GWPF optimism on climate sensitivity is ill-founded
Posted on 10 March 2014 by dana1981
The UK anti-climate policy advocacy group Global Warming Policy Foundation (GWPF) has published a report written by Nic Lewis and Marcel Crok claiming “the IPCC hid the good news” regarding climate sensitivity. Lewis is an amateur researcher and retired financier who has published a few papers estimating climate sensitivity, and Crok is a freelance science writer.
GWPF asked climate scientist Judith Curry to write the Foreword to the report, presumably to lend it more credibility. However, Curry has no publications or expertise in this area, and once said that the global equilibrium climate sensitivity could fall anywhere between 0 and 10°C for doubled CO2. This comment is totally incompatible with the body of climate sensitivity research, and also with the GWPF report.
The report itself is essentially a commentary and includes no new information. It boils down to Lewis and Crok trying to make the case that climate sensitivity is on the lower end of the IPCC estimated range. In the report, they find reasons to dismiss the many studies and varying approaches that arrive at higher climate sensitivity estimates, and fail to discuss the shortcomings of the lower sensitivity studies that they prefer. In short, it’s a very selective and biased review of the scientific literature on the subject. Recent papers by Gavin Schmidt and Drew Shindell at NASA GISS, not considered in the GWPF report, entirely contradict its conclusions, for example.
There are a few main methods to estimate the global climate sensitivity; Lewis and Crok focus on three of these and present their case for why each should be considered valid (when yielding low sensitivity results) or disregarded (when yieliding moderate or high sensitivity results). Here we’ll look at each, including evidence the GWPF report failed to consider, and show that their conclusions are not supported when the full body of research is considered. As climate scientist Steven Sherwood described it,
"The report is standard cherry-picking. It offers no new evidence not already considered by the IPCC, relying very heavily on a few strands of evidence that seem to point toward lower sensitivity while ignoring all the evidence pointing to higher sensitivity.
It relies heavily on the estimate by Forster and Gregory, which was an interesting effort but whose methodology has been shown not to work; this study did not cause the IPCC to conclude that sensitivity had to be low, even though both Forster and Gregory were IPCC lead authors and were obviously aware of their own paper."
However, the good news is that the report is consistent with the 97 percent expert consensus on human-caused global warming. It acknowledges that global warming will continue as long as humans continue increasing the greenhouse effect, and merely suggests that future warming will be toward the lower, slower end of the IPCC estimates. As climate scientist Ed Hawkins at the University of Reading also noted,
"Remarkably for a report published by the GWPF, the authors agree with mainstream climate scientists that significant further warming is expected ... It is great to see the GWPF accepting that business-as-usual means significant further warming is expected. Now we can move the debate to what to do about it."
Paleoclimate studies attempt to estimate climate sensitivity based on the forcings and temperature responses from climate change events in the geologic record. The most robust study of this type was done by the PALEOSENS team, published in Nature in 2012. This study evaluated past climate changes over the previous 65 million years, considering nearly two dozen investigations of many different geological time periods.
The study estimated with 68 percent probability that the equivalent equilibrium climate sensitivity is between 2.2 and 4.8°C for a doubling of CO2, generally consistent with IPCC estimates, and inconsistent with the lower estimates preferred by GWPF. The 95 percent confidence range in this study was between about 1 and 7°C equilibrium sensitivity, so very low and very high climate sensitivities could not be ruled out, but are relatively unlikely, based on the historical record. Additionally, the GWPF report uses 68 percent confidence ranges throughout, so the 2.2 and 4.8°C PALEOSENS paleoclimate estimate is inconsistent with the GWPF low sensitivity conclusions.
Various paleoclimate-based equilibrium climate sensitivity estimates from a range of geologic time periods. Adapted from PALEOSENS (2012) Figure 3a by John Cook.
The GWPF report has very little discussion of paleoclimate sensitivity estimates. They just say that these studies don’t tightly constrain the possible climate sensitivity range, and past climate states are different than current and future climate states, so “little weight can be put on the palaeoclimate estimates.” While there is some truth to these critiques, entirely disregarding the results of these studies is simply not justifiable.
In summary, paleoclimate studies provide one line of evidence that supports an equilibrium climate sensitivity between about 2 and 4.5°C, and the GWPF justification for dismissing these estimates is weak.
General Circulation Models
Climate models (general circulation models or GCMs) provide another method by which to estimate climate sensitivity. The physics of the climate system are input into very detailed climate models, which can then estimate how the global temperature will respond to various forcings. The results can give us projections of future global warming under a variety of scenarios, and also give us an estimate of the global climate sensitivity. Most GCM equilibrium climate sensitivities range between 2 and 4.5°C (average 3.2°C in GCMs used in IPCC AR5). This range is consistent with paleoclimate estimates.
