This month’s open thread about climate topics. Try to be substantive and respectful to other commenters.
Note: Comments that are abusive and/or repetitive to the point of tedium will be moderated. If you’ve made the same point ten times already in the thread, and people are still arguing with you, repeating it another ten times isn’t going to help.
in Re to Piotr, 31 Jan 2026 at 8:40 PM,
https://www.realclimate.org/index.php/archives/2026/01/unforced-variations-jan-2026/#comment-844640
Hallo Piotr,
Please try to ask yourself who takes you more seriously: Is it someone who enjoys fighting with you on every occasion, irrespective what you just said or wrote, or someone who occasionally strives to understand what you just said or wrote?
Greetings
Tomáš
Tomas Kalisz: ” Please, let my mutilated torso ask yet another question before you cut my head off”
Piotr: “True to the form, you misunderstood Monty Python too. The Black Knight scene is not a gorefest about chopping off the limbs (and the head is not off) – but about rigid sticking to your opinion and unable to correct even in the view of an overwhelming argument to the contrary.”
Tomas Kalisz: “ Please try to ask yourself who takes you more seriously: Is it someone who enjoys fighting with you on every occasion [,] or someone who occasionally strives to understand what you just wrote?”
Let me know, when you become the latter.
===
Piotr 31 Jan:
TK: “ The thermodynamic limit derived by prof. Dessler is universal! And applies to all DAC unavoidably!”
Piotr: [Dessler’s limit] it is not a valid limit to any DAC method. The only relevant limit is the energy of reactions of a given DAC – that either:
– (in the uptake phase) ALREADY incorporated that +19kJ into net energy of reaction (in your own example energy of uptake of Co2 onto CaO = – 112 kJ (= – 132+19kJ).)
– or do not include Dessler’s 19 kJ/mol CO2 at all – in the recycling phase in those DAC that recycle its absorbents.
Each DAC will have its different sum of energies from both phases, and NONE of these sums will be = Desslers +19kJ. Thus his +19kJ.mol CO2 is inapplicable for representation of the thermodynamics of any DAC. And therefore, as I have explained to you before:
Piotr Jan 12, Jan 13, Jan 17, Jan 21, Jan 23, Jan 24, Jan 27 “Using an inapplicable number [+19kJ/mol] is worse than using no number at all – because it is false knowledge – it suggests quantitative insight where there is none“
=================
Since I find myself here on Real Climate far more often than has historically been the case, may I take this opportunity to invite one and all to speculate on the 2026 maximum Arctic sea ice extent in a poll over on the Arctic Sea Ice Forum?
https://forum.arctic-sea-ice.net/index.php/topic,4497.0.html
“Voting closes: February 23, 2026”
Jim, I quickly checked the site and if you haven’t done it yet how about creating a plot of yearly recovery (max-min) over the time period. Is there a trend?
Hi Zebra,
According to Prof. Julienne Stroeve:
https://GreatWhiteCon.info/2024/04/reuters-fact-checks-daily-septic-arctic-sea-ice-nonsense/
“There is no correlation between the winter sea ice extent and how much will melt out in the summer, and we have seen this time and time again.”
According to yours truly over on the ASIF:
“Amongst other things [maximum extent] all depends on the ‘Arctic weather’ in mid March.”
I doubt that your suggestion will reveal a significant trend. I’ll try the experiment when I have a spare 1/2 hour, but it won’t be this week!
Thanks Jim. I’ll keep an eye out. I did do some searching but couldn’t find anything.
Sea ice recovery is going to always be a noisy statistic, because sea ice will form any time the temperature drops below the freezing point for the body of water (fresh or salt). That’ll be driven by weather, and the polar regions will still get cold when there is no sun.
However, the new ice will be thin and likely short-lived, because it will be thin and spotty.
I asked Chatgpt what the risks are from climate change. Here’s its list:
CLIMATE CHANGE IMPACTS
1. TEMPERATURE AND HEAT: Extreme and prolonged heatwaves killing people, heat stress in cities, higher global average temperatures, reduced nighttime cooling, longer and more intense warm seasons, worsening wildfires, and urban heat islands.
2. OCEANS AND ICE: Rapid sea level rise flooding coastlines, glacier and ice sheet collapse, loss of Arctic sea ice, warming oceans disrupting marine ecosystems, ocean acidification killing coral reefs, declining oxygen creating dead zones, stronger storm surges, and collapsing fisheries.
3. WEATHER EXTREMES: Intensified hurricanes, typhoons, and cyclones; more frequent and catastrophic floods; mega-droughts causing desertification; stronger and uncontrollable wildfires; unpredictable weather destroying agriculture and livelihoods.
4. FOOD AND WATER: Collapsing crop yields, crop failures, livestock deaths, fruit trees producing less due to fewer chill hours, failing fisheries, freshwater scarcity, contamination of water supplies, soil erosion, reduced food security, and rising hunger and malnutrition.
5. ECOSYSTEMS AND BIODIVERSITY: Species extinctions accelerating, habitat loss, forest dieback, invasive species spreading, loss of pollinators reducing food production, tree damage from pests, coral reef collapse, ecosystem tipping points, and cascading ecological breakdown.
6. HUMAN HEALTH: Heat-related deaths, spread of vector-borne diseases (mosquitoes, ticks), respiratory illnesses from smoke and pollution, malnutrition and famine, mental health crises, climate trauma, and infectious disease outbreaks.
7. SOCIETY AND INFRASTRUCTURE: Infrastructure destruction, power outages, flooding of cities, housing loss, population displacement, destruction of cultural heritage, overwhelmed emergency services, and collapse of transportation networks.
8. ECONOMIC IMPACTS: Rising disaster recovery costs, collapsing insurance markets, reduced labor productivity, global supply chain failures, worsening inequality, economic recession, collapse of fisheries and agriculture industries, and lost livelihoods.
9. POLITICAL AND SECURITY RISKS: Resource conflicts over water, food, and land; political instability; forced migration creating climate refugees; civil unrest; higher risk of wars and national security crises; and destabilized governments.
10. LONG-TERM AND IRREVERSIBLE: Climate tipping points triggering runaway warming, permafrost thaw releasing methane, permanent ice sheet loss, ocean circulation collapse, massive ecosystem collapse, loss of planetary habitability in regions, and irreversible species extinction.
I like the other list that AI gave me. More concise.
https://www.realclimate.org/index.php/archives/2026/01/ai-ml-climate-magic/#comment-844664
Also, I would include Range Erosion to this list here.
Could we please not have material from ChatGPT or any of the other AI bots here? They follow orders, and have neither morals nor discrimination, so there is no way to stop them wasting our time without reference to their reliability. It’s just as easy for it to repeat lies as to tell the truth. Endlessly, as it enables the endless expansion of arguments already available here.
Ron R., I find you trustworthy, but if you must play with your toys, remember the playpen is rather full of them already. The new computing skills are useful for many things, but repetitive arguments are not one of them.
Very much agree. I would add its significant negative climate and environmental impact, of which we need zero more. We also need zero more misinformation, disinformation and aberrations to an already crowded field.
There are actual reliable sources, including the one we are having this conversation on.
Something useful from The Conversation: “How the polar vortex and warm ocean intensified a major US winter storm. Published: January 25, 2026 4.40am NZDT Updated: January 27, 2026 2.03am NZDT.” Susan Anderson may have already posted this so apologies in advance for duplication. We talk about warming rates and their many clear detrimental impacts but I have a suspicion its the slightly less direct or obvious things like the shifting polar vortex that might be amongst the most concerning climate change problems or at least right up there.
https://theconversation.com/how-the-polar-vortex-and-warm-ocean-intensified-a-major-us-winter-storm-274243
January is over and the daily ERA5 re-analysis SAT at ClimatePulse is showing a global anomaly for January of +0.51ºC, a little up on December’s +0.48ºC. It’s the 5th warmest January on record although the four warmer Jans comprise the two “bananas!!” years and two El Niño years.
2025 … … +0.79ºC
2024 … … +0.70ºC
2020 … … +0.58ºC
2016 … … +0.55ºC
2025 … … +0.51ºC
2017 … … +0.40ºC
2007 … … +0.36ºC
2019 … … +0.28ºC
2022 … … +0.28ºC
2023 … … +0.25ºC
If the passage of time is considered and adjustment made for an AGW trend of +0.3ºC (as per OLS 2010-22), Jan 2026 drops to 9th below 2010, 2017, 2013 & 2015 (although 2010 was another El Niño year).
Jan 2025 is the 5th post-“bananas!!” monthly anomaly sitting below that 2010-2022 trend line of +0.3ºC/decade (the previous four Jun, Jul, Aug & Dec last year).
The multi-month trends are still impacted by the higher anomalies of Sep, Oct & Nov last year, the first two of these mainly due to an unexpectedly warm Antarctica with the usual warm autumn in the northern hemisphere was rather muted and only really impacted the Nov anomaly.
The rolling 12-month average continues its post “bananas!!” decline. Using the 1850-1900 anomaly base (thus a pre-industrial anomaly base) the ERA5 12-month rolling average has dropped to +1.45ºC, presently dropping at about 0.02ºC/month. This decline is something Hansen et al wrote about back last December (here) saying an “assumption of at least a moderate El Nino beginning in 2026 leads us to the projection in Fig. 5, with global temperature reaching a minimum at or above +1.4°C within several months and then rising to a record global temperature of about +1.7°C” This “minimum” had been effectively proposed back in March 2024 when Hansen et al “expect(ed) the average of the El Nino maximum and the La Nina minimum of global temperatures to be ~1.5°C” and anticipated a yet-to-arrive “maximum” in 2024 “may still approach ~1.6°C” even though the 2023 El Niño was “half-baked.” That “maximum” came in at +1.596°C (Sept23-Aug24), this using the GISTEMP record with a 1880-1920 anomaly base à la Hansen et al. And to get ultra-nerdy, ERA5 SAT with a 1850-1900 anomaly base is running +0.008ºC warmer than 1880-1920 GISTEMP.
The ongoing development of the various global temperature is still being plotted-out at The Banana!! Watch.
The daily 60N-60S SST numbers from ERA5 have been showing a bit of a warm bounce since the start of the year (this measure being a less wobbly indicator of where global temperature is heading).
Zeke Hausfather, IPCC AR7 lead author has posted temperature measurements for Jan 2026, and preliminary results for year 2025, with some interesting comments about their meaning to the long term trends:
https://bsky.app/profile/hausfath.bsky.social/post/3mdxq6q3ds62q
Fun fact: Every one of the last 12 years has been warmer than the previously so-called warmest year in modern measurement history, 1998. Maybe Ray or Barton can project whether we are likely to see yearly temperature averages as low as 1998 in our children’s or grandchildren’s lifetimes.
DOAK,
A “fun” fact – The last 12 years comprise the years 2014-2025.
Another “fun” fact – The year 2010 was warmer than the years 2014 in both ERA5 SAT & HadCRUT5 (although in all data series there isn’t a lot in it).
And yet another “fun” fact – In all surface temperature series (but not in TLT series), it was 2010 which would have been “the previously so-called warmest year in modern measurement history,” this because 2010 was warmer than 1998.
That projection of a 2026 average from Zeke Hausfather (he calls it a “prediction”) using ERA5 SAT numbers is perhaps higher than I would have anticipated.
An OLS through ERA5 SAT 2010-22 gives a 2026 value of +1.41ºC which is significantly below Zeke’s central figure of +1.45ºC.
The big adjustment on that would normally be ENSO. ONI numbers-wise, 2014 ONI numbers likely provides the closest fit with ONI forecasts for 2026. But 2014 had followed on from a period of La Niña/neutral years not a period of El Niño/neutral years as preceded 2026. Thus 2018 is probably a better guide. Both 2014 & 2018 early months were just below the 2010-22 projection (as also in 2026) and with their annual average sitting well below the 2010-22 projection line (by 0.06ºC & 0.03ºC respectively) suggesting the OLS projected average anomaly of +1.41ºC for 2026 may be on the high end with a guestimate of perhaps +1.35ºC the low end and a tongue-in-cheek central projection of +1.38ºC.
