But why have emissions been flat, and what does it mean for the future?
In January, we published a paper which addressed those exact questions. We developed a nested framework of key indicators that can track progress towards the goals of the Paris Agreement, at the global and country level.
A new framework
The problem with many studies is that they put in various factors that may affect emissions into a box, shake the box around, and then try and see what factor is causing what.
We took an alternative approach, or at least a variation on an old approach. Think of it as learning to walk before you run!
We first focus on the key overarching variables, and then gradually “zoom in” to get more detail on specific factors. This makes it easier to separate out interactions and focus on the main factors influencing emissions, which may vary by country.
The framework is therefore ideal for tracking progress under the Paris Agreement.
Keeping it simple
In principle, the framework can be iteratively followed deeper and deeper, but we focus only on the main factors.
Within this first level, one could also track emissions from different sources, such as coal, oil, gas, and cement. The recent slowdown in emissions growth is dominated by a slowdown in the growth in coal consumption, driven primarily by China and the USA.
This simple form of tracking is important, but not sufficient to describe what factors are causing the changes.
The Kaya Identity
The Kaya Identity, named after the Japanese energy economist Yoichi Kaya, decomposes emissions into key driving factors.
We used a simplified form of the Kaya Identity, where emissions growth is decomposed into the growth in economic activity (Gross Domestic Product, GDP), energy use per unit GDP (energy intensity), and emissions per unit energy (carbon intensity),
ΔCO2 = ΔGDP + Δ(Energy Intensity) + Δ(Carbon Intensity) + ΔInteractions
where each term is measured in percentage changes per year. The equation is shown in a simplified form, where the annual growth rates of each term can be added together. The interaction term is small in our method, but is included for completeness.
A variation we use in our analysis is to smooth the input data using a 11-year smoothing window (every data point is averaged with ten adjacent data points). This removes interannually variability, and makes interpretation of trends much easier. Though, care is then required in interpreting annual data points.
We further extend the Kaya Identity to “zoom in” on the carbon intensity. This is because emission scenarios show that this term has a rather rapid deviation from baseline scenarios without climate policy.
The carbon intensity, in turn, can be broken into the share of fossil fuels in total energy use and the amount of carbon emissions per unit fossil fuel use. These terms can be further broken down into technological solutions.
The following set of figures show a simplified Kaya Identity at the global level, and for the top-four emitters: China, USA, EU28, and India. In each figure, the coloured bars give percentage annual changes that add to the growth in global CO2 emissions (black line).
The input data is smoothed over an 11-year period, and this period gets smaller as the last year (2015) is approached (indicated by the grey shaded vertical bars). Because of the smoothing, the focus should be on trends and not annual data points.
At the global level, GDP growth (green) has played a strong role in pushing emissions up over the last decade. Since the global financial crisis in 2008/2009 (not evident in the smoothed data), there has been slightly lower economic growth, though this is starting to return to pre-crisis levels in recent years.
There have been improvements in the energy intensity (purple) and carbon intensity (orange) in recent years, returning to levels not seen since the 1990’s. The energy and carbon intensity have tended to reduce emissions, and this longer-term pattern appears to be returning after the rapid emissions growth in the 2000’s.
These three effects combined – slightly lower economic growth, improved energy intensity, improved carbon intensity – have all led to the slower growth in carbon dioxide emissions in the last few years.
At the global level, it is not clear that one factor is stronger than any other. Teasing apart these factors, requires separating the developments for individual countries.
The top four
The challenge at the global level is that we are analysing the sum of the drivers at the national level, which may add in different ways.
It is easier to understand national level trends, and these are further much more relevant for tracking progress of the emission pledges to the Paris Agreement.
China is the world’s biggest emitter of carbon dioxide emissions, accounting for about 30% of global emissions. China had rapid emissions growth in the 2000’s at around 8% per year, and this has gradually declined to zero growth in 2015. Both the 2000’s growth and the 2010’s slowdown were not expected by analysts.
The recent slowdown in Chinese emissions growth (black line) is due to lower economic growth (green) and improvements in both energy intensity (purple) and carbon intensity (orange). The improvements in carbon intensity are largely due to an increase in non-fossil energy sources, like wind, solar, and hydropower.
China has pledged to peak its carbon dioxide emissions before 2030 and improve its emission intensity by 60-65% in 2030 relative to 2005 levels. The emission intensity improvements correspond to about 3-4% per year improvements.
China’s emissions will peak when economic growth balances with improvements in emission intensity. Since economic growth is lower than expected, and emission intensity (energy intensity plus carbon intensity) exceeds the required 3-4% per year declines, it is quite possible that China will meet its emissions pledge much earlier than expected.
Further details on the potential peak in China’s emissions can be found in dedicated blog post.
The USA emits about 15% of global carbon dioxide emissions, but emissions have been trending downwards since about 2007. Since 2005, carbon dioxide emissions from fossil fuels have decreased about 12%.
The lower emissions growth (black line) is driven by lower economic growth since the global financial crisis (green), continual improvements in energy intensity (purple), and the emergence of improvements in carbon intensity (orange).
In recent years, the improvements in carbon intensity have been driven by a shift from coal to gas, with the rapid growth in wind and solar making an important contribution.
