Last year was the warmest since records began in the mid-1800s – and likely for many thousands of years before.
It was the first year in which average global temperatures at the surface exceeded 1.5C above pre-industrial levels in at least one global temperature dataset.
Here, Carbon Brief examines the latest data across the oceans, atmosphere, cryosphere and surface temperature of the planet.
Noteworthy findings from this 2023 review include…
- Global surface temperatures: It was the warmest year on record by a large margin – at between 1.34C and 1.54C above pre-industrial levels across different temperature datasets.
- Exceptional monthly temperatures: Global temperatures set a new record each month between June and December. September smashed the prior record for the month by a “gobsmacking” 0.5C.
- Warmest over land: It was the first year the global average land temperature was more than 2C above pre-industrial levels.
- Warmest over oceans: It was the first year that global average ocean surface temperatures exceeded 1C compared with pre-industrial levels.
- Ocean heat content: It was the warmest year on record for ocean heat content, which increased notably between 2022 and 2023.
- Regional warming: It was the warmest year on record in 77 countries – including China, Brazil, Austria, Bangladesh, Germany, Greece, Ireland, Japan, Mexico, the Netherlands, South Korea and Ukraine – and in areas where 2.3 billion people live.
- Unusual warmth: 2023 was much warmer than scientists estimated it would be at the start of the year and there remain open questions about what precise factors have driven the exceptional warmth. Even El Niño – the usual suspect behind record warm years – does not clearly explain 2023 temperatures.
- Comparison with climate models: Observations for 2023 are above the central estimate of climate model projections in the Intergovernmental Panel on Climate Change (IPCC) sixth assessment report, but well within the model range.
- Warming of the atmosphere: It was the warmest year in the lower troposphere – the lowest part of the atmosphere. The stratosphere – in the upper atmosphere – is cooling, due in part to heat trapped in the lower atmosphere by greenhouse gases.
- Sea level rise: Sea levels reached new record-highs, with notable acceleration over the past three decades.
- Shrinking glaciers and ice sheets: Cumulative ice loss from the world’s glaciers and from the Greenland ice sheet reached a new record high in 2023, contributing to sea level rise.
- Greenhouse gases: Concentrations reached record levels for CO2, methane and nitrous oxide.
- Sea ice extent: Arctic sea ice saw its sixth-lowest minimum extent on record, while Antarctic sea ice saw a new record low extent for almost the entire year, much of it by an exceptionally large margin.
- Looking ahead to 2024: Carbon Brief predicts that global average surface temperatures in 2024 are most likely to be slightly warmer than 2023 and set a new all-time record. However, large uncertainties remain given how exceptionally and unexpectedly warm 2023 was.
Use the links below to navigate between the article’s sections.
- Warmest year on record for the Earth’s surface
- Pushing up against the 1.5C target
- Highest ocean heat content on record
- A year of climate extremes
- Explaining 2023’s unusual heat
- Observations broadly in-line with climate model projections
- Record atmospheric temperatures
- Greenhouse gas concentrations reach new highs
- Accelerating sea level rise
- Shrinking glaciers and ice sheets
- Record-low Antarctic sea ice levels
- Looking ahead to 2024
Warmest year on record for the Earth’s surface
Global surface temperatures were exceptionally hot in 2023, exceeding the prior record set in 2016 by between 0.14C and 0.17C across different surface temperature datasets. It was unambiguously the warmest year since records began in the mid-1800s.
The figure below shows global surface temperature records from five different datasets: NASA; NOAA; the Met Office Hadley Centre/University of East Anglia’s (UEA) HadCRUT5; Berkeley Earth; and Copernicus ERA5.
Other surface temperature datasets not shown, including JRA-55, the AIRS satellite data and the Japanese Meteorological Agency, also show 2023 as the warmest year on record.
Annual global average surface temperatures over 1850-2023. Data from NASA GISTEMP, NOAA GlobalTemp, Hadley/UEA HadCRUT5, Berkeley Earth and Copernicus ERA5. Temperature records are aligned over the 1981-2010 period and use the average of NOAA, Berkeley and Hadley records to calculate warming relative to the pre-industrial baseline. Chart by Carbon Brief.
Global surface temperature records can be calculated back to 1850, though some groups such as NASA GISTEMP choose to start their records in 1880 when more data was available.
Prior to 1850, records exist for some specific regions, but are not sufficiently widespread to calculate global temperatures with high accuracy (though work is ongoing to identify and digitise additional records to extend these further back in time).
These longer surface temperature records are created by combining ship- and buoy-based measurements of ocean sea surface temperatures with temperature readings of the surface air temperature from weather stations on land. (Copernicus ERA5 and JRA-55 are an exception, as they use weather model-based reanalysis to combine lots of different data sources over time.)
Some differences between temperature records are apparent early in the record, particularly prior to 1900 when observations are more sparse and results are more sensitive to how different groups fill in the gaps between observations. However, there is excellent agreement between the different temperature records for the period since 1970, as shown in the figure below.
Annual global average surface temperatures as in the prior chart, but showing the period from 1970-2023. Chart by Carbon Brief.
Global temperatures in 2023 clearly stand out as much warmer than anything that has come before. This can be seen in the figure below from Berkeley Earth. Each shaded curve represents the annual average temperature for that year. The further that curve is to the right, the warmer it was.
The width of each year’s curve reflects the uncertainty in the annual temperature values (caused by factors such as changes in measurement techniques and the fact that some parts of the world have fewer measurement locations than others).
The year 2023 was the warmest on record for both the world’s land and ocean regions.
It was also the first year where global average land temperatures exceeded 2C and the first year in which global ocean temperatures exceeded 1C relative to pre-industrial levels.
The figure below shows land (red) and ocean (blue) temperatures along with their respective confidence intervals, relative to pre-industrial levels, in the Berkeley Earth surface temperature record.
Global land regions – where the global human population lives – has been warming around 70% faster than the oceans – and 40% faster than the global average in the years since 1970.
While 2023 as a whole has been exceptionally warm, it started off a bit cooler, with the first few months of the year failing to set any new records. However, from June onward each month was warmer than the same month in any prior year since records began. September was particularly “gobsmacking”, shattering the prior September record by a full 0.5C.
The figure below shows each month of 2023 in black, compared to all prior years since 1850. Each year is coloured based on the decade in which it occurred, with the clear warming over time visible as well as the exceptional margin by which 2023 exceeded past years between July and December.
Pushing up against the 1.5C target
In the 2015 Paris Agreement, the world agreed to work to limit global temperatures to well-below 2C and to pursue efforts to “limit the temperature increase to 1.5C above pre-industrial levels”.
While the exceedance of these climate targets was not specifically defined in the agreement, it has since been widely interpreted (for example, by the IPCC) as a 20-year average period.
Crucially, the limits refer to long-term warming, rather than an individual year that includes the short-term influence of natural fluctuations in the climate, such as El Niño.
However, a single year exceeding 1.5C still represents a grim milestone, a sign that the world is quickly approaching the target. And, in the Berkeley Earth dataset, 2023 was the first year above 1.5C.
It came in a hair’s width below 1.5C in the Copernicus and Hadley datasets, at 1.48C and 1.46C, respectively, and was lower on NOAA and NASA datasets as shown in the table below.
| Temperature record | 2023 temperatures relative to preindustrial |
|---|---|
| NOAA GlobalTemp | 1.34C |
| NASA GISTEMP | 1.39C* |
| Hadley/UAE HadCRUT5 | 1.46C |
| Copernicus/ECMWF | 1.48C |
| Berkeley Earth | 1.54C |
Global temperature anomalies for 2023 relative to preindustrial temperatures (1850-1899). *Note that GISTEMP uses a 1880-1899 baseline as it does not cover the 1850-1879 period.
As noted earlier, these datasets are nearly identical over the past 50 years. Differences in warming relative to pre-industrial levels emerge earlier in the record, particularly prior to 1900 when observations are more sparse and the choice of how to fill in the gaps between observations has a large impact on the resulting temperature estimate.
The figure below shows how different temperature records look if each is calculated relative to its own pre-industrial baseline, rather than using an average pre-industrial baseline as shown in the prior section. Focusing on warming since pre-industrial – rather than more recent warming – magnifies differences between groups, with the variation in warming across groups largely due to the most uncertain early part of the record.
