Climate Sensitivity Estimates: Declining or Not?

Guest “climate sensitivity explaining by David Middleton

  • ECS: Equilibrium Climate Sensitivity (AKA equilibrium sensitivity)
  • TCR: Transient Climate Response (AKA transient sensitivity)
  • AKA: Also Known As

Here’s how the folks a the Geophysical Fluid Dynamics Laboratory (GFDL) explain the difference:

Transient and Equilibrium Climate Sensitivity

Projections of the severity of anthropogenic climate change are strongly dependent on our estimates of climate sensitivity, traditionally defined as the global average warming at the Earth’s surface due to a doubling of the carbon dioxide from pre-industrial levels. This importance arises not because global temperature change directly causes all of the impacts of major concern, but because many effects of climate change are predicted to increase in severity with larger global warming.

An important distinction is made between the equilibrium sensitivity — the temperature change realized after allowing the climate system to equilibrate with a higher value of CO2 — and the response on shorter time scales, before the deep oceans have had time to equilibrate, that is of more direct relevance to the changes we are likely to see in the 21st century. The latter is often quantified by raising the carbon dioxide in a model at the rate of 1% per year and examining the response at the time when carbon dioxide concentration has doubled, referred to as the transient climate sensitivity or response. (At a rate of 1% per year, doubling requires 70 years.)

Equilibrium sensitivities in global climate models typically range from 2 to 5K, while the transient climate responses are smaller, in range of 1.0-2.5 K, due to the cooling influence of ocean heat uptake. 

[…]

The ratio of transient to equilibrium sensitivity varies from 1/3 to 1/2 in this group of GFDL models indicating significant variation of the transient cooling influence of the ocean. The relationship of the ocean’s cooling influence to ocean heat uptake and circulation changes has been an ongoing thread of GFDL research. For example, He et al (2017) noted that stronger deep ocean circulation prior to forcing reduced the magnitude of transient warming in a GFDL model.

Cloud feedbacks are widely considered to contribute the largest uncertainty to climate sensitivity. Simulated climate sensitivity varies considerably with choices made about cloud parameterizations that are not well constrained by observations (Zhao et al 2016). Simulated cloud responses depend on the pattern of surface temperature change, not just its global magnitude (Silvers et al 2018). Because of the importance and complexity of the interactions of clouds and climate GFDL is focusing effort on a cloud climate initiative.

[…]

GFDL

Here’s a graphical illustration from IPCC TAR, 2001:

Figure 1. ECS vs TCR (IPCC AR1, 2001)
  • IPCC: Intergovernmental Panel on Climate Change
  • TAR: Third Assessment Report

TCR occurs simultaneously with the rise in atmospheric CO2. While, the difference between ECS and TCR occurs over the next several hundred years.

I am not claiming that this is correct or empirically verifiable. I’m just explaining the concept. These are the sensitivities of the GFDL climate models:

Model Transient Climate Response  Equilibrium Climate Sensitivity
CM2.1 1.5 K (Randall et al 2007) 3.4 K (Stouffer et al 2006)
ESM2M 1.3 K (Flato et al 2013) 3.3 K (Paynter et al 2018)
ESM2G 1.1 K (Flato et al 2013) 3.3 K (Krasting et al 2018)
CM3 2.0 K (Flato et al 2013) 4.8 K (Paynter et al 2018)
CM4 2.1 K (Winton et al submitted) 5.0 K (Winton et al submitted)
ESM4 1.6 K (Dunne et al in prep) 3.2 K (Dunne et al in prep)
Table 1. GFDL model climate sensitivities.

In the ESM4 model, the average temperature would rise 1.6 °C over a period of about 70 years as atmospheric CO2 and then another 1.6 °C over the subsequent 430 years. This doesn’t strike me as particularly catastrophic, considering that the first 1.0 °C or so has already occurred.

This brings me to the reason I wrote this post. Andy May’s post from yesterday featured this graph from Scafetta et al., 2017:

Figure 2. The decline in estimates of ECS from 2000 to 2015. Source: Scafetta, Mirandola, and Bianchini, 2017.

A couple of commentators posted rebuttals of the decline in climate sensitivities featuring this graph from Carbon Brief:

Figure 3. Stable climate sensitivities from Carbon Brief.

