The amount of CO2 being released by human activity each day fell by as much as 17% during the height of the coronavirus crisis in early April, a new study shows.
This means daily emissions temporarily fell to levels last seen in 2006, the study says. In the first four months of the year, it estimates that global emissions from burning fossil fuels and cement production were cut by 1,048m tonnes of CO2 (MtCO2), or 8.6%, compared with 2019 levels.
The research projects a decline of up to 2,729MtCO2 (7.5%) in 2020 as a whole, depending on how the crisis plays out. It is the first to have been through the peer-review process and is broadly in line with an early estimate for China published by Carbon Brief in February, as well as separate global estimates published last month by Carbon Brief and the International Energy Agency.
Today’s study also marks the first-ever attempt to quantify CO2 emissions on a daily basis, for the world and for 69 individual countries, in close to real time. Until now, annual CO2 emissions data has typically been published months or even years later.
A publicly available daily estimate of global or national CO2 emissions would be “incredibly useful, particularly for motivating policy action and pressure”, another researcher tells Carbon Brief.
The ongoing coronavirus crisis has claimed the lives of hundreds of thousands of people around the world and seen the introduction of severe restrictions on movement in many countries.
These lockdowns have included “stay at home” orders, border closures and other measures that have had direct effects on the use of energy and, consequently, on the release of CO2 emissions.
As the crisis has unfolded, so too have attempts to quantify its impact on CO2 emissions. These efforts have been challenging, however, because real-time CO2 emissions data does not exist.
The annual emissions inventories that countries submit to the UN take years to compile – and even these are estimates rather than direct measurements.
Greenhouse gas emissions are estimated using a variety of methods, often based on “activity data”. This might be the number of miles being driven, the amount of electricity generated or even – in the case of nitrous oxide, which is used as a propellant – via cream consumption.
Today’s study, published in Nature Climate Change, combines activity data for six sectors with a “confinement index” of lockdown measures in each country or region over time.
This allows for an estimate of changes in daily global CO2 emissions in January-April 2020, relative to the 100MtCO2 released on an average day in 2019.
During peak confinement in individual countries, daily CO2 emissions fell by 26% on average, the paper says. However, the size of this effect is reduced at a global level, because not all countries were under the most severe type of lockdown at the same time.
At the peak of the crisis in early April, regions responsible for 89% of daily CO2 emissions were under some form of lockdown, the paper says. Daily global CO2 emissions fell to 83MtCO2 (-17%, with a range of -11 to -25%) on 7 April, equivalent to levels last seen in 2006.
In a press release, lead author Prof Corinne Le Quéré, professor of climate change science at the University of East Anglia’s Tyndall Centre (who will be a panelist at Carbon Brief’s webinar on 21 May), says:
“Population confinement has led to drastic changes in energy use and CO2 emissions. These extreme decreases are likely to be temporary, however, as they do not reflect structural changes in the economic, transport, or energy systems.”
In order to estimate daily global CO2 emissions, the researchers use a novel approach that combines sectoral activity data with a country-by-country confinement index.
The paper looks at six sectors, shown in the chart below according to their share of global CO2 emissions from fossil fuels and cement. These are electricity and heat (44%); industry (22%); surface transport (20%); homes (6%); public buildings and commerce (4%); and aviation (3%).
Notably, this split highlights the limited potential for individual actions to radically reduce global emissions, in contrast to the societal choices that govern CO2 from electricity and industry.
The split in global CO2 emissions, shown above, is then broken down further for each of 69 countries, 50 US states and 30 Chinese provinces, which account for 97% of the global total. This gives industrial CO2 emissions in Italy, for example, on an average day in 2019.
The paper then uses 669 datasets, covering each of these sectors over time, and classified according to the level of confinement in place at each point. For example, this might be daily reports on mobility, traffic and congestion to measure “activity” for surface transport.
This daily data is then adjusted to remove effects unrelated to coronavirus, such as the mild northern hemisphere winter or the day of the week.
