Leading Coral Expert Calls Impending Demise of the Great Barrier Reef a “Fairy Tale”

According to climate change activists, Australia’s Great Barrier Reef is in imminent peril of becoming extinct. But recent evidence shows that belief to be a “fairy tale,” says Australian scientist and leading coral reef authority, Dr. Peter Ridd.

The Great Barrier Reef is the world's largest coral reef system, consisting of perhaps 3,000 individual reefs, 2,300 km (1,400 miles) long in total and visible from outer space. Labeled by CNN as one of the seven natural wonders of the world, the reef is a constant delight to tourists, who can view the colorful corals from a glass-bottomed boat or by snorkeling or scuba diving.  

The belief that its days are limited has been fostered for years by other coral reef scientists, especially those at Australia’s JCU (James Cook University) in Queensland. One reason is that in the early 2000s, average coral cover on the Great Barrier Reef took a sudden tumble from 20% to about 10% in 2012, as illustrated in the figure below. This data comes from AIMS (the Australian Institute of Marine Science), who have been measuring over 100 reefs every year since 1986.

The rock-bottom cover in 2012, which was actually the result of two major cyclones, set alarm bells ringing in the coral reef scientific community and the mainstream media. However, such pessimism was unwarranted, as you can see from the figure showing a long recovery of the average coral cover in 2012 to a record high in 2022. That high level has been essentially maintained in the four years since then – leading to Ridd’s designation of the earlier narrative as a fairy tale.

There is some variation in coral cover over the 3,000 or so reefs that constitute the Great Barrier Reef. This can be seen in the next figure,depicting cover for three separate regions of the reef; current cover varies from approximately 27% in the south to 40% in the north. A reef expedition in 1928-29 discovered that coral cover at that time on a reef island known as Low Isles ranged from 8% to 42%.

But corals are susceptible to overheating and undergo bleaching when the water gets too hot, losing their vibrant colors. The photo on the left in the next figure shows a healthy section of the Great Barrier Reef, while the photo on the right shows a different section of the reef as it recovers from a bleaching event. The colors in the bleached section are noticeably less vibrant.

What is not widely known, however, is that corals are capable of much more rapid recovery from bleaching events than most reef scientists thought. It was once believed that the recovery process, if it occurred at all, took 10 years or longer. But recent studies (see here and here) have found that corals can return to pre-bleaching levels in as little as two years, even when seawater temperatures are higher than normal.

Ridd explains the adaptability of corals to heating as a survival mechanism, in which the multitude of polyps that constitute a coral exchange the microscopic algae that normally live inside the polyps and give coral its striking colors.

Hotter than normal water causes the algae to poison the coral that then expels them, turning the polyps white. But to survive, the coral needs resident algae which supply it with energy by photosynthesis of sunlight. So from the surrounding water, the coral selects a different species of algae better suited to hot conditions, a process that enables the coral to recover within a few years, says Ridd.  Coral death and regrowth are a normal, natural phenomenon.

That corals worldwide are not on the verge of extinction was first revealed in a 2021 study by four JCU researchers. The study completely contradicted previous apocalyptic predictions of the imminent demise of coral reefs, predictions that included an earlier warning from three of the same authors and others of ongoing coral degradation from global warming.

The JCU study included data on more than 900 coral reefs across the Pacific, from Indonesia to French Polynesia. To estimate abundances, the researchers used a combination of coral reef habitat maps and counts of coral colonies. They estimated the total number of corals in the Pacific at roughly half a trillion, similar to the number of trees in the Amazon or birds in the world. This colossal population is for a mere 300 species, a small fraction of the 1,619 coral species estimated to exist worldwide by the IUCN (International Union for Conservation of Nature).

The fact that the Great Barrier Reef’s coral cover has been at record high levels for four years now is still minimized by climate activists, who insist that the new growth is the wrong kind of coral. But Ridd says that is irrelevant, and that failure to recognize the remarkable recovery of this wonderful treasure is merely an example of misguided groupthink.

Next: Deception in Science: The Sad Story of Elizabeth Holmes

No Evidence That Weather Extremes Becoming More Frequent or Intense: (6) Wildfires

This post on wildfires completes the present series on weather extremes. Just as with all the other examples of extreme weather presented in this series, there’s no scientific evidence that wildfires today are any more frequent or severe than anything experienced in the past. Although wildfires can be exacerbated by other weather extremes such as heat waves and drought, we’ve already seen that those extremes are not on the rise either.

Together with tornadoes, wildfires are probably the most fearsome of the weather extremes commonly blamed on global warming. Both can arrive with little or no warning, making it difficult or impossible to flee, are often deadly, and typi­cally destroy hundreds of homes and other structures.

The worst wildfires occur in naturally dry climates such as those in Australia, Cali­fornia or Spain. The next three figures show the annual area burned by wildfires in Australia, the U.S. and the Mediterranean, respectively. Clearly, the area burned annually is declining in all three regions, despite hysterical claims to the contrary by the mainstream media.

The top figure above presents a recent estimate of the percentage of Australia’s land surface burned every decade from 1900 to 2000, plus actual satellite measurements from 1997 to 2020. Also shown is the 1974-75 bushfire season, when a record total of 117 million hec­tares (300 million acres) were consumed. The annual average during a 30-year period of global cooling from 1940 to 1970 was about 80 million hectares (200 million acres).

Although the number of acres burned annually in the U.S. (middle figure above) has gone up slightly over the last 25 years or so, the present burned area is still only a small fraction of what it was back in the 1930s. Because modern global warming was barely underway in the 1930s, climate change clearly has noth­ing to do with the incineration of U.S. forests. The recent uptick reflects the introduction of wildfire suppression that has led to increases in forest density and the buildup of undergrowth, both of which greatly enhance the potential for bigger fires.

The bottom figure above reveals no such uptick in recent Mediterranean wildfires, at least until 2010. This contrast between the Mediterranean and the U.S., where wildfires are becoming fewer but larger in area, has been attrib­uted to different forest management policies on the two sides of the Atlantic – despite the protestations of U.S. politicians and some firefighting officials in western states that climate change is responsible for the recent increase in fire size.

The figure below illustrates the timeline from the year 1600 onward of fire occurrence at more than 800 different sites in western North America.

The sudden drop in wildfire occurrence around 1880 has been ascribed to the expansion of American livestock grazing in order to feed a rapidly growing population. The depletion of fuel by grazing cattle, together with the firebreaks created by the ar­rival of railroads, drastically reduced the incidence of wildfires. And once mechanical equipment for firefighting such as fire engines and aircraft became available in the 20th century, more and more emphasis was placed on wildfire prevention.

Intentional burning, long used by native tribes and early settlers, has become a thing of the past in the U.S. Only now, after several dev­astating wildfires in California, is the idea of controlled burning being revived. In Europe, on the other hand, prescribed burning has been supported by land managers for many years.

Globally, the wildfire trend is the same as in Australia, the U.S. and the Mediterranean. Wildfires have diminished globally as the planet has warmed, as depicted in the next figure showing the estimated worldwide area burned by wildfires by decade from 1900 to 2010.

Over the century from 1915 to 2015, the global burned area declined steadily at a rate of approximately 7% per decade, calculated from the data in the figure above and a 2017 study. The study attributes the declining trend to the dominance over higher temperatures of heavier precipitation and increased population density: while warming enhances wildfires by drying out vegetation, population increases lead to a reduction in vegetation through clearing of land.

The downward global trend has continued in 2026, in spite of extensive wildfires in southern Europe several months ago:

The ever-rising level of CO₂ in the atmosphere would be expected to increase wildfire risk, since CO₂ promotes plant growth. The concentration of CO₂ in the atmosphere is currently rising by about 2.6 parts per million (ppm) per year and the global average reached 426 ppm in 2025. But at the same time, higher CO₂ levels reduce plant transpira­tion, meaning that plants’ stomata or breathing pores open less, the leaves lose less water and more moisture is retained in the soil. In­creased soil moisture has led to a worldwide greening of the planet.

Next: Leading Coral Expert Calls Impending Demise of the Great Barrier Reef a “Fairy Tale”

No Evidence That Weather Extremes Becoming More Frequent or Intense: (5) Tornadoes

Tornadoes, which occur predominantly in the U.S., are smaller and claim fewer lives than hurricanes. But they form rapidly and frequently demolish homes and buildings in their often-narrow path. Like hurricanes, tornadoes are categorized according to wind speed, using the enhanced Fujita Scale going from EF0 to EF5 (F0 to F5 before 2007); EF5 tornadoes attain wind speeds up to 480 km per hour (300 mph).

The figure below, compiled by NOAA (the U.S. National Oceanic and Atmospheric Administration), shows where all tornadoes in the U.S. from January 1, 1950 to May 31, 2026 touched down. As you can see, tornadoes cluster primarily in the eastern half of the country, especially in so-called “Tornado Alley”, an area extending northward from central Texas through the plains states.

The upper panel of the next figure shows the annual count of all EF1 or stronger tornadoes in the U.S. between 1950 and 2021, revealing no meaningful trend. This is a pe­riod that included both warming and cooling spells, with net global warming of approximately 1.0 degrees Celsius (1.8 degrees Fahrenheit) during that time. If EF0 tornadoes are included, the same plot shows an apparent increase in annual number since 1990, but that is simply due to the introduction of Doppler weather radar in the early 1990s, which led to the reporting of more tornadoes than before.

However, as seen in the lower panel, the annual number of strong U.S. tornadoes (EF3 or greater) has declined dramatically over the 72 years from 1950 to 2021. In fact, the average number of strong tor­nadoes annually from 1986 to 2017 was 40% less than from 1954 to 1985. Global warming, which might be expected to increase strong tornado activity, cannot have played any role.

U.S. tornadoes typically cause about 80 deaths and more than 1500 injuries per year. The deadliest episode of all time in a sin­gle day was the “tri-state” outbreak in 1925, which killed over 700 peo­ple and resulted in the most damage from any tornado outbreak in U.S. history. The photo montage on the right shows one of the 12 or more tornadoes observed in Missouri, Illinois and Indiana approaching a farm (top); some of the 154 city blocks obliterated in Murphysboro, Illinois (middle); and the wreckage of Murphysboro’s Longfellow School, where 17 children were killed (bottom). 

Unlike the narrow path of most tornadoes, the swath of destruction wrought by the main F5 tornado was up to 2.4 km (1.5 miles) wide. Amazingly, the ferocious storm persisted for a distance of 353 km (219 miles) in its 3 ½-hour lifetime. Together with smaller F2, F3 and F4 tornadoes, the F5 tri-state tornado destroyed or almost destroyed numerous towns. Another 33 schoolchildren died in De Soto, Illinois when their school collapsed. De Soto’s deputy sheriff was sucked into the funnel cloud, never to be seen again.

