Scientists Report Global Temperature Records Being Broken at Unprecedented Rate and Magnitude

For most of the instrumental temperature record, new annual heat records edged past their predecessors by hundredths of a degree Celsius — increments so small they required decimal places to detect. In 2023 and 2024, that pattern broke. The records weren't just broken; they were, in the words of multiple climate agencies, "smashed" [1][2].

The Copernicus Climate Change Service confirmed that 2024's global mean surface air temperature reached 15.10°C — 1.60°C above the 1850–1900 pre-industrial baseline and 0.12°C above 2023, itself a record [1]. Berkeley Earth placed the figure slightly higher at 1.62°C (±0.06°C) [3]. The World Meteorological Organization, using a consolidated analysis of six independent datasets, estimated the 2024 anomaly at 1.55°C (±0.13°C) [2]. By any measure, 2024 became the first calendar year to exceed the 1.5°C threshold that the Paris Agreement designated as a boundary beyond which climate risks escalate sharply.

One hundred and four countries recorded their warmest annual averages on record in 2024 [3]. Roughly 24% of Earth's surface set local temperature records, affecting 3.3 billion people — 40% of the global population [3]. From June 2023 through August 2024, every single month set a new monthly record: 15 consecutive months of records, a streak NASA called "unprecedented" [4].

Source: NASA GISS / Copernicus / NOAA

Data as of Jan 1, 2025CSV

The Margins Are the Story

In prior decades, when a new warmest year was declared, it typically exceeded the previous record by 0.01 to 0.05°C. The year 2014 beat the prior record by roughly 0.02°C. Then 2015 exceeded 2014 by about 0.10°C, and 2016 topped 2015 by 0.12°C — both boosted by a strong El Niño [1].

What happened in 2023 broke that mold. According to NOAA, 2023 beat 2016's longstanding record by 0.15°C, the largest margin by which any year had surpassed the previous record in the modern era [5]. Then 2024 added another 0.12°C on top of that [1]. In two years, the global temperature record jumped by roughly 0.27°C — a shift that previously required a decade or more of gradual warming to accumulate.

Source: Copernicus Climate Change Service

Data as of Jan 1, 2025CSV

Land temperatures moved even faster. Berkeley Earth measured the 2024 land surface average at 2.28°C (±0.12°C) above the pre-industrial baseline, exceeding 2023 by 0.20°C [3]. Ocean surface temperatures exceeded 2023 by 0.05°C [3].

Disentangling El Niño From the Background Signal

A natural question follows: how much of the 2023–2024 spike was El Niño — the recurring Pacific Ocean warming pattern that temporarily boosts global temperatures — and how much reflects something new?

Three studies highlighted by the American Geophysical Union's journal Eos point to El Niño as a key driver of the 2023 record [6]. Research published in Communications Earth & Environment found that the "atypical warming pattern" of the 2023–24 El Niño, which matured with two distinct spatial peaks of sea surface temperature anomaly in the equatorial Pacific, amplified global temperatures beyond what a typical El Niño of comparable magnitude would produce [7].

But the residual — the warming left over after subtracting El Niño's estimated contribution — is where the picture becomes more complicated. The WMO's update on 2023–2024 anomalies found that 2024's residual was the highest in decades, despite the El Niño event being only moderate in strength [8].

A May 2026 study from the University of Maryland, published in Earth System Dynamics, identified the Indian Ocean Dipole (IOD) — a separate ocean-atmosphere oscillation — as an overlooked factor. When the IOD was excluded from their model, researchers could account for only 69% of the temperature spike in 2023 and 77% in 2024 [9]. The 2023 IOD event was one of the two most extreme since modern recordings began around 1850 [9].

Gavin Schmidt, director of NASA's Goddard Institute for Space Studies, described the 2023 overshoot as "humbling" and "confounding," acknowledging that neither greenhouse gases nor El Niño fully explained the magnitude [10]. Schmidt initially placed the odds of 2023 becoming a record year at just one in five [10].

Is Warming Accelerating? Scientists Disagree

Whether the 2023–2025 spike signals a structural acceleration or a temporary extreme within known variability is an active scientific dispute with credentialed researchers on both sides.

A March 2026 study led by Grant Foster, published in Geophysical Research Letters, reported a "strong and statistically significant acceleration of global warming since around 2015" [11]. After removing El Niño, volcanic eruptions, and solar cycles from the temperature record, the study found the warming rate had increased from approximately 0.2°C per decade (1970–2015) to approximately 0.35°C per decade since 2015 — a 75% increase [11]. The result held across all five major global temperature datasets through 2024, and remained consistent even after accounting for the 2023–2024 records as outliers [11].