Under the highest level of confinement, surface transport “activity” fell by 50% on average, the paper finds. This is shown in green in the chart, below, where each dot represents a single data point, open circles show the average and the horizontal lines show the variability between datasets. The chart also shows changes in activity for electricity, industry, homes and aviation.
For electricity, the paper looks at total daily demand in Europe, the US and India, finding an average 15% reduction in demand under strict lockdown. In industry, the paper looks at daily coal use in China reported by Carbon Brief and weekly reports on steel production in the US.
For homes, the paper draws on figures from UK smart meters. And for aviation – the most strongly affected sector – it uses data on domestic and international departures around the world.
As the chart above shows, the analysis relies on relatively sparse information for industry, whereas activity levels in transport draw on a wider range of datasets.
The team then uses the average change in activity, for each sector and level of confinement, to build up an estimate of daily CO2 emissions around the world.
For example, on days when Turkey is under the strictest lockdown, the analysis assumes that its power-sector CO2 emissions would fall by 15% compared with the average in 2019 – and those from surface transport by 50%.
When Turkey shifts from “confinement index three”, the strictest controls, down to level two, its power-sector emissions would be 5% below usual levels and transport 40% lower. For each confinement level, the same percentage reductions are assumed to apply to all countries.
This approach means that the team only needed to know when each country, state or province changed its coronavirus lockdown from one “confinement level” to another, as well as the daily average level of CO2 emissions from each sector in 2019.
Putting all of these countries and lockdown levels together, the paper finds that the cut in daily global CO2 emissions peaked at -17% on 7 April, shown in the figure, below. Across the first four months of 2020, emissions fell by 1,048MtCO2 (8.6%), compared with 2019 levels.
Within this global total, the largest impacts were in China, where emissions fell by an estimated 242MtCO2 in the first four months of the year, followed by the US (-207MtCO2), Europe (-123MtCO2) and India (-98MtCO2).
Dr Glen Peters, research director at Norwegian climate institute Cicero and one of the study authors, tells Carbon Brief that while the approach was designed around the current crisis, the team has gathered the “raw material” to make daily CO2 estimates on an ongoing basis. He says:
“We have discussed more ‘real-time’ estimates for sometime and there are many advantages. We are illustrating one advantage with our paper to see the consequences of particular policy interventions in near real time.”
But Peters notes that some of the daily data they used – the urban congestion index series from satnav maker TomTom, for example – is only being made publicly available during the current crisis and might be made private again in the future. He also asks whether daily data is truly needed, or whether weekly or even monthly estimates might be sufficient for scientists and policymakers.
“I think daily CO2 estimates would be incredibly useful, particularly for motivating policy action and pressure…Climate change already has the classic long-termism problem, but this is exacerbated by the fact that we get a figure on CO2 emissions published once a year, as a marker of how each country is doing.”
If daily CO2 estimates were publicly available for all countries, it would become possible to actively track progress, she says, adding: “You can have a counter on the news, or an app or dashboard on your phone – just like we do with other metrics like stock markets.”
Today’s research is not the first to analyse the CO2 impacts of the coronavirus crisis, although it is the first to have completed its passage through peer review.
Another paper, which is currently in review, also attempts to estimate daily global CO2 emissions in close to real time. This work finds the coronavirus crisis cut global emissions by -542MtCO2 below 2019 levels in the first quarter of 2020, similar to the -530MtCO2 figure from today’s paper.
In mid-February, Carbon Brief published an analysis showing that emissions in China were temporarily cut by 200MtCO2 (25%) over a four-week period, during the height of the restrictions. The new study finds that the cut in Chinese emissions peaked at 24%.
Today’s research also includes estimates of the emissions impact in 2020 as a whole, based on three scenarios for the length of lockdowns around the world. These entail CO2 emissions falling by between -4% and -8%, depending on how the crisis plays out. This range is consistent with estimates published in April byCarbon Brief (-6%) and the International Energy Agency (-8%).