Nature’s wrath was on display once more in the most ferocious tornado outbreak ever recorded, spawning a total of 30 F4 or F5 tornadoes – the so-called Super Outbreak – in April 1974. A total of 148 tornadoes of all strengths struck 13 states in Tornado Alley and the Canadian province of Ontario over two days; their distribution and approximate path lengths are depicted in the left panel of the figure below.

The photos on the right illustrate the massive F5 tornado, the worst of the 148, that bore down on Xenia, Ohio (population 29,000, top) and the resulting damage (middle and bottom). The Xenia tornado was so powerful that it tossed freight trains on their side and even dropped a school bus onto a stage where students had been practicing just moments before. In all, the Super Outbreak caused 335 fatalities and over 6,000 injuries.

As with hurricanes, the IPCC (Intergovernmental Panel on Climate Change) in its Sixth Assessment Report has dismissed any connection between global warming and tornadoes. While it concedes that escalating temperatures and humidity may create atmospheric in­stability conducive to tornadoes, it also points out that other fac­tors governing tornado formation, such as wind shear, diminish in a warming climate.

U.S. tornado activity, however, can vary drastically from year to year. The next figure depicts the average number of tornadoes per state annually from 2006 to 2025.  

The average for Illinois, which at 66 per year is less than that for several other plains states, has already been far surpassed in 2026. By August 29, the number of confirmed tornadoes in Illinois was 210, by far the most the state has documented in any year in reliable records dating to the 1950s. The adjacent state of Indiana had also experienced a record number of tornadoes by the same date. This unusual behavior has been attributed to a sharp southward plunge of the jet stream, bringing excessive warm and humid air to the Midwest.

Next: No Evidence That Weather Extremes Becoming More Frequent or Intense: (6) Wildfires

No Evidence That Weather Extremes Becoming More Frequent or Intense: (4) Hurricanes and Typhoons

This fourth post in our series on the spurious links between climate change and extreme weather examines the incidence of hurricanes and typhoons – powerful tropical cyclones that cause devastation around the globe and have claimed hundreds of thousands of lives. But, contrary to public opinion, hurricanes and typhoons show a decreasing trend worldwide, and their landfalling frequency has not changed for at least 50 years.

Hurricanes are classified by their sustained wind speeds on the Saffir-Simpson scale, ranging from Category 1, the weakest, to Category 5, the strongest. A major hurricane is defined as one in Category 3, 4 or 5, corresponding to a top wind speed of 178 km per hour (111 mph) or greater. Hurricanes occur in the Atlan­tic and northeastern Pacific Oceans, especially in and around the Gulf of Mexico; their cousins, typhoons, occur in the northwest­ern Pacific.

Hurricanes can be hundreds of kilometers in extent, with wind speeds up to 240 km per hour (150 mph) or more and often exact a heavy toll in human lives and personal property. The deadliest U.S. hurricane in record­ed history struck Galveston, Texas in 1900, killing an estimated 8,000 to 12,000 people. In the Caribbean, the Great Hurricane of 1780 killed 27,500 and winds exceeded an estimated 320 km per hour (200 mph).

How often hurricanes have occurred globally since 1981 is depicted in the figure below. You can see that the frequency of hurricanes as a whole has slightly diminished on average over this period; major hurricanes show no apparent trend.

The next figure plots the frequency of typhoons since 1951, showing no trend. However, a 2022 study demonstrates a decline in both total and major typhoons for the 32-year period from 1990 to 2021. The same study reinforces the recent decrease in total global hurricanes seen in the first figure.

Data for the North Atlantic basin, which has the best quality data available in the world, during the period from 1851 to 2024 are shown in the following figure. It’s clear that hurricane frequency has been essentially unchanged since 1851. The apparent heightened activity over the last 20 years, particu­larly in 2005 and 2020, simply reflects improvements in observational capabilities since 1970, and is unlikely to be a true climate trend, say a team of hurricane experts.

Aircraft reconnaissance flights to gather data on hurri­canes only began in 1944, while satellite coverage dates only from the 1960s. The relative lack of observations in early years of the record reflects the fact that eyewitness accounts on land excluded most non-landfalling hurricanes, while reports from ships at sea were sporadic.

Another 2022 study reveals that tropical cyclones in general, which include hurricanes, typhoons and tropical storms, are letting up as the globe warms. Over the period from 1900 to 2012, the study authors found that the annual number of tropical cyclones declined by about 13% compared with the period between 1850 and 1900, when such powerful storms were actually on the rise.

The lack of any long-term trend in global hurricanes and typhoons is borne out by the number of tropical cyclones that make landfall, il­lustrated in the next figure; the solid blue line represents all tropical cyclones, the solid red line major ones (Category 3, 4 or 5). The apparent slight uptick in both total and major landfalling tropical cyclones is statistically insignificant and could reflect the improvements in observation of North Atlantic hurri­canes since 1970 mentioned above.      

Finally, the figure below depicts the frequency by decade of all Florida landfalling hurricanes and major hurricanes since the 1850s. Florida has long been a favorite target of hurri­canes. While there is no trend in major hurricanes over 170 years, the trend in hurricanes overall is downward – even in a warming world.

The downward trend in hurricanes and typhoons in general is somewhat counterintuitive, since one would think that higher sea surface temperatures combined with higher atmospheric moisture levels due to increased evaporation would provide extra fuel for cyclones. However, tropical cyclone formation is inhibited by wind shear, which also increases as sea surface temperatures rise.

The IPCC (Intergovernmental Panel on Climate Change) in its Sixth Assessment Report mentions two aspects of tropical cyclones that have been observed recently, namely more rapid intensification and slower forward speed. Both phenomena result in heavier rain following landfall. Nevertheless, the IPCC does not attribute them to global warming, and indeed their origins are not yet understood.

For Amer­icans in and around the Gulf of Mexico, hurricanes have been a fact of life since Galveston and before. The death toll has fallen over time though, with improvements in planning and engineering to safeguard structures, and the development of early warning systems to allow evacuation of threatened communities.

Next: No Evidence That Weather Extremes Becoming More Frequent or Intense: (5) Tornadoes

How Much of a Threat Are Melting Ice Sheets and Glaciers?

The recent tragic and catastrophic Himalayan flash flood has brought to the fore an important question: how much do we need to be concerned about melting ice, from both glaciers and ice sheets?

Glaciers are found all over the planet. They number over 200,000 excluding the massive ice sheets of Antarctica and Greenland. Unsurprisingly, most glaciers are retreating as the world warms. However, what is not well known is that the retreat began early in the 19th century as global temperatures started to rise following the Little Ice Age – long before human CO₂ emissions became significant. This is illustratedin the figure below.

It has been estimated that complete melting of all Earth's glaciers and ice sheets will take several thousand years, but smaller mountain glaciers could vanish during this century. Nevertheless, not all glaciers are retreating. About 10% are actually advancing or periodically advancing and retreating. Recent surveys have found advancing glaciers in Alaska, Norway and Switzerland, for example.

Like ice sheets, melting glaciers – including those in Antarctica and Greenland – contribute to rising sea levels. At present, this contribution is 25-30% of the observed increase in global sea levels, corresponding to an annual rise of approximately 1 mm as shown in the next figure. The mass change is measured in billions of tonnes (gigatonnes, Gt where 1 gigatonne = 1.102 U.S. gigatons).

But for mountain glaciers, a bigger threat than their contribution to sea level rise is the flash flooding caused by melting ice, especially from those glaciers that are not in remote areas but close to human settlements.

Despite the claims of climate alarmists that the devasting flood in Nepal and Tibet was caused by global-warming induced glacial collapse, scientists soon figured out that the source of the calamitous event was actually a landslide of the underlying bedrock. This resulted in a massive chunk of the glacier breaking off and falling about 2 km (7,000 feet) into a river, the shattered ice gathering rock debris, sediment and river water as it rushed downstream at speeds approaching 150 km per hour (90 mph).

Regardless of the cause, the flood was devastating in terms of loss of life and property. Many commentators have pointed out, however, that the flood would have been less deadly had an early warning system been in place, allowing people to escape to higher ground or evacuate – as is found elsewhere in the world in areas subject to natural disasters. Improvements in planning and engineering to safeguard structures may have minimized property damage too.

The mass loss from glaciers currently dominates that from the ice sheets of Antarctica and Greenland, which together store about 99% of all the fresh water on Earth. The following figure shows the annual rate of ice mass loss from all three sources between 2000 and 2019.

Greenland’s ice sheet holds only about 10% as much ice as Antarctica’s. We hear more about it because the Greenland ice sheet is melting at a faster rate and has contributed more to sea level rise. But the melt rate is no faster today than it was 90 years ago and has slowed over the last few years. The next figure shows the total Greenland ice sheet mass going all the way back to 1840. The increase in ice loss that began around 2000 has since reversed due to above-average winter snowfall; the historical data exhibits similar fluctuations.

Just as with glaciers, the Greenland ice sheet started melting around 1900, long before so-called man-made global warming began. Its contribution to sea level rise is illustrated in the figure below, showing the cumulative ice mass change from 1972 to 2025. The rise in sea level over that period from melting Greenland ice averages 16/53 or 0.3 mm per year; the current rate of rise is approximately 0.4 mm per year.

The contribution of the Antarctic ice sheet to sea level rise is illustrated in the next figure, which shows the cumulative ice mass change from 1979 to 2024. The corresponding average annual rise in sea level over that period from melting Antarctic ice is 13.5/45 or 0.3 mm, the same as from Greenland’s ice sheet; again, the current rate of rise is approximately 0.4 mm per year.

It can be seen that most Antarctic ice loss comes from West Antarctica, where melting of glaciers is significant and partly caused by active volcanoes underneath the continent. The figure below graphs changes in the total Antarctic ice sheet mass between 2002 and 2025. You can see that while the loss rate accelerated during the two decades after 2002, it appears to have stabilized since 2021.

The total contribution to sea level rise of melting ice from glaciers and ice sheets is approximately 1.8 mm per year. The remainder of the current average rise rate of 4 mm per year is due to the expansion of seawater as it warms; sinking of the seafloor under the weight of extra meltwater; and land subsidence from groundwater depletion, or rebound after melting of the last ice age’s heavy ice sheet.