Highest ocean heat content on record
Last year was the warmest on record for the heat content of the world’s oceans. Ocean heat content (OHC) has increased by around 473 zettajoules – a billion trillion joules – since the 1940s. The heat increase in 2023 alone compared to 2021 – about 15 zettajoules – is around 25 times as much as the total energy produced by all human activities on Earth in 2021 (the latest year in which global primary energy statistics are available).
Human-emitted greenhouse gases trap extra heat in the atmosphere. While some of this warms the Earth’s surface, the vast majority – around of 93% – goes into the oceans. About two-thirds of this accumulates in the top 700 metres, but some also ends up in the deep oceans.
The figure below shows annual OHC estimates between 1950 and present for both the upper 700 metres (light blue shading) and 700-2,000 metres (dark blue) of the ocean.
Annual global ocean heat content (in zettajoules – billion trillion joules, or 10^21 joules) for the 0-700 metre and 700-2,000 metre layers. Data from Cheng et al. (2024). Chart by Carbon Brief.
In many ways, OHC represents a much better measure of climate change than global average surface temperatures. It is where most of the extra heat ends up and is much less variable on a year-to-year basis than surface temperatures. It shows a distinct acceleration after 1991, matching the increased rate of greenhouse gas emissions and other radiative forcing elements over the past few decades.
This year saw a substantial update to the OHC dataset provided by the Institute for Atmospheric Physics (IAP) that Carbon Brief features in its State of the Climate reports. The transition from version 3 to version 4 introduced a new quality control system to detect and remove spurious measurements across different instrument types.
As the figure below highlights, this results in a notable increase in OHC over the past decade (red lines and shading) relative to the prior version of the dataset (black lines).
A year of climate extremes
While media coverage of 2023 temperatures has largely focused on the global average, there have been many different regions of the planet experiencing climate extremes.
The figure below shows global temperature anomalies in 2023 across the world, with red areas warmer than the baseline period (1951-80) used by Berkeley Earth and blue areas experiencing cooler temperatures.
In 2023, 77 countries saw their warmest year on record, including: Afghanistan, Albania, Antigua and Barbuda, Argentina, Austria, Azerbaijan, Bangladesh, Bhutan, Bolivia, Bosnia and Herzegovina, Brazil, Bulgaria, Cape Verde, Cameroon, China, Comoros, Costa Rica, Croatia, Cuba, Czechia, Dominica, Dominican Republic, Ecuador, El Salvador, Federated States of Micronesia, Gambia, Germany, Greece, Grenada, Guatemala, Guinea, Guyana, Haiti, Honduras, Hungary, Ireland, Ivory Coast, Jamaica, Japan, Kazakhstan, Kiribati, Kosovo, Kyrgyzstan, Liechtenstein, Macedonia, Mexico, Moldova, Montenegro, Morocco, Myanmar, Netherlands, Nicaragua, Nigeria, North Korea, Oman, Panama, Paraguay, Peru, Republic of the Congo, Romania, Saint Kitts and Nevis, Saint Lucia, Saint Vincent and the Grenadines, San Marino, Senegal, Serbia, Slovakia, Slovenia, South Korea, Tajikistan, The Bahamas, Trinidad and Tobago, Turkmenistan, Ukraine, Uzbekistan, Venezuela and Yemen.
Approximately 2.3 billion people, or around 29% of Earth’s population, live in places that observed their locally warmest year during 2023.
The figure below highlights regions of the planet that experienced their top-five warmest (red shading) or coldest (blue) temperatures on record in 2023. Overall, around 17% of the planet set a new record, including 23% of the land and 14% of the ocean. No location on the planet experienced record cold temperatures (or even top-5 record cold temperatures) for the year as a whole.
Explaining 2023’s unusual heat
Scientists did not expect 2023 to be all that exceptional at the start of the year. As Carbon Brief reported at the start of 2023, four different groups provided temperature predictions for the year prior to any data being collected – the UK Met Office, NASA’s Dr Gavin Schmidt, Berkeley Earth and Carbon Brief’s own estimate.
Temperature predictions for 2023 from the UK Met Office, NASA’s Dr Gavin Schmidt, Berkeley Earth and Carbon Brief relative to pre-industrial (1880-99) temperatures. Chart by Carbon Brief.
As Carbon Brief noted in January 2023:
“La Niña conditions are expected to persist for at least the first three months of 2023. Because there is a lag of a few months between when El Niño or La Niña conditions peak in the tropical Pacific and their impact on global temperatures, these La Niña conditions will likely have a lingering cooling influence on 2023 temperatures.”
Carbon Brief estimated that 2023 was “very likely to be between the third and ninth warmest year on record, with a best estimate of being the fifth warmest on record – similar to 2022”, and suggest that if an El Niño develops in latter half of 2023 it would make it likely that 2024 will set a new record.
This estimate, alongside all the other groups predicting 2023 temperatures, was wrong. Not only did 2023 turn out to be the warmest year on record, but it fell well outside the confidence intervals of any of the estimates. And while there are a number of factors that researchers have proposed to explain 2023’s exceptional warmth, scientists still lack a clear explanation for why global temperatures were so unexpectedly high.
Over the longer-term, human emissions of CO2 and other greenhouse gases alongside planet-cooling aerosols are the main driver of global temperatures. Global temperatures have risen by approximately 1.3C since pre-industrial times as a result of human activity. However, on top of long-term warming, global temperatures vary year to year by up to 0.2C.
These variations are primarily driven by El Niño and La Niña events that redistribute heat between the atmosphere and oceans. However, other factors such as volcanic eruptions, the 11-year solar cycle and changes in short-lived climate forcers can influence year-to-year temperature changes.
The figure below, created by Dr Robert Rohde at Berkeley Earth, explores some of the main drivers of temperature change over the past decade.
These include continued accumulation of greenhouse gases, the evolution of El Niña and La Niña, and the 11-year solar cycle. It also includes two new factors that emerged during the decade: the 2022 eruption of the Hunga Tonga volcano and the 2020 phase-out of sulphur in marine fuels. Both of these are estimated to have relatively modest effects at present – less than 0.05C each – but with large scientific uncertainties.
However, both the Tonga eruption and the phase-out of sulphur in marine fuel are problematic explanations of extreme temperatures in 2023.
There is still a vigorous debate in the scientific literature about whether the eruption cooled or warmed the planet based on estimates of both sulphur dioxide and water vapour in the atmosphere, with some papers arguing for warming and others for cooling. Some modelling suggests that the largest impacts of the eruption would be in winter months, which does not match the timing of extreme summer temperatures experienced in 2023.
Similarly, the phase-out of sulphur in marine fuels occurred in 2020. If it had a large climate impact, it would show up in 2021 and 2022 rather than suddenly affecting the record in 2023. While it definitely has had a climate impact – alongside the broader reduction in aerosol emissions over the past three decades – the timing suggests that its likely not the primary driver of 2023 extremes.
Even El Niño – the usual suspect behind record warm years – does not clearly explain 2023 temperatures. Historically global temperatures have lagged around three months behind El Niño conditions in the tropical Pacific; for example, El Niño developed quite similarly in 1997, 2015 and 2023. But it was the following year – 1998 and 2016 – that saw record high temperatures.
The figure below shows the El Niño (red shading) and La Niña (blue) conditions over the past 40 years (collectively referred to as the El Niño-Southern Oscillation, or “ENSO”). While not unprecedented, the extended La Niña conditions since the latter part of 2020 have extended for an unusually long period of time.
Carbon Brief has used this historical relationship between ENSO conditions and temperature to effectively remove the effects of El Niño and La Niña events from global temperatures, as shown in the figure below.
However, this approach – which has worked well for prior years – indicates that there would be almost no effect of El Niño on temperatures in 2023. This is because the lingering global temperature impact of La Niña conditions on the first half of the year would approximately cancel out the influence of El Niño on the second half. This model would suggest that the current El Niño event would primarily affect 2024 temperatures, analogous to what occurred in 1998 and 2016.
Annual global average surface temperatures from Berkeley Earth, as well as Carbon Brief’s estimate of global temperatures with the effect of El Niño and La Niña (ENSO) events removed using the Foster and Rahmstorf (2011) approach. Figures are shown relative to a 1981-2010 baseline. Chart by Carbon Brief.