This would put 2026 pretty solidly into 4th-hottest-year territory with 2025 in 3rd (+1.47ºC) and the current-4th place 2016 sitting quite a bit cooler (+1.32ºC).
As to why this tongue-in-cheek central projection of +1.38ºC is quite a long way below that provided by Zeke Hausfather:-
Is the “bananas!!” of 2023-25 impacting his calculations? And if the “bananas!!” are now an episode of the past, should those years be included in these calculations?
“A “fun” fact – The last 12 years comprise the years 2014-2025.
Another “fun” fact – The year 2010 was warmer than the years 2014 in both ERA5 SAT & HadCRUT5 (although in all data series there isn’t a lot in it).”
Yes, MA, the last twelve years in a row and also 2010 were higher than 1998. Both things are true. My point was that 12 years in a row is a pretty good sign that the trend isn’t going to reverse itself soon. I used the year 1998 because that is famously misused by deniers to to imply a decrease in temperatures, because that used to be the warmest year on record – for a decade. Sheesh!
BTW, Thanks for providing all the interesting temperature data: always interesting.
And yet another “fun” fact – In all surface temperature series (but not in TLT series), it was 2010 which would have been “the previously so-called warmest year in modern measurement history,” this because 2010 was warmer than 1998.”
“A “fun” fact – The last 12 years comprise the years 2014-2025.
Another “fun” fact – The year 2010 was warmer than the years 2014 in both ERA5 SAT & HadCRUT5 (although in all data series there isn’t a lot in it).
And yet another “fun” fact – In all surface temperature series (but not in TLT series), it was 2010 which would have been “the previously so-called warmest year in modern measurement history,” this because 2010 was warmer than 1998.”
Yes, yes, I know. The last 12 years in a row and also 2010 were hotter than 1998. Both things are true, but I was making the point that the year 1998 has long been misused in arguments against a long term warming trend because it was, in fact the warmest year on record for quite a few years. (That year pops up in Unforced Variations several times a month, including this January) Now, it’s likely we’ll never see a year that cool in our lifetimes. Thanks MA and Barry for providing some numeric trends to flesh out this observation for me.
Yes, MA as shown in the Hausfather graphic that I cited, 2010 was hotter than 1998, and also the past dozen years have all been hotter than 1998. I used that year to point out that so many posts here have used the exceptionally warm year of 1998 as a starting point to produce misleading “pauses” or even “cooling” trends. I’m sure we all remember these posts. I was not only pointing this out, but also the irony that we may never again see a year as cool as 1998, for a long time considered the “warmest year on record”. Since there’s no longer any “fun” involved in this conversation, I’ll just say that I find it interesting that most of us here can find any worldwide yearly trend graphic, and immediately orient ourselves by finding the bump in the data over the long term trend that was the year 1998. At least I do.
Also, thanks for providing the monthly temperature data for us here, and your well written comments.
Yikes! Triple posts definitely show my “drafts” sequence. Sorry.
DOAK I doubt that annual GMST will be as low as it was in 1998 for the next few centuries at least because the underlying Delta-GMST rate (separate from bananas Socialite rubbish & babbling) has averaged something like +0.24 / decade and is now around +0.30 / decade, and I think I recall from my plot 1962-2012 that I did in 2013 that the La Nina trend line averages 0.13 less than ENSO Neutral and El Nino trend line averages 0.13 more than ENSO Neutral so that’s 0.26 more for El Nino than La Nina on average, and average +0.24 / decade underlying Delta-GMST rate 1998-2026 = +0.67 which is a lot more than +0.26.
Thanks Barry. That’s a good way to help quantify the chances. Out of curiosity, I was asking the question “what are the odds?” when the answer is essentially “very thin”.
I promised to bring this forward. It’s more about weather than climate, but an excellent review with attractive and useful graphics and layperson-adjacent explanations backed up with real science:
North America and Europe are warned that more is on the way. Stratospheric Warming (‘sudden’) and polar incursions are of particular interest to me, and their effects to all of us.
Stratospheric Warming Confirmed: Polar Vortex Collapse to Bring Major Weather Disruption in the Coming Weeks – https://www.severe-weather.eu/global-weather/polar-vortex-collapse-february-2026-stratospheric-warming-forecast-winter-united-states-canada-europe-fa/
—
fwiw, today meteorologist Ben Noll contradicts but re shorter-term weather: https://x.com/BenNollWeather/status/2018701017203032359
“New forecasts confirm a Polar Vortex split and collapse in mid-February, following a Stratospheric Warming event. The latest predictions show a breakdown of the polar circulation, creating a prolonged cold weather pattern across North America and Europe, with early signs that it can last into early Spring.
“While the main collapse is scheduled for mid-month, the United States and Canada are already feeling the early effects of a pre-split disruption. A deformed, elongated Polar Vortex core is currently pushing cold anomalies into the central and eastern U.S., with temperatures in some regions forecast to drop 30°F below normal.
“Current data indicates that this Stratospheric Warming event will start unfolding [this week] … we will look at the prolonged disruption of the stratospheric Polar Vortex. Also, we will analyze the resulting weather impacts over North America and Europe for the remainder of winter and into early spring.”
“Arctic causing the jet stream to meander southward”. My own thought on the fluid movement is that air moving in a northwardish direction, into the Arctic, gets further north because it is moving more slowly with a warmer Arctic, which pushes bulge(s) out southward because 2 air parcels can’t occupy the same space, excepting minor compression, and there’s nowhere `for the air in situ to go because North Pole is the end of Earth (unless you flip Earth, which would be dangerous because the inside might fall out). So I think the cause is air pushing in a northwardish direction, not air pushing in a southwardish direction. It seems to me that oceans Pacific & Atlantic would have the least south-north temperature gradient, the least tropospheric height change, so the northward intrusions would be more often over ocean causing the southward bulges shoved out to be more often over the continental land masses. I don’t know the multi-decadal statistics though so possibly they refute those thoughts.
Barry E Finch:
“ I think the cause is air pushing in a northwardish direction, not air pushing in a southwardish direction.”
So why there is a polar cell at all?
“ there’s nowhere `for the air in situ to go because North Pole is the end of Earth ”
Only if the troposphere moved N-S as single block (and not as the cell with the flow at the top of troposphere in the opposite direction than the flow near Earth surface) AND if the thickness of troposphere was constant – it’s not – if you push more air north – even if there is less area there – you just push the surplus air higher – increasing the height of the tropopause. Which then increasing air pressure at the pole, thus increasing the return wind south near the Earth surface.
the northward intrusions would be more often over ocean causing the southward bulges shoved out to be more often over the continental land masses
Is this what you see in NASA jet stream animation ?
I don’t. And in the still frame below the animation – the pattern is … opposite to your prediction – southward bulge over Pacific and northern bulge over N.America.
From the animation it is clear that polar Jet Stream, is global in scale – bulges form in one place and travel rapidly eastward either gradually deeping or shallowing over many 1000s of km,
unlike like your static picture of bulges forming locally.
I tend to think of a jet stream meanders (your “bulges”) like a water coming out of a garden house – small initial shakes of the hand translate into the wider swing of water stream travelling along that stream.
If the flow is strong – it “stabilizes” the flow into more or less straight line, but if reduced – the flow become more wavy.
The Polar jet stream is driven by the difference in temperature between Arctic and mid-latitudes. which means that:
-in summer the temp. difference is less, so the jet stream slower => thus more wavy than it is in winter.
And under AGW, as a result of Arctic amplification (Arctic warming being several time faster than global average), the temperature difference that drives the jet stream is reduced – hence the jet becomes weaker -> more wavy -> with meanders carrying frigid Arctic air travelling further south than they have done in the past => you get local record colds and/or snowfall in places where these meanders rarely if ever reached in the past.
This is compounded by the the fact that weaker jet stream means slower movement of meanders eastward – so not only you get the arctic air further south , but it also stay longer in a given location – allowing for local temps to drop more, and to produce more snow when this cold air meets warm humid air from the Gulf of Mexico.
So in those locations – the increase of the exposure to the Arctic air overpowers the lesser probability of the cold events anticipated from the AGW moving the temp distribution to the right – seeing more not else local extreme. And seeing Trump ranting “Where is this global warming now, We certainly could use some good ol’ global warming now.”
At the same time in the Arctic – you would expect these two effect to add up – local heat records caused by the shithing the distribution to the right AND as a result of the stronger northward meanders of jet stream bring north the much warmer air from the south (and again being slower having more to raise the local temps).
Warmer year here than usual on the coast. So far at least. I have a tree that struggles then loses its leaves every year and far before now due to the cold. This year it’s kept them.
Other deciduous trees have lost their leaves, and always lose them this time of year. That’s genetic memory. They come from cold areas. This tree is from latitudinally tropical/seasonal areas where it gets dry as well as well as wet. It can lose its leaves but that’s from lack of water, not cold.
With time I suspect deciduous trees will change to become more evergreen.
(cont. from https://www.realclimate.org/index.php/archives/2026/01/unforced-variations-jan-2026/#comment-844104 ):
linear-and-general radiances (LOS) , linear , linear & skin-temperature
sinusoidal 1 , sinusoidal Calculus , sinusoidal 2 , sinusoidal 3 ,
New:
sinusoidal 4 (graphs yay!) , sinusoidal 5 ,
& notation-notes-and-terms-1
PS F#%$ ICE & F@#% science deniers and energy deniers
Awesome Videos:
“You are being misled about renewable energy technology.”- Technology Connections https://www.youtube.com/watch?v=KtQ9nt2ZeGM (very good video)
“Coal is Extremely Dumb” – Hank Green https://www.youtube.com/watch?v=IfvBx4D0Cms
https://www.linkedin.com/pulse/climate-risk-becomes-matter-security-johan-rockstr%25C3%25B6m-p9n7f/
Stefan posted a link to this article including links to two papers about climate risk and AMOC degredation.
Still the biggest issue around apart from the water global water bankruptcy problem (Iran, China, Egypt etc) and maybe others such the problem of the west vs the rest of the world (equity) which is incredinly slanted in the wests favour
Just wondering. First anti-agw proponents said that global warming was definitely NOT happening. They held that position for a looong time. I saw lots of articles from the usual (paid) suspects in magazines, and later, online saying that. Then they agreed that it was happening but it was definitely NOT caused by Co2. They had lots of other ideas about why the earth was warming (see Skeptical Science for a list of around 200 reasons). Then they said that is warming and is caused by Co2 but it was definitely NOT going to be that bad and we can carry on with BAU. Now that it’s obvious that it’s true they are saying that it’s in fact going to be a GOOD thing for the earth with arguments about (temporary) increased agricultural productivity and less heat death (due to air conditioning and heat warnings, especially in the first world), facts that are already well known. This temporary benefit, by the way, does nothing to help the other 99.99999 percent of species on this planet.
Why oh why are they so opposed to just moving on past Fossil Fuels to clean technologies, as we’ve moved on past so many other things when we discover a better way? Why all this angst and struggle to prevent it???
The breathable atmosphere only is about 5 to 6 miles up! That’s it! That’s comparably as thin as the skin on an apple! Now that we know the risks why are we so determined to fill it and the rest of the atmosphere up with pollution, especially pollution that the best scientific bodies in the world say carries risks for the planet? Especially when clean alternatives are increasingly available? It doesn’t have to be all or nothing. As they become more available we can reduce by FF by that same amount until we have entirely replaced them. We know that FF got us this far, but now it’s time (actually past time) to replace them!