Next: No Evidence That Weather Extremes Becoming More Frequent or Intense: (4) Hurricanes and Typhoons

Measles Surging Worldwide, Thanks to Vaccine Hesitancy

Failure to vaccinate children against measles has led to a skyrocketing number of cases around the world this year. In the U.S., the number of confirmed cases had reached 2,318 by July 30, a 35-year high surpassing the 2,289 total cases recorded during 2025. Even higher numbers have been reported in other countries such as Bangladesh, Guatemala and Yemen.

The figure below compares the annual number of confirmed U.S. measles cases since 2000. The surge in 2025 is obvious, while the present partial number for 2026 will undoubtedly increase further.

The next figure depicts the current 2026 number of cases in most countries of the Americas. You can see that the incidence of measles is even higher in Canada, Mexico, Guatemala and Peru than it is in the U.S.

The primary reason for all these outbreaks is that more and more parents are choosing not to vaccinate their children. In 2019, the WHO (World Health Organization), for the first time, listed vaccine hesitancy as one of the top 10 global threats of 2019.

Highly contagious, measles is not a malady to be taken lightly. One in 1,000 people who catch it die of the disease; most of the victims are children under five. Even those who survive are at high risk of falling prey to encephalitis, an often-debilitating infection of the brain that can lead to seizures and mental retardation. Other serious complications of measles include blindness and pneumonia.

While some parents oppose immunization on religious or philosophical grounds, by far the most objections come from those who insist that all vaccines cause disabling side effects or other diseases – even though the available scientific data doesn’t support such claims. As discussed in a previous post, there’s absolutely no scientific evidence for the once widely held belief that vaccination with the MMR (measles-mumps-rubella) vaccine results in autism, for example.

Anti-vaccinationists, when accused of exposing their children to unnecessary risk by refusing immunization because of unjustified fears about vaccine safety, rationalize their stance by appealing to herd immunity. Herd immunity is the mass protection from an infectious disease that results when enough members of the community become immune to the disease through vaccination, just as sheer numbers protect a herd of animals from predators. Once a sufficiently large number of people have been vaccinated, viruses and bacteria can no longer spread in that community.

For measles, herd immunity requires up to 95% of the populace to be immunized. That the threshold is lower than 100%, however, enables anti-vaccinationists to hide their children in the herd. By not vaccinating their offspring but choosing to live among the vaccinated, anti-vaxxers avoid the one in one million risk of their children experiencing serious side effects from the vaccine, while simultaneously not exposing them to infection – at least not in their own community.  

But hiding in the herd takes advantage of others and is morally indefensible. Certain vulnerable groups can’t be vaccinated at all, including those with weakened immune systems such as children undergoing chemotherapy for cancer or the elderly on immunosuppressive therapy for rheumatic diseases. If too many people choose not to vaccinate, the percentage vaccinated will fall below the threshold, herd immunity will break down and those whose protection depends on those around them being vaccinated will suffer.

The figure below shows MMR vaccination rates for U.S. kindergarteners for the 2024-25 school year, the latest period for which data is available. It can be seen that only 10 states (including DC) met the 95% threshold for herd immunity, a drop from just 21 states during the 2019-20 pre-pandemic school year. (Montana and West Virginia didn’t report data for the 2024-25 school year).

MMR coverage rates among kindergarteners for the latest school year varied widely from state to state: from a low of 78.5% in Idaho to a high of 98.2% in Connecticut. In 2024-25 alone, over half the states experienced declines in vaccination rates for all state-required vaccines, including MMR, DTP (diphtheria-tetanus-pertussis), polio, and chicken pox.

A meeting of international health officials is scheduled in November to decide whether the U.S. and Mexico should lose their measles-free status. The analysis period is one year from the onset of recently reported outbreaks: January 20, 2025 in the United States and February 1, 2025 in Mexico. In 2000, the U.S. achieved elimination of local measles transmission by maintaining vaccination rates above 95%, but coverage has clearly declined since then.

It’s true that serious adverse reactions to a vaccine shot do occur occasionally – typically about once in every million vaccinations. But this is a minuscule risk compared with the risk of contracting the disease itself in an unvaccinated community during an epidemic.

Next: How Much of a Threat Are Melting Ice Sheets and Glaciers?

No Evidence That Weather Extremes Becoming More Frequent or Intense: (3) Floods

You’d think that floods would proliferate in a warming world, since precipitation increases due to enhanced evapo­ration from tropical oceans. Yet there is no evidence that flooding is becoming more prevalent.

A 2017 study found that, despite average rainfall getting heavier as the planet warms, excessive precipitation is not the only cause of flooding. What is less obvious is that alterations to the catchment area – such as land-use changes, deforestation and the building of dams – also play a major role. This had even been recognized by the IPCC (Intergovernmental Panel on Climate Change) in its Sixth Assessment Report, which stated that:

In addition to precipitation, flooding also depends on basin and river characteristics such as permeability, antecedent soil moisture, and antecedent flow levels for river flooding, so projections of extreme precipitation and flooding are not always closely linked.

Even so, the 2017 study concluded that the biggest influence on flood trends is none of these factors, but rather the size of the catchment area. While smaller catchments do show a trend in flood risk that’s increasing over time, larger catchments exhibit a decreasing trend.

Globally, larger catchments dominate, so the trend in flood risk is actually decreasing rather than increasing in most parts of the world. This is illustrated in the figure below, the data coming from 1,907 different locations over the 40 years from 1966 to 2005.

Another 2017 study, this time restricted to North America and Europe, found “no compelling evidence for consistent changes over time” in the occurrence of major floods from 1930 to 2010. Like the first study described above, this research included both small and large catchment areas. But the only catchments studied were those with minimal alterations, so as to focus on any trends driven by climate change.

The next figure shows the likelihood of a 100-year flood occurring in North America or Europe in any given year, during two slightly different periods toward the end of the 20th century. Blue dots represent observed values, lines a logistic regression fit to the data. A 100-year flood is a massive flood that occurs on average only once a century and has a 1 in 100 or 1% chance of occurring or being exceeded in any given year – although the actual interval between 100-year floods is often less than 100 years.

You can see that for both periods studied, the probability of a 100-year flood in North America or Europe hovers around the 1% (0.01) level or below, implying that 100-year floods were no more or less likely to occur during those intervals than at any time. The straight lines drawn through the data points show no significant trend. Similar results were obtained for 50-year floods.

Although the study authors concluded that major floods in the Northern Hemisphere between 1931 and 2010 weren’t caused by global warming and were no more likely than expected from chance alone, they did find that floods were influenced by the climate. The strongest influence is the naturally occurring Atlantic Multidecadal Oscillation, an ocean cycle that causes heavier than normal rainfall in Europe and lighter rainfall in North America during its positive phase – leading to an increase in major European floods and a decrease in North American ones.

Floods in the modern era are no more common nor deadly or disruptive than any of the thousands of floods in the past. The figure below shows flood levels of Germany’s Danube River recorded in the town of Passau over more than 500 years. As the caption states, the highest flood level over this period was back in 1501, and only two of the nine severe floods occurred during the last 70 years.

Germany in fact has been battered by devastating floods many times during the last few centuries, including a 1717 Christmas Eve flood that killed 13,700 people. More recently, the Ahr valley in Germany, which was struck in July 2021, experienced major floods in the same locations on 12 June 1910, when at least 52 people were killed.

These numbers pale in comparison with the number of fatalities caused by historical floods in Asia, notably in China, India, Pakistan and Japan. The death toll from past flooding in these countries has often been in the millions, either by drowning or from subsequent famine or disease.

In the more developed world, the illusion that major floods are becoming more frequent is due in part to the growing population and the appeal of living near water. This has led to people building their dream homes in harm’s way, on river or coastal floodplains, where rainfall-swollen rivers or storm surges result in intermittent flooding and subsequent devasta­tion. It is changing human wants rather than climate change that are responsible for disastrous floods.

Next: Measles Surging Worldwide, Thanks to Vaccine Hesitancy

No Evidence That Weather Extremes Becoming More Frequent or Intense: (2) Droughts

Droughts have been a continuing feature of the earth’s climate for millennia. Both short-term and long-term trends show no signs of any increase in frequency or intensity.

In fact, recent droughts have slightly declined on a global scale, as warmer oceans release more moisture into the atmosphere. This is illustrated in the figure below, in which drought conditions from 1982 to 2012 are denoted by D0 (abnormally dry); D1 (moderate); D2 (severe); D3 (extreme); and D4 (exceptional).

The next figure depicts observational data showing the percentage of the contiguous U.S. in drought from 1895 up until 2015. There’s clearly no trend over this period, apart from pronounced drought during the record heat waves of the 1930s.

A 2007 U.S. study was able to reconstruct the drought pattern in western North America over the last 1,200 years, using tree rings as a proxy. The width and color of the rings consti­tute a record of past climate, including droughts. The reconstruc­tion, illustrated in the following figure, reveals that several unprecedently long and severe “megadroughts” have occurred in North America since the year 800, droughts that the study authors re­mark have never been experienced in the modern era.

Comparison of the previous figure with the yearly data in this figure (indicated in gray; the thick black line is a 60-year mean), shows that the long-term pattern of overall drought in North America is featureless, despite global warming during both the Mediaeval Warm Period and today. A similar conclusion was reached by a 2021 study that compared the duration and sever­ity of U.S. hydrological droughts between 1475 and 1899 to those from 1900 to 2014. A hydrological drought refers to drought-induced decreases in streamflow, reservoir levels and groundwa­ter.

That some histori­cal droughts were even longer and more severe than those of today can be seen in another tree-ring reconstruction, of drought in California where I live. The next figure shows the pattern of dry and wet periods in my drought-prone state over the past 1,200 years. Although the third-driest period in the 1100s and the fifth driest period in the 1200s both occurred during the Mediaeval Warm Period, the driest (1500s) and fourth driest (800s) periods of drought occurred during rela­tively cool epochs.

Just like North America, Europe has also experienced mega­droughts over the past millennium, although in different peri­ods. The figure below illustrates a 2021 European reconstruction, again from tree ring proxies, of the drought pattern in central Europe from 1000 to 2012, with observational data from 1901 to 2018 ­superimposed. Black in the figure depicts the PDSI or Palmer Drought Severity Index that measures both dryness (negative values) and wetness (positive values); red denotes the so-called self-calibrated PDSI (scPDSI); and the blue line is the 31-year mean.

The authors of the study point out that droughts from 1400 to 1480 and from 1770 to 1840 were much longer and more severe than those of the 21st century. Their conclusions are reinforced by the results of another recent study, which failed to find any statistically significant trend in meteorological droughts in western Europe during the last 170 years. A meteorological drought describes one arising from a precipitation deficit alone.