It is possible that this El Niño event is behaving differently and that the rapid switch from a rare and extended triple-dip La Niña event from late 2020 to the start of this year into strong El Niño conditions is resulting in a more rapid global temperature response.
But this remains speculative at this point and researchers are just starting to disentangle the causes of the unexpected extreme global heat the world experienced in 2023.
Observations broadly in-line with climate model projections
Climate models provide physics-based estimates of future warming given different assumptions about future emissions, greenhouse gas concentrations and other climate-influencing factors.
Here, Carbon Brief examines a collection of climate models – known as CMIP6 – used in the 2021 science report of the IPCC’s sixth assessment. In CMIP6, model estimates of temperatures prior to 2015 are a “hindcast” using known past climate influences, while temperatures projected after 2015 are a “forecast” based on an estimate of how things might change.
Using the CMIP6 ensemble is a bit more challenging than past generations of models because a subset of models have unrealistically high climate sensitivity and poorly reproduce historical observations. Rather than simply average all the models – as had been done in prior assessments – the IPCC used an approach that effectively weights models by their performance. This means the models align better with the range of climate sensitivity derived from multiple different lines of evidence.
The figure below shows the range of individual CMIP6 model projections since 1950, as well as future projections through to 2100 under the middle-of-the-road SSP2-4.5 emissions scenario.
The black line shows the average of 22 different models whose transient climate response (TCR) falls within the IPCC’s “likely” range (which results in temperature projections nearly identical to the IPCC assessed warming). The grey area shows the 95% (two standard deviation) range of the TCR-screened model projections. Observational temperatures are plotted on top of the climate model data, with individual observational records represented by coloured lines.
Annual global average surface temperatures from CMIP6 models and observations between 1950 and 2030 (through to 2023 for observations). Models use the SSP2-4.5 scenario after 2015. They are screened to only include those models with a transient climate response (TCR) in-line with the IPCC’s “likely” range as discussed in Hausfather et al (2022). Anomalies plotted with respect to a 1981-2010 baseline. Chart by Carbon Brief.
Using the TCR-screened subset, the findings show that 2023 temperatures are above the model average, but still well within the range of climate model projections. While some individual months have approached the upper end of the model range, the year as a whole is less anomalous.
The figure below shows the observed warming since pre-industrial (1850-99) in the five different observational datasets compared with climate model projections. The box plot below shows the 25th to 75th percentile of models and observations with a solid box, while the whiskers show 1.5 times the interquartile range.
Temperatures in 2023, relative to pre-industrial, were in the upper quartile (>75th percentile) of climate models assessed. However, this is not unexpected, as models do not have internal variability such as El Niño or La Niña events that occur at the same time as they do in the real world. (Scientists would expect observations to be on the high end of the model range during El Niño years and on the low end in La Niña years even if climate models were perfectly accurate.)
Finally, the figure below shows how observations compare to the full ensemble of 37 CMIP6 models, with no screening for transient climate response. Here, observations are generally below the model average over the past two decades and are roughly equal to the model average in 2023.
Annual global average surface temperatures from CMIP6 models and observations between 1950 and 2030 (through 2023 for observations). Models use the SSP2-4.5 scenario after 2015. Anomalies plotted with respect to a 1981-2010 baseline. Chart by Carbon Brief.
Record atmospheric temperatures
In addition to surface measurements over the world’s land and oceans, satellite microwave sounding units have been providing estimates of temperatures at various layers of the atmosphere since 1979.
The lowest layer they estimate – the lower troposphere – reflects temperatures a few kilometres above the surface and shows a pattern of warming similar – though not identical – to surface temperature changes. The records produced by Remote Sensing Systems (RSS) and the University of Alabama, Huntsville (UAH) both show 2023 as the warmest year on record in the lower troposphere. The chart below shows the two records – RSS in red and UAH in blue – for the lower troposphere (TLT).
Global average lower-troposphere temperatures from RSS version 4 (red) and UAH version 6 (blue) for the period from 1979-2023, relative to a 1981-2010 baseline. Chart by Carbon Brief.
The lower troposphere tends to be influenced more strongly by El Niño and La Niña events than the surface. Therefore, satellite records show correspondingly larger warming or cooling spikes during these events. This is why, for example, 1998 shows up as one of the warmest years in satellites, but not in surface records.
The two lower-tropospheric temperature records show large differences after the early 2000s. RSS shows an overall rate of warming quite similar to surface temperature records, while UAH shows considerably slower warming in recent years than has been observed on the surface. Both have seen large adjustments in recent years that have warmed RSS and cooled UAH compared to prior versions of each record.
Overall, there is more agreement with RSS to other records that include satellite data (such as reanalysis products ERA5 and JRA-55, as well as the shorter AIRS satellite-based surface temperature record), while the UAH record is a bit of an outlier compared to the others.
In addition to a temperature record of the lower troposphere, RSS and UAH also provide measurements of the lower stratosphere – a region of the upper atmosphere around 18km above the surface. The figure below shows lower stratospheric temperature (TLS) records for both RSS (red) and UAH (blue) from 1979 through 2023.
Global average lower stratospheric temperatures from RSS version 4 (red) and UAH version 6 (blue) for the period from 1979-2023, relative to a 1981-2010 baseline. Chart by Carbon Brief.
The stratosphere has been cooling for the past few decades in a clear fingerprint of human greenhouse gases, which warm the lower part of the atmosphere by trapping heat while cooling the upper atmosphere as less heat escapes. If other factors – such as changing solar output – were causing climate change, both the stratosphere and troposphere would be warming.
Greenhouse gas concentrations reach new highs
Greenhouse gas concentrations reached a new high in 2023, driven by human emissions from fossil fuels, land use and agriculture.
Three greenhouse gases – CO2, methane (CH4) and nitrous oxide (N2O) – are responsible for the bulk of additional heat trapped by human activities. CO2 is by far the largest factor, accounting for roughly 42% of the increase in global surface temperatures since the pre-industrial era (1850-1900).
Methane accounts for 28%, while nitrous oxide accounts for around 5%. The remaining 25% comes from other factors including carbon monoxide, black carbon and halocarbons, such as CFCs.
Human emissions of greenhouse gases have increased atmospheric concentrations of CO2, methane and nitrous oxide to their highest levels in at least a few million years – if not longer.
The figure below shows concentrations of these greenhouse gases – in parts per million (ppm) for CO2 (blue line) and parts per billion (ppb) for methane (orange) and nitrous oxide (red) – from the early 1980s through to October 2023 for CO2 and September 2023 for CH4 and N2O (the most recent data currently available).
Global concentrations of CO2, methane (CH4) and nitrous oxide (N2O). Based on data from NOAA’s Earth Systems Research Laboratory. Note that the y-axes do not start at zero. Chart by Carbon Brief.
Methane concentrations, in particular, have seen a sharp rise over the past decade after a plateau in the 2000s. This appears to be driven by both increased emissions from agriculture and fossil fuels, in roughly equal measure. Thawing Arctic permafrost and other sources of methane in high-latitude regions are growing, but emissions above a latitude of 60 degrees north are only around 4% of the total global methane emissions.
Accelerating sea level rise
Modern-day sea levels have risen to a new high, due to a combination of melting land ice (such as glaciers and ice sheets), the thermal expansion of water as it warms and changes in land water storage.
In recent years, there have been larger contributions to sea level rise from melting ice sheets and glaciers, as warmer temperatures accelerate ice sheet losses in Greenland and Antarctica.
Since the early 1990s, the increase in global sea level has been estimated using altimeter data from satellites. Earlier global sea levels have been reconstructed from a network of global tide gauge measurements. This allows researchers to estimate how sea level has changed since the late 1800s.
The chart below shows five different modern sea level rise datasets (coloured lines), along with satellite altimeter measurements from NASA satellites as assessed by the University of Colorado (in black) after 1993. (As sea level rise data has not yet been released for the whole year, the 2023 value is estimated based on data through October.)
Global average sea level rise reconstructed from tide gauge data between 1880 and 2023 from Frederikse et al 2020, Dangendorf et al 2019, Hay et al 2015, Church and White 2011, and Palmer et al 2021. Satellite altimeter data from 1993 (black) to present is taken from the University of Colorado. Chart by Carbon Brief.