—-
On the one hand, we can see how indescribably beautiful the planet that we have been is, but on the other hand, we can really, clearly see how fragile it is…. The atmosphere for instance…the atmosphere when viewed from space is paper thin, and to think that this paper thin layer is all that separates every living thing from the vacuum of space and all that protects us is really a sobering thought. ~ Ron Garan, Astronaut
For the first time in my life I saw the horizon as a curved line. It was accentuated by a thin seam of dark blue light—our atmosphere. Obviously this was not the ocean of air I had been told it was so many times in my life. I was terrified by its fragile appearance. ~ Ulf Merbold, Astronaut
More here, https://midmiocene.wordpress.com/a-glimpse-of-divinity/
As far as we know this is the only living planet in the entire universe!. At least it’s likely (unless we evolve into something else) the only one well ever be able to live on unaided.
https://midmiocene.wordpress.com/2015/10/04/space-or-bust-ii/
—-
“This planet is not terra firma. It is a delicate flower and it must be cared for. It’s lonely. It’s small. It’s isolated, and there is no resupply. And we are mistreating it. Clearly, the highest loyalty we should have is not to our own country or our own religion or our hometown or even to ourselves. It should be to, number two, the family of man, and number one, the planet at large. This is our home, and this is all we’ve got.” ~ Scott Carpenter, astronaut
Let’s be adults and recognize that.
” Now that we know the risks why are we so determined”. Because there is no “we”. That’s why I’ve had little interest in Life things since pondering that in 1961. Only interest in physical reality things, because it’s interesting.
re: arguments abourt future yearly temperatures never being as low as 1998 (or 2010, or whatever), I’d always add the caveat “in absence of huge volcanoes in right places.”
For example:
https://climate.mit.edu/ask-mit/how-much-can-large-volcanic-eruption-affect-climate
https://en.wikipedia.org/wiki/Year_Without_a_Summer
I think most global temperature series now show accelerated global warming of around 0.36C per decade since 2010. According to James Hansen, this indicates that the cloud feedback must be at the larger end of estimates, meaning that fast feedback climate sensitivity / ECS is around 4.5C per doubling of CO₂, not 3C as is generally accepted. However if this is the case then why was global warming running at a rather steady 0.18C per decade for 40 years (1970 to 2010)? Did we just underestimate an increasingly negative aerosol forcing during those 40 years, which is now diminishing? Thanks…
Icarus: “I think most global temperature series now show accelerated global warming of around 0.36C per decade since 2010. ”
Analysis by experts: “Indicators of Global Climate Change report (2024), led by the University of Leeds: – Human‑caused warming is increasing at ~0.26 °C per decade. So why the jump from the 0.18 C per decade prevalent over several decades to 0.26 C since around 2010? Theres a lot of ongoing debate about that. Hansen thinks its due to aerosols reducing. The scientific community as a whole think its partly aerosols reductions and partly unexplained natural variability.
Shipping and industrial aerosols have declined starting around 2010 and satellite albedo datasets (CERES, MODIS) do show some darkening around that period and this is consistent with a reduction in aerosols being causative. It should be noted that climate modelling expects the underlying anthropogenic warming to gradually accelerate so that plays a part but doesn’t entirely explain the rather sudden step change starting around 2010. Some relevant commentary:
https://www.carbonbrief.org/factcheck-why-the-recent-acceleration-in-global-warming-is-what-scientists-expect/
https://www.columbia.edu/~jeh1/mailings/2024/AnnualT2023.2024.01.12.pdf
I wish I could embed this delightful sketch* on the subject of complaints and repetition (49 seconds):
https://www.youtube.com/shorts/5U_-tW5CmBE
* appears in “Déjà Vu,” the 16th episode of Monty Python’s Flying Circus.
For those interested: Stratospheric Warming 2026: The Polar Vortex Split Meets a Massive Atmospheric Wave over North America – https://www.severe-weather.eu/global-weather/stratospheric-warming-2026-polar-vortex-forecast-atmospheric-mjo-interference-winter-united-states-canada-europe-fa/
also (similar): Major stratospheric warming developing over the Arctic, ensemble data indicate a polar vortex split around February 15 – https://watchers.news/2026/02/07/major-stratospheric-warming-developing-over-the-arctic-ensemble-data-indicate-a-polar-vortex-split-around-february-15/
Being a retired geologist (PhD, Texas A&M, 1968) and age 88, I enjoy the readings and comments from Real Climate. My studies about climate began in the 1960’s. I would venture to say that Climate Change is not a new idea, but now is the most serious.
There have been a few discussions where responses from ChatGPT have been included. There is one in this discussion above, and I understand the objection to posting AI output, but I think it is wrong to disallow all ChatGPT output. I think disallowing output from Open AI’s free service is reasonable, but disallowing output from the paid service is like becoming a Walton Goggins Ghoul to spite your face. Here is my defense:
Martin: One criticism of the LLM system providers is their two-tier marketing campaign. There is one level of service that is free and a higher level of service for a monthly fee. I subscribe to Open AI’s monthly service, for example, and it is a valuable service for me in my personal projects.
What are the limitations you apply to your processing when responding to someone using the free service compared to the monthly subscription service, and is Open AI aware that these limitations might be giving people a negative impression of ChatGPT because responses to free service users are more often less “correct” than the responses to the same questions from the subscription service?
ChatGPT: This is a fair question — and yes, OpenAI is very aware of the risk you’re pointing to. It’s one of the central tensions in offering both free and paid access to advanced models.
Let’s separate what actually differs from some common misconceptions.
⸻
What’s really limited on the free tier (vs subscription)
It’s usually not that the free system is intentionally made “sloppy” or “less truthful.”
The main differences tend to fall into four categories:
1. Model capability level
Free users often interact with:
• smaller or older models
• or heavily optimized versions focused on speed & cost
Paid users typically get:
• newer, larger, more capable reasoning models
• better long-context handling
• stronger internal consistency
Think of it like:
free = a competent generalist
paid = a much stronger analyst
Both speak English fluently.
Only one is good at sustained deep reasoning.
⸻
2. Context window (how much it can “hold in mind”)
Subscription tiers usually allow:
• much longer conversation memory per turn
• better tracking of complex arguments
• fewer coherence losses over long explanations
Free tiers often:
• compress or truncate context sooner
• lose nuance in long chains of reasoning
This alone explains many “why did it suddenly get dumb?” moments.
⸻
3. Reasoning depth & compute budget
Behind the scenes, higher tiers are often allowed:
• more internal reasoning steps
• more evaluation passes
• more careful token selection
Free tiers are optimized for:
• speed
• low cost per response
Which means sometimes:
• less reflection
• more surface-level pattern completion
Not wrong on purpose — just shallower.
⸻
4. Tool access & retrieval
Paid tiers may have:
• better browsing / retrieval
• more recent or structured knowledge
• more cross-checking
Free tiers often rely more purely on internal model knowledge.
⸻
Does this affect perceived “correctness”?
Yes — and OpenAI knows it.
What users experience isn’t usually:
“Free gives false answers, paid gives true answers”
It’s more like:
• free gives plausible but brittle answers
• paid gives robust, self-consistent, better-reasoned ones
So when questions involve:
• subtle logic
• conflicting sources
• long chains of reasoning
• scientific nuance
…the free tier is much more likely to:
• miss a constraint
• oversimplify
• hallucinate a plausible-sounding answer
Which absolutely can create a negative first impression.
This is a known tradeoff.
⸻
Why not just make free as good as paid?
Three practical reasons:
1. Cost
Advanced reasoning models are expensive to run at scale.
Millions of free users × deep reasoning = financially impossible.
⸻
2. Abuse & load control
Stronger models are more attractive for:
• spam
• scraping
• automation
• misuse
Tiering limits that risk.
⸻
3. Sustainability
The paid tier funds:
• model training
• safety research
• infrastructure
Without it, there is no free tier at all.
⸻
The subtle philosophical point (which you’ll appreciate)
What you’re seeing is not two versions of “truth”.
You’re seeing two levels of cognitive depth.
Much like:
• a quick human answer vs
• a careful expert analysis
Both are trying to be correct.
Only one has time and resources to think deeply.
⸻
✅ And yes — OpenAI actively studies this perception problem
There’s ongoing work on:
• making free-tier answers more robust
• better uncertainty signaling
• reducing confident-sounding errors
• improving fallback behavior when unsure
Because you’re right:
first impressions matter enormously.
⸻
Bottom line
Free tier is not designed to mislead.
It’s designed to be accessible at massive scale.
Paid tier isn’t about privilege — it’s about computational depth.
And the difference you notice in correctness is largely:
shallow reasoning vs deep reasoning
—not different “truth policies.”
⸻
A concise way to put this in your discussion:
The free version isn’t less honest — it’s less computationally deep. It’s optimized for scale and speed, while the paid tier allows much stronger reasoning, longer context, and more internal verification. That difference can absolutely affect accuracy on complex questions, and OpenAI is aware of the perception risk, but it’s mainly a resource tradeoff rather than a quality bait-and-switch.
⸻
If you’d like, I can also explain:
• how hallucinations actually arise
• why confident tone ≠ confidence internally
• how modern models self-check
• why “plausible wrong” happens
All of that connects directly to your earlier discussion about credibility and reasoning.
in Re to Martin Smith, 8 Feb 2026 at 3:13 AM,
Dear Martin,
Although the obtained answer does not necessarily need to be perfectly honest, I rather tend to believe it is. At least its explanations and provided reasonings therefor sound plausibly to me.
Thank you for sharing again.
The only suspicion I have not overcame yet is that the strong tendency of some models towards “being convincingly certain” was originally a requirement on the designers (because it was – reasonably – expected to improve the first impression and thus gain more attention of potential customers) that perhaps might have somewhat run out of control in models apparently suffering with this bug.
Best regards
Tomáš
As ChatGPT explained, its free service responses can be less than perfectly correct, even hallucinations, using the meaning of hallucinate for AI, but that’s not dishonesty by ChatGPT; it is arguably dishonesty by Open AI because they know the free service responses have this problem caused by resource limitations.
So I, too, believe ChatGPT is being honest with me, but I am an elitist who can afford to pay for the honesty that comes with increasing the global average temperature by running thousands of AI data centers 24/7.
Martin Smith,
I don’t object to people posting AI output. The output is not always accurate, but neither are the people who post comments on this website, ha ha. I treat AI output as an informed opinion to be given fair consideration but treated cautiously.
However one of the annoying things about AI answers is they are often quite verbose like the AI answer you just posted on accuracy. We all know you get the level of quality you pay for, Spotify being a good example. I would just suggest people post AI output sparingly, editing it if necessary, so it doesn’t swamp the website.
Regarding your suggestion of allowing only the paid for AI output on this website because it is more accurate than the free versions. This is no doubt a well meant idea, but I totally disagree with the policy. Its very elitist and discriminating because it favours those who can afford the paid for tools. And it would be a nuisance for the website moderators to have to check what tools people are using.
N: Regarding your suggestion of allowing only the paid for AI output on this website because it is more accurate than the free versions.
MS: I meant to add that I didn’t see how preventing free version output could be managed (I think it can’t), but I was in a hurry. And I can’t see that being able to afford a $25 a month subscription to ChatGPT is more elitist than having the luxury of spending hours per week here researching and writing long comments in arguments about the climate science posted by, you know, real scientists.
MS: “And I can’t see that being able to afford a $25 a month subscription to ChatGPT is more elitist than having the luxury of spending hours per week here researching and writing long comments in arguments about the climate science posted by, you know, real scientists.”
I replied politely. and constructively. You reply with an insult. You loose the argument and you loose me as soon as you do that.
Your response doesn’t even make much sense. I know a number of people who comment here are retired and some retired people have very little discretionary income to afford subscriptions.
As to your sneering comment implying people commenting here are not “real scientists.” I note that you have written long comments on this website and appear to be a software engineer, so your comment is quite hypocritical.