What stands out in all these studies is the lack of any long-term trend in drought worldwide over at least a millennium. There is no evidence of the warming that began in the late 19th century, after the Little Ice Age ended, having played any role so far.

Indeed, ice cores from Antarctica demonstrate that much more dust – a sign of a dry climate – was deposited during the ice ages than during warmer interglacial periods. Even the UNDRR (UN Office for Disaster Risk Reduction), in its Special Report on Drought 2021 which drew a link between drought and global warming, acknowledged that one of the main sources of episodic droughts globally is the natural El Niño Southern Oscillation.

There are regional variations, however. A 2021 research pa­per found that, from 1901 to 2017, the risk of meteorological droughts increased in the southwestern and southeastern U.S., while it decreased in northern states. Indeed, the UNDRR report found that such regional differences in drought are not restricted to the U.S. but occur worldwide.

Although generally caused by a severe fall-off in precipitation, droughts can be aggravated by other factors such as elevated temperatures, soil erosion and overuse of available groundwater. The consequences of drought, which can be catastrophic for human and animal life, include crop failure, starvation and mass migration. A major exodus of early humans out of Af­rica about 135,000 years ago is thought to have been driven by drought.

Next: No Evidence That Weather Extremes Becoming More Frequent or Intense: (3) Floods

No Evidence That Weather Extremes Becoming More Frequent or Intense: (1) Heat Waves

Climate reporting of the last few years has routinely promoted the mistaken belief that weather extremes are worsening be­cause of climate change.

But the perception that weather extremes are increasing in frequency and intensity is false. The perception actually results from modern technology – the Internet and smart phones – which has revolutionized communication and made us much more aware of extreme weather than we were 50 or 100 years ago.

To demonstrate how extreme weather has not changed or even declined over time, this and subsequent posts will present observational evidence showing long-term trends over the past century or so for heat waves, droughts, floods, hurricanes, tornadoes and wildfires.

We’ll start with heat waves. It’s commonly argued that heat waves today are hotter than ever before. That is possible, but only because global warming has raised the baseline temperature by 1.3 degrees Celsius (2.3 degrees Fahrenheit) since the Little Ice Age ended around 1850 – so we would expect heat waves to be that much hotter too.

Even so, there’s no evidence that current heat waves are hotter on average than they were in the past. The figure below shows the frequency and magnitude of heat ­waves in the U.S. from 1901 to 2018. The frequency (top panel) is defined as the annual number of calendar days the average U.S. maximum temperature exceeded the 90th percentile for 1961–1990 during at least six consecutive days, in a window centered on that calendar day; it repre­sents the average duration of all heatwaves of six days or longer in that year.

It’s clear that there were far more frequent and/or longer U.S. heat waves, and they were hotter, in the 1930s than in the present era of global warming. The average annual heat wave or warm spell duration (top panel) is seen to have dropped from 11 days during the 1930s to about 6.5 days during the 2000s. The peak heat wave index (bottom panel) in 1936 was a full three times higher than in 2012 and up to nine times higher than in many other years.

In addition, the aver­age maximum temperature during any particular U.S. heat wave has declined slightly from 38 degrees Celsius (101 degrees Fahrenheit) in the 1930s to 37 degrees Celsius (99 degrees Fahrenheit) since the 1980s. This can be seen indirectly in the next figure, which plots the average number of days per weather station with maximum temperatures above 35 degrees Celsius (95 degrees Fahrenheit) for the conterminous U.S. (bars) and regions (lines, 11-year averages).

With the exception of the Pacific southwest (dashed blue line) and the Four Corners states (Arizona, Utah, Colorado and New Mexico, dashed black line), all other regions exhibit a declining trend. The absence of any trend in the U.S. as a whole is clearly evident. 

As I’ve documented in a GWPF report, the hottest years of the 1930s in the U.S. were 1934 and 1936. In the summer of 1934, Fort Smith, Arkansas recorded an unbelievable 53 consecutive days with maximum temperatures of 38 degrees Celsius (100 degrees Fahrenheit) or higher. Topeka, Kansas, had 47 days, Oklahoma City had 45 days and Columbia, Missouri had 34 days when the mercury reached or passed that level. Approximately 800 deaths were attributed to the widespread heat wave, at a time when the U.S. population was about 60% smaller than today.

In comparison, the U.S. heat wave during July 2023, which was falsely trumpeted by the main­stream media as the hottest month in history, did not outmatch the scorching heat of 1934. El Paso, Texas did experience 44 consecutive days with maximum temperatures above 38 degrees Celsius (100 degrees Fahrenheit), somewhat shorter than Fort Smith’s 53 days in 1934 just mentioned. And Phoenix, Arizona saw the maximum there exceed 43 degrees Celsius (109 degrees Fahrenheit) – a comparable baseline for a city with a hotter climate than El Paso – but only for 31 days in a row.

Heat waves lasting a week or longer in the 1930s were not confined to North America; the Southern Hemisphere baked too. Adelaide on Australia’s south coast experienced a heat wave at least 11 days long in 1930, and Perth on the west coast saw a 10-day hot spell in 1933. Not to be outdone, 1935 saw heat waves elsewhere in the world, including India, France and Italy.

That today’s heat waves are nothing extraordinary is apparent from the following table, which depicts all-time record high tem­peratures for the seven continents. Three records date from the 1930s or before, while only Europe and Asia have set new records in the 21st century.

The record high temperatures during the deadly recent heat wave in Europe, which ranged up to 41.7 degrees Celsius (107 degrees Fahrenheit), are still lower than the continent’s all-time record in the table, set on the Italian island of Sicily five years ago. And the recent European heat wave’s duration pales in comparison with the U.S. heat waves described above.

Next: No Evidence That Weather Extremes Becoming More Frequent or Intense: (2) Droughts

The Pseudoscience behind Extreme Weather Attribution

As I wrote in a previous post, newly popular extreme event attribution studies are deeply flawed, with fundamental logical and meth­odological errors. Here I examine these failings in more detail.

To begin with, attribution studies rely on computer climate models, the weaknesses of which I’ve described multiple times in these pages (see for example here and here). The models have a dismal track record in predicting the future, or indeed of hindcasting the past. Not only do the majority of models overestimate the warming rate, but they also wrongly predict a hot spot in the upper atmosphere that isn’t there and are unable to accurately reproduce sea surface temperatures and sea-level rise.

Most importantly for attribution studies, the models are poor at hindcasting. That matters because the models are used to compare the present climate with anthropogenic CO₂ emissions to a preindustrial climate without extra CO₂.

In particular, climate models underestimate the warming during the global warming spell from 1910 to 1940, seen in the figure below. If the same deficiency applies to hindcasted temperatures before that period, including those in 1850 which is the baseline preindustrial year for attribution studies, then such studies will underestimate the magnitude of any counterfactual, comparison warming (and probably any other weather characteristic) in preindustrial times.

Another weakness of attribution method­ology is that, in the absence of CO₂, we simply do not know what the “natural” climate would have been. Although we have reliable historic mete­orological data for a few countries such as the UK and the U.S., we do not for most of the other countries on the planet. And historical data for variables such as cloud cover and wind speeds is lacking worldwide.

A further deficiency in event attribution studies is lack of attention to the uncertainty involved in temperature or other measurements, as I mentioned in the earlier post. An example of this is a Grantham Institute attribution study claiming that 2024’s Hurricane Helene that struck the U.S. was 100%, or 2 times, more likely than would have occurred in a preindustrial climate. Helene was a large Category 4 (top wind speed 250 km per hour or 156 mph) hurricane when it made landfall and caused significant damage across Florida and the southeastern U.S.

The next figure shows the path of Hurricane Helene, together with the maximum speed locations of historically observed and simulated hurricanes that followed paths within a 2° radius of Helene landfall. It also includes the attribution study’s estimates of landfall maximum wind speed as a function of “return period”; the return period is the expected interval between successive hurricanes at the same location, which is the inverse of hurricane frequency.

However, the methodology of the Hurricane Helene study is fundamentally flawed. Beyond the general limita­tions of attribution studies, hurricane behavior in particular is poorly reproduced by climate models. This contrasts with heatwaves, for which models are relatively skillful. Furthermore, very few North Atlantic hurricanes have made landfall in the same region of Florida, limiting the amount of observational data available.

The study’s solution to this problem was to include data from hurricanes that passed nearby but never actually made landfall in the same region – historic “near misses,” which are depicted in the top panel of the figure. Such practice is highly questionable at best, deceptive at worst. The blatant dishonesty is exemplified in the bottom panel, where it can be seen that the uncertainty (colored bands) associated with the partly simulated 2024 Helene curve (orange line) embraces the preindustrial curve (blue line); in other words, Helene may have been no more likely than normal.

The lack of attention to uncertainty is evident in the same data. For example, the uncertainty range (blue band) for the preindus­trial curve includes the observational data (black line) for all North Atlantic hurricanes since 1900 – suggesting that there has been nothing very different about Florida landfalling hurricanes, including Helene, over the whole period of observation.

After accounting for this uncertainty, therefore, the Helene study’s conclusions that the return period decreased from 130 years in the preindustrial era to 52 years, and the wind speed increased by 6.1 meters per second or 11%, are meaningless.

That there was indeed nothing exceptional about Hurricane Helene is reinforced by the plot of all Florida landfalling hurricanes since 1850, presented below. Neither hurricanes overall nor major hurricanes, of which Helene was the latest, exhibit any long-term trend. Taken as a whole, the conclusions of this Hurricane Helene attribution study are faulty.

Extreme event attribution studies, as currently conducted, fail to reinforce the mistaken belief that weather extremes are rising due to global warming – as I will demonstrate in subsequent posts.

Next: No Evidence That Weather Extremes Are Becoming More Frequent or Intense: (1) Heat Waves

RIP RCP8.5: A Hopeful Sign That the Attack on Science Is Letting Up

Nowhere is the attack on science more evident than in the debate over the human contribution to current global warming. But the recent publication of brand-new 2026 emissions scenarios that finally killed off the older RCP8.5 – which was behind the highly exaggerated notion of a “climate crisis” and other alarmist proclamations – raises hopes that reason may be returning to science.

The introduction of RCP8.5 in the early 2010s was probably well-intentioned. As the highest of standard low-, medium- and high-emissions pathways created by climate scientists, RCP8.5 anticipated annual CO₂ emissions roughly tripling by 2100, as seen in the figure below. Even though this was unlikely, it was included as part of an effort to capture all supposedly plausible outcomes – a worst case scenario around the 90th percentile.

