Climate sensitivity and visual rhetoric.

 

I found another visual way to depict an aspect of climate science in the course of a discussion. The issue was what value of climate sensitivity can be used to inform policy. The amount the global average surface temperature will rise for a doubling of CO2 is often chosen for its policy implications rather than for its scientific credibility. The problem is there are a range of values, and no good reason to prefer one over another.


The context

Climate sensitivity comes in a variety of versions defined in relation to the timescale and input.

Skeptical science do a fair job of description –

Or wiki or the IPCC
there are three main sources of research that inform our knowledge of these values;-
1)- Energy balance calculations based on observations and simple adjustments for the changing thermal capacity of the air/land/ocean system over time.
2)- Complex physical simulations of the climate system using GCMs from which values for the climate sensitivity emerge.
3)- Inferring from the proxy paleoclimate record how the climate responded in the past to comparable inputs.

All the methods have flaws, which the proponents and enthusiasts for one that may give values they prefer, are always eager to point out.

Energy balance calculations tend to give lower values. problems with the observations used and the adjustment for ‘thermal inertia’ are invoked.
GCMs produce a well constrained range of values. But with the suspicion that the outliers may have been discarded because they did not conform to expectations.
Values derived from paleoclimate data are often caught in uncertainty about the accuracy of the temperature or forcing that is derived from the proxy data. sometimes the data can have a degree of co-dependency.

The bottom line is that climate sensitivity, of whichever flavour, has been somewhere around 3degC since the first guess in 1903.

There is an assumption that a lower climate sensitivity will mean less warming (and/or) more slowly which alters the urgency and scale of policy choices that may have to be made.

This assumption is almost certainly mistaken – an issue that may be discussed at greater length *

However the published research despite over a hundred years of effort has narrowed down climate sensitivity to…
something between 1.8degC ~ 4.2degC with 90% confidence. (!)

ecs-range3Reto Knutti & Gabriele C. Hegerl 2008

One of the main reasons that climate sensitivity has proven so difficult to define as a tightly constrained number is because the metric is inherently uncertain in the definition.

The outcome is there is no information available that can exclude any value within the range. Without special pleading for the superiority of one method of determining ECS there is also no way of preferring one value over another. Because all three methods define the range you cannot argue that one method is better if another method is excluded because any judgement of how good a method is can only be made in the context of comparison with the results from other methods.

So in an attempt to explain why I think all numbers in the range of ECS are equally likely, rather like throwing a dice, I made the following motion graphic. The curves and error ranges are illustrative rather than accurate. It is a variation on the more accurate information in the graphic above. This is a piece of visual rhetoric, which I find easier to construct than a lot of words, but possibly just as misleading!

ecs_pdf3b

In the interests of full disclosure; the peaks chosen and the error ranges are not derived from any actual values. They should be closer and the error ranges should overlap much more. But it looked better with this spacing. There was also an intentional ordering of the layers and colours of the three lines and error regions. While the graphic may be making a case for the equal probability of any value in the range, the higher values just happen to be more brightly coloured and in front of all the others. Visual rhetoric to imply subconsciously at least that such higher valued are predominant!

*Climate sensitivity
It should be a subject of a longer post, but I will leave here some earlier comments made elsewhere for future reference.
I think the reliance on climate sensitivity as a means of making policy on adaption may be equally misleading. It will be the local and seasonal peaks in the climate response that are most damaging and costly. Perhaps seen already with the last 2 years just a 1degC rise above the average, but large increases in flooding, wildfires and hurricane intensity.
The global average temperature rise may be the least of our problems.
Climate sensitivity is not a simple metric of damage. While climate sensitivity may be hard to define, or confine to a narrow range, the input, the extra energy absorbed by the land/sea/air system is much better constrained. Direct measurement, and the expansion of the oceans makes the amount of thermal energy added much less uncertain.
The most effective way of re-establishing energy equilibrium is by a global rise in temperatures because of the T^4 relationship. Whatever the ESS or TCR eventually temperature will rise somehow to restore equal energy flows. But there are other mechanisms and patterns of regional temperature rise that can balance the energy flows with different effects on the global average.
It is unwise to assume that a process that reduces the climate sensitivity metric by a process that offsets the global rise, (sequester more energy into the oceans, or increases polar amplification) would therefore be less damaging to our civilisation’s infrastructure than another degree rise in global temperatures.
Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s