The USA has pledged to reduce greenhouse gas emissions in 2025 to be 26-28% below the level in 2005. If this target translates equally to carbon dioxide emissions from fossil fuels only, then the USA would need to reduce emissions by about 1.5-2% per year for the next decade.
Particularly if economic growth returns to levels seen before the global financial crisis, this would require the USA to double down on efforts to improve energy and carbon intensity in the coming decade.
The European Union has been consistently reducing emissions over the last decades, but particularly since the global financial crisis. Carbon dioxide emissions from fossil fuels are now 20% below the levels in 1990, already meeting its 2020 emissions pledge.
The emission reductions in Europe have many parallels with the reductions in the USA. The lower emissions growth (black line) is driven by lower economic growth since the global financial crisis (green), and continual improvements in energy intensity (purple) and carbon intensity (orange).
The carbon intensity has improved due to an increased share of renewables, but a slight shift back to less efficient use of fossil fuels has tempered those gains.
The EU28 has pledged to reduce greenhouse gas emissions in 2030 to be 40% below the level in 1990. If this target translates equally to carbon dioxide emissions from fossil fuels only, then the EU28 would need to reduce emissions by about 1.5% per year for the next 15 years.
Particularly if economic growth returns to levels seen before the global financial crisis, the EU needs to continue strong efforts to improve energy and carbon intensity in the coming decade.
India has had robust emissions growth of around 5% per year in the last decade. This is driven by strong growth in economic activity (green), only tempered slightly by improvements in energy intensity (purple).
India has pledged to reduce its emission intensity by 33-35% in 2030 relative to 2005 levels, which translates to about 1.5% per year. This equates roughly to the current improvements in energy intensity, and achieving this target should be well within reach.
Because of the strong growth in economic activity, much larger than improvements in the emission intensity, it is expected that India’s emissions will grow strongly in the next decades.
Tracking progress of emission trends in the last decade is important to understand the short-term trajectory of emissions, but over the longer term, the deployment of existing technologies and the development of new technologies is needed.
Emissions scenarios consistent with limiting global warming to less than 2°C degrees above pre-industrial levels can give insight into potential pathways in the next decade or two, and we have drawn on the scenarios used in the IPCC Fifth Assessment Report.
The deployment of wind and solar power generation has received a lot of media attention in recent years, and these are key technologies moving forward.
Models have been critiqued for not sufficiently incorporating the cost declines of solar and wind technologies, particularly the International Energy Agency.
However, the scenarios assessed by the IPCC tend to be consistent with the historical evolution of wind and solar generation, assuming globally uniform climate policies starting in 2010, 2020 or 2030.
To continue the rapid rates of deployment of wind and solar over the next decades requires an acceleration of efforts as the annual increase in generation needs to become larger and larger each year to continue near exponential growth.
While there has been positive progress on wind and solar power generation, the same is not true for the deployment of carbon capture and storage.
In the 2000’s, there was high hopes of carbon capture and storage, but many projects never eventuate. Emission scenarios tend to be very optimistic towards carbon capture and storage, and this remains true in the latest generation of emission scenarios.
In the scenarios assessed by the IPCC in the Fifth Assessment report, carbon capture and storage could be as high 1 to 4 billion tonnes CO2 in 2020, depending on global climate policies. This could require the construction of 1,000 to 4,000 facilities with carbon capture and storage over a decade, compared to the tens currently in operation or planned.
If carbon capture and storage is not available at such scales, the costs of climate mitigation become higher, there is a risk society will place too little emphasis on near-term mitigation, and there needs to be a much more disruptive decline in fossil fuel consumption.
But wait, CO2 concentrations are still growing fast?
While the growth of carbon dioxide emissions have slowed, the concentration of carbon dioxide in the atmosphere has continued to grow at record high levels. This has led many to conclude that the emission statistics must be wrong. However, other explanations are more likely.
Even though emissions from fossil fuels and industry were flat in the last three years, it still means that emissions are at record high levels. Thus, a large increase in the atmospheric concentration should be expected.
In 2015, there was additionally a large El Niño event. This caused hotter and dryer conditions, reducing the uptake of carbon dioxide on land and increasing the risk of large fire events.
Even though carbon dioxide emissions from fossil fuels and industry have not grown in the last few years, when including emissions from land-use change the total emissions have grown. Particularly in 2015, there was a large increase in land-use emissions in Indonesia due to the El Niño event.
Combining the equal record high emissions from fossil fuels, industry, and land-use change, the extra emissions from fires, and the reduced uptake on land due to the hotter and dryer El Niño conditions, it is expected that atmospheric concentrations will have rapid growth.
Taking it a step further
In this blog, and indeed the article that forms the basis of the blog, I have only described the tracking framework with a few illustrative examples.
The only limitation to broadening the analysis is time and resources. The tools are available to provide deeper country level analysis, with closer links to the emission pledges and country-level climate policies.
Two key areas of science do, however, require further development.
First, the lower layers of the nested framework require the developments of methods and data to anticipate the future deployment of technologies based on indicators measured today, such as investments in research and development.
Second, emission scenarios consistent with limiting global warming to less than 2°C degrees have not adequately built in real world constraints, and new scenarios are needed to explore more realistic policy and technological configurations.