But RCP8.5 quickly became adopted by climate alarmists and many in the climate science community as a “business as usual” scenario, which assumed emissions were on an ever-increasing trajectory. It ignored true science that was reflected in the lower emissions pathways.

That RCP8.5 was in fact implausible is evident in its underlying assumptions. First, the scenario predicted a five to sevenfold increase in global coal use that included fueling cars with coal-based liquids, displacing gasoline. As climate writer Roger Pielke Jr. has pointed out, the scenario extrapolated then rapidly rising coal growth rates in China and assumed that such vast amounts of coal would still be available by the end of the century, regardless of geological reality or other real-world constraints.      

The next figure shows that actual coal use started to decline as early as 2015. The dashed line depicts the erroneous, steadily increasing projection of RCP8.5, a false assumption baked into the scenario from the outset.

Another implausible assumption in RCP8.5 was massive population growth, from approximately 7 billion in 2010 to between 11 and 12 billion by 2100. Combined with purported high energy demand and slow income growth, as well as scorching heat, the assumed population gain would result in a lowering of global GDP. However, current demographic models peg the global population at under 10.4 billion by 2080, and socioeconomic trends suggest a much more prosperous world with substantial gains in GDP by then.

A third mistaken assumption behind RCP8.5 was future stagnation in improvements to non-coal technologies such as solar, wind and nuclear power, due to the dominance of coal energy. That hasn’t happened either, with major strides in all three areas.

The figure below compares CO₂ emissions and projected warming between 2000 and 2100 for three different pathways: the now obsolete RCP8.5; the mean of the baseline scenarios published by the IPCC (Intergovernmental Panel on Climate Change) in its Sixth Assessment Report; and one of the new 2026 scenarios.  As can be seen, projected 2100 global warming falls from 4.5 degrees Celsius (8.1 degrees Fahrenheit) in RCP 8.5 to 2.8 degrees Celsius (5.0 degrees Fahrenheit) in the latest representation.

The widespread adoption of RCP8.5 greatly distorted the climate change debate. Hundreds if not thousands of published research papers and government reports that assumed RCP8.5 was realistic took the underlying science in an entirely wrong direction, by coming up with results or predictions that must now be regarded with suspicion.

On the policy side, central banks used it to test whether financial institutions were prepared for the effects of climate change – effects that RCP8.5 exaggerated. Governments around the world used it for adaptation planning, and insurance companies used it to set rates. A whole host of past actions now need to be reexamined in the light of the new emissions scenarios.

But perhaps the greatest damage was to the public perception of climate change. Fed by media hype, people were persuaded that not only was global warming man-made and dangerous, but we were also in a climate crisis. The UN secretary-general was among many duped leaders, hysterically describing the IPCC’s Sixth Assessment Report as “code red for humanity” and proclaiming that “the era of global boiling has arrived.”

Some, like Roger Pielke Jr. and his colleague Justin Ritchie, were not fooled and began questioning conventional wisdom and arguing against the continued use of RCP8.5 almost 10 years ago. Yet these early voices of reason were largely dismissed until now. That’s a sad reflection on the state of science today.

But, as I said at the outset, this major body blow to climate alarmism may represent a turning point in the ongoing attack on science.

Next: The Pseudoscience behind Extreme Weather Attribution

A New Threat to Science: AI Making Up References in Research Papers

It’s not a new phenomenon. AI has been “hallucinating” – making things up – for some time, in areas ranging from law to medicine. But a new example of this affliction has recently emerged: fabricating citations in scientific research papers.

A recent article in the prestigious medical journal The Lancet reported the first systematic evaluation of reference authenticity. The audit of 2.5 million biomedical papers spanning 3 years showed that fabricated references are deeply embedded in the peer-reviewed literature. And the rate of fabrication is accelerating, as demonstrated in the figure below.

The integrity of references is fundamental to the integrity of scientific literature. Each reference implicitly asserts that a verifiable source exists and bolsters the claims being made. When references cite non-existent studies, readers, reviewers and policy makers are unable to evaluate the evidence.

Fabricated references can arise from paper mill activity and intentional deception, as well as the indiscriminate use of AI writing tools. Large language models (LLMs) are capable of generating plausible sounding but fictitious references. Previous studies have estimated that between 30% and 69% of LLM-generated references in biomedical research are fabricated, say The Lancet authors. The references are often correctly formatted, attributed to real researchers and include credible publication dates, making them difficult to detect by conventional peer review.

The authors developed an automated reference verification system by scanning PubMed Central’s Open Access subset from January 1, 2023, to February 18, 2026. That effort examined 2,471,758 papers and 125,615,773 structured references. Of the 125.6 million references, 77% carried a PubMed identifier and were verified; the remaining 23%, predominantly non-indexed references (from sources such as websites and books), were excluded.

Among the verified references, which came from 2810 papers, the authors identified 4046 fabricated citations. In 2023, approximately one in 2828 papers contained at least one fabricated reference. By 2025, this had gone up to one in 458, and in the first 7 weeks of 2026, a staggering one in 277 papers had at least one fabricated reference. As shown in the figure above, the fabrication rate increased by more than 10 times from 2023 to 2026.

It was found that 91% of affected papers contained just one or two fictitious references; 9% contained three or more. The sharp inflection in mid-2024 coincides with the expected publication lag following widespread LLM adoption, although increased paper mill activity and changes in journal indexing practices might also have contributed. LLMs became widely available in late 2022 and 2023.

The article’s authors qualify their findings by stating that the exclusion of 23% of references with no PubMed identifier could bias their estimates. Furthermore, PubMed Central’s Open Access does not cover all biomedical literature, and the early 2026 data span only 7 weeks. And, they say, their approach identifies the problem but not the underlying cause.

Nonetheless, the authors do have some recommendations for publishers. These include integrating automated reference verification into submission workflows before peer review; having indexing services add integrity metadata to article records so that users can assess the reliability of references; and retroactively screening existing publications and issuing corrections or retractions when fabricated references compromise a paper’s conclusions.

On the broader question of why such fabrication is occurring, the answer lies in the training of LLMs. A 2025 article in Science magazine argues that AIs hallucinate because they’re trained to fake answers when they don’t know how to answer a question. A team from OpenAI and the Georgia Institute of Technology is more specific:

They learn to bluff because their performance is ranked using standardized benchmarks that reward confident guesses and penalize honest uncertainty.

Although AIs could be modified to admit that “I don’t know,” this could simply turn off users, who would then seek an alternative LLM. A better solution, say several computer scientists, would be to rework benchmarks to penalize an AI for guessing incorrectly. Current benchmarks don’t penalize incorrect guesses any more than nonanswers.

The need for AI to clean up its act is especially important in medicine. In deciding on patient treatment, medical professionals are often influenced by clinical guidelines in the research literature. Doctors and practitioners have no way of knowing whether the evidence they are relying on actually exists.

In a comment accompanying The Lancet article, two pediatrics professors place the onus on manuscript authors themselves, saying that:

Authors must take responsibility and be held accountable for the entire content of a manuscript, including the references.  

The professors believe that a published manuscript in which a fabricated reference is detected should be retracted. While this may seem harsh, such a move would go a long way toward meeting the AI challenge and restoring the integrity of scientific research.

Next: RIP RCP8.5: A Hopeful Sign That the Attack on Science Is Letting Up

Sea Level Rise Dominated by Subsidence, Not Global Warming

Of all the topics I’ve covered in this blog, none have attracted as many readers as sea level rise (see for example here, here and here). And for good reason, since rising seas threaten the livelihood of tens of millions of coastal dwellers around the world.

What is not generally known is that the main culprit is land subsidence resulting from groundwater extraction and urban development that causes more sea level rise than any human-induced global warming. The importance of subsidence has been brought to light by two recent research papers: a Dutch study finding that current model-based estimates of sea levels are much too low, and a study of river delta subsidence by an international team of environmental scientists and others.

Several different processes can affect vertical motion of land. Long-term glacial rebound after melting of the last ice age’s heavy ice sheets is causing land to rise in high northern latitudes. Yet in many regions, the ground is sinking because of sediment settling and aquifer compaction caused by human activities, especially groundwater depletion resulting from rapid urbanization and population growth.

Land subsidence artificially amplifies any local rise in sea level, which can be measured by either tide gauges or satellite altimetry. Tide gauges measure the height of the sea relative to the land to which the gauge is attached, the so-called RSL (Relative Sea Level) metric. Satellite observations of absolute sea level measure the height of the sea, which is the distance of its surface to the center of the earth.

Subsidence of the land independently of sea level makes satellite-measured sea levels higher than tide gauge RSLs. That’s why the average rate of rise from 1900 to 2020 determined by tide gauges was only 1.75 mm (0.07 inches) per year, while NASA’s satellite measurements say the rate was more like 3.3 mm (0.13 inches) per year from 1993 to 2024.

But neither tide gauges nor satellites measure sea levels at all points on every coastline in the world, which is why so-called geoid models, rather than actual sea level measurements, are used to make estimates. A geoid is an equipotential surface that approximates mean sea level based on gravity and the earth’s rotation.

However, because the actual height of the sea surface is not determined only by the gravity and rotation of the earth, but also by other factors such as ocean currents, winds, tides and seawater temperature, the time-average sea-surface height can deviate strongly (up to several meters) from a geoid. The discrepancy between geoid and actual sea level has led to widespread underestimation of coastal sea level rise.

This occurs predominantly in parts of the world where the most rapid land subsidence is happening – in southeastern and eastern Asia, for example. The Dutch study examined 40 river deltas worldwide subject to severe subsidence such as the Vietnamese Mekong Delta, which is one of the most densely populated coastal landscapes exposed to rising sea levels worldwide.

The figure below shows how much land subsidence has already occurred in the 40 deltas. Yellow and red circles denote subsidence, the circle size indicating the percentage of the delta area subsiding faster than satellite-measured sea level rise, which averages about 4 mm (0.16 inches) per year globally. 

In 13 of the deltas (Brahmani, Brantas, Ceyhan, Chao Phraya, Ciliwung, Godavari, Mahanadi, Mekong, Nile, Po, Red, Vistula and Yellow River), the subsidence rate exceeds 4 mm (0.16 inches) per year. Among these, the Chao Phraya (Thailand), Brantas (Indonesia) and Yellow River (China) deltas show an average sinking rate of more than twice the current rate of global sea level rise.

