crust and upper mantle.

How large are they?

  • Typical amplitudes are only a few kilopascals (kPa), often 1-10 kPa, sometimes up to ~100 kPa when both solid- Earth tides and ocean loading are combined.
  • For comparison: atmospheric pressure is about 100 kPa, and the pressure under a gentle hand press is a few kPa. Lithostatic stresses at depth are measured in megapascals (MPa) to hundreds of MPa, so tidal stresses are extremely small relative to the background forces acting on faults.

Why they still matter

Even though the stresses are tiny, they are:

  • Continuous and perfectly periodic (mainly ~12- hour and ~24- hour cycles, plus longer fortnightly and monthly variations).
  • Applied globally and predictably.

On most faults these stresses are negligible. But on certain faults that are already close to failure and have the right frictional properties (especially the near-critically stable, velocity- weakening patches studied in the spring- block model), the periodic tidal stresses can be amplified through resonance.

When the tidal period matches the fault’s natural response timescale, even a few-kilopascal “nudge” can trigger measurable slip, the slow earthquakes (tremor, low- frequency earthquakes, slow- slip events) that are observed to correlate with the tides in places such as Cascadia and southwest Japan.

In short: the gravitational forces of the Moon and Sun really do continuously stress Earth’s crust by tiny amounts, and under the right conditions those tiny stresses can trigger slow earthquakes.

_____________________________________________________________________________________

Theoretical Constraints on Tidal Triggering of Slow Earthquakes

Even very small tidal stresses (a few kilopascals, comparable to the pressure of a gentle hand press) caused by the Moon and Sun can trigger slow earthquakes on certain faults through a resonance mechanism. When the period of the tidal stress matches the fault’s natural response timescale, the effect is amplified, much like pushing a swing at the right rhythm.

The authors use a classic spring -block model with rate- and- state friction focused on velocity-weakening faults that are stably sliding (stiffness slightly above the critical value).

They first apply idealized step and finite- duration (“boxcar”) normal- stress perturbations to demonstrate resonance- like amplification of slip velocity for specific durations. They then examine realistic harmonic (tidal- like) perturbations and identify the controlling dimensionless parameters through nondimensional analysis and numerical simulations.

Key controlling parameters

Triggering is governed mainly by two normalized quantities:

  • Normalized perturbation period (PTP_T)
  • Normalized perturbation amplitude (PσP_\sigma)
  • Increasing the normalized period shifts the timing of slip events from the peak of tidal stress toward the peak of the stress rate.
  • Increasing the normalized amplitude promotes a transition from slow to faster slip events.

Even small perturbations can produce both periodic and temporally complex slip events on otherwise stable sliding faults.

Implications for fault properties

The parameter space that allows triggering implies:

  • The instantaneous frictional strength parameter aσa\sigma is typically only tens to hundreds of kilopascals (much lower than typical lithostatic values).
  • The characteristic slip distance for frictional weakening (dcd_c) is likely on the order of micrometers.

Matching observed tidal correlations (e.g., the common 12- and 24- hour peaks of tremor in Cascadia and southwest Japan) with the model predictions can therefore help constrain these frictional properties of the plate interface.

Broader context

Slow earthquakes (including low-frequency earthquakes, tremor, and slow-slip events) are widely observed around the Pacific Rim, often updip or downdip of megathrust seismogenic zones.

Because they are sensitive to tiny stress changes, they provide useful information about stress accumulation and the potential size and extent of future large earthquakes.

The resonance framework offers a physically grounded explanation for why some faults respond strongly to tidal forcing and supplies a practical way to interpret those observations.

Plain- language takeaway: Tiny, continuous gravitational tugs from the Moon and Sun can set off slow, imperceptible earthquakes when their timing matches a fault’s natural “rhythm”. The study maps the conditions under which this happens and shows how those conditions can be used to learn about the mechanical properties of real faults.

Journal information: Journal of Geophysical Research: Solid Earth (2026)

Doi: 10.1029/2026jb033983 (also available as arXiv:2602.06703)

Authors: Yishuo Zhou, Ankit Gupta, Hideo Aochi, Alexandre Schubnel, Satoshi Ide, Pierpaolo Dubernet, Harsha S. Bhat

Abstract

Tidal stress is a globally acting perturbation driven primarily by the gravitational forces of the Moon and the Sun. Understanding how tidal stresses can trigger seismic events is essential for constraining tectonic environments that are sensitive to small stress perturbations. Here, employing a spring–block model with rate-and-state friction, we investigate tidal triggering on velocity-weakening stable sliding faults with stiffness slightly exceeding the critical stiffness. We first apply a step and a boxcar with finite duration normal stress perturbation to demonstrate a resonance-like amplification of slip velocity for specific boxcar durations. Next, we perform nondimensional analyses and numerical simulations with harmonic perturbations to identify the key parameters controlling tidal triggering and their admissible ranges. Triggered slip events are further characterized using physically observable quantities, including radiation efficiency and tidal phase. Our results show that even small stress perturbations can trigger periodic as well as temporally complex slip events on stable sliding faults. The triggering behavior is primarily controlled by the normalized perturbation period and the normalized perturbation amplitude. An increase in the normalized period shifts event timing from the peak of tidal stress toward the peak of stress rate, whereas increasing the normalized amplitude promotes a transition from slow to fast events. This framework helps explain the period-dependent sensitivity and the observed phase preference between tidal stress and maximum slip velocity. Comparison between observed and model-predicted tidal correlation patterns may therefore help constrain the instantaneous frictional strength of the interface, as well as the characteristic slip distance for frictional weakening.