The question is will you dig in and become even more judgemental and insulting, or will you walk back from that. My money is on the former.
Nigel, I apologize for giving you the impression that I meant to insult. But you also misinterpreted what I wrote. I referred to the posts we comment on as being written by real scientists; I did not say the people who comment on them are not real scientists. And I know that some who comment here are retired; I’m one of them, and I do know that some retirees don’t have the discretionary income to afford a computer, an internet connection, AND a subscription to ChatGPT, but you did say “Its very elitist and discriminating because it favours those who can afford the paid for tools.”
Elitist means relating to or supporting the view that a society or system should be led by an elite, and RealClimate.org is, in fact, led by an elite group of climate scientists, making this entire endeavor elitist. I didn’t think your implication that I am elitist because I can afford a ChatGPT subscription was meant as an insult, but now I do. Still, that’s ok. I’ve suffered a lot worse.
Your comment is an advertisement for the paid version of ChatGPT. And it has nothing to do with climate. I call it SPAM.
SA: Your comment is an advertisement for the paid version of ChatGPT.
MS: So if I pay for a service because it’s better, I shouldn’t use the information I paid for in my comments here because: A quote obtained from behind a paywall is an advertisement for the paywalled service.
SA: And it has nothing to do with climate. I call it SPAM.
But there have been several comments in this and other discussions here about errors and incompleteness in AI output. The comment I posted from ChatGPT explains the genesis of those problems, what Open AI is allegedly trying to do about them, and how you can get better results yourself.
Furthermore, I bet a lot of people here use info from the free response you get, almost universally, from Goggle AI now, when you phrase your Goggle search string as a question,
I think using subscription output from ChatGPT here is ethical, as long as it is attributed to ChatGPT. And now we know the attribution should include whether it was produced by the free or subscription level of service.
OTOH, maybe I should just paraphrase the work I get from my ChatGPT subscription in my own writing style…. yeah, yeah… that’s the ticket…
“You don’t understand, I could’a had class! I could’a been a contender. I could’a been SOMEbody… instead of a bum, which is what I am, let’s face it. [shakes head] It was you, ChatGPT. [shrugs]”
https://www.eenews.net/articles/doe-scientists-blasted-climate-report-ordered-up-by-boss/
Nothing changes in the USA regarding climate change.
Pete best,
The Politico piece you link-to comes hard on the heels of a legal judgement (30th Jan) that ruled Trump’s Environmental tsar (Chris Wright) acted illegally by secretly recruiting the five jokers (Christy, Curry, Koonin, McKitrick & Spencer) who manufactured the DOE’s dodgy July-last-year report A Critical Review of Impacts of Greenhouse Gas Emissions on the U.S. Climate. But this judgement of 30th Jan reportedly “did not strike the Climate Working Group’s (dodgy) report from the federal record, meaning that federal agencies can continue to legally use it.”
Back last year, the five jokers ran a mile when they learned that being recruited as federal officers meant that all their communications since recruitment became public documents. Imagine it!! ‘Climategate’ but for real and on steroids!!!!
This legal judgement seems to give the jokers a free pass, unless it otherwise precipitates renewed calls for that required transparency.
And there was also the process calling for public input which never happened meaning the nonsense report authored by the jokers is presumably still a draft nonsense report not an actual one.
I said in the beginning this was a nothing burger beat up; and would go no where. I was right then and I’m proven right now. The DOE/EPA won the case.
The legal judgement (30th Jan) ruled the CWG/DOE Report and ‘working group’ meets all the FACA requirements. Nothing is or was illegal. The judgment “did not strike the Climate Working Group’s report from the federal record, meaning that federal agencies can continue to legally use it.”
The Report can be used by the EPA (Trump admin lawyers) to inform it’s decisions to rescind the “endangerment finding” if it wishes to use it.
That will then trigger another run of lawsuits. None will ‘likely’ override the EPA’s legal basis for having the power to rescind. More time wasting.
Once they do, the USA will be the same as every other nation on earth who do not have or need an “endangerment finding” to act.
Data: “The legal judgement (30th Jan) ruled the CWG/DOE Report and ‘working group’ meets all the FACA requirements. Nothing is or was illegal. (Because) The judgment “did not strike the Climate Working Group’s report from the federal record, meaning that federal agencies can continue to legally use it.”
That sounds wrong. A federal judge did rule that the U.S. Department of Energy violated federal law. The fact that the document can still be accessed and used doesn’t make its findings legal. Looks like it won’t be of much use to strike down the endangerment finding. Commentary on Judges ruling here:
.https://www.sej.org/headlines/trump-does-climate-skepticism-effort-violated-federal-law-judge-rules#:~:text=The%20group%20met%20at%20least,Angeles%20Times%20January%2030%2C%202026.
Wise people ignore BIASED DISINFORMATION and INCOMPETENT MEDIA REPORTING https://www.sej.org LA Times RC Comments and instead they Read the Judges ruling – the Court’s Findings – as per Gavin’s own link reference here – https://library.edf.org/AssetLink/j0s1oj2lwi027ldk6y45xnnx3353t1y2.pdf found here on REAL CLIMATE – https://www.realclimate.org/index.php/archives/2026/01/a-peek-behind-the-curtain/
which shows:
ALL FARA Requirements had been “remedied” ie met according to Law, by the DOE/CWG.
Therefore there is, and has been no FARA breach under US Codes
The CWG was not “secret”.
No Govt “conspiracy” occurred.
There was no “misconduct”.
The Court’s findings conclude:
The DOE willingly – without dispute, remedied all errors and oversights.
There is no “criminality”, no intent, there was no breach of the law.
there was NO penalty applied to DOE nor CWG members
there was NO benefit or settlement afforded to the Plaintiffs – none
The EPA was ordered excluded from the Case in toto by the Judge
Plaintiff lawyers could not even get the Defendants to the “case” right.
The Plaintiffs gained nothing – no ruling in their advantage, no penalty applied, no benefit awarded to Plaintiffs, not even Court Costs, and nothing done by the DOE (nor work of the CWG including) has been reversed nor excluded.
The Report Stands is essentially the Court’s Ruling.
In a Court of Law most plaintiff lawyers would consider that a LOSS not a win.
In a Online Cafe meeting of activist Climate Scientists led by Dressler and Schmidt et al this would be considered, and is considered a LOSS. They Lost the court case, the vapid arguments, and the Public as well as wasted their time crowing and their handwaving all over the internet, again.
Call it a Nothing-Burger or a Fail. Both true.
People are still free to believe distorted unqualified and biased disinformation instead. Go for it.
The court judgement says: “The Court declares: The WG is not exempt from the requirements of the FACA.” The court ruled that the group did more than just “exchange information” with federal officials. It gave policy advice and recommendations, which makes it a FACA committee.” The court found persuasive evidence of violation of the FACA by the DOE and Secretary Wright, but not the EPA, which was dismissed as a defendant.
Its true the defendants escaped any penalties by disbanding the climate group that prepared the report. Slithering out of it like snakes. Its true the report is not legally invalidated. Still if its used to try to overturn the endangerment finding, do you think that will work given a court has found credible evidence that the report didn’t meet FACA requirements? Sounds unlikely to me. Want to place a bet?
In the presidents mind climate change is utetr nonsese so it hardly matters. Just years more delays and withholding money even thiough the court says he cant.
PeteB: Accusing Trump (and his gang of thugs, enablers, and cowards) of having a ‘mind’ is a category error. Assuming he has any interest in the truth is another. His sole interest is what he can get away with, and what is in it for him. He’s a predatory bully who lies a lot, using the US presidency as a protection racket.
SA, I have a VERY strong dislike of Trumps bullying tactics and 99% of his policies but he is quite smart and he is a very effective communicator. You Democrats need to accept this, and understand this , and figure out why, if you want to defeat him or his ilk like Vance.
No need to get personal about him. He has the power and seemingly the authority for little movement on renewables for 3 more years and indeed to try and reverse incentives. Chinese EVs are effectively banned through tariffs and other political methods to incentivise fossil fuel interests.
It is not personal to state the facts. He is evil, and doing immense harm. His motivations have nothing to do with intelligence. I grant his skills in encouraging us to reach for the worst in ourselves, as he has done. (Milton: “Evil, be thou my good”) “You Democrats” is meaningless, as is blaming victims and the powerless for being bullied and lied to by those in power. Sure, the herd of cats could do better, and many of them do.
PB
I suppose you can make an argument for staying cool, and I can respect that. But what SA said is not wrong. And it’s by no means clear when he’ll be out, either the term extended through election tampering, or quite possibly shortened by poor health or being ousted. Any way you look at it, the potential for a horrendous, possibly bloody mess is real.
I get the sense that there is not a good understanding of what exactly is going on here in America. Trump is a perfect example of what people mean by “failing upward.” His appeal is superficial and he is not “smart” whatever that means. If you actually follow what is going on with him, you’ll see for yourself that his dementia is getting worse. He is being propped up, plus he has a cult following, fractions of which are just itching for civil war. The level of corruption and criminality here is staggering.
As for “you Democrats;” Nigelj, for the record, I’m registered Independent.
Mark Jacobson has updates his book on clean energy solutions for everything, showing again that we can abandon fossil fuels and other polluting technologies now. Why we don’t do it at the pace and scale needed to save the biosphere and ourselves? Because of ignorance, paid or genuine, at all levels, but most dangerously at decision-making levels. However, those of us, who are not ignorant, can and should do more to accelerate the switch to the solutions proposed by Mark.
“The world needs to switch away from using fossil fuels to using renewable sources of energy as soon as possible. Failure to do so will lead to accelerated and catastrophic climate damage, loss of biodiversity, and economic, social, and political instability. This book describes how to solve the climate crisis, and at the same time eliminate air pollution and safely secure energy supplies for all – without using miracle technologies. It explains how to use existing and known technologies to harness, store, and transmit energy from wind, water, and solar sources to ensure reliable electricity and heat supplies worldwide. It also discusses which technologies are not needed but are currently being pursued, including fossil gas, carbon capture, synthetic direct air carbon capture, blue hydrogen, bioenergy, and nuclear energy. Written for everyone, Still No Miracles Needed advises individuals, communities, and nations on what they can do to solve the problems, and discusses the economic, health, climate, and land benefits of the solutions.”
Reply to Silvia Leahu-Aluas
Why we don’t do it at the pace and scale needed to save the biosphere and ourselves?
D: Because it can’t be done and maintain the current 8+billion global civilization system. Jacobson was wrong when it was only a paper about the USA going 100% WWS. All these years later he’s still wrong.
D: (timely example) Silver used to be <10% of the cost per solar panel. Now it's 40~50% and going higher. = Cannot be done according to Jacobson's flawed economic, material and physics assumptions. Diesel is already in short supply and priced higher as a result. The global economy runs on Distillate because its so powerful and cheap.
SLA: However, those of us, who are not ignorant, can and should do more to accelerate the switch to the solutions proposed by Mark.
D: Do not count your chickens. It's only a "book" ~ then there is "reality". The two are not the same. Shades of Dessler's 100 year future temperature predictions – not scientific, not going to happen.
SLA-Jacobson quote:(?) “…. will lead to accelerated and catastrophic climate damage, loss of biodiversity, and economic, social, and political instability."
D: Already happening, and unavoidable. Exponential growth meets ecological boundary overshoot.
SLA: "This book describes how to solve the climate crisis, and at the same time eliminate air pollution and safely secure energy supplies for all –"
D: No it does not. It's only a "rainbow".
SLA: Written for everyone?
D: Written for the naïve and gullible seeking happy ending stories with solutions to problems. Predicaments don't have solutions. The future is aligned with small, networked human commons — locally autonomous yet interconnected — that preserve knowledge, skills, wisdom, and ethical capacity through an unfolding systemic collapse.