Overall, more than 90% of the delta area is affected by subsidence in 19 of the 40 deltas studied, while more than 50% of the delta area is subsiding in 38 of the 40. Coastal cities in the deltas are experiencing subsidence rates even greater than the deltas themselves. In Jakarta (Ciliwung), for example, parts of the city are subsiding at alarming rates surpassing 30 mm (1.2 inches) per year, dwarfing average global sea level rise by almost an order of magnitude. 

The effect of land subsidence in river deltas on local populations is demonstrated in the next figure. The upper left panel (a) compares rates of local sea level rise and subsidence; current average subsidence rates exceed sea level rise in 18 of the 40 deltas, including the Mekong and Nile. The lower left panel (b) makes the same comparison, but only for populations living within 1 meter of sea level. Panel c on the right is a bar plot of subsidence vs sea level rise for 30 of the 40 deltas.

Of the various possible contributions to subsidence, groundwater storage has the strongest relative influence in 10 of the 40 deltas. The other drivers are sediment flux and urban expansion.

Next: A New Threat to Science: AI Making Up References in Research Papers

Uncertainty in Ocean Heat Content May Be Grossly Underestimated

According to measurements made by Argo profiling floats, the amount of heat stored in upper ocean layers has been rising rapidly over the past few decades. But a new paper casts doubt on this assertion, claiming that uncertainty in the measurements is so large that any increase in stored heat is statistically indistinguishable from zero.

Argo floats have been used since 2004 to study ocean temperatures and other properties at depths up to 2 km (1.2 miles), by means of a global array of over 4,100 robotic cylinders. As illustrated in the figure below, these free-drifting floats patrol the oceans, taking a deep dive every 10 days to probe the temperature, salinity and pressure of the watery depths, and then transmitting the data to a satellite within hours of bobbing up to the surface again. A recent map of the Argo array is shown in the second figure.

One of the principal sources of uncertainty in the float measurements is evident from the first figure above – namely, the large volume of seawater sampled during a float’s traverse to record just a single reading. That reading is an average of measurements made approximately every second during the float’s 6-to-10-hour ascent from depth to surface.

During the ascent stage, the exact horizontal position when the temperature is measured is highly uncertain because of depth-varying currents, mesoscale eddies and other turbulent motions. The study authors estimate the horizontal uncertainty to be as much as 5-50 km (3-30 miles), stating that the float’s path as it ascends is “an unknown and complex, meandering curve through a spatially and temporally varying flow field.”

Furthermore, the average reading during ascent is considered to be representative of the whole volume encountered by the float since its initial descent. The positional uncertainty in a single temperature measurement, therefore, can cause an associated uncertainty in temperature of 0.5-3 degrees Celsius (0.9-5.4 degrees Fahrenheit), say the researchers.

The uncertainty in Argo float temperature measurements leads to even more uncertainty in estimating OHC (ocean heat content), which is the mathematical integral of absolute temperature (usually expressed as an anomaly from some mean) multiplied by seawater density multiplied by ocean specific heat capacity. Density and heat capacity vary with depth and other factors such as large-scale currents.

Another major source of uncertainty is sampling error. The total global volume sampled by all Argo floats in one year is about nine orders of magnitude smaller than the volume of the upper ocean (to a 2 km depth). The associated uncertainty in OHC estimates is approximately 0.1-0.5 watts per square meter, the researchers say – a level from sampling error alone comparable to the total uncertainty in OHC of 0.2 watts per square meter cited by the IPCC (Intergovernmental Panel on Climate Change) in its 2021 AR6 (Sixth Assessment Report).

For the deep ocean, sampled by a paltry 315 Argo floats, they estimate the sampling error uncertainty to be as large as 0.35 watts per square meter from 1971 to 2018, which is half the IPCC’s actual OHC value of 0.7 watts per square meter.

The uncertainty in temperature measurement of 0.5-3 degrees Celsius (0.9-5.4 degrees Fahrenheit) results in an OHC uncertainty of up to 0.1 watts per square meter. Additional sources of uncertainty include unresolved mesoscale variability, altimetry uncertainty and baseline date range, among others.

By mesoscale variability, the researchers are referring to uncertainties in boundary currents and eddy-rich regions. Certain areas of the global ocean such as the Gulf Stream and the Antarctic Circumpolar Current exhibit intense fluctuations, on a scale of 10-200 km (6-125 miles), in temperature and salinity driven by eddies and other instabilities. The researchers estimate that these fluctuations contribute 0.6-1.2 watts per square meter of uncertainty to the OHC.

Altimetry uncertainty arises from the discrepancy between changes in sea level estimated from OHC, and changes determined by satellite altimetry. Altimetry deduces absolute sea level by measuring the height of the sea, which is the distance of its surface to the center of the earth. The much less direct OHC method relies on Argo float measurements of seawater temperature and salinity, the latter of which is related to density. The researchers assign an OHC uncertainty of 0.29-0.37 watts per square meter to this cause.

And baseline date range uncertainty refers to possible errors in the calculated OHC due to any shift in the reference period, for example from 2004-18 to 2005-19. Such a shift can obscure the influence of different phases of ocean cycles such as the PDO (Pacific Decadal Oscillation) or AMO (Atlantic Multidecadal Oscillation). This can produce an OHC uncertainty of 0.1-0.3 watts per square meter, say the paper’s authors.

Collectively, these and other sources can contribute a total uncertainty of 1 watt per square meter, which exceeds the IPCC’s estimate of 0.7 watts per square meter for the OHC itself.

Next: Sea Level Rise Dominated by Subsidence, Not Global Warming

The Evidence Shows We Are Nowhere Near a Climate Tipping Point, As Alarmists Claim

A recent paper raises the specter of the earth’s climate being on a “hothouse trajectory” – a pathway in which self-reinforcing feedbacks push the climate system past a point of no return, an irreversible disaster beyond which the planet would become unbearably hot. But a careful look at the evidence reveals this claim to be absurd, with no sign that we are currently anywhere close to such a tipping point.

This is not a new form of alarmism. In fact, the paper’s lead author published another paper over six years ago titled “World Scientists’ Warning of a Climate Emergency,” perhaps the beginning of the recent obsession with the erroneous notion of a climate crisis caused by global warming. And a new NGO (non-governmental organization), Global Tipping Points, has in 2023 and 2025 produced fearmongering reports on tipping points.

That we are approaching a tipping point, insists the new paper, is evident from an imagined surge in extreme weather that the paper claims is becoming “more frequent, intense, and costly.” However, the observational evidence shows that most forms of extreme weather are becoming neither more frequent nor more intense, as I’ve demonstrated numerous times in these pages (see Category “Weather extremes”). The increasing costs of natural disasters are simply a result of population gain and the ever-escalating value of property in harm’s way.

The NGO’s 2025 report features the following summary of supposed global-warming-enhanced tipping points faced by our earth. The vertical bars represent the range of anticipated temperature increases that would trigger various tipping points. The report’s purported nearness of tipping points is reflected in the lower limits of all the bars being in the current level of warming band.

Nevertheless, it’s not difficult to show that none of these tipping points – nor several others cited in the report – are imminent. I’ll discuss just three here: coral reefs, ice sheets, and the AMOC (Atlantic Meridional Overturning Circulation).

According to the figure above, die-off of low-latitude coral reefs has already begun, with an estimated tipping point of 1.2 degrees Celsius (2.2 degrees Fahrenheit) above preindustrial temperatures. This shortsighted claim, probably based on temporary losses in global coral cover during the recent period of elevated sea surface temperatures, is irrational.

Australian physicist and leading coral reef authority, Professor Peter Ridd has explained in a 2023 report that most corals that bleach due to higher temperatures do not die, but are capable of rapid recovery in a decade or less. This is exemplified by studies of Australia’s Great Barrier Reef, which has the most reliable long-term record of large-area coral cover. Despite four supposedly catastrophic bleaching events in the six years prior to 2022, the reef’s coral cover reached a record high in 2024, as depicted in the image on the left below.

The image on the right shows the estimated global average cover of hard coral (solid line) and its associated uncertainty (shaded areas) since the late 1970s. Note that data before the late 1990s is of little value, says Ridd, because of small sample sizes; but the data since then reveals little overall variation – certainly nothing suggestive of a tipping point, either already passed or impending.

As for likely collapse of the West Antarctic ice sheet, there’s no evidence that such a catastrophic event is just around the corner either. As I discussed in a 2025 post, the Antarctic ice sheet overall is growing, and no longer melting, for the first time in decades. This is illustrated in the figure below, which shows changes in Antarctic ice sheet mass from April 2002 to December 2023, measured in billions of tonnes (gigatonnes, Gt where 1 gigatonne = 1.102 U.S. gigatons).

Although the ice sheet does appear to be growing in East Antarctica, the ice loss there in the form of melting glaciers is partly caused by active volcanoes underneath the continent. There’s no evidence that the East Antarctica portion of the ice sheet is anywhere near collapse.

Finally, I also discussed the very unlikely slowing down, let alone collapse, of the AMOC in a very recent blog post. All claims of impending doom rely on computer climate models, which have a generally poor history of making predictions. Although some models do indeed support the existence of a weakened AMOC, such cherry picking is highly unscientific and many of the ignored models in fact simulate a strengthened AMOC.

The next figure, from the new paper, graphs the mean global temperature over past millennia, with future projections based on so-called SSPs (Shared Socioeconomic Pathways) that range from low- to high-emission scenarios. As climate writer Roger Pielke Jr. has emphasized on many occasions (see here, for example), high-emission scenarios such as SSP5-8.5 are implausibly extreme. More realistic scenarios such as SSP1-2.6 or even SSP2-4.5 will produce only modest warming in the near future, with no likelihood of triggering tipping points.

Next: Uncertainty in Ocean Heat Content May Be Grossly Underestimated

Abuse of Science: Extreme Event Attribution Studies

As its header suggests, this blog’s purpose is to combat today’s abuse and rejection of true science – the hallmarks of which are empirical evidence and logic. These two pillars are being outrageously abused in climate science by the ever-increasing use of extreme event attribution studies, which attempt to describe the extent to which specific extreme weather events are influ­enced by anthropogenic climate change.

Both the mainstream media and government reports have latched on to event attribution to fan the flames of climate alarmism and the associated narrative that we are in a climate crisis. But such studies are deeply flawed, with errors in both science and inter­pretation, having been created for legal and political rather than scientific reasons.

One of the fundamental failings of attribution studies is the logical fallacy of “begging the question.” To estimate the effect of a particular extreme event, the studies need to replicate the climate without the impact of global warming.