Sea levels have risen by over 0.2 metres (200mm) since 1900. While sea level rise estimates mostly agree in recent decades, larger divergences are evident before 1980. There is also evidence of accelerating sea level rise over the post-1993 period when high-quality satellite altimetry data is available. (See Carbon Brief’s explainer on how climate change is accelerating sea level rise.)
Shrinking glaciers and ice sheets
A portion of this sea level rise is being driven by melting glaciers on land. Scientists measure the mass of glaciers around the world using a variety of remote-sensing techniques, as well as through GRACE measurements of the Earth’s gravitational field. The balance between snow falling on a glacier and ice loss through melting and the breaking off – or “calving” – of icebergs determines if glaciers grow or shrink over time.
The World Glacier Monitoring Service is an international consortium that tracks 164 different glaciers in 19 different regions around the world. The figure below shows the change in global average glacier mass from 1950 through to the end of 2022 (2023 values are not yet available). Note that glacier melt is reported in metres of water equivalent, which is a measure of how much mass has been lost on average.
Global average glacier melt over the 1950-2022 period from the World Glacier Monitoring Service, in metres of water equivalent. Carbon Brief.
Greenland ice sheets have become a larger contributor to sea level rise in recent years due to accelerating loss of mass. The year 2023 was the 27th in a row where Greenland lost ice overall, with 196Gt of ice lost over the 12 months from September 2022 to August 2023. Greenland last saw an annual net gain of ice in 1996.
The figure below shows the cumulative mass balance change – that is, the net ice loss – from Greenland between 1970 and October 2021. The different coloured lines indicate estimates from different studies. The authors find that Greenland has lost over 6tn metric tonnes of ice over the past 50 years – more than 700 tonnes lost per person for every person on the planet.
Cumulative ice loss from Greenland in billion metric tonnes (gigatonnes) between 1970 and 2023 from Mankoff et al 2021, updated through December 2023. Chart by Carbon Brief.
Record-low Antarctic sea ice levels
Arctic sea ice was at the low end of the historical (1979-2010) range for most of 2023, but did not set any new all-time low records apart from a few individual days in February and March. The summer minimum extent – the lowest recorded level for the year – was the sixth-lowest since records began in the late 1970s.
Antarctic sea ice, on the other hand, saw new record low levels for virtually the entire year – except for periods in April, November and December of the year.
Antarctic sea ice was particularly low between June and November, shattering prior records by a substantial margin. While long-term trends in Antarctic sea ice have been ambiguous in the past (unlike in the Arctic where there is a consistent long-term decline), there is increasing evidence that human-driven warming is starting to drive significant loss of sea ice in the region.
The figure below shows both Arctic (red line) and Antarctic (blue line) sea ice extent for each day of the year, along with how it compares to the historical range (corresponding shading).
Arctic and Antarctic daily sea ice extent from the US National Snow and Ice Data Center. The bold lines show daily 2023 values, the shaded area indicates the two standard deviation range in historical values between 1979 and 2010. The dotted black lines show the record lows for each pole. Chart by Carbon Brief.
Looking ahead to 2024
The unusual global temperatures experienced in 2023 makes it more challenging to predict what 2024 might have in store.
If history is any guide here, the year after the development of an El Niño event tends to be considerably warmer than the year in which it develops. For example, strong El Niño events developed over the latter half of both 1997 and 2015 and peaked early the following year, and 1998 and 2016 both set new records by a large margin.
However, 2023 was so exceptionally warm that it suggests that this El Niño might be behaving differently, with global surface temperatures responding more rapidly than in the past. If this is the case, 2024 would not necessarily follow the pattern of past El Niño events and is less likely to be substantially warmer than 2023.
There have been four published predictions – from the UK Met Office, NASA’s Dr Gavin Schmidt, Berkeley Earth and Carbon Brief (in this article) – of what temperatures might look like in 2024.
The figure below shows the four different 2024 predictions compared to the average of six different temperature records (NASA, NOAA, Hadley, Berkeley, Copernicus, and the Japanese JRA-55 reanalysis) used by the World Meteorological Organization (WMO). These have been “normalised” to show 2024 warming relative to 2023 in the WMO dataset. This is to remove any differences in predictions due to divergences in the baselines used by different temperature records.
Carbon Brief’s prediction of likely 2024 temperatures is based on a statistical model using the year, the average temperature of the past year, and projections of El Niño/La Niña conditions over the first three months to predict the temperatures of the year.
Temperature projections for 2024 from the UK Met Office, NASA’s Dr Gavin Schmidt, Berkeley Earth, and Carbon Brief, relative to pre-industrial (1880-99) temperatures and compared to the historical average of six different datasets produced by the WMO. Chart by Carbon Brief.
The Met Office, Dr Schmidt, Berkeley Earth and Carbon Brief estimates all have 2024 as more likely than not to be warmer than 2023 – but only by a small margin. In all estimates it is close to a coin flip which will end up as the warmer year.
Against a 1880-99 pre-industrial baseline, the central estimate of all four forecasts is just below 1.5C of warming, with ranges suggesting that temperatures could top 1.5C next year.
However, the poor performance by all of these groups in accurately predicting 2023 temperatures mean these should be treated with some caution. Until scientists have a clearer understanding of the drivers of the unusual warmth the world experienced in 2023, it will not be possible to fully predict how temperatures will evolve in 2024 as the current El Niño event peaks and fades away.
Ultimately, what matters for the climate is not the leaderboard of individual years. Rather, it is the long-term upward trend in global temperatures driven by human emissions of greenhouse gases. Until the world reduces emissions down to net-zero, the planet will continue to warm.
As the chart below highlights, if the warming trend over the past 40 years continues, average surface temperatures are likely to pass 1.5C in the early 2030s and 2C by around 2060.
The post State of the Climate: 2023 smashes records for surface temperature and ocean heat appeared first on Carbon Brief.
State of the Climate: 2023 smashes records for surface temperature and ocean heat
The past three years have been exceptionally warm globally.
In 2023, global temperatures reached a new high, after they significantly exceeded expectations.
This record was surpassed in 2024 – the first year where average global temperatures were 1.5C above pre-industrial levels.
Now, 2025 is on track to be the second- or third-warmest year on record.
What has caused this apparent acceleration in warming has been subject to a lot of attention in both the media and the scientific community.
Dozens of papers have been published investigating the different factors that could have contributed to these record temperatures.
In 2024, the World Meteorological Organization (WMO) discussed potential drivers for the warmth in a special section of its “state of the global climate” report, while the American Geophysical Union ran a session on the topic at its annual meeting.
In this article, Carbon Brief explores four different factors that have been proposed for the exceptional warmth seen in recent years. These are:
- A strong El Niño event that developed in the latter part of 2023.
- Rapid declines in sulphur dioxide emissions – particularly from international shipping and China.
- An unusual volcanic eruption in Tonga in 2022.
- A stronger-than-expected solar cycle.
Carbon Brief’s analysis finds that a combination of these factors explains most of the unusual warmth observed in 2024 and half of the difference between observed and expected warming in 2023.
However, natural fluctuations in the Earth’s climate may have also played a role in the exceptional temperatures, alongside signs of declining cloud cover that may have implications for the sensitivity of the climate to human-caused emissions.
An unusually warm three years
Between 1970 and 2014, average surface temperatures rose at a fairly steady rate of around 0.18C per decade.
Set against this long-term trend, temperature increases during the period from 2015 to 2022 were on the upper end of what would be expected.
The increases seen in 2023, 2024 and 2025 were well outside of that range.
The high temperatures of the past three years reflect a broader acceleration in the rate of warming over the past decade.
However, the past three years were unusually warm, even when compared to other years in the 2010s and 2020s.
Record-breaking warmth in 2023 meant that it beat the prior warmest year of 2016 by 0.17C – the largest magnitude of a new record in the past 140 years.
The year 2024 then swiftly broke 2023’s record, becoming the first year where average global temperatures exceeded 1.5C above pre-industrial levels.
The 10 months of data available for 2025 indicates that the year is likely to be slightly cooler than 2023 – though it is possible it may tie or be slightly warmer.
The figure below shows global surface temperatures between 1970 and 2025. (The figures for 2025 include uncertainty based on the remaining three months of the year.)
