Climate policies energy policies Our papers

How much global warming should we expect under “business-as-usual” policies?

In this post, we summarise the key findings of our 2020 paper “How much human-caused global warming should we expect with business-as-usual (BAU) climate policies? A semi-empirical assessment” published in the scientific peer-reviewed journal, Energies.

Photo of Eiffel Tower and Pont Alexandre III at night by “Getfunky Paris” taken from Wikimedia Commons (Creative Commons Attribution 2.0 Generic)

This summary of our 2020 “How much human-caused global warming should we expect with business-as-usual (BAU) climate policies? A semi-empirical assessment” paper is also available at the CERES-science website.



The 2015 international agreement known as “the Paris Agreement” (or “Paris Accord”) states that all signed-up governments agree to limit the amount of human-caused global warming that will occur to less than 2°C (and ideally less than 1.5°C). But, what exactly does this mean? How much human-caused global warming should we expect to occur without this agreement, i.e., if everybody continued “business-as-usual” (BAU)? This was a question we attempted to answer in our 2020 paper “How much human-caused global warming should we expect with business-as-usual (BAU) climate policies? A semi-empirical assessment” published in the scientific peer-reviewed journal, Energies.

In this post, we summarise the key findings from that paper. We find that the Paris Agreement doesn’t have any practical meaning in terms of actual climate change. Rather, it is an arbitrary paper exercise over hypothetical numbers that don’t reflect the current scientific opinions on climate change. For instance, depending on which scientific studies you base your analysis on, the Paris Agreement could plausibly be kept without any governments doing anything different from business-as-usual until the 22nd century or later.

The 2015 Paris Agreement on climate change

In 2015, an international agreement was reached at an UN-organised meeting in Paris, France, that all governments would try to do their part to keep the amount of human-caused global warming below 2 °C (3.6 °F) above pre-industrial levels, and ideally below 1.5°C. This is known as the “Paris Agreement“. Reaching this agreement required a lot of international diplomacy and negotiation. And the stated goals initially seem very specific and impressive. Therefore, most people would assume that this agreement provides a very solid and precise set of actions and requirements for individual governments should implement. Meanwhile, superficially, you might (wrongly!) assume that anybody who objects to the Paris Agreement must somehow be wishing for the world to experience dramatic and risky global warming.

However, in a recent paper which we published in the scientific peer-reviewed journal, Energies, (How much human-caused global warming should we expect with business-as-usual (BAU) climate policies? A semi-empirical assessment), we showed that, when you actually consider the full scientific evidence available, the Paris Agreement is remarkably vague. And to meet the stated goals, governments would not actually need to carry out any change in their policies. Moreover, depending on how sensitive the Earth’s climate is to atmospheric greenhouse gases, the time at which the Paris Agreement would be “broken” under business-as-usual conditions could be anything from a few decades away to the 22nd century or later.

The approach of our paper

In our paper, we pointed out that there is a fundamental problem underlying the entire premise of the Paris Agreement. The Paris Agreement says that governments should try to implement national and international policies to keep the total amount of human-caused global warming below 2 °C, and ideally 1.5 °C. But, how much human-caused global warming would occur if the world continued “business-as-usual”? In other words, what is the “human-caused global warming baseline” against which efforts to meet the Paris Agreement are to be assessed?

This is the question we attempted to answer in this paper. However, while the question might initially seem fairly reasonable and straightforward, we found that it is remarkably challenging to answer satisfactorily. In essence, it depends on the answers to four separate questions:

  • Question 1. What would future greenhouse gas emissions be over the coming century under “Business-As-Usual” conditions?
  • Question 2. For each of the greenhouse gases, what is the relationship between emissions and actual changes in atmospheric concentrations?
  • Question 3. How different would global average temperatures be at present if greenhouse gases were still at “pre-industrial concentrations”? In other words, how do we define the “pre-industrial levels” of global average temperatures to which the Paris Agreement refers?
  • Question 4. How “sensitive” are global average temperatures to increases in the atmospheric concentrations of greenhouse gases?

With each of these questions, we found that there is considerable debate in the scientific literature. Therefore, in our paper, we systematically reviewed what answers different scientists have reached for each of the four questions. We then established plausible ranges for each of the answers that accurately reflected the different views among the scientific community. We then combined the results to generate plots of how much human-caused global warming we should expect under business-as-usual (BAU) conditions up to 2200 A.D.

It turns out that the politicians who negotiated the Paris Agreement were not aware of this debate and (mistakenly) assumed that each of these questions had already been largely resolved, and that all that was needed was to establish how much “global warming” was “too much”. Their conclusion apparently was 2 °C (and ideally 1.5 °C). But, as we will see, depending on the true answers to the four questions above, the time at which this “too much” target would be surpassed without doing anything could be as early as 2050 or as late as the 22nd century (or even later)!

What is the “climate sensitivity” and why is it so important?