But this ignores the role of natural variability, which is neglected in climate models used in attribution studies on the assumption that only anthropogenic CO₂ – with no contribution from natural sources – accounts for all current warming. If natural sources play no role in warming today, then their contribution to a preindustrial climate is unknown.

Other scientific shortcomings are a lack of appropriate peer review, faulty statistics and the ignoring of important evidence. Defective methodological practices include the improper use of temperature datasets, a lack of sufficient attention to uncertainties in the data, and the neglect of historical records.

In a new GWPF report, I’ve discussed several recent attribution studies in detail. One example is a heatwave concentrated in the U.S. southwest, Mexico and Central America in May and June 2024, which generated alarmist media headlines proclaiming the event “35 times more likely” than before. The media reports were fed by an attribution study conducted by the Grantham Institute at Imperial College, London.

Apart from the general limitations of attribution studies, the absurdity of such a claim can be seen in the conclusions drawn from the temperature datasets employed. The figure below shows the May to June maximum of 5-day maximum daytime temperatures in the study region from 1950 to 2024, for all three datasets; the dashed lines are 10-year moving averages.

To begin with, the study’s estimate of the region’s 2024 heatwave being 35 times more likely than in the preindustrial past is an average estimate for all three temperature datasets. But its estimate for only the ERA5 data (the European Centre for Medium-Range Weather Forecasts 5th generation reanalysis product), which includes observations going back to 1950 rather than only 1979 in the other two datasets, is a much lower 13 times more likely.

Another reason this attribution study is in error is that it does not emphasize the uncertainty involved. It can be seen from the figure that when uncertainties in the temperature measurements are taken into account, a heatwave possibly com­parable to 2024 occurred in the early 1990s; the temperature uncertainty is indicated by the height of the boxes for each data point.

A second example is a devastating flash flood that swept through holiday camps and homes in central Texas on July 4, 2025. An attribution study conducted by ClimaMeter, a French counterpart of the UK’s Grantham Institute, concluded that “Natural variability alone cannot explain the changes in precipitation associated with this very exceptional meteorological condition.”

ClimaMeter attributed the catastrophic flooding to two factors: a purported temperature increase of up to 1.5 degrees Celsius (2.7 degrees Fahrenheit) in the flood-affected area, from the period of 1950–86 in the past to the more recent period of 1987–2023; and present-day rainfall up to 2 mm (0.08 inches) per day, or up to 7% wetter than in the past, in parts of central Texas.

But these assertions don’t stand up to scrutiny. First, the reference period of 1950–86 includes 25 of approximately 35 years of global cooling from 1940 to 1975, so that the estimated temperature increase for central Texas between 1950 and 1986 is most likely inflated.

Second, heavier rainfall does not necessarily lead to an increase in extreme flooding probability, which depends on other factors such as rainfall duration, landscape and the type of river basin. In fact, disastrous floods in Texas’ Flash Flood Alley are nothing new and have caused devastation for more than a century, as shown in the following table.

As my report discusses, extreme event attribution was developed because of the inability of the IPCC (Intergovernmental Panel on Climate Change) to achieve high or even medium confidence in the detection and attribution of most types of extreme weather events. This has been politi­cally problematic for climate activists, who have therefore pushed for rapid extreme event attribution in order to pursue climate litigation against fossil fuel companies.

The origins of this climate lawfare date back to a 2012 meeting of U.S. environmental advocates, climate scientists and others, which attempted to mimic strategies from the 1960s campaign against smoking tobacco.

More details of several studies and of the history can be found in the report itself.

Next: The Evidence Shows We Are Nowhere Near a Climate Tipping Point, As Alarmists Claim

Proxy Evidence Shows Early Holocene Was Warmer than Today

A persistent claim in the climate science literature, one that is used to bolster the narrative of catastrophic anthropogenic climate change, is that current temperatures are the highest the world has seen in the 125,000 years since the interglacial period between the last two ice ages. But a close look at the evidence reveals it was warmer than today as recently as the early Holocene – the geological epoch since the last ice age ended.

Underlying the preposterous claim is an erroneous temperature graph featured in the 2021 Sixth Assessment Report of the IPCC (Intergovernmental Panel on Climate Change). As shown in the figure below, the report displays the IPCC’s latest version of the so-called “hockey stick,” which exhibits no change or a slight decline in temperature for the past 2020 years and a sudden, rapid upturn since 1900.

The solid grey line from 1 to 2000 is a reconstruction of global surface temperature from paleoclimate proxies, while the solid black line from 1850 to 2020 represents direct observations. Both are relative to the 1850–1900 mean and averaged by decade.

Among other features missing from the spurious hockey stick are two previously well-documented aspects of our past climate: the MWP (Medieval Warm Period) around the year 1000, a time when warmer than normal conditions were reported in many parts of the world, and the cool period centered around 1650 known as the LIA (Little Ice Age).

But the IPCC report also excludes early Holocene warming by its statement “Warmest multi-century period in more than 100,000 years” in the figure above. That such warming took place is evidenced by numerous paleoclimate proxies such as tree rings, marine sediments, ice cores, boreholes and leaf fossils.

A 2009 research paper provides proxy evidence from Greenland ice cores of a period known as the Holocene Thermal Maximum that occurred from approximately 10,000 to 6,000 years ago. The researchers used the isotopic ratio of ¹⁸O to ¹⁶O, or δ¹⁸O, in ice cores from six different locations as a proxy for past surface temperatures in Greenland; the locations are indicated in the image on the left below.

The lower part of the image on the right shows the proxy temperature at two of the locations, Agassiz and Renland, over the period from the present back to 11,700 years ago. (Note that the time scale is reversed relative to the IPCC figure.) Of the six Greenland sites, only these two were suitable for analysis, because the ice cores were retrieved from icecap domes and were not influenced by ice flow that occurred at the other four sites. The existence of elevated temperatures during the Holocene Thermal Maximum is abundantly clear.

On the other side of the globe, further proxy evidence for the Holocene Thermal Maximum comes from marine sediment cores in Indonesia. A 2013 paper deduced that temperatures at both the sea surface and at intermediate depths in passages between the Pacific and Indian Oceans were higher during the early Holocene than in the past century.

The proxy in this case was the ratio of magnesium to calcium in the sediment cores, which is a measure of the seawater temperature at the time the sediment was deposited; a timeline can be established by radiocarbon dating. Temperatures dating back 16,000 years derived from well-dated sediment cores in several locations are depicted in the next figure. Part A denotes sea surface temperatures, while part B denotes sub-surface temperatures at the thermocline that divides warmer surface water from cooler water below.

Again, it can be seen that temperatures were higher during the Holocene Thermal Maximum than at present, more so at depth than at the surface. The magnitude of the Maximum relative to the 1850-80 mean was estimated to be 2.2 degrees Celsius (4.0 degrees Fahrenheit) and 1.5 degrees Celsius (2.7 degrees Fahrenheit) at depth and the surface, respectively.

Both the Greenland and Indonesian (Makassar Strait at 500 meters depth) proxy temperature data during the Holocene are displayed together in the figure below, compiled by petrophysicist Andy May. You can see that both datasets are closely matched and that the Holocene Thermal Maximum (Climatic Optimum) was more than 1 degree Celsius (1.8 degrees Fahrenheit) warmer than today.

Likewise, sea surface temperatures near the southern Australian coast are estimated to have been as much as 4.0 degrees Celsius (7.2 degrees Fahrenheit) higher during the mid-Holocene than today. This was deduced from δ¹⁸O and radiocarbon measurements in mollusc fossils, which were reported in a recent paper by Australian scientists.

It could be argued that all of the assertions I’ve made here rely on proxy, not actual data. But so does the shaft of the hockey stick. Reliable global instrumental measurements of temperature and other climate variables date back only about 150 years.

Next: Abuse of Science: Extreme Event Attribution Studies

The Atlantic “Cold Blob” – Cause for Alarm or Just a Curiosity?

In contrast to global warming, a little-known cooling trend has existed in the subpolar North Atlantic Ocean for the past century. Designated the “cold blob” or occasionally the “warming hole” (although this moniker actually refers to a different phenomenon), the trend appears on maps as a cooling blue dot surrounded by a sea of red heat, as shown in the figures below.

Even as sea surface temperatures in the Atlantic as a whole have risen considerably in recent years, the blob’s average temperature has steadily fallen by as much as 0.3 degrees Celsius (0.5 degrees Fahrenheit). Why?

The popular explanation among climate scientists is a weakening of the AMOC (Atlantic Meridional Overturning Circulation). The AMOC forms part of the ocean conveyor belt that redistributes seawater and heat around the globe. Exactly how global warming is changing the AMOC, if at all, is currently unknown. But alarmist proclamations in both the scientific literature and the mainstream media have suggested the AMOC may be slowing or even approaching a tipping point when it would shut down altogether.

However, all claims of impending doom rely on computer climate models, which have a generally poor history of making predictions. Apart from correctly predicting a 20-year pause in loss of Arctic sea ice, the models greatly exaggerate future warming, overestimate low cloud feedback, and are unable to reproduce observed sea surface temperatures and sea-level rise.

Two recent papers, published almost simultaneously in 2025, attempt to link the cold blob to a weakened AMOC using model simulations. In the first paper, two scientists from the University of California, Riverside compared long-term records of the Atlantic Ocean to various climate models. The second paper, by a team of researchers from Pennsylvania State University, focused on both atmospheric and oceanic contributions to model behavior.

Yet what’s striking about both papers is that climate models do not universally support either the existence of a cold blob in the North Atlantic or a weakened AMOC. The Pennsylvania State paper finds that of 32 CMIP6 models, only 11 exhibit long-term cooling in the subpolar North Atlantic, while 9 models actually “simulate greater warming in the subpolar basin compared to elsewhere in the North Atlantic”; the authors have no comment on the other 12 models.

The Riverside researchers observe that a majority of older CMIP5 models simulate a weakened AMOC historically, but a majority of newer CMIP6 models simulate a strengthened AMOC during the 20th century. Of 94 CMIP5/6 models that they utilized, 51 hindcasted a weakened AMOC from 1900 to 2005, while the other 43 models did just the opposite by hindcasting a strengthened AMOC.

Both papers ignore the models that don’t simulate the cold blob or a weakened AMOC in drawing their conclusions – cherry picking already dubious models in order to bolster their erroneous assertions of an AMOC slowdown.