It includes a smoothed average based on temperature data for 1970-2022 that takes into account some acceleration of warming – and then extrapolates that smoothed average forward to 2023-25 to determine what the expected temperature for those years would have been. (This follows the approach used in the WMO’s “state of the global climate 2024” report.)
This approach calculates how much warmer the past three years were than would be expected given the long-term trend in temperatures.
It shows that 2023 was around 0.18C warmer than expected, 2024 was a massive 0.25C warmer and 2025 is likely to be 0.11C warmer.
Researchers have identified a number of potential drivers of unexpected warmth over 2023-25. Here, Carbon Brief looks at the evidence for each one.
A weirdly behaving El Niño event
El Niño is a climate pattern of unusually warm sea surface temperatures (SSTs) in the tropical Pacific that naturally occurs every two to seven years. Strong El Niño years generally have warmer global temperatures, with the largest effect generally occurring in the months after El Niño conditions peak (when SSTs reach their highest levels in the tropical Pacific).
A relatively strong El Niño event developed in the latter half of 2023, peaking around November before fading in the spring of 2024.
This event was the fourth-strongest El Niño ever recorded, as measured according to SSTs in the Niño 3.4 region in the central tropical Pacific. However, it was notably weaker than the El Niño events in both 1998 and 2016.
This can be seen in the chart below, which shows the strength of El Niño events (red shading) since the 1980s. (The blue shading indicates La Niña events – the opposite part of the cycle to El Niño, which results in cooler SSTs in the tropical Pacific.)
(It is worth noting that measuring the strength of El Niño events is not entirely straightforward. Other tools used by scientists to monitor changes to El Niño – such as the US National Oceanic and Atmospheric Administration’s (NOAA’s) multivariate ENSO index – show the 2023-24 event was much weaker than indicated in the Niño 3.4 dataset.)
Global surface air temperatures tend to be elevated by around 0.1-0.2C in the six months after the peak of a strong El Niño event – defined here as when SSTs in the Niño 3.4 region reach 1.5C above normal.
The figure below shows the range of global temperature change for the 12 months before and 22 months after the peak of all 10 strong El Niño events since 1950. The light line represents the average of past strong El Niño events, the dark blue line the temperature change observed during the 2023-24 event and the shaded blue area the 5-95th percentile range.
The figure shows the 2023-24 El Niño was quite unusual compared to other strong El Niño events since 1970. Global temperatures rose to around 0.4C above expected levels – which is on the high side of previous El Niños.
The heat also came early, with high temperatures showing up around four months before the El Niño event peaked. This early heat is unlike any other El Niño event in modern history and is one of the reasons why 2023’s global temperatures were so unexpectedly warm.
Global temperatures remained elevated for a full 18 months after the El Niño peaked, well after conditions in the tropical Pacific shifted into neutral conditions – and even after mild La Niña conditions developed at the end of 2024 and into early 2025.
This figure does not explain how much of this unusual heat was actually caused by El Niño, compared to other factors, but it does suggest that El Niño behaviour alone does not fully explain unusually high temperatures in recent years.
Based on the historical relationship between El Niño and global temperatures, Carbon Brief estimates that El Niño contributed a modest 0.013C to 2023 temperatures and a more substantial 0.128C to 2024 temperatures, albeit with large uncertainties. (See “methodology” section at the end for details.)
However, it is possible that this 2023 estimate is too low. There are some suggestions in the literature that 2023-24 El Niño’s early warmth may have been caused by the rapid transition out of a particularly extended La Niña event. There are indications that temperatures have spiked in similar situations further back in the historical temperature record.
Falling sulphur dioxide emissions
Sulphur dioxide (SO2) is an aerosol that is emitted into the lower atmosphere by the burning of coal and oil. It has a powerful climate cooling effect – Carbon Brief analysis shows that global emissions of SO2 have masked about one-third of historical warming.
Global SO2 emissions have declined around 40% over the past 18 years, as countries have increasingly prioritised reducing air pollution, including through the installation of scrubbers at coal plants.
These declines have been particularly concentrated in China, which has seen a 70% decline in SO2 emissions since 2007. In addition, a rule introduced for international shipping in 2020 by the International Maritime Organization (IMO) has resulted in an 80% decline in the sulphur content of shipping fuel used around the world.
The decline of SO2 emissions is shown in the figure below.
Shipping in particular has been suggested as a potential culprit for recent temperatures, given that ships emit SO2 over oceans where the air tends to be cleaner and so emissions have a bigger effect.
Seven of the eight studies that have explored the temperature impact of the IMO regulations have suggested a relatively modest effect, in the range of 0.03-0.08C. However, one study – led by former NASA scientist Dr James Hansen – calculated a much stronger effect of 0.2C that would explain virtually all the unusual warmth of recent years.
The figure below shows Carbon Brief’s estimate of the global average surface temperature changes caused by the low-sulphur shipping fuel rules, using the estimates produced by all eight studies. The central estimate (dark blue line) is relatively low, at around 0.05C, but the uncertainty range (light blue shading) across the studies remains large.
Overall, Carbon Brief’s analysis finds that around 0.04C of warming over 2020-23 and 0.05C of warming over 2020-24 can be attributed to SO2 declines from shipping and other sources.
However, this approach might slightly overstate the effects of SO2 on the exceptional temperatures of the past three years, as shipping and other SO2 declines would have had some effect on 2021 and 2022 as well.
It is also worth noting that the total effects of SO2 declines on global temperatures have been considerably larger and are estimated to be responsible for around one-quarter of all warming since 2007.
However, these SO2 decreases occurred over a long period of time and do not clearly explain the recent spike in temperatures.
An unusual volcanic eruption in Tonga
In early 2022, the Hunga Tonga-Hunga Ha’apai underwater volcano erupted spectacularly, sending a plume 55km into the atmosphere. This was by far the most explosive volcanic eruption since Mount Pinatubo erupted in 1991.
This was a highly unusual volcanic eruption, which vaporised vast amounts of sea water and lofted it high into the atmosphere. Overall, around 146m metric tonnes of water vapour ended up in the stratosphere, which is the layer of the atmosphere above the troposphere.
Water vapour is a powerful greenhouse gas. While it is short-lived in the lower atmosphere, it can stick around for years in the stratosphere, where it has a significant warming effect on the climate.
The figure below shows the concentration of water vapour in the stratosphere between 2005 and mid-2025. It shows how the 2022 eruption increased atmospheric concentrations of the greenhouse gas by around 15%. More than half the added water vapour has subsequently fallen out of the upper atmosphere.
Most early studies of the Hunga Tonga-Hunga Ha’apai volcano focused specifically on the effects of stratospheric water vapour. These tended to show strong warming in the lower stratosphere and cooling in the middle-to-upper stratosphere, but only a slight warming effect on global surface temperatures of around 0.05C.
Hunga Tonga-Hunga Ha’apai had much lower sulphur emissions than prior explosive eruptions, such as Pinatubo and El Chichon. However it put 0.5–1.5m tonnes of sulphur into the stratosphere – the most from an eruption since Pinatubo.
Studies that included both sulphur and water vapour effects tend to find that the net effect of the eruption on surface temperatures was slight global cooling, concentrated in the southern hemisphere.
By using the estimates published in a 2024 study published in Geophysical Research Letters, which used the FaIR climate emulator model, Carbon Brief estimates that the Hunga Tonga-Hunga Ha’apai eruption cooled global surface temperatures by -0.01C in 2023 and -0.02C in 2024.
This suggests that the eruption was likely only a minor contributor to recent global surface temperatures.
A stronger-than-expected solar cycle
The source of almost all energy on Earth is the sun. Over hundreds of millions of years, variations in solar output have a big impact on the global climate.
Thankfully, over shorter periods of time the sun is remarkably stable, helping keep the Earth’s climate habitable for life. (Big changes – such as ice ages – have more to do with variations in the Earth’s orbit than changes in solar output.)
However, slight changes in solar output do occur – and when they do, they can influence climate change over shorter periods of time. The most important of these is the roughly 11-year solar cycle, which is linked with the sun’s magnetic field and results in changes in the number of sunspots and amount of solar energy reaching Earth.
The figure below shows a best-estimate of changes in total solar irradiance since 1980, based on satellite observations. Total solar irradiance is a measure of the overall amount of solar energy that reaches the top of the Earth’s atmosphere and is measured in watts per metre squared.