In the 1960s and 1970s, early climate computer models assumed that changes in the atmospheric concentrations of the various greenhouse gases were the primary drivers of climate change. In particular, they assumed that carbon dioxide (CO2) was the most influential of the greenhouse gases. Because it had been recognised since the early 1960s that the concentration of CO2 in the atmosphere was slowly rising and that human activities were the most likely explanation, several modelling groups began speculating on how the climate would change if CO2 continued to increase at the same rate over the coming century or so.

The models all predicted that increasing CO2should cause substantial global warming. This was embedded into the models from the beginning. But, there was considerable disagreement over exactly how much global warming would occur for a given increase in CO2, e.g., a doubling of atmospheric CO2. The model predictions were hypothetical and at the time, global temperatures were actually falling.

That is, during the 1960s and 1970s, despite the computer model predictions that CO2 should have been causing global warming, the data itself was showing global cooling! See, e.g., Murray Mitchell Jr. (1961) (abstract; Sci-hub link); Kukla et al. 1977 (abstract; Sci-hub link). Still, the computer modellers argued that the global cooling which was occurring at the time was only a temporary thing, and that eventually their predictions would come true.

The ”equilibrium climate sensitivity” (ECS) definition

At any rate, most of the early computer climate models weren’t able to study gradual changes in climate over a few decades. Instead, the modellers would usually simulate the climate with one set of parameters. They would run the simulation until the climate stopped changing much and reached an equilibrium. Then they would change some of the parameters, e.g., doubling the concentration of CO2, and re-run the simulation until equilibrium. They would then compare the modelled climate of both runs.

By taking this approach, modellers would try to estimate the so-called, “equilibrium climate sensitivity” (ECS), i.e., how much global warming should we expect from a doubling of atmospheric CO2? However, the results were quite inconsistent. For instance, Manabe & Stouffer (1980) [abstract; pdf] suggested that doubling CO2 would cause 2 °C of global warming. But, an earlier study by the same group, Manabe & Wetherald (1975) [open access] had suggested that doubling CO2 would cause 3 °C of global warming.

Indeed, as can be seen from Figure 1, taken from Knutti et al. (2017) [abstract;Google Scholar], to this day, different studies suggest widely-different estimates for this ECS value. For their most recent 5th Assessment Report (2013), the IPCC concluded that the ECS could be anywhere from 1.5°C to 4.5 °C. However, other studies have suggested that the ECS is probably lower, e.g., Ziskin & Shaviv (2012) [abstract; Google Scholar]; Bates (2016) [open access]. Meanwhile other studies have suggested that it might even be higher, e.g., Marvel et al. (2016) [abstract; Google Scholar]; Zelinka et al. (2020) [open access].

Figure 1. Various estimates of the “transient climate response” (TCR) climate sensitivity. Adapted from Figure 1 of Knutti et al. (2017) [abstract; Google Scholar].

This has major implications that the groups deciding on the Paris Agreement didn’t seem to consider. If the climate sensitivity to a doubling of CO2 is 5 or 6°C then if CO2 concentrations increased by 30-40% relative to pre-industrial concentrations, the Paris Agreement would already be effectively broken. According to Antarctic ice core estimates, the pre-industrial CO2 concentration was 0.027-0.029%. Currently, it is 0.041%. So, by those estimates, CO2 concentrations is already 40-50% over pre-industrial concentrations. And the Paris Agreement would already have effectively been broken before it had been signed!

On the other hand, what if the equilibrium climate sensitivity is at the lower end of the IPCC’s current estimates, i.e., 1.5°C, or even lower, e.g., 1°C? Well, if it is 1°C, then presumably we could afford to quadruple pre-industrial concentrations before the 2°C limit would be reached. In that case, the Paris Agreement wouldn’t be particularly urgent. We could continue business-as-usual for several decades before even beginning to think about doing anything about it.

In other words, depending on the true value of the ECS, the Paris Agreement could be anything from irrelevant because it is already effectively broken to irrelevant because the climate sensitivity to CO2 is relatively low. It is only if the ECS is somewhere in between those two extremes that the Paris Agreement could have any importance. And even then, the urgency (or lack of urgency) of the Agreement would depend on where exactly the true ECS lies.

The ”transient climate response” (TCR) definition

You might be wondering, “Why don’t people just look at how much CO2 has increased to present and then how much global warming has already occurred over the same period? Then, you could extrapolate how much global warming to expect for a doubling of CO2 and the problem would be solved.”

It’s a good question, but there are several problems:

  1. As mentioned above, during the 1950s-70s, as CO2 was increasing, temperatures were decreasing. The modellers justified this contradiction by arguing that their predicted warming is a long-term prediction, and so couldn’t be evaluated by a few decades of data.
  2. What if some (or all) of the global warming up to present was due to other factors than CO2, e.g., natural climate cycles? Then that fraction of the global warming wouldn’t be “CO2-driven warming”. Some supporters of lower values of the climate sensitivity argue for this.
  3. Similarly, what if some of the warming was due to artificial biases in the data, such as failing to adequately correct for urban heat islands in the temperature records? Again, that fraction wouldn’t be “CO2-driven warming”.
  4. On the other hand, some supporters of higher climate sensitivity values argue that human emissions of air pollution (specifically “stratospheric sulphur aerosols”) have been adding a temporary “global cooling” effect. They argue that this is partially masking the CO2-driven warming, but only temporarily.