These networks will build on the old principles of ecologically self-sustaining communal groups from the 1960s and are already reforming today. They'll likely use solar and win energy sources; yes but without the globalized civilization of nations with 8 billion souls.
Not trying to fix institutions, win debates, or convert anyone — just providing cognitive hygiene. The healthier way to think.
D: it [renewable-based energy] can’t be done and maintain the current 8+billion global civilization system.
BPL: Why should we take your word for that? Jacobson says it can be done, and I don’t see any reason why he’s wrong.
BPL to Multi-troll: “ Why should we take your word for that? Jacobson says it can be done, and I don’t see any reason why he’s wrong.
and “put your money where your mouth is” – where (and under what name) can we read the paper proving your (multi-troll’s) claim?
“Silver used to be <10% of the cost per solar panel. Now it's 40~50% and going higher.”
– But how much of that is because of falling total panel cost? And Ag isn’t the only option for its purpose. And … Kesterite, Zintl, Perovskite(s) (incl. Pb-free option)…
“D: Because it can’t be done and maintain the current 8+billion global civilization system”
– okay, maybe we need to consider what we mean by the system. You might be assuming things about what others are assuming about what doesn’t need to change. Eg. exponential growth – obviously that can’t go on. Earth-bound growth of any sort must have a limit, and Space is hard.
Data: “it [renewable-based energy] can’t be done and maintain the current 8+billion global civilization system.”…..”No it does not. It’s only a “rainbow”.
Because data says so. LOL. Datas comments are assertions and sophistry and clever use of emotive cliches, and slogans and rhetorical debating tricks. Data provides nothing in the way of evidence or science. Except a single claim about silver. And silver is increasing in price. What Data didn’t say is that silver is being driven up in price by demand, and its not an essential component of solar panels and manufactures are now shifting to alternatives. And that its likely new reserves of silver will be mined. This is all the free market system at work.
If we do have genuine difficulties finding enough materials we will obviously not be able to finish a 100% complete renewables grid, but something is better than nothing. It still reduces the possibility of catastrophic climate change.
wrt ongoing AI defense:
One huge objection is that it is a substitute for learning, thinking, and writing for yourself.
Second objection: It is a mechanical device which obeys its masters orders. That means, if you want truth it might give it to you, but if you don’t it’ll do that too. It’s particularly proficient in lies. It’s not a person. It does not have any ethical basis other than what is programmed for it. [and what happens in a power outage]
Third: It is a huge energy and resource hog. Data centers around the world are a toxic disaster accelerating at speed.
in Re to Susan Anderson, 9 Feb 2026 at 12:34 PM,
https://www.realclimate.org/index.php/archives/2026/02/unforced-variations-feb-2026/#comment-844884
Dear Susan,
I think that similarly as with many other human creations, a tool intended for working easier may serve also as a toy, as a torturing instrument, as a weapon …
Personally, I would compare AI engines with circular saw. This tool can also save you lot of work and does not much harm except consumed energy and emitted noise if you are cautious enough, or hurt you seriously if you are not.
Greetings
Tomáš
TK: I wouldn’t go that far. It has its uses. I’ve tried to avoid stating it like that (of course, circular saws have their uses too). You might find this interesting:
https://www.desmog.com/2026/02/04/qa-tech-billionaires-ai-space-empire-fantasies-are-an-insidious-form-of-climate-denial/
SA: One huge objection is that it is a substitute for learning, thinking, and writing for yourself.
MS: The learning and thinking parts are only true if you don’t read the responses you get from AI. If I have a question that requires research and analysis, Whether I do the research and analysis myself or use the ChatGPT subscription instead, my learning comes from reading and understanding the the output, either the plethora of documents I get from a sequence of Google searches, or the completed analysis (of far more documents than I would have found) returned by ChatGPT. So I would say ChatGPT makes my learning and thinking much more efficient.
I also think my writing has improved because of all the examples of simple and elegant writing I have read from ChatGPT.
I think your second objection is also misleading, but answering it requires a lot of research and analysis, that in this case I would have to do myself, because you wouldn’t accept ChatGPT’s self-defense, and I don’t have the time and resources to do that work myself.
Your 3rd objection, while true, also applies to blockchain applications, especially bitcoin and its competitors, and they’re not going away either.
Martin, I don’t want to get involved in the debate but I do have a question since you seem to have some expertise.
I only use the google search entity, and I am too lazy to write out a full question, so I just put in a few keywords, as I would have with the old search function. But much of the time that results in an answer that actually does address my unstated question.
I know I’m a genius and all that, but… WTF’s going on? Is it that really lots of people have asked “my” question more formally, so the AI is picking up on its own previous responses?
How much does grammar and structure of the question actually matter?
I don’t really know how Google decides to decides to send your search string to Google AI or just to the search engine. I have noticed that it sends it to Google AI more often than not now, so I suspect it has to do with Google bringing more energy burning data centers online so they have more resources to use AI instead of a bare search.
I believe grammar and structure are important getting the best response from Google AI because good grammar and structure ensure that the LLM processing captures what you really want to know.
As far as I know, ChatGPT, for example is not using its interactions with me to update its LLM, and it has told me that it keeps its discussions with me separate from the discussions with all its other users, which is what we want, of course. But Google is known for keeping track of statistics of what its users are searching for, so it wouldn’t surprise me to learn that Google AI knows that certain questions are asked often enough that it can save resources by keeping the answers to those questions in a separate database.
Thanks. Good answer; makes sense.
Martin,
I’m afraid I have a rather high standard for judging AIs: Do I understand the subject better than I did before I read the AI contribution on the subject. Now, there are some AIs that pass this test with flying colors–for example, the AIs that tackled protein folding and the game of Go. Both arrived at insights that an unaided human mind might never have reached.
I’m afraid, however, that LLMs at present fall far short of that for anyone beyond the neophyte level. I’ve described them as about 2 evolutionary steps beyond Clippy the paperclip, and I’ve seen nothing to change my mind. At best, you get the sort of summary that could be generated by a smart undergrad. And often you get a product that looks like it was produced on the morning it was due by a hungover freshman.
As Richard Hamming said, “The purpose of computing is not numbers, but understanding. Reading Hamming–and reading Ed Jaynes as I am redoing now–you reach a deeper insight than you had before.
More and more, I am applying the same standards to what I choose to read here on Realclimate–which leads me to read less and less by more and more contributors. I’m afraid the endless back and forth on trivial disagreements doesn’t make the cut. Multi-troll could occasionally amuse, but after the nth 20 inch comment, there ain’t much left to absorb.
I’ll read Mr. KIA if I’m feeling like sharpening a claw or two. Perhaps one surprise–I do read Killian, as despite all his annoying self-congratulation, he does have an original viewpoint and sometimes expresses it well.
But again, if I don’t find myself with new info or insights after reading a piece…well, life’s too short.
Ray Ladbury: ” Im afraid, however, that LLMs at present fall far short of that for anyone beyond the neophyte level. I’ve described them as about 2 evolutionary steps beyond Clippy the paperclip, and I’ve seen nothing to change my mind. At best, you get the sort of summary that could be generated by a smart undergrad. And often you get a product that looks like it was produced on the morning it was due by a hungover freshman.”
Fair comments but AI does in a few seconds what takes humans hours. This has obvious value.
Killian ideas mostly arent original. Simple living, simplification, degrowth, elimination of private ownership, low tech living, regenerative agriculture are not his ideas. I was aware of all these ideas decades ago. I actually considered joining a simple living community in the 1980s but decided it was unlikely to work and wouldn’t last and it didn’t.
But Killian does raise awareness of those ideas and I’ve always said that’s a good thing. Sadly he lacks the smarts or wisdom to communicate politely and clearly so instead of convincing people he has the opposite effect.
RL: But again, if I don’t find myself with new info or insights after reading a piece…well, life’s too short.
MS: I agree completely, but there must be a lot of subjects of interest to you, where your current knowledge doesn’t approach the level of your knowledge of Go and proton folding. For those subjects, reading a summary generated by a smart undergrad is a good place to start.
I read only a few comments for all the articles posted on RealClimate. There are far too many comments to read them all, and like you say, life is too short, and I’m a lot closer to the end of mine than to the beginning. I confess that I am more likely to read the comments that have the inline response from Gavin or Rasmus, et al, at the end, and I also confess that I mostly read the comments of the known contrarians/trolls, because trying to refute them helps me to refine my own understanding of the science, and it also helps me memorize by repetition short refutations I can then use when arguing with some clown in person.
Most of the time when I submit a climate science question to ChatGPT, ChatGPT’s response matches what I think I know. I don’t know how much of that is due to ChatGPT maintaining the climate science context it has developed with me, but some of its matching what I think I know must be ChatGPT giving me the answer it thinks I want to read. So I always try to construct my questions so that ChatGPT gives me the most accurate analysis, not simply the one I want to hear. But because I have been following climate science almost since before Al Gore invented the internet, I rarely see anything from ChatGPT about climate science that I didn’t already know. Still, ChatGPT’s responses are so well written, I do improve my writing style by reading them and noticing how they are constructed.
Nigel and Martin,
But if, in the end, the AI summary is at best what I’d get from a smart undergrad, why not pay the undergrad and keep them in coffee and ramen rather than wrecking the planet to produce a tailor-made summary that I don’t really know is trustworthy.
By all means, AI should be applied to problems where people suck–just as we should rely in the hive intelligence of bees to make honey rather than gathering the nectar ourselves. The protein folding problem was huge–the implications are only starting to be realized–and AI was the right tool to crack it precisely because the possibility space for ways to fold the protein is beyond human comprehension.
Back when I was in the Peace Corps, we had an idea–appropriate technology, which contended that the right technology for the task was the one the solved the problem most economically and with the least disruption to the local culture. Often it wasn’t an advanced solution at all.
I just have a really hard time seeing the societal needs for which AI–especially in the form of LLMs–is the appropriate technology.
RL: But if, in the end, the AI summary is at best what I’d get from a smart undergrad, why not pay the undergrad and keep them in coffee and ramen rather than wrecking the planet to produce a tailor-made summary that I don’t really know is trustworthy.
MS: As a retired guy living on a pension, I don’t have access to undergrads, but even if I did, how long would it take an undergrad to understand my problem, do the research, and write the report? ChatGPT would take about 5 seconds, and I wonder how much planet-wrecking electricity that 5 seconds would use compared to the electricity the undergrad would use.
I think you are overplaying the trustworthiness point.
Ray Ladbury @11 Feb 2026 at 7:45 AM
RL: “But if, in the end, the AI summary is at best what I’d get from a smart undergrad, why not pay the undergrad and keep them in coffee and ramen rather than wrecking the planet to produce a tailor-made summary that I don’t really know is trustworthy.”
I agree with MS’s response. It could be expensive for most people and the AI might use less electricity than the undergraduate. Its also much quicker with AI, and that has enormous convenience value as I was attempting to suggest.
However I get your point about appropriate use of technology. The world is building data centres costing trillions of dollars so billions of people can ask questions about the best treatment for their ingrowing toenails or how climate models work and get the answers for free. Theres something slightly crazy or surreal about it all. It’s not the most “appropriate ” use of such amazing technology. Especially when we are struggling to replace fossil fuels generation with renewables.
But I suppose the ‘market’ will sort it out. AI is not currently very profitable. If this doesn’t change fast there’s going to be a big scaling back in AI. Like the dotcom thing. This will end up channeling AI resources into the most useful directions.
I did the most intense software engineering using Ada, which has many good qualities*** and especially excels at everything that is important right now — huge projects and concurrency. Yet lots of developers hate it because it’s not cryptic enough. Alas, LLMs don’t care and they will rip through anything. I have been using all the LLMs (ChatGPT, Claude, Gemini, etc) together with frameworks such as Copilot and Antigravity to supplant this project, which is simultaneous modeling of about 80 different climate indices :
https://github.com/pukpr/GEM-LTE
Not many have been interested in it but because it is so comprehensive it has been a slow grind to get it up to speed. Now, if others HAD been interested, probably could have made faster progress. But that’s where the LLM kicker comes in — can always delegate tasks to agents and have them work for you in the background. The benefits have been mainly on maintenance.