The 11-year solar cycle is relatively modest compared to the sun’s total output, varying only a few watts per metre squared between peak and trough – amounting to around 0.01% of solar output. However, these changes can result in variations of up to 0.1C in global temperatures within a decade.
The most recent solar cycle – solar cycle 25 – began around 2020 and has been the strongest solar cycle measured since 1980. It was stronger than most models had anticipated and likely contributed to around 0.04C global warming in 2023 and 0.07C in 2024.
Putting together the drivers
By combining earlier estimates of different factors contributing to 2023 and 2024 global surface temperatures, about half of 2023’s unusual warmth and almost all of 2024’s unusual warmth can be effectively explained.
This is illustrated in the figure below, which shows the five different factors discussed earlier – El Niño, shipping SO2, Chinese SO2, the Hunga Tonga-Hunga Ha’apai volcano and solar cycle changes – along with their respective uncertainties.
The sum of all the factors is shown in the “combined” bar, while the actual warming compared to expectations is shown in red.
The upper chart shows 2023, while the lower one shows 2024.
It is important to note that the first bar includes both El Niño and natural year-to-year variability; the height of the bar reflects the best estimate of El Niño’s effects, while the uncertainty range encompasses year-to-year variability in global temperatures that may be – at least in part – unrelated to El Niño.
The role of natural climate variability
Large natural variability to the Earth’s climate is one of the main reasons why the combined value of the different drivers of expected warmth in 2023 has an uncertainty range that exceeds the observed warming – even though the best-estimate of combined factors only explains half of temperatures.
Or, to put it another way, there is up 0.15C difference in global temperatures year-on-year that cannot be explained solely by El Niño, human-driven global warming, or natural “forcings” – such as volcanoes or variations in solar output.
The figure below shows the difference between actual and expected warming in the global temperature record for every year in the form of a histogram. The vertical zero line represents the expectation given long-term global warming and the other vertical lines indicate the warming seen in 2023, 2024 and 2025.
The height of each blue bar represents the number of years over 1850-2024 when the average global temperature was that far (above or below) the expected level of warming.
Based on the range of year-to-year variability, temperatures would be expected to spike as far above the long-term trend as they did in 2023 once every 25 years, on average. The year 2024 would be a one-in-88 year event, whereas 2025 would be a less-unusual, one-in-seven year event.
These likelihoods for the past three years are sensitive to the approach used to determine what the longer-term warming level should be.
In this analysis, Carbon Brief used a local smoothing approach (known as locally estimated scatterplot smoothing) to determine the expected temperatures, following the approach used in the WMO “state of the climate 2024” report.
This approach results in a warming of 1.28C in 2023 and 1.30C in 2024, against which observed temperatures are compared.
Other published estimates put the longer-term warming in 2024 notably higher.
Earlier this year, the scientists behind the “Indicators of Global Climate Change” (IGCC) report estimated that human activity caused 1.36C of recent warming in 2024. They also found a slightly lower overall warming level for 2024 – 1.52C, as opposed to the WMO’s 1.55C – because they looked exclusively at datasets used by IPCC AR6. (This meant estimates from the Copernicus/ECMWF’s ERA5 dataset were not included.)
Based on climate simulations, the IGCC report finds the likelihood of 2024’s warmth to be a one-in-six year event and 2023’s a one-in-four event.
Using the same assumptions as the IGCC, Carbon Brief’s approach calculates that 2024 would be a less-common, one-in-18 year event.
However, the IGCC estimate of current human-induced warming is based on the latest estimates of human and natural factors warming the climate. That means that it already accounts for additional warming from low-sulphur shipping fuel, East Asian aerosols and other factors discussed above.
Therefore, the results from these two analyses are not necessarily inconsistent: natural climate variability (including El Niño) played a key role – but this came in addition to other factors. Natural fluctuations in the Earth’s climate alone would have been unlikely to result in the extreme global temperatures seen in 2023, 2024 and 2025.
A cloudy picture
Even if unusual recent global warmth can be mostly attributed to a combination of El Niño, falling SO2 emissions, the Hunga Tonga-Hunga Ha’apai volcano, solar cycle changes and natural climate variability, there are a number of questions that remain unanswered.
Most important is what the record warmth means for the climate going forward. Is it likely to revert to the long-term average warming level, or does it reflect an acceleration in the underlying rate of warming – and, if so, what might its causes be?
As explained by Carbon Brief in a 2023 article, climate models have suggested that warming will speed up. Some of this acceleration is built into the analysis presented here, which includes a slightly faster rate of warming in recent years than has characterised the period since 1970.
But there are broader questions about what – beyond declining SO2 and other aerosols – is driving this acceleration.
Research recently published in the journal Science offered some potential clues. It found a significant decline in planetary reflectivity – known as albedo – over the past decade, associated with a reduced low-level cloud cover that is unprecedented in the satellite record.
The authors suggest it could be due to a combination of three different factors: natural climate variability, changing SO2 and other aerosol emissions and the effects of global warming on cloud reflectivity.
Natural climate variability seems unlikely to have played a major role in reduced cloud cover, given that it was relatively stable until 2015. However, it is hard to fully rule it out given the relatively short satellite record.
Reductions in SO2 emissions are expected to reduce cloud reflectivity, but the magnitude of the observed cloud reflectivity changes are much larger than models simulate.
Models might be underestimating the impact of aerosols on the climate. But, if this were the case, it would indicate that climate sensitivity might be on the higher end of the range of model estimates, because models that simulate stronger aerosol cooling effects tend to have higher climate sensitivity.
Finally, cloud cover might be changing and becoming less reflective as a result of warming. Cloud responses to climate change are one of the largest drivers of uncertainty in future warming. One of the main reasons that some climate models find a higher climate sensitivity is due to their simulation of less-reflective clouds in a warming world.
The Science study concludes that the 2023 heat “may be here to stay” if the cloud-related albedo decline was not “solely” caused by natural variability. This would also suggest the Earth’s climate sensitivity may be closer to the upper range of current estimates, it notes.
Methodology
Carbon Brief built on work previously published in the IGCC 2024 and WMO state of the global climate 2024 reports that explores the role of different factors in the extreme temperatures in 2023, 2024 and 2025.
The impact of El Niño Southern Oscillation (ENSO) on the temperatures was estimated using a linear regression of the annual mean global temperature anomaly on the Feb/Mar Niño 3.4 index. This resulted in an impact of −0.07C, 0.01C and 0.13C for 2022, 2023 and 2024 respectively (with a 95% confidence interval of ±0.13 ºC).
It is important to note that the uncertainties in the ENSO response estimated here also incorporate other sources of unforced internal (modes of variability in other basins such as AMV), and potentially some forced variability. The bar in the combined figure is labelled “El Niño and variability” to reflect this.
For details on calculations of the temperature impact of shipping and Chinese SO2 declines, see Carbon Brief’s explainer on the climate impact of changing aerosol emissions.
Solar cycle 25 was both slightly earlier and slightly stronger than prior expectations with a total solar irradiance anomaly of 0.97 watts per metre squared in 2023 relative to the mean of the prior 20 years. This resulted in an estimated radiative forcing of approximately 0.17 watts per metre squared and an estimated global surface temperature increase of 0.07C (0.05C to 0.10C) with a one- to two-year lag based on a 2015 study. Thus, the impact on 2023 and 2024 is around 0.04C and 0.07C, respectively (+/- 0.025C). This is a bit higher warming than is given by the FaIR model, as the 2015 study is based on global models that have ozone responses to the UV changes, which amplifies the temperature effects a bit.
The Hunga Tonga-Hunga Haʻapai volcanic eruption added both SO2 and water vapour to the stratosphere (up to 55km in altitude). The rapid oxidation of SO2 to sulphate aerosol dominated the radiative forcing for the first two years after the eruption. As a result, the net radiative forcing at the tropopause was likely negative; −0.04 watts per metre squared and −0.15 watts per metre squared in 2022 and 2023, respectively, implying a temperature impact of -0.02C (-0.01C to -0.03C) calculated using the FaIR model.
The post Analysis: What are the causes of recent record-high global temperatures? appeared first on Carbon Brief.
Analysis: What are the causes of recent record-high global temperatures?
Local officials are often viewed as relatively weak actors in China’s governance structure, largely implementing policies issued from the central level.