I think, but I’m not certain, that the climate sensitivities in the Scafetta paper were all based on instrumental data (observations). The climate sensitivities in the Carbon Brief article were derived by multiple different methods:

  1. Models (GIGO (Garbage In Garbage Out))
  2. Constrained models (higher quality GIGO)
  3. Instrumental (actual observations)
  4. Paleo (reconstructions of past climate based on proxies)
  5. Combined approaches (whatever)

The author, Zeke Hausfather, very kindly provided a link to the Excel file in the article. I downloaded the Excel file and plotted only the instrumental estimates of ECS, the real data.

ECS study year  min   max   ECS mean 
Harvey and Kaufmann 2002 2002   1.0          3.0              2.0
Gregory et al. 2002 2002   1.6        10.0              2.1
Kaufmann and Stern 2002 2002   2.0          2.8              2.6
Knutti et al. 2002 2002   2.0          9.2              4.8
Frame et al. 2005 2005   1.2          5.2              2.3
Tsushima et al. 2005 2005   3.1          4.7              3.8
Forster and Gregory 2006 2006   1.0          4.1              1.6
Forest et al. 2006 2006   2.1          8.9              4.1
Stern et al. 2006 2006   4.4          4.5              4.4
Chylek et al. 2007 2007   1.1          1.8              1.6
Schwartz 2007 2007   0.9          2.9              1.9
Lindzen and Choi 2009 2009   0.4          0.5              0.5
Murphy et al. 2009 2009   0.9        10.0              3.0
Lin et al. 2010 2010   2.8          3.7              3.1
Lindzen and Choi 2011 2011   0.5          1.1              0.7
Aldrin et al. 2012 2012   1.2          3.5              2.0
Schwartz 2012 2012   1.5          6.0              3.0
Lewis 2013 2013   1.0          3.0              1.6
Otto et al. 2013 2013   0.9          5.0              1.9
Bengtsson and Schwartz 2013 2013   1.5          2.5              2.0
Otto et al. 2013 2013   1.2          3.9              2.0
Skeie et al. 2014 2014   0.9          3.2              1.8
Loehle 2014 2014   1.8          2.3              2.0
Lewis 2014 2014   1.2          4.5              2.2
Kummer and Dessler 2014 2014   1.6          4.1              2.3
Lovejoy 2014 2014   2.5          3.7              3.1
Donohoe et al. 2014 2014   3.1          3.2              3.1
Urban et al. 2014 2014   2.1          4.6              3.1
Monckton et al. 2015 2015   0.8          1.3              1.0
Loehle 2015 2015   1.5          1.6              1.5
Lewis and Curry 2015 2015   1.1          4.1              1.6
Cawley et al. 2015 2015   1.8          4.4              2.0
Johansson et al. 2015 2015   2.0          3.2              2.5
Johansson et al. 2015 2015   1.6          7.8              3.1
Bates 2016 2016   1.0          1.1              1.0
Lewis 2016 2016   0.7          3.2              1.7
Loeb et al. 2016 2016   0.8        10.0              2.0
Forster 2016 2016   1.1          5.3              3.0
Armour 2017 2017   1.7          7.1              2.9
Lewis and Curry 2018 2018   1.2          3.1              1.8
 Average               2.3
 σ               0.9
 -2σ               0.4
 +2σ               4.2
Table 2. Instrumental ECS estimates

The average ECS was 2.3 °C. This would translate to a TCR of 1.2-1.6 °C.

Then I graphed it up two ways:

  1. A stock chart, similar to a candlestick plot.
  2. A simple scatter plot.
Figure 4. Declining ECS from instrumental estimates, stock plot.

Multiple estimates published in the same year is why several years appear multiple times.

Figure 5. Declining ECS from instrumental estimates, scatter plot.

We can clearly see that ECS estimates derived from instrumental observations have indeed declined from around 3.0 °C in 2001 to around 2.0 °C in 2018.

Scatter plots of the other methods demonstrate that the problem is models, “all the way down”…

Figure 6. Garbage In, Garbage Out

Paleo methods produce a higher sensitivity because, almost invariably, the CO2 proxies are of much lower resolution than the temperature proxies. This is particularly true of ice core data.

Like this:

Like Loading…

Related

via Watts Up With That?

https://ift.tt/35vcpFl

November 14, 2020 at 04:29AM

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 )

Google photo

You are commenting using your Google 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 )

Connecting to %s