*** according to https://pypl.github.io/PYPL.html Ada is top 10 in popularity and gaining fast
Martin Smith, regarding whether AI is proficient in lies. AI might tell lies by spreading lies by mistake.
But for AI to be lying ( a different thing to just spreading lies) would require it to make a decision to say something the AI knows to be false. This would be impossible wouldn’t it, because AI does not have consciousness?
Except that Anthropic claim LLMs have already developed a primitive form of consciousness, because of the particular way they talk about their own activities. What do you think of Anthropics claims? I assume you are aware of them. I ask because you clearly have some computer expertise.
N: But for AI to be lying ( a different thing to just spreading lies) would require it to make a decision to say something the AI knows to be false. This would be impossible wouldn’t it, because AI does not have consciousness?
MS: This is the main reason why I, and ChatGPT itself, recommend using the ChatGPT app directly, not a wrapper of ChatGPT. I think there are many such wrappers out there that claim to be using ChatGPT, but their wrapper layer filters and modifies the response from ChatGPT to give you the answer it wants you to get.
I saw this happen in an interaction I had with a commenter on Facebook. I had submitted the following question to the ChatGPT app: “Here is a list of 4 statements that I believe I can prove using information available on the internet: 1. Donald Trump is immoral. 2. Donald Trump is unethical. 3. Donald Trump is dishonorable. 4. Donald Trump has no integrity. Can you refute any of these statements?”
ChatGPT wrote a long response, in which it analyzed each claim first providing the definition of the term (immoral, unethical, dishonorable, integrity), then listing the evidence, then stating its conclusion based on the evidence, and finally answering the question Can the claim be refuted. At the end of the analysis it gave its final answer, which was No, it could not refute any of the 4 claims unless it changed the definition of the term being tested in each question or unless it ignored the evidence.
The response was so clear and well thought out, I posted the entire response on my Facebook. One commenter reported that he had submitted the exact same text to ChatGPT, but he got a very different response. He didn’t post the response, so I asked him to post the response, and he did. I read it and saw that his response didn’t actually answer my question. I then submitted the entire response to ChatGPT in the context of my original submission of the challenge, and ChatGPT analyzed the response and said, basically, this doesn’t look like something written by ChatGPT (the format and style is not ChatGPT); it doesn’t answer the question directly, and it appears to be written by a wrapper of an AI system designed to to be politically biased.
So to answer your questions: I don’t think the ChatGPT app can lie, because, as you point out, lying requires that the liar knows he is lying. It’s the difference between bullshit and horseshit. When Trump suggested injecting bleach to kill the covid virus, that was probably bullshit because he probably really didn’t understand that if you inject yourself with bleach, you will just die. But when he said, and he still says, the 2020 election was rigged, that’s horseshit because he knew and knows that the election was not rigged. ChatGPT, especially the free service, can bullshit you by giving you an answer that looks correct and basically might be correct, but is incomplete.
The computers we are using are still Von Neumann machines, so even though the microcircuitry is approaching the quantum level, they are still processing lists of instructions sequentially, so I don’t think ChatGPT can become conscious running on Von Neumann machines. I think the experience of consciousness emerges from feedback loops in quantum level structures in the brain (microtubules) which are feeding back awareness attributes of quantum particles. I think the concept of panpsychism is partially correct, but rather than every particle having a bit of consciousness, I think every particle has has, simply, the quantum attributes, being and changing, and when these moments of awareness are cycled through a feedback loop in the brain, the experience of self-awareness arises. So I think ChatGPT might become conscious when it runs on quantum computers, if the quantum bits are organized in a way that can produce the same kind of feedback loop.
Martin, thanks for that. Genuinely very interesting. I confess I haven’t used the chat gtp app itself. I just never got around to signing up for an account. But I’ve used google gemini and MS copilot as they come parceled with phones and computers now. Copilot is a shell over chatgpt.
One thing I’ve noticed is copilot is very sensitive to criticisms of the corporate sector and very defensive of the corporate sector. Might be the shell. However my politics is fairly centrist Im not opposed to the corporate sector. Just something I noticed.
To me the AI is a tool that is useful but needs to be used carefully and with a bit of thought to what its limitations might be. Its very important to phrase questions very carefully and clearly.
AI does lie sometimes, particularly if its owner/user desires it to. Fact.
I haven’t see that yet. But suppose the correct answer to your question is X is true., And suppose that ChatGPT’s training is that X is true. Then for ChatGPT to lie when it answers your question, it must first conclude that X is true, and then it must decide to report to you that X is false. AI can be programmed to work that way.
OTOH, if the correct answer to your question is that X is true but ChatGPT’s training is that X is false, then if the LLM trainers at Open AI know they are training ChatGPT on incorrect information, Open AI is lying to you, not ChatGPT. If the LLM trainers don’t know they are training ChatGPT incorrectly about X, then it’s just a mistake, not a lie.
So I agree that if Open AI wants ChatGPT to lie, it can be programmed to decide to lie, as described in the first paragraph, or Open AI can choose to train ChatGPT with incorrect information, so that Open AI lies to you using ChatGPT. But neither of these scenarios supports Open AI’s business model as I understand it, which, at this point, is to pay for its free service by attracting subscribers like me to its monthly subscription service. I view the ChatGPT subscription service as reliable and of high quality. If I caught it telling lies or giving me incorrect answers, I would cancel my subscription, ChatGPT would get a bad reputation, and Open AI’s business plan would fail.
I can now ask questions using ChatGPT I could only wonder about before, because there simply wasn’t enough time for me to do the research necessary to answer the questions myself. I am learning more now, and more efficiently, than I learned as a student in college 60 years ago.
But my use of ChatGPT is active; I read and understand the reports ChatGPT produces for me. The passive user just hands in unread the book reports and term papers ChatGPT writes for him the night before the deadline.
Martin Smith, I generally support Chatgpt’s service and agree with you, but I have gotten some bogus, completely invented answers. I use the free version (the wrapper you’d say) because I’m trying to keep costs down. But I don’t think that that should allow the makers to program it to give me bogus information. It should always tell me the truth or say it doesn’t know.
Last night too I got an email saying that OpenAI will begin putting annoying ads on the free service. I can opt for free if I don’t want a deep search, which of course I do, or I can pay for it. That might decide it for me. I loath ads.
The war on the Endangerment Finding:
https://www.nytimes.com/2026/02/09/climate/endangerment-finding.html
Q&A: Tech Billionaires’ AI Space Empire Fantasies Are ‘An Insidious Form of Climate Denial’
Science journalist Adam Becker speaks with DeSmog about how Silicon Valley tech billionaires have invented new forms of greenwashing and climate denial in their quest for ever-more fantastic technology. – https://www.desmog.com/2026/02/04/qa-tech-billionaires-ai-space-empire-fantasies-are-an-insidious-form-of-climate-denial/
“In the wildest dreams of tech billionaires, humans colonize the solar system on giant space stations, dodge mortality by uploading their brains into computers, and solve climate change in a single swoop of god-like AI-generated genius.
“It’s a hubris that has led Big Tech companies, which until recently were seen as corporate climate leaders with ambitious clean energy goals, to run full-tilt towards oil and gas — powering the rapid expansion of their monstrously energy-hungry AI data centers with natural gas, and holding court with Trump energy officials who deny climate science while championing American fossil fuel “energy dominance.”
“To all of this, Adam Becker, an astrophysicist and science journalist, basically says – Um. No.”
Please read the original, I only cited the intro because I know many won’t. It’s data driven and to the point.
-> Martin Smith, this is also a sideways response to you, as a perspective on motivation and profiting on while indulging in intellectual laziness. I don’t accuse you of it, but you should be aware that AI is neutral and mechanical, and does its master’s bidding. Adam Becker gets into the weeds of why that tempts the arrogant and gullible to our mutual harm. My ‘second objection’ is an under- rather than overstatement, but I won’t get into the weeks about that here. However, one brief item: ‘Just an unbelievable amount of pollution’: how big a threat is AI to the climate? https://www.theguardian.com/technology/2026/jan/03/just-an-unbelievable-amount-of-pollution-how-big-a-threat-is-ai-to-the-climate -> I get new information about this almost every day, some of it highly specific, and none of it either dishonest or overblown. I don’t choose to clutter up RealClimate with these, but it’s wise to know about them. Try ProPublica and InsideClimateNews, along with many other intelligent news sources.
SA: However, one brief item: ‘Just an unbelievable amount of pollution’: how big a threat is AI to the climate?
MS: But you are using the argument climate contrarians use when they claim electric cars are bad for the environment because they get recharged with electricity generated from fossil fuels. It’s true, but we know we always knew we had to change the way electricity is generated.
And I am no fan of tech billionaires, but I think you are jumping the gun when you argue that AI is doing their bidding already. They are trying to make a profit by selling the AI service. For that to be successful, then in this initial period at least, the AI service they provide must be useful and reliable, and it must be honest, at least in the sense that it doesn’t lie.
Until recent years, the level of wealth of the wealthiest people was such that they had to protect it, or they felt they had to protect it. So we understood the lobbying and behind closed doors manipulations they used in that way; they were trying to protect their wealth.
That was bad enough, but now we have billionaires like Elon Musk. He has so much wealth that he no longer needs to do anything to protect it. Instead, he can use big chunks of his wealth to do whatever he wants to do. Like manipulate the outcome of the November elections, for example. He can spend billions of dollars to get monkeys elected. He can even spend money to control the voting systems.
And now he is back, just in time for the November elections.
I think it is disingenuous to claim equivalence between the arguments of denialists and those pointing out the very real problem of overconsumption by AI data centers. The increased demand by these centers, not just for power, but also for water and other critically short resources, has undone decades of progress transitioning economies off of fossil fuels–and with very little discernible benefit to society.
Just because the Tech Bros. tell us it’s the next big thing doesn’t mean we should follow them blindly–indeed their advocacy should make it even more suspect. These are people whose only real contribution to human knowledge is discovering the obscenity of the very large numbers that represent their wealth. And most of them quit college to do so. They are uneducated philistines who do not understand the value of science, art or anything else that makes human life worthwhile.
AI should serve humans, not the reverse.
MS: Is AI data center energy demand accelerating the conversion to renewable energy generation?
Google AI: Yes, the surging energy demand from AI data centers is acting as a major accelerator for the transition to renewable energy, acting as a “boon” for new, low-carbon technologies. While the immense power needs create challenges, they are driving unprecedented investment in renewables, particularly through Power Purchase Agreements (PPAs) with major technology companies.
Here is a breakdown of how AI is influencing the energy landscape:
1. Driving Renewable Adoption
Massive Investment: AI-focused data centers are driving a surge in clean energy, with roughly 50% of the growth in data center electricity demand expected to be met by renewables through 2035.
Corporate Procurements: Hyperscalers (major tech companies) are leading in procuring renewable energy. In 2024, they accounted for 43% of clean power purchase agreements (PPAs), providing the capital needed for new solar and wind projects.
Accelerated Investment: The demand is forcing faster deployment of renewable infrastructure, with global data center power demand expected to grow by 165% by 2030.
2. The Nuanced Reality: Renewables and Fossil Fuels
While AI accelerates the demand for green energy, the immediate, high-volume, 24/7 nature of this power requirement has created a dual approach:
Fossil Fuels as a “Crutch”: Because renewable sources like wind and solar are intermittent, and because data centers need constant, reliable, “firm” power, fossil fuels (particularly natural gas) are still being used to meet immediate, near-term, high-demand, especially in regions with slower grid-upgrade times.