However, a new book – “Implementing a low-carbon future: climate leadership in Chinese cities” – argues that these officials play an important role in designing innovative and enduring climate policy.
The book follows how four cities – Shenzhen, Zhenjiang, Xiamen and Nanchang – approached developing low-carbon policies over the course of almost a decade.
It identifies “bridge leaders” – mid-level local bureaucrats who have a strong interest in a specific policy area and who are unlikely to move often between different posts – as key to effective local climate policymaking.
Carbon Brief interviews author Weila Gong, non-resident scholar at the UC San Diego School of Global Policy and Strategy’s 21st Century China Center and visiting scholar at UC Davis, on her research.
The interview has been edited for length and clarity.
- Gong on why cities are important: “Over 85% of China’s carbon emissions come from cities. The majority of Chinese people live in cities, so the extent to which cities can become truly low-carbon will also influence China’s climate success.”
- On what motivates local policymakers: “Mid-level bureaucrats need to think about how to create unique, innovative and visible policy actions to help draw attention to their region and their bosses.”
- On cities as a way to test new policies: “Part of the function of local governments in China is to experiment with policy at a local level, thereby helping national-level officials develop responses to emerging policy challenges.”
- On how local policymakers get results: “Even though we tend to think that local officials are very constrained in terms of policy or financial resources, they can often have the leverage and space to build coalitions.”
- On uneven city-level engagement: “To begin with, all regions received political support if they joined the [low-carbon city] pilot programme. But over the years, different regions have engaged very differently.”
- On the need for ‘entrepreneurial bureaucrats’: “China will always need local officials willing to introduce new legislations or try new policy instruments…For that, it needs entrepreneurial bureaucrats who are willing to turn ideas into actions.”
- On international cooperation: “Even with how geopolitics is really complicating things, many cities continue to have common challenges. For example, collaboration between Shanghai and Los Angeles on green shipping corridors is still ongoing”.
- On the effectiveness of mid-level bureaucrats: “They are creative, they know how to convince their boss about the importance of climate action and they know how that can bring opportunities for themselves and their boss. And because of how long they have worked in one area, they understand the local politics, policy processes and the coalitions needed to provide solutions.”
Carbon Brief: You’ve just written a book about climate policy in Chinese cities. Could you explain why subnational governments are important for China’s climate policy in general?
Weila Gong: China is the world’s largest carbon emitter, so the extent to which global efforts to address climate change can actually reach their goal is largely influenced by China’s efforts.
If you look at the structure of China’s carbon emissions, over 85% of China’s carbon emissions come from cities. The majority of Chinese people live in cities, so the extent to which cities can become truly low-carbon will also influence China’s climate success. That’s why I started to look at this research area.
We tend to think of China as a centralised, big system and a unitary state – state-run and top-down – but it actually also has multi-level governance. No climate action or national climate targets can be achieved without local engagement.
We also tend to think subnational level [actors], including the provincial, city and township levels, are barriers for environmental protection, because they are focused on promoting economic growth.
But I observed these actors participating in China’s low-carbon city pilot programme [as part of my fieldwork spanning most of the 2010s]. I was really surprised to see so many cities wanted to participate in the pilot, even though at the time there was no specific evaluation system that would reward their efforts.
We think of local governments just as implementers of central-level policy. When it comes to issues like climate change and also low-carbon development – in 2010 [policymakers found these concepts] very vague…So I was curious why those local officials would want to take on this issue, given that there was no immediate reward, either in terms of career development or in terms of increasing financial support from the central government.
CB: Could you help us understand the mindset of these bureaucrats? How do local-level officials design policies in China?
WG: The role of different local officials in promoting low-carbon policy is not very well understood. We tend to focus on top political figures, such as mayors or [municipal] party secretaries, because we see them as the most important policymakers.
But that is not entirely true. Those top local politicians are very important in supporting efforts to tackle problem areas…but the focus in my book is the mid-level bureaucrats.
Unlike mayors and party secretaries, mid-level officials tend to stay in one locality for their entire career. That helps us to understand why climate policy can become durable in some places and not others.
Mayors and party secretaries are important for [pushing through policy solutions to problem] issues, but they can also be key barriers for ensuring continuation of those policies – particularly when they change positions…as they tend to move to another locality every three to five years.
Therefore, these top-level officials are not the ones implementing low-carbon policies. That’s why I looked at the mid-level bureaucrats instead.
The conventional understanding of these bureaucrats is that they are obedient and only follow their bosses’ guidance. But actually, when low-carbon policies emerged as an important area for the central government in 2010, opportunities appeared for local governments to develop pilot projects.
Mid-level local officials saw this as a way to help their bosses – the mayors and party secretaries – increase their chances of getting promoted, which in turn would help the mid-level bureaucrats to advance their own career.
Impressing central government officials isn’t really a consideration for these officials…but their bosses need visible or more reliable local actions to show their ability to enforce low-carbon development.
As such, mid-level bureaucrats need to think about how to create unique, innovative and visible policy actions to help draw attention to their region and their bosses.
Secondly, mid-level bureaucrats are more interested in climate issues if it is in the interest of their agency or local government.
For example, Zhenjiang [a city in east China] came to be known as a leader in promoting low-carbon development due to a series of early institutional efforts to establish low-carbon development. In particular, in part because of this, it was chosen for a visit by president Xi Jinping in 2014.
As a result, the city created a specialised agency [on low-carbon development]. This made it one of the first regions to have full-time local officials that followed through on low-carbon policy implementation.
This increased their ability to declare their regulatory authority on low-carbon issues, by being able to promote new regulations, standards and so on, as well as enhancing the region’s and the local policymakers’ reputations by building institutions to ensure long-term enforcement.
Another motivation for many local governments is accessing finance through the pilot programmes. If their ideas impress the central-level government, local policymakers could get access to investment or other forms of financial resources from higher levels of government.
In the city of Nanchang, for example, officials were trying to negotiate access to external investment, because the main central government fund for low-carbon initiatives only provided minimal finance.
Nanchang officials tried to partner with the Austrian government on sustainable agriculture, working through China’s National Development and Reform Commission (NDRC).
It didn’t materialise in the end, but they still created a platform to attract international investment, and gathered tens of millions of yuan [millions of dollars] in central-level support because the fact they showed they were innovating allowed them to access more money through China’s institutional channels.
CB: Could you give an example of what drives innovative local climate policies?
WG: National-level policies and pilot programme schemes provide openings for local governments to really think about how and whether they should engage more in addressing climate change.
The national government has participated in international negotiations on climate for decades…but subnational-level cities and provinces only joined national efforts to address climate issues from the 2010s – starting with the low-carbon city programme.
So we can see that local responses to addressing climate change have been shaped by the opportunities provided by the national government, [who in turn] want more local-level participation to give them successful case studies to take to international conferences.
Local carbon emission trading systems (ETSs) are an example of giving local governments opportunities to experiment.
In my book, I look at the case of Shenzhen, which launched China’s first local ETS. [Shenzhen was one of seven regions selected to run a pilot ETS, ahead of the national ETS being established in 2018.]
Part of the function of local governments in China is to experiment with policy at a local level, thereby helping national-level officials develop responses to emerging policy challenges.
I remember a moment during my field research in 2012, when I was with a group of officials from both the national and local government.
The national government officials asked the local officials to come up with some best practices and solutions, to help them envision what could be done at the national level.
Then there are drivers at the international level, which I think is very interesting.
I observed that the officials particularly willing to take on climate issues usually had access to international training.
During the early stages of subnational climate engagement, organisations such as the German Agency for International Cooperation (GIZ) worked with the NDRC and other national-level agencies to train local officials across the country.
This created more opportunities to help local officials understand what climate change and carbon markets were, and how to use policy instruments to support low-carbon development.
In Shenzhen, local bureaucrats also turned to their international partners to help them design policy.
The city created a study group to visit partners working on the EU ETS and learn how it was designed. They learned about price volatility in the EU ETS and pushed legislation through the local people’s congress [to mitigate this in their own system].
One thing that made the Shenzhen ETS so successful is what I call “entrepreneurial bureaucrats” [who have the ability to design, push through and maintain new local-level climate policies].
Shenzhen’s vice mayor worked with the local people’s congress to push the ETS legislation through. This was the first piece of legislation in China to require compulsory participation by more than 600 local industrial actors. It also granted the local government authority to decide the quotas and scope of the ETS.