Those are just a few of the main problems, but it should be enough to see why after more than half a century of speculation, different researchers still have such widely different views over what the actual “equilibrium climate sensitivity” (ECS) is.

Nonetheless, since establishing the true climate sensitivity is so fundamental to deciding how big a concern CO2-driven warming is, a number of researchers have proposed a more practical definition of the “climate sensitivity”. This is the so-called “transient climate response” (TCR).

The various TCR definitions are usually quite technical, e.g., “the expected global temperature change at the time of CO2 doubling after 70 years of CO2 increasing at an average of 1% per year”.

But, roughly speaking, the TCR definition means, “how much actual global warming would we expect to physically observe for a doubling of CO2 in real-time?”.

In principle, this should be easier to calculate from looking at how much global warming has already occurred. But, the problems mentioned above still apply. So, even for this “easier” and “more pragmatic” definition, different studies give different estimates. This can be seen by studying Figure 2 – again taken from Knutti et al. (2017) [abstract;Google Scholar].

Figure 2. Various estimates of the “equilibrium climate sensitivity” (ECS) climate sensitivity. Adapted from Figure 2 of Knutti et al. (2017) [abstract; Google Scholar]

Our findings

Figure 3 shows the main results of our analysis in the paper. Figure 3a shows the results of our analysis for a TCR of 0.5, 1.0, 1.5, 2.0, 2.5, and 3.0 °C, while Figure 3b shows the results for an ECS of 1, 2, 3, 4, 5, and 6 °C. If the ECS is 5 °C or higher, or the TCR is 2.5 °C or higher, then under Business-As-Usual, we are projected to have broken the 2 °C target by 2045–2053. On the other hand, if the ECS is 2 °C, we are not projected to break the 2 °C target until the 22nd century. Similarly, if the TCR is 1.5 °C, then we are not projected to break the 2 °C target until the 22nd century (or 2095 at the earliest). Meanwhile, if the ECS or TCR is 1 °C or less, then we are not projected to break either of the two targets in the 21st century under BAU.

Figure 3. Projected human-caused global warming up to 2100 under “business-as-usual” (BAU) conditions for various estimates of the “climate sensitivity”, i.e., the expected global warming caused by a doubling in atmospheric carbon dioxide (CO2). The top panel shows the projections for “TCR” climate sensitivity estimates and the bottom panel shows the equivalent “ECS” estimates. For comparison, the two stated goals of the Paris Agreement are shown with dashed horizontal lines at 2 and 1.5°C.


The urgency (or otherwise) of the Paris Agreement depends critically on what the actual value of the “climate sensitivity” is, i.e., how much global warming would we expect for a doubling of atmospheric CO2.

There are two main definitions of the “climate sensitivity”: (1) the equilibrium climate sensitivity (ECS) and (2) the transient climate response (TCR). [Estimates of the ECS tend to be larger than those for the TCR]

According to the UN IPCC’s 5th Assessment Report, the ECS is “likely” to be any value in the range 1.5–4.5 °C and the TCR is “likely” to be any value in the range 1.0–2.5 °C. However, as we saw from Figure 3, those upper and lower bounds lead to dramatically different conclusions over how urgent the Paris Agreement is. That is, the results of the latest IPCC Assessment Report still do not tell us whether the Paris Agreement is trying to solve a problem for the next few decades or the 22nd century.

Moreover, several studies have suggested that the climate sensitivity may be either higher or lower than the IPCC’s “likely” ranges. For instance, Zelinka et al. (2020) [open access] has suggested that the ECS might be greater than 4.5 °C. On the other hand, Ziskin & Shaviv (2012) [abstract; Google Scholar] and Bates (2016) [open access] have both suggested that the ECS is probably only about 1°C, i.e., less than the IPCC’s “likely range”.

Meanwhile, we have argued elsewhere in Soon et al. (2015) [abstract; pre-print pdf] that the TCR is probably less than 0.44 °C. Further, we showed that it is possible to explain all of the observed warming since at least 1881 in terms of natural climate change, i.e., that no CO2 role is needed to explain any of the warming to present.

Therefore, the Paris Agreement’s stated goal of “Holding the increase in the global average temperature to well below 2 °C above pre-industrial levels and pursuing efforts to limit the temperature increase to 1.5 °C above pre-industrial levels” might sound decisive and impressive on paper. But, in terms of actual science or actual climate change or actual real-life policies, the agreement is utterly meaningless.

Overview of the paper

The review was published as an open-access paper in the scientific peer-reviewed journal, Energies, and can be downloaded for free from the following URL: (pdf).

The full citation is as follows: Connolly, R.; Connolly, M.; Carter, R.M.; Soon, W. How Much Human-Caused Global Warming Should We Expect with Business-As-Usual (BAU) Climate Policies? A Semi-Empirical Assessment. Energies 2020, 13, 1365. (pdf).