Stefan Rahmstorf of the Potsdam Institute for Climate Impact Research called the acceleration "a bit unexpected" but noted it aligned with some IPCC model trajectories [11].

Michael Mann of the University of Pennsylvania disagreed. Mann stated there is "no evidence of any acceleration in the rate of warming over the past 10 years," attributing recent heat spikes to El Niño and stating that "the planet is warming at a roughly constant rate and that's bad enough" [12].

A study published in Nature examining the record sea surface temperature jump in 2023–2024 concluded it was "unlikely but not unexpected" — an extreme event after which ocean temperatures are expected to revert toward the long-term warming trend [13]. This statistical argument holds that while 2023–2024 represents an exceptional departure, it does not by itself revise century-scale projections any more than the 1998 spike (followed by a relative plateau) did.

The year 2025 provides a partial test case. Copernicus confirmed it as the third-warmest year on record, only 0.01°C cooler than 2023 and 0.13°C below 2024's peak [14]. The partial cooling aligns with the transition to ENSO-neutral and weak La Niña conditions. But 2025 remained far above the pre-2023 trend line, and the three-year average for 2023–2025 exceeded 1.5°C — a first [14].

The Shipping Fuel Factor: An Accidental Experiment Ends

One contributor gaining increased scientific attention is the International Maritime Organization's 2020 regulation (IMO 2020), which cut the maximum sulfur content of shipping fuel from 3.5% to 0.5%. This reduced sulfur dioxide emissions from ships by approximately 80%, from over 10 million tonnes of SO₂ per year to roughly 2.5 million tonnes — about 10% of the global total [15].

Sulfur dioxide emissions from ships had been producing aerosols that brightened clouds over shipping lanes and reflected sunlight, effectively masking some warming. When those emissions dropped, the masking effect weakened. Estimates of the resulting radiative forcing vary: Carbon Brief placed it at 0.079 W/m², while a Nature study estimated 0.2 (±0.11) W/m² averaged over the global ocean [15][16]. A study in Earth's Future found that IMO 2020 may have accelerated global warming by up to three years in one climate model simulation [17].

However, a 2025 study in Atmospheric Chemistry and Physics concluded that the surface temperature effects of the shipping sulfur reduction "are within internal variability," suggesting the signal is too small to clearly distinguish from natural fluctuations at the global scale [18]. Carbon Brief's analysis stated that IMO 2020 is "not a primary driver" of recent record ocean temperatures, with only hundredths of a degree having materialized globally by 2023 [15].

The broader implication matters regardless of the precise magnitude: sulfur dioxide emissions globally are estimated to mask between 0.3 and 0.6°C of human-caused warming — roughly 20–40% of the total [15]. As air quality regulations reduce this pollution, a clear public health benefit, the full extent of accumulated warming becomes more visible.

Station Records and Data Integrity

NASA assembles its global temperature record from tens of thousands of meteorological stations plus ship- and buoy-based sea surface temperature measurements. The methodology accounts for uneven station spacing and urban heat island effects [4].

The geographic breadth of the 2023–2024 records — spanning ocean surfaces far from any urban area, polar regions, and tropical zones simultaneously — makes it difficult to attribute the signal primarily to urban heat island growth. In the United States, data from 2024 showed approximately three times more daily heat records than cold records, with 88% of 247 tracked locations (217) setting more heat records than cold records in the 2020s [19]. Cities like Tampa (114 heat records, zero cold records), Phoenix (110), and Miami (102) illustrate the skew, though these are urban locations [19].

Satellite-based temperature measurements, which are independent of surface stations, have broadly corroborated the surface record's trajectory. Some discrepancy exists: the University of Alabama in Huntsville's satellite dataset shows somewhat less warming than surface records. This gap is a subject of ongoing investigation but has not altered the broad conclusion across all major datasets that 2023–2025 rank as the three warmest years on record [3][14].

The Human Toll: 178,000 Excess Deaths in a Single Year

A study published in The Innovation found approximately 178,486 excess deaths attributable to heatwaves globally in 2023, accounting for 0.73% of global deaths — 23 deaths per million people [20]. Of those deaths, 54.29% were attributed specifically to human-induced climate change [20]. The study analyzed data from 2,013 locations across 67 countries.

Asia bore the largest burden, accounting for 48% of global heat deaths (85,611), followed by Europe at 37% (66,443) [20]. In Europe, the toll worsened in 2024: a study in Nature Medicine estimated 62,775 heat-related deaths, significantly exceeding the 2023 burden of approximately 48,000 [21]. European adaptation efforts — including public cooling centers, modified work schedules, and early-warning alerts — have reduced heat-related mortality by as much as 80% where fully implemented [21].