These 600 entities also included Shenzhen’s public building sector[, a powerful local interest group].
This shows that, even though we tend to think that local officials are very constrained in terms of policy or financial resources, they can often have the leverage and space to build coalitions – even in China’s more centralised political system – and know how to mobilise political support.
CB: You chose to look at the effectiveness of four cities – Shenzhen, Zhenjiang, Xiamen and Nanchang – in climate policymaking. Why did you choose these cities and how representative are they of the rest of China?
WG: We tend to believe that only economically-advanced areas or environmentally-friendly cities will become champions for low-carbon development…But I was surprised, because Zhenjiang and Nanchang are not known for having an advanced economy, but [they nevertheless built impactful climate] institutions – regulations, standards and legislation that shape individual and organisational behaviours in the long term. I thought they were interesting examples of how local regions can really create those institutions.
Then there was Xiamen, which is seen as an environmentally-friendly city and economically is comparable to Shenzhen when you look at GDP per capita. Xiamen actually did not turn its low-carbon policy experimentation into long-term institutions, instead randomly proposing new initiatives [that were not sustained].
I conducted more than 100 interviews, talking with policy-practitioners inside and outside of government about specific policies, their processes and implementation.
I found that, over the course of eight years, these [cities] showed very different levels of engagement.
Some I categorised into substantive engagement, where the local government delivered on their climate goals. [Shenzhen falls into this category.]
Then there is performative engagement – such as in the case of Nanchang – where the local government was more interested in [using climate policies to] attract external investment and access projects from higher levels of government.
But they were not able to enforce the policies, because impressing higher levels of government became the primary motivation.
Zhenjiang was a case of symbolic engagement. It actually created a lot of institutions, such as a specialised agency and a screening system to ensure new [low-carbon] investment. When I was observing Zhenjiang, from 2012 to 2018, officials recognised they needed to be carbon-constrained.
The problem was that Zhenjiang has a very strong power sector – mainly coal power – which supplies the whole eastern coast. That meant, even though the government was very determined to promote low-carbon policies, they faced [opposition from] very strong local actors – meaning the government could only partially implement the targets they set.
Then there is sporadic engagement, as seen in Xiamen. [The city’s approach to climate policy was incremental and cautious] because of a lack of political support [from officials in Xiamen], as well as local coalitions between key actors. So instead, we find random initiatives being promoted.
This explains the uneven policy implementation in China. To begin with, all regions received political support if they joined the pilot programme. But over the years, different regions have engaged very differently, in terms of the regulations, standards and legislation they have introduced, and whether those were paired with enforcement by a group of trained personnel to follow through on those initiatives.
CB: What needs to be done to strengthen sub-national climate policy making?
WG: It’s very important to have groups of personnel trained on climate policy. Since 2010, when I started studying the low-carbon pilot programme, there were no provincial-level people or agencies fully responsible for climate change. Back then, there was only the [central-level] department of climate change under the NDRC.
By the time I finished the book, provincial-level departments of climate change had been created across all provinces. But almost nothing has been established at the city level, so most city-level climate initiatives are being managed under the agencies responsible for air quality.
That means climate change is only one of those local officials’ day-to-day responsibilities. Only a handful of cities have dedicated staff working on climate issues: Beijing, Shanghai, Zhenjiang, Shenzhen and Guiyang.
Nanchang devised some of China’s first legislation to include an annual [financial] budget for low-carbon development. But when I revisited the city, officials were not actually sure about how and whether that budget was being used, because there wasn’t a person responsible for it.
Therefore, even if there are resources available, they can go unused because local officials at the city level are so busy. If climate policy is not prioritised, or written into their job responsibilities, that can be a challenge for sustaining implementation.
In China’s governance structure, the national government comes up with ideas, and the provincial level transfers these ideas down to local-level governments. City-level governments are the ones implementing these ideas.
So we need full-time staff to follow through on policies from the beginning right up to implementation.
Secondly, while almost all cities have now made carbon-peaking plans, one area in which the Chinese government can make further progress is in data.
China has recently emphasised the need to strengthen carbon-emissions data collection and monitoring. But when I was conducting my research, most Chinese cities had not yet established regular carbon-accounting systems.
As such, inadequate energy statistics and insufficient detail remain key barriers to effective climate-policy implementation.
In addition, the relevant data usually is owned by China’s National Bureau of Statistics (NBS), which does not always share it with other agencies. Local agencies can’t always access detailed data.
When I visited Xiamen, officials told me the local government is now improving emissions monitoring systems. But there should be more systematic and rigorous data collection, covering both carbon emissions and non-CO2 greenhouse gases. Also, much of the company-level data is self-reported, which could affect the accuracy of carbon-emissions statistics.
For continued climate action, it’s also important that the central government ensures that local officials have the institutional support needed to experiment and propose new ideas.
…China will always need local officials willing to introduce new legislations or try new policy instruments – like Shenzhen with its ETS, or establishing new carbon-monitoring platforms.
For that, it needs entrepreneurial bureaucrats who are willing to turn ideas into actions. Ensuring that local governments have the right set of conditions to do this is very important.
CB: What did you find most surprising when researching this book?
WG: That international collaboration is still very important. I found that many officials learnt about climate change through international engagement.
In the current situation, I think international engagement is still very important – particularly given how, even with how geopolitics is really complicating things, many cities continue to have common challenges. For example, collaboration between Shanghai and Los Angeles on green shipping corridors is still ongoing.
That can bring opportunities for continuing climate action at the city level in the face of rising international tensions, as long as national governments give them space to be involved in international climate action.
Another surprise was the factors of what exactly made climate action durable. I was really surprised that many of the cities that I revisited were still involved in the pilot programmes, despite the central government restructuring that shifted the climate change portfolio from the NDRC to the Ministry of Ecology and Environment – which created challenges for the local governments who had to navigate this.
I also thought that the change in mayors for all four cities would lead to climate initiatives falling off the agenda.
But actually, Zhenjiang, Xiamen and Nanchang all maintained their low-carbon initiatives, despite these changes. This showed it isn’t only strong mayors that bring success, but rather a group of trained personnel building and enforcing regulations and standards. So the importance of bureaucrats and bureaucracy in making climate action durable was actually way beyond my initial expectations.
I was also surprised that bureaucrats can be entrepreneurial, even though they work in a centralised system. They are creative, they know how to convince their boss about the importance of climate action and they know how that can bring opportunities for themselves and their boss. And because of how long they have worked in one area, they understand the local politics, policy processes and the coalitions needed to provide solutions.
The post Interview: How ‘mid-level bureaucrats’ are helping to shape Chinese climate policy appeared first on Carbon Brief.
Interview: How ‘mid-level bureaucrats’ are helping to shape Chinese climate policy
Nuevo Laredo was dumping millions of gallons of sewage a day into the Rio Grande. The U.S. and Mexico worked together to find a solution.
For years, raw sewage has flowed into the Rio Grande from a beleaguered wastewater treatment plant in Nuevo Laredo, Mexico.
How Belém launched the Just Transition mechanism
EU weakening of corporate sustainability rules ‘jeopardises’ climate action, critics say
Analysis: What are the causes of recent record-high global temperatures?
Guest post: Why China is still building new coal – and when it might stop
Guest post: Why China is still building new coal – and when it might stop
Spanish-language misinformation on renewable energy spreads online, report shows
-
Climate Change4 months ago
Guest post: Why China is still building new coal – and when it might stop
-
Greenhouse Gases4 months ago
Guest post: Why China is still building new coal – and when it might stop
-
Climate Change2 years ago
Spanish-language misinformation on renewable energy spreads online, report shows
-
Greenhouse Gases2 years ago嘉宾来稿：满足中国增长的用电需求 光伏加储能“比新建煤电更实惠”
-
Climate Change Videos2 years ago
The toxic gas flares fuelling Nigeria’s climate change – BBC News
-
Climate Change2 years ago嘉宾来稿：满足中国增长的用电需求 光伏加储能“比新建煤电更实惠”
-
Carbon Footprint2 years agoUS SEC’s Climate Disclosure Rules Spur Renewed Interest in Carbon Credits
-
Climate Change2 years ago
Why airlines are perfect targets for anti-greenwashing legal action