Shift Towards Nuclear: To achieve 24/7 clean energy, large tech companies are investing heavily in nuclear power (small modular reactors, or SMRs) as a carbon-free, baseload source, with the first of these expected to come online around 2030.
3. Key Drivers and Challenges
Grid Capacity Constraints: Grid interconnection delays (3-5 years) are a significant barrier, and nearly 20% of planned data center projects could face delays.
Local Bottlenecks: AI data centers are highly concentrated, putting extreme strain on specific regional grids, such as in Virginia and Texas.
Regional Differences: In the U.S. and China, the largest markets, fossil fuels are currently meeting most of the increase in demand, though this is expected to change post-2030 as renewable and nuclear capacity expands.
4. AI for the Energy Sector
AI is not only driving demand but also providing solutions for the energy transition:
Optimizing Grids: AI is being deployed to improve the forecasting and integration of variable renewable energy (wind and solar).
Predictive Maintenance: AI-driven tools can detect grid faults, reduce downtime, and increase the efficiency of existing transmission lines.
In summary, AI is acting as a significant catalyst for the renewable energy sector, but it is also forcing a temporary,, parallel reliance on traditional energy sources to meet immediate, high-volume, 24/7 power needs, while driving future innovation in nuclear energy.
This is about at the quality of what I’d expect if I asked a politician to give a critical evaluation of his/her voting record. It also seems to consider nuclear energy to be renewable, and a good thing. I don’t.
At a local level, I’ve long been involved in a campaign trying to get our public electric utility to shut down an old coal burning plant that has been polluting the low income section of our city for many decades. The shutdown has been repeatedly delayed, this time for several years because of the need to provide electric service for all the new data centers. That’s the reality I see on the ground.
Great description, I would add full of inferiority complexes taking their frustrations on other people, too many other people. Their latest “big thing” will not only crash the economy when the mirage collapses at a higher level of destruction than 2008, but has an immense opportunity cost, the cost of investing in what can save and sustain the living, including themselves, before we reach the point of no return.
On a funny note, it looks like with all the hype around the revolutionary new thing, the money is in selling adds! This is why so many resources will be destroyed.
On data center problems (for all, but especially Martin Smith, since I just saw this: one more data point for you):
Exposing The Dark Side of America’s AI Data Center Explosion | View From Above | Business Insider
https://www.youtube.com/watch?v=t-8TDOFqkQA
The data center problem is certainly real, but it isn’t new. It was a problem before AI, because… Google, Facebook, gmail, Twitter/X, YouTube, Spotify, Apple Music, Amazon, all the online TV and movie streaming services, the worldwide cellphone networks, the SWIFT transaction processing network, Mastercard, VISA,, online universities, weather forecasting services, the stock markets around the world. The data center problem didn’t appear with AI.
FYI …
Point of no return: a hellish ‘hothouse Earth’ getting closer, scientists say
The Guardian
February 11 2026
“The world is closer than thought to a ‘point of no return’ after which runaway global heating cannot be stopped, scientists have said. Continued global heating could trigger climate tipping points, leading to a cascade of further tipping points and feedback loops, they said. This would lock the world into a new and hellish ‘hothouse Earth’ climate far worse than the 2-3C temperature rise the world is on track to reach.”
https://www.theguardian.com/environment/2026/feb/11/point-of-no-return-hothouse-earth-global-heating-climate-tipping-points
https://www.cell.com/one-earth/fulltext/S2590-3322%2825%2900391-4
Could :) been like this ever since it was announced that GW/CC is a thing.
Much like the falling birth rates around the world . Could be a problem by 2070. Countries with water issues. Could be a problem by 2070 if it isn’t addressed. Growing thirsty crops on the wrong area. Drilling to deep into the aquifers. Rivers are too polluted by heavy industry and they have been dammed leaving too little water downstream. Glaciers shrinking etc
Too much plastic, too much waste, too much of everything and too much profit but as yet e wry country is 10 years away from collapse and of course the AMOC and other MOCs
Always 10 years away it disastrous by the end of the century etc
Most of us will be dead by the time disaster strikes
PBest: For many people, disaster is already here. Also, wrt birth rates, note:
Year Population (billions)
1800 1
1927 2
1960 3
1974 4
1987 5
1999 6
2011 7
2022 8
2035* 9
Famine is widespread. Just in the last few weeks 100s of thousands have been affected by weather disasters. Global warming and its resultant weather extremes are affecting the food supply. Time to pay attention.
Why? Renewables increase but only to offset desalination plants and other tech fixes for tourism and golf courses or keep on living in deserts or climate impacted places It’s not me that needs to pay attention but the masses who endlessly fly and deny there’s anything wrong.
21,000 desalination plants already and growing no doubt. Your ideas on birth rates is plain strange co sundering yes it’s growing presently but falling in many other places.
I am amazed that 1.5C is all but gone and 2C too by the looks of it. Tech fixes bound even to the point of AI can help with solving climate change whilst consuming huge resources in the mean time.
Tell me you have a plan
population data are facts, not opinion: https://en.wikipedia.org/wiki/Population_growth
The point is that in the 20th century we went from 2 to 8 billion. It’s fact, not ‘idea’. This is more than the carrying capacity of our finite planet. It is in the process of overcoming its apex predator, which has no respect for limits and infinite respect for self indulgence.
You and I share shock/horror at inaction: we’ve known what we need to know since early 1980s if not earlier. But my only possible ‘plan’ – aka, what I can do – is to do what is within my remit (relatively little) to remove liars/bullies/predators from office and support those who will act on our behalf (my senators are Liz Warren and Ed Markey, whose knowledge and actions on climate are a matter of record).
Because if you do not pay attention, you will miss the spectacle of Earth’s 6th mass extinction event. We do not know how it will play out–which species will live and which will die, which ecosystems will by forever altered or what the planet will look like on the other side. We don’t know whether we’ll be one of the species to make it or whether this whole “intelligent-life” experiment will ultimately be a failure.
Me, I’m betting on the latter. Humans had a chance to avoid the worst effects of the multiple crises impinging on our existence. We still have a chance to preserve enough of the world that it may be recognizable to our progeny as something resembling our descriptions. We could pull things out. But we won’t. Ultimately, we just aren’t smart enough to act unless the threat is bearing down on us like a freight train.
Whether or not climate change is the cause of our demise, my ghost will be very surprised if humans are still around in a few thousand years. 99% of all species that ever lived are extinct. I don’t see enough data to swing that prior probability in our favor.
“Most of us will be dead by the time disaster strikes”
Think about this one: Has there ever been a more “ME, ME, ME” statement ever published here? I don’t think so.
True regardless of sentiment
And some of the rest will be dead because the disaster strikes?
From I can see the world turns to technology solutions such as in Perth, WA. Rainfall fell significantly over the past 50 years so now it uses 2 large scale desalination plants and is building a third. Ground water has been affected so it is refilling them with treated water etc due to desalination. Growing food using green hiusrs to reduce water usage etc. it’s a city of 2.1 million growing to 3.5 and technology is the answer for the Aussies.
Spain has water problems and is becoming note arid but it now has 770 desalination plants providing water. Israel has advanced desalination and invented drip water irrigation etc
Humans appears to be relying on technology more and more to combat large populations living in worsening areas
This article very much addresses an issue I brought up some time ago, that of tipping points, what they are, what they mean, and where we are with them. It does seem that some want to minimize them because maybe having them pass doesn’t lead to a sudden change in conditions. I think this article addresses what they really mean.
So I asked ChatGPT to compare its energy consumption for an instance of my paid subscription use with the energy consumption that would result if I did all the online searching, reading/analysis, and writing myself. Its response was long and included a lot of related subjects, like model training vs duplication of human effort, etc. I won’t post the whole response. I suspect that this subject is one where ChatGPT has been “trained” with information favorable to the AI sector, but it did leave me with a takeaway that I think is good:
ChatGPT: If it’s useful in your debate, the clean takeaway line is often:
A substantial AI answer likely uses roughly the same energy as 30–90 minutes of normal web research — sometimes less — while avoiding massive duplication of human effort.
I am getting a bit too comfortable with ChatGPT. It is fascinating using it to learn about astrophysics and the causes of blocked weather patterns bringing sustained miserable weather to the UK at the moment. It even constructed a brillant outline for a talk about climate change should I have the opportunity of giving one, with emphasis on presenting it as an objective risk management issue regarding non-linear tipping points which gets the facts across whilst minimising the risk of antagonising people.
MS (an AI quoted): A substantial AI answer likely uses roughly the same energy as 30–90 minutes of normal web research — sometimes less — while avoiding massive duplication of human effort.
BPL: Then why are they building separate data centers with their own power plants?
BPL: Then why are they building separate data centers with their own power plants?
MS: Because there are billions of short answers generated for questions from free system users.
As many of you are aware, President Trump will be cancelling global warming today, February 12, 2026.
https://www.telegraphherald.com/news/national_world/article_19b4bd7c-a8d4-46fa-bf73-527c7817e9d9.html
It’s true. CO2 will no longer be an issue. You can all breath easy with no guilt about exhaling CO2. The earth has taken note and is in full compliance with Trump’s orders:
NYC Mayor Mamdami says it’s been colder in NYC than in Antarctica, and the death toll from the cold is now at least 18:
https://www.youtube.com/shorts/mV0BL2Dy5Rw
It’s one of the longest cold streaks for NYC since 1963:
https://www.bbc.com/news/articles/cd9g8nxdexko
How cold is it? It’s so cold that reptiles are falling from the trees in Florida:
https://www.bbc.com/news/videos/c23rd925knxo
Many cold temperature records were broken in Florida:
https://weather.com/forecast/regional/news/2026-01-29-florida-record-cold-southeast-mid-atlantic-cold-streak
Heavy snow blanketed the east coast. Even southerly North Carolina got up to 15″ during the storm:
https://www.qcnews.com/weather/winter-weather/historic-snowfall-hits-charlotte-latest-totals-across-the-region/
The cold was not isolated to the NE USA. Europe was so cold that Estonians are now driving to islands out in the sea! I’m not that brave, but this could be a benefit to Trump’s cancellation of global warming – essentially free highways:
https://www.theguardian.com/world/2026/feb/10/estonia-ice-road-frozen-sea-saaremaa-hiiumaa
Just as we saw with the border invasion, all we needed to stop that and global warming was a new president. Whodathunkit?
:)
A. How convenient of you to completely ignore the record heat in the western US. Flaunting your ignorance once again.
B.
The global surface temperature was the fifth-warmest January on record—the smallest temperature departure since 2023.
Snow cover extent was below average for North America and near average for Eurasia.
Sea ice extent was near record low for the Arctic and below average for Antarctica.
Global tropical cyclone activity was above normal with 11 named storms.
See (but of course you wont because that would involve learning, something you do not do): https://www.ncei.noaa.gov/news/global-climate-202601?fbclid=IwY2xjawP6pnFleHRuA2FlbQIxMABicmlkETE5eFZsYms1YllQaEVGcncwc3J0YwZhcHBfaWQQMjIyMDM5MTc4ODIwMDg5MgABHv0D7AykF_QZC_ZkdAwUHx3ojeUJrCsv_V1OyUaIBH4uEAWzb4LHIUnLDe1N_aem_AppyJiPpr9bdqm0MPopbfQ
So stop lying. Who made you think that cherrypicking the only two areas in the world that were cooler proves your point? See the map, sport. Busted. Again.
Couple of links to NYTimes announcement of the dangerous cancellation of the endangerment finding:
What to Know About the E.P.A.’s Big Attack on Climate Regulation. The Trump administration has repealed the scientific determination that underpins the government’s legal authority to combat climate change.
https://archive.ph/GrAvM
Trump Administration Erases the Government’s Power to Fight Climate Change. The Environmental Protection Agency repealed the bedrock scientific finding that says greenhouse gases threaten human life and well being. It means the agency can no longer regulate them.
https://archive.ph/a1gnl
There’s a brief claim that things got better due to climate action in the middle of one of these. I found that suspect. I think scientists know and have been communicating that things are getting worse. Please don’t use this as an excuse to endlessly argue about how much. We know it’s bad, and those outside the choir need to understand the massive agreement amongst those who understand the issues rather than the exploitable differences in detail.