Source: The Innovation / Cell Press

Data as of Jun 1, 2025CSV

The mortality gap between rich and poor nations is stark. Death rates from extreme weather events in vulnerable regions over the last decade were 15 times higher than in less vulnerable regions [22]. Climate Impact Lab projections estimate roughly 391,000 annual deaths in lower-income countries versus 39,000 in higher-income countries due to shifting temperatures — a 10:1 ratio [23]. By mid-century, heat-related deaths in Africa's Sahel could increase by more than 60 per 100,000 people per year in countries like Niger and Burkina Faso — a rate exceeding the continent's current malaria death rate [24].

As of 2024, only 21 of 38 European Economic Area countries had formal heat-health action plans in place [25]. In much of South Asia and sub-Saharan Africa, where vulnerability is highest, such plans remain partial or absent, limited by unreliable electricity, lack of cooling infrastructure, and resource constraints [22].

Economic Damage: Billions in Losses, Uneven Burdens

In the United States alone, drought, excessive heat, and wildfires caused over $16.59 billion in crop losses in 2023, with total crop and rangeland losses from major disasters reaching $21.94 billion [26]. Texas cotton damage reached $2.3 billion; Kansas wheat losses exceeded $1 billion [26]. Nearly $10 billion in losses were uninsured [26]. Globally, extreme heat imposed an estimated $6 billion impact on fisheries and aquaculture [27].

Research published in Scientific Reports quantifies the sensitivity: for wheat, each 1°C of temperature increase reduces yields by 6.1% when total warming remains below 2.38°C, and by 8.2% per degree above that threshold — a nonlinear acceleration of losses [28].

These costs fall disproportionately on countries least responsible for emissions. According to the OECD, developed nations mobilized $132.8 billion in climate finance in 2023 and $136.7 billion in 2024, exceeding the long-delayed $100 billion annual target for the third consecutive year [29]. But the distribution remains sharply skewed: low-income countries received only 7% of total climate finance, while loans dominated the financial instruments used, raising debt sustainability concerns for the most climate-vulnerable nations [29]. The new climate finance framework agreed at COP29 calls for $300 billion per year by 2035 [30], while UNEP estimates developing countries will need $310 billion annually for adaptation alone by that date [22].

The 1.5°C Threshold: When Does Temporary Become Permanent?

The Paris Agreement's 1.5°C limit refers to long-term averages, typically over 20-year periods, not individual years. A single calendar year exceeding 1.5°C does not constitute a formal breach [1][2].

But a study published in Nature Climate Change in 2025 examined what 12 consecutive months above 1.5°C — which first occurred from June 2023 to June 2024 — historically signifies within climate model ensembles. In 76% of CMIP6 simulations (CMIP6 being the latest generation of coupled climate models used for IPCC assessments), the long-term 1.5°C threshold had already been crossed by the time a 12-month streak was observed [31]. The median gap between the first 12-month streak and the permanent crossing was negative 33 months — meaning the long-term breach typically preceded the streak by nearly three years [31].

The Copernicus analysis projects the 20-year running average benchmark could be reached by May 2029 [31]. The IPCC's Sixth Assessment Report, published in 2021, had projected a mid-2030s crossing under most emissions scenarios. The WMO now estimates there is a 70% chance the five-year average warming for 2025–2029 will exceed 1.5°C, up from 47% projected in 2024 and 32% in 2023 [32].

Source: OpenAlex

Data as of Jan 1, 2026CSV

What Remains Unexplained

Despite the convergence of known factors — rising greenhouse gas concentrations, El Niño, the Indian Ocean Dipole, reduced aerosol masking, and a solar cycle peak — several scientists have noted that these drivers together still leave a portion of the 2023–2024 warming unexplained. NASA's Schmidt identified changes in low cloud cover as an additional candidate, though the mechanisms remain uncertain [10].

The proposed explanation for the acceleration detected by Foster et al. centers on the reduction in atmospheric aerosols since the early 2000s, driven by air quality improvements that inadvertently removed a warming brake [11]. If correct, this would mean the underlying rate of warming had been partially hidden for decades — and is now emerging as pollution declines.

Whether the 2023–2025 sequence marks the onset of a steeper warming curve or a temporary extreme within a known trajectory is a question that only continued observation can definitively answer. What is not in dispute across any major scientific institution is the direction: upward, and faster than most models anticipated even five years ago.