That the cold blob actually exists is not in doubt, given the empirical evidence. The figure at the top of this post shows the changes in observed sea surface temperature from 1901 to 2022; at the blob’s center, the temperature can be seen to have dropped by more than 0.3 degrees Celsius (0.5 degrees Fahrenheit) over that time.

The figure immediately above shows the temperature trend per century over the area of the blob (blue dots) for three different temperature datasets. The trends for the HadISST, ERSSTv5 and Kaplan datasets are -0.04 degrees Celsius/century, -0.14 degrees Celsius /century, and -0.26 degrees Celsius /century, respectively.

Attempts in both recent papers to correlate the century-long trend of the mean observed sea surface temperature in the cold blob with the corresponding modeled trend in the strength of the AMOC are unimpressive. The Pennsylvania State researchers find an R² correlation of only 0.50, while the Riverside team estimates a correlation from 0.42 to 0.57.

To explain the low correlation, the Pennsylvania State authors point to atmospheric influences on the AMOC. They say that heat losses from the ocean in the subpolar North Atlantic induce cooling and drying in the lower troposphere, both of which decrease downward longwave radiation and further intensify sea surface cooling.

The figure above depicts the modeled effect of seven different physical processes, both radiative and oceanic, on the long-term temperature trend in the cold blob. The processes are, from left to right: surface albedo feedback; longwave cloud radiative forcing; shortwave cloud radiative forcing; shortwave clear-sky radiation, mostly from aerosols; longwave clear-sky radiation, which includes the effect of water vapor and other greenhouse gases; ocean heat transport; and surface latent and sensible heat fluxes.

Partial cancellation of both radiative and oceanic effects is evident, but overall the authors propose that radiative processes from the atmosphere contribute about two thirds of cold blob cooling, while oceanic processes contribute the other third.

Nevertheless, neither study should set the alarm bells ringing, given the selective choice of climate models discussed above.

Next: Proxy Evidence Shows Early Holocene Was Warmer than Today

AI Proves Its Mettle in Reconstruction of Antarctic Temperatures

In a 2025 post, I described the largely unsuccessful attempt of an AI to spearhead research in climate science. Now, however, another AI appears to have succeeded in the more technical task of accurately reconstructing surface air temperatures across Antarctica – something that standard temperature datasets have been unable to achieve. The work is reported in a recent paper by a team of Chinese researchers.

The figure below illustrates the geographical distribution of available observational data in Antarctica from 1979 to 2023. The data comes from a number of manned and automatic weather stations, together with meteorological observations over the ocean collected from ships and buoys. As can be seen, the majority of observations are in coastal or near-coastal regions, precluding full spatial coverage of the continent.

To overcome this shortfall, the limited station observations have traditionally been interpolated using various reanalyses. But it’s difficult for reanalysis datasets to capture complex spatial patterns, say the researchers, and such datasets often contain significant uncertainties. Moreover, deriving surface air temperatures from reanalysis datasets depends in part on model simulations rather than actual instrumental measurements.

In the light of these limitations to an accurate reconstruction of Antarctic temperatures, the Chinese research team has applied deep learning methods. This approach has already been utilized successfully to reconstruct Arctic temperatures.

Antarctic temperatures were reconstructed using daily surface air temperature data from the various sources depicted in the figure above. Daily average temperatures were calculated from observations made at 3-hour intervals for some sources, 1-hour intervals for others. Training data for the deep learning model was provided by surface air temperatures from the three reanalysis datasets that showed the best agreement with observed temperatures.

The training data, which covered the period from 1979 to 2005, totaled 29,211 daily temperature samples. Using reanalysis data from different time periods, such as 1995 to 2021, as the training set had little impact on the temperature reconstruction. Validation of the training data and testing of the reconstructed temperature set employed reanalysis data from 2006 to 2012 and 2013 to 2018, respectively.

Testing of the reconstructed Antarctic temperatures was conducted for three specific days in 2015: January 1, July 1 and November 1. For these days, the reconstructed temperatures were found to be highly correlated with their reanalysis counterparts, with spatial correlation coefficients >0.99. The researchers say this correlation shows that the trained deep learning model is capable of accurately reproducing Antarctic surface air temperatures, even with the limited observational data available.

Just how different the reconstructed surface temperatures are from global observational temperature datasets for Antarctica is depicted in the next figure. The figure shows linear trends in annual Antarctic surface air temperatures from 1979 to 2023, measured in degrees Celsius per decade. The datasets are: (a) this reconstruction; (b) Berkeley Earth; (c) ERA reanalysis; (d) NOAAGlobalTemp5.1; (e) GISTEMPv4; and (f) HadCRUT5.

You can see that none of the standard datasets exhibit the pronounced cooling trend in East Antarctica in (a), something that was inferred earlier from the ERA reanalysis dataset by a different group of Chinese researchers. Nevertheless, all datasets show warming in the Antarctic Peninsula (on the left of the continent in the maps above).

Differing from the annual trends is the pattern for the summer months (November to April) only, presented in the figure below for the period from 1989 to 2022. Although the cooling trend still dominates in East Antarctica, warming is no longer prominent in the Peninsula but is found in West Antarctica and the southern portion of East Antarctica.

East Antarctica actually experienced a summer heat wave in 2022, when the temperature soared to -10.1 degrees Celsius (13.8 degrees Fahrenheit) at the Concordia weather station, located at the 4 o’clock position from the South Pole, on March 18. This balmy reading was the highest recorded hourly temperature at that station since its establishment in 1996, and 20 degrees Celsius (36 degrees Fahrenheit) above the previous March record high there. Remarkably, the temperature remained above that record for three consecutive days, including nighttime.

But Antarctica is nothing if not unpredictable. Despite the 2022 heat wave, the mercury dropped to -51.2 degrees Celsius (-60.2 degrees Fahrenheit) on January 31, 2023. This marked the lowest January temperature recorded anywhere in Antarctica since the first meteorological observations there in 1956. Two days earlier on January 29, the nearby Vostok station, about 400 km (250) miles closer to the South Pole, registered a low temperature of -48.7 degrees Celsius (-55.7 degrees Fahrenheit), that location’s lowest January temperature since 1957.

Such swings from record highs to record lows remain a puzzle, but the present reconstruction at least helps to characterize long-term trends.

Next: The Atlantic “Cold Blob” – Cause for Alarm or Just a Curiosity?

Challenges to the CO2 Global Warming Hypothesis: (13) Global Warming Entirely from Declining Planetary Albedo

One of my early posts in this series on challenges to the CO₂ global warming hypothesis featured a paper claiming that the greenhouse effect doesn’t exist. Now the same authors appear to have sidestepped that claim in proposing instead that current global warming can be explained entirely by the observed decrease in planetary albedo over the last few decades.

In a new paper, U.S. research scientists Ned Nikolov and Karl Zeller compare empirical observations of albedo made by CERES (Clouds and the Earth’s Radiant Energy System) satellites to the output of a semiempirical mathematical model used in their earlier work. In their model, the radiative effects integral to the greenhouse effect are replaced by a thermodynamic relationship between air temperature, solar heating and atmospheric pressure, analogous to compression heating of the atmosphere.

Nikolov and Zeller conclude that the observed decrease of planetary albedo, together with variations of solar output or TSI (total solar irradiance), account for 100% of the global warming trend as well as 83% of the interannual variability in global surface air temperature. They find that temperature changes are driven primarily by albedo changes, with only a marginal contribution from TSI.

Albedo is a measure of the earth’s ability to reflect incoming solar radiation. Melting of light-colored snow and sea ice due to warming exposes darker surfaces such as soil, rock and seawater, which have lower albedo. The less reflective surfaces absorb more of the sun’s radiation and thus push temperatures higher.

The same effect occurs with low-level clouds, which are the majority of the planet’s cloud cover. Low-level clouds such as cumulus and stratus clouds are thick enough to reflect 30-60% of the sun’s radiation that strikes them back into space, so they act like a parasol and normally cool the earth’s surface. But less cloud cover lowers albedo and therefore results in warming.

The figure below shows CERES measurements of global albedo from 2000 to 2024, expressed as a percentage anomaly of the 2001 to 2022 mean. The albedo decrease since 2000 is approximately 0.79%, which implies an increase in absorbed shortwave solar radiation of 2.7 watts per square meter over that period. The two researchers point out that this absorption is very close to the total anthropogenic forcing of 2.72 watts per square meter from 1750 to 2019 estimated by the IPCC (Intergovernmental Panel on Climate Change) in its AR6 (Sixth Assessment Report). 

The decline in albedo corresponds to a similar reduction in low cloud cover studied by a trio of German environmental scientists, which I discussed in a 2025 post. Their data for total (red curve) and low cloud (lower black curve) cover, calculated by the study authors from two separate sets of satellite data, is presented in the next figure. The lowered cloud cover is most prominent in northern mid-latitudes and the tropics.

The next figure shows CERES measurements of absorbed shortwave radiation from 2000 to 2024 along with global surface air temperature, measured as the average of six temperature datasets (HadCRUT5, GISTEMP4, NOAA Global, BEST, RSS and NOAA STAR). The temperature anomaly tracks changes in the absorbed solar flux with an average lag of approximately 4 months. Nikolov and Zeller say this close empirical relationship suggests that temperature is directly controlled by the absorbed solar flux.

The researchers quantify the individual effects of albedo and TSI on surface temperatures by using their semiempirical mathematical model. The following figure compares modeled temperatures based on each of these two factors with observations. It’s clear that variations in TSI play only a minimal role while variations in albedo dominate the temperature trend.

This conclusion is also supported by the next figure, which depicts the modeled contribution to albedo and TSI separately going back to 1981. According to Nikolov and Zeller’s model, the low contribution from TSI variation indicates that the earth’s climate is 5.6 times more sensitive to changes in solar absorption than to TSI variations.

The estimate is in good agreement with a conclusion reached by Israeli physicist Nir Shaviv, who has investigated the effect of TSI cycles on several climatic variables. Shaviv found that total solar forcing of the earth’s climate, which includes indirect effects, is 5 to 7 times larger than that associated with direct solar warming.  Such a large amplification factor would imply that the climate is much more sensitive to the sun than to CO₂. Of the various possible indirect amplification mechanisms, Shaviv favors the influence of cosmic rays on low cloud cover.

Nonetheless, the observation of declining albedo or low cloud cover begs the chicken-and-egg question: is decreased albedo the cause or effect of current warming? While Nikolov and Zeller argue for the former, they also call for large-scale interdisciplinary research into the physical mechanisms controlling the earth’s albedo and cloud physics, in order to resolve the question.

Next: AI Proves Its Mettle in Reconstruction of Antarctic Temperatures