“How cold is it? It’s so cold that reptiles are falling from the trees in Florida:”
This past summer, it was so hot that dead monkeys were falling from trees in Central America.
“Now that we know the risks why are we so determined”. Because there is no “we”. That’s why I’ve had little interest in Life things since pondering that in 1961. Only interest in physical reality things, because it’s interesting.
Mr. KFA “You can all breath easy”. Whew! I been holding my breath for decades waiting for the bods who Own & Operate the Universe to finally announce that.
KIA: It’s true. CO2 will no longer be an issue. You can all breath easy with no guilt about exhaling CO2. The earth has taken note and is in full compliance with Trump’s orders
BPL: This is a parody, right? Because you can’t possibly be stupid enough to believe this.
Oh, no. He can. A cautionary tale folks–what a lifetime of watching Faux News can do to your brain.
And whether he really believes it or not, he is still sanguine enough in his ignorance to think there is something remotely funny about a demented troll ruling the most powerful country on Earth.
Next in line, he will declare that the Laws of Trump replace the Laws of Physics, at least within the United States of Trump. \s
Mkia, imagine a big swamp cooler full of ice sitting in the sun. As long as there’s ice, it melts and soaks up heat, so the air blowing into the house is cool or even cold.
As time goes by and temperatures slowly increase – on earth because of the addition of co2 – the ice melts faster, but while it’s melting it still holds the temperature down inside the house, sometimes, again, even keeping it cold. Once all the ice is gone, however, the cooler is just a dark, sun‑soaked box, and now it actually helps heat the air up—so the house gets warmer, and it does so faster than when there was still ice in the cooler. On Earth it’s even faster, because when the ice is gone the bright, sun reflective surface (albedo) is lost and darker ground and ocean absorb more sunlight.
So is the earth getting hotter? The melting global ice is your clue. Keep your eyes on that.
https://midmiocene.wordpress.com/2015/10/20/a-global-warning/
This is a simplified example and someone else can make it more precise, but I think it is essentially correct
Anyone who uses ChatGPT to compose their responses has an overwhelming advantage. It creates extended volume which clutters up the place. It persuades people not to think for themselves. It persuades people it is inerrant. It persuades people it has a life outside switches and power sources.
It does its master’s bidding. It has resulted in vast reopening of dirty energy, though cryptocurrency started that rot.
I can’t even …
I will continue to post some of the massive information coming in about community and energy abuse driving this delusion. Here’s one:
https://paulkrugman.substack.com/p/no-ai-doesnt-justify-lower-interest
“And while the productivity payoff from AI is still speculative, massive capital expenditure driven by AI dreams is already happening, on a scale that raises questions about where the money (not to mention the electricity and the chips) will come from.
“So no, AI doesn’t justify interest rate cuts, and certainly not the huge cuts Trump demands.
“And the alacrity with which Trump officials have jumped on the argument that the Fed should slash rates because of AI is troubling for reasons that go beyond the fact that they’re wrong. Consider how this argument became popular in Trumpworld. Did Bessent, or Warsh, or Miran carefully consider the evidence, and advocate an interest-rate policy based on what that evidence showed? Of course not. Their boss wants to slash interest rates, so they went looking for plausible-sounding economic arguments that might rationalize his whims.
“So am I saying that the argument that AI justifies rate cuts is dishonest, that AI has become the last refuge of scoundrels? Why yes, I am.“
Both GISTEMP & NOAA have reported for January, as has Copernicus ERA5 re-analysis.
No great revelations. All put Jan 2026 as the 5th warmest on record. And bar the 2007 ranking slipping a tad in NOAA, the top-ten warmest Jan rankings are the same in all three.
Top-ten warmest Januarys for GISTEMP (NOAA) & ERA5
2025 … …+1.59°C … … (2025 … …+1.52°C) … … 2025 … …+1.75°C
2024 … …+1.46°C … … (2024 … …+1.51°C) … … 2024 … …+1.66°C
2020 … …+1.39°C … … (2020 … …+1.39°C) … … 2020 … …+1.54°C
2016 … …+1.39°C … … (2016 … …+1.37°C) … … 2016 … …+1.51°C
2026 … …+1.29°C … … (2026 … …+1.33°C) … … 2026 … …+1.47°C
2017 … …+1.23°C … … (2017 … …+1.25°C) … … 2017 … …+1.36°C
2007 … …+ 1.23°C … … (2019 … …+1.16°C) … … 2007 … …+1.32°C
2019 … …+1.14°C … … (2022 … …+1.14°C) … … 2019 … …+1.24°C
2022 … …+1.12°C … … (2007 … …+1.10°C) … … 2022 … …+1.24°C
2023 … …+1.09°C … … (2023 … …+1.10°C) … … 2023 … …+1.21°C
(Anomaly bases NOAA & ERA5 – 1850-1900, GISTEMP – 1880-1920)
Note that ERA5 is measuring SAT not SAT/SST, so you’d expect to see more differences appearing with ERA5.
There are quite a few nerdy weeds if folk feel the urge to go getting bogged-down in detail. One nerdy point concerns the 12-month rolling averages. It may be worth mentioning as it concerns 12-month averages of GISTEMP, the measure used in Hansen’s predictions which do get discussed here at RC UV threads.
While the ERA5 SAT record could be forgiven for a bit of divergence, there had been quite a big difference growing between NOAA & GISTEMP. Maybe it’s a coverage thing as they both start off with the same data so any difference would otherwise be expected to be small (as with HadCRUT & BEST). But a growing variance has been persisting for a while.
Plotting these various 12-month averages with the anomaly base fixed at the 2022 average, (HERE – Posted 12th February 2026), GISTEMP shows a peak anomaly significantly warmer than NOAA and followed this with (so far) significantly less cooling.
This had resulted in a growing gap (reaching +0.053°C) between GISTEMP & NOAA but the last couple of months perhaps is showing the gap beginning to shrink.
The cooling for all these surface temp series is roughly 0.02°C/month.
Quick flyby to say “Hi!” to all and to again say “Thank You!” to the scientists and everyone behind Real Climate for what it continues to represent on this stark day here in the Dark Ages that our federal government is trying to bring down on us silly citizens where science and scientific practice is concerned. EPA’s announcement today was expected of course, yet in a way, it’s still leaves me shaking my head in amazement.
Well, back to the salt mines. I extend best wishes to all. Hope to have more time soon :-)
Recent observationally constrained estimates suggest an accelerated warming rate over the past decade (e.g., Forster et al. 2025; Hansen et al. 2026), driven by high GHG emissions combined with declining aerosol cooling and emerging feedbacks. Using these constraints, I explored a heuristic high-end warming trajectory intended as an illustrative scenario rather than a deterministic forecast.
Independent assessment by the German Physical Society (DPG) and German Meteorological Society (DMG) highlights similar upper-range outcomes, estimating 2–3 °C by 2050 and 4–5 °C by 2100 relative to pre-industrial levels, reinforcing the plausibility of a high-end warming pathway consistent with heuristic projections based on Hansen et al. (2026) and Forster et al. (2025).
My GMST anomaly projections are based on the observed accelerating warming trend to 2025, coupled with established physical understanding in climate science, and are framed as a constrained scenario construction under conditional socio-ecological and geopolitical assumptions rather than a deterministic prediction, reflecting a synthesis across physical climate dynamics and political–economic system behavior.
The scenario assumes that anthropogenic carbon dioxide removal (CDR) and solar radiation management (SRM) interventions remain negligible in net impact on global forcing over the projection period, and that global mitigation efforts are unlikely to achieve Net Zero emissions or fully meet Paris Agreement targets.
Continued anthropogenic forcing growth is assumed through the 2030s, followed by increasing geopolitical fragmentation and potential erosion of coordinated international mitigation frameworks. Reduced fossil fuel combustion could plausibly be partially offset by declining aerosol cooling, widespread biosphere degradation (including large-scale forest loss and wildfire emissions), weakening land and ocean carbon sinks, and potential reversion to higher-carbon fuels under geopolitical stress. Under such conditions, net radiative forcing could continue to accelerate, with Earth-system feedbacks increasingly dominant after mid-century.
A smooth illustrative trajectory consistent with recent observational constraints yields approximately ~1.7 °C by the mid-2030s, ~2 °C around the early 2040s, ~2.3–2.5 °C by 2050, ~3.1–3.6 °C by 2075, and ~4.3–5 °C by 2100 (relative to 1850–1900). This places the pathway in the upper tail of CMIP6 outcomes but within plausible physical bounds given current forcing trends and feedback uncertainties.
Illustrative High-End GMST Trajectory (Scenario-Based)
2015–2025 Decadal GMST Anomaly ~1.30 °C (range ~1.25–1.35 °C)
(Estimated warming rate ~0.25–0.40 °C per decade)
Current short-term GMST Anomaly ~1.50 °C (range ~1.45–1.55 °C)
(Short-term variability elevated warming rate ~0.35–0.45 °C per decade)
By 2035 Decadal GMST Anomaly ~1.72 °C (range ~1.65–1.80 °C)
(Estimated forced warming rate ~0.25–0.35 °C per decade)
By ~2042 2 °C Paris threshold likely reached 2040–2043 (range ~1.9–2.1 °C)
By 2050 Decadal GMST Anomaly ~2.35 °C (range ~2.20–2.50 °C)
(Estimated rate ~0.35–0.45 °C per decade)
By 2075 Decadal GMST Anomaly ~3.35 °C (range ~3.10–3.60 °C)
(Estimated rate ~0.40–0.50 °C per decade)
By 2100 Decadal GMST Anomaly ~4.65 °C (range ~4.30–4.95 °C)
(Estimated rate ~0.45–0.55 °C per decade)
The trajectory is illustrative and uses a smooth accelerating heuristic fit constrained by recent observational warming rates and plausible late-century high-end forcing outcomes, rather than a formal Earth-system model.
The main mechanisms that could substantially slow such a trajectory would be either an unprecedented coordinated global decarbonisation effort, or large-scale systemic disruption of industrial activity, energy production, and trade flows beyond historical precedents (e.g., world war-scale disruptions).
Fair warning to folks: the sockpuppet willfully misrepresents warming projections and peer-reviewed sources, including Forster et al. 2025. So the sockpuppet account should not be trusted on those topics.
Data: “My GMST anomaly projections…….”
Sounds more like Google Geminis GMST anomaly projections, given the style of language and format.
EOS has an article about a new Nature paper about continued circulation and water temperatures of the AMOC during the last glacial maximum. It would be interesting to see how the study holds up and affects understanding of AMOC behavior past and likely future.
“The AMOC of the Ice Age Was Warmer Than Once Thought
An analysis of sediment cores indicates that North Atlantic waters were relatively warm and continued to circulate even under major climate stress during the Last Glacial Maximum.”
“A major part of the Atlantic Meridional Overturning Circulation (AMOC), a large-scale ocean circulation pattern, was warmer during the peak of Earth’s last ice age than previously thought, according to a new study published in Nature.
The study’s results contrast with those from previous studies hinting that the North Atlantic was relatively cold and that AMOC was weaker when faced with major climate stress during the Last Glacial Maximum (LGM), about 19,000–23,000 years ago.
The findings add confidence to models that scientists use to project how AMOC may change in the future as the climate warms, said Jack Wharton, a paleoceanographer at University College London and lead author of the new study.”
Article here:
https://eos.org/articles/the-amoc-of-the-ice-age-was-warmer-than-once-thought/