Climate Sensitivity - Explanation

[WYorksWeather]

I'm going to do a bit of a write up on climate sensitivity, which is one of the most important numbers in climate science. I think it is also misunderstood quite often, so hopefully I can clear up a few misconceptions.

Definitions

Broadly speaking, there are two types of climate sensitivity figures that are commonly quoted.

The first is the Transient Climate Response (TCR). Put simply, this is the expected change in temperature with an increase in CO2 concentration of 1% per year, at the point at which atmospheric CO2 has doubled (approximately 560ppm). This is a good estimate of the relatively short-term warming over decades.

The second is the Equilibrium Climate Sensitivity (ECS). This measure is the expected change in temperature after a doubling in atmospheric CO2, and after full equilibrium is reached. This includes slow feedback processes such as the melting of ice sheets, which may not reach equilibrium for centuries, or even millenia.

In almost all cases, ECS > TCR. Where I use the term climate sensitivity from here on in, I am referring to ECS.

Radiative forcing is the impact of an agent on the Earth's energy balance, measured in watts per square metre (W/m^2).

The Planck response

The simplest way of lower-bounding climate sensitivity is to simply look at temperature change implied by a doubling of CO2 with no further feedback responses. The forcing caused by a doubling of CO2 is roughly 3.7W/m^2 (compared to an estimated net anthropogenic forcing of 2.7W/m^2 in the IPCCs AR6 report). Derived from basic physical laws, warming to equilibrium if Earth behaved as a black body is estimated at 0.27C/W/m^2. This therefore implies an absolute lower bound of around 1.0C warming for a doubling of CO2. 

Of course, that isn't the end of the matter. The Earth has of course warmed by around 1.2C already compared to pre-industrial, and we are only just past half a doubling (logarithmically, since 280 * sqrt(2) = 396mb). There is a need to consider both positive and negative feedbacks.

Positive feedbacks

Ice-albedo feedback is the first to consider. For those not familiar, the albedo of an object is a value between 0 and 1 which determines its propensity to reflect sunlight. An object with an albedo of 0 absorbs all incoming solar radiation, whilst an object with an albedo of 1 reflects all incoming solar radiation. In the case of a warming Earth, this results in a positive feedback. A warmer planet has a lower average albedo, as sea ice and snow (albedo typically 0.5 or higher) is replaced by open ocean (albedo <0.1) or bare ground (albedo ~0.2), and hence the Earth absorbs more incoming solar radiation. This also largely explains the phenomenon of Arctic amplification, which is the fact that the Arctic has warmed much faster than the global average.

Another fairly trivial feedback to understand is the water vapour feedback. As the Earth warms, the maximum water vapour capacity of the atmosphere tends to increase, by about 7% per degree Celsius of warming. This is one of the strongest feedbacks, and most sources quote it as leading to roughly a doubling of the initial Planck response.

Negative feedbacks

Aerosol emissions are perhaps the most well known negative feedback. An increase in SO2 emissions as a result of air pollution tends to reduce temperatures by increasing albedo. The IPCCs AR6 estimate of the aerosol forcing is -1 W/m^2, albeit with wide confidence intervals. The size of this effect makes up the majority of the uncertainty in estimates of climate sensitivity.

Natural CO2 removal is another negative feedback. Unfortunately for us, this is very slow on a human timescale. The majority of anthropogenic CO2 will be absorbed by the oceans over centuries, whilst chemical weathering and biological processes would eventually remove what remains, but only over millennia.

Other feedbacks

There are other feedback processes that may act, and are somewhat less certain in terms of size and effect. Potential positive feedbacks include those relating to the carbon cycle (melting permafrost releasing methane, forest dieback due to increased wildfires). Potential negative feedbacks include those relating to uptake of CO2 by the biosphere, or increases in SO2 (which causes a negative forcing) as a result of dimethyl sulphide produced by phytoplankton (known as the CLAW hypothesis).

There are many other feedbacks I have not mentioned here, otherwise this post would turn into a book!

Overall impact

Overall, the feedback responses amplify warming beyond that predicted by the Planck response. The latest IPCC AR6 report states, based on the available research, that ECS has a 66% probability of being between 2.5C and 4.0C, and a 90% probability of being between 2.0C and 5.0C. This estimate has changed remarkably little from the first IPCC report in 1990, which suggested a likely range of 1.5C to 4.5C, and a best estimate of 3C.

IPCC scenarios vary on if or when a CO2 doubling is expected to occur.

In SSP1-2.6 (an ambitious scenario likely to limit warming to <2C), CO2 does not double.

In SSP2-4.5 (often considered a good current policy analogue) , doubling occurs around 2070.

In the pessimistic scenarios of SSP3-7.0 and SSP4-8.5, doubling occurs around 2050. In the latter scenario, a second doubling occurs by 2100! Fortunately, these scenarios are generally considered highly unlikely by most climate scientists.

One of the things people often fail to appreciate is that any targets to keep warming below a particular level are inherently probabilistic.  SSP1-2.6, which is commonly described as a 2C compliant pathway, still shows a lot of variation by climate sensitivity. The central estimate is 1.8C, and the 90% likely range is 1.3C to 2.4C.  Optimistically, this scenario, which can be roughly described as net zero by 2075, could limit warming to below 1.5C. Pessimistically, it may fail to limit warming to 2C.

Uncertainties for each scenario are shown in graphical form below.

Happy to provide more links or discuss further.

Hope this late Friday night post provides some useful insights, if nothing else.