Lewis and Crok make the following argument.
"Between the Fourth and Fifth [IPCC] Assessment Reports the best estimate of the cooling effect of aerosol pollution was greatly reduced. That necessarily implies a substantially lower estimate for climate sensitivity than before. But the new evidence about aerosol cooling is not reflected in the computer climate models. This is one of the reasons that a typical climate model has a substantially higher climate sensitivity than would be expected from observations: if a model didn’t have a high climate sensitivity, its excessive aerosol cooling would prevent it matching historical warming."
However, according to climate modeler Gavin Schmidt of NASA GISS, this is incorrect.
“Their logic is completely backwards. Climate model sensitivity to a doubling of atmospheric CO2 is intrinsic to the model itself and has nothing to do with what aerosol forcings are. In CMIP5 there is no correlation between aerosol forcing and sensitivity across the ensemble, so the implication that aerosol forcing affects the climate sensitivity in such 'forward' calculations is false ... The spread of model climate sensitivities is completely independent of historical simulations.”
Climate scientist Kevin Trenberth also notes that the change in the estimated aerosol forcing is mainly associated with indirect aerosol effects, but half of GCMs don’t include these indirect effects, and those that do actually tend to simulate less warming.
“Some other models like CESM1 did include microphysics and an indirect aerosol effect, and had slightly lower 20th Century warming than observed … yet its climate sensitivity is higher than for [some other models that don’t include the indirect aerosol effect] … the [GWPF] comment presumes that models have been tuned to reproduce the 20th Century temperature record, but this is mostly not true”
This point was also made by Schmidt et al. (2014), which additionally showed that incorporating the most recent estimates of aerosol, solar, and greenhouse gas forcings, as well as the El Niño Southern Oscillation (ENSO) and temperature measurement biases, the discrepancy between average GCM global surface warming projections and observations is significantly reduced. This approach also accounts for the previously underestimated volcanic aerosol forcing, demonstrated by Santer et al. (2014), but not included in the GWPF report.
GCM mean (dark blue #1) and envelope (lighter blue) range of global surface temperature projections vs. HadCRUT4 (red #1) and Cowtan & Way (red #2) global surface temperature instrumental estimates. The GCM mean result incorporating changes in ENSO and updated solar and aerosol forcings (blue #2 and #3) are also shown. Adapted from Schmidt et al. (2014) by Kevin Cowtan.
In summary, GCMs provide another line of evidence that generally supports an equilibrium climate sensitivity between about 2 and 4.5°C, and the GWPF justification for dismissing these estimates is incorrect.
The method preferred by the GWPF report, and that which Lewis has used in his own papers, involves estimating climate sensitivity using a combination of recent instrumental temperature data (including ocean heat content data), less complex climate models, and statistics. A few studies using this approach since about 2012 have begun yielding lower climate sensitivity estimates. In their report, GWPF cite Ring et al. (2012), Aldrin et al. (2012), Lewis (2013), and Otto et al. (2013) as all yielding central equilibrium climate sensitivity estimates between 1.76 and 2.00°C. However, the GWPF report only references the “main results” of Aldrin et al. (2012), whose study actually estimated equilibrium climate sensitivity of about 2.5 or 3.3°C when accounting for cloud and indirect aerosol effects. Aldrin et al. wrote,
“Thus, the estimate from our original analysis should be interpreted with care. In further work, the uncertainty of the cloud lifetime effect should also be taken into account.”
The GWPF report, however, did not interpret their estimate with care. It simply used the result that was convenient for their argument, and left out the cloud uncertainties.
As for Lewis (2013), it's not without its own red flags. As documented at And Then There's Physics, when using data up to 1995, the method yields an estimated climate sensitivity range of 2.0–3.6°C, but incorporating an additional 6 years of data reduces the estimate approximately 33 percent, to 1.2–2.2°C. Climate sensitivity is a relatively constant parameter; if adding just 6 years of data changes the result so dramatically, one should really question the method being used. Instead, Lewis argues that it's the only reliable method for estimating climate sensitivity.
The challenge with this 'instrumental' method of estimating equilibrium sensitivity is that it’s based on transient instrumental measurements. There is currently a global energy imbalance, and reaching a new equilibrium state will take over a century. Therefore, estimating equilibrium climate sensitivity based on measurements of a climate that’s out of equilibrium requires making some significant assumptions, for example that feedbacks will remain constant over time. However, several recent studies have suggested that these assumptions may not be correct. For example, Armour et al. (2013),
"Time-variation of the global climate feedback arises naturally when the pattern of surface warming evolves, actuating regional feedbacks of different strengths. This result has substantial implications for our ability to constrain future climate changes from observations of past and present climate states."
“Results imply that global and regional warming rates depend sensitively on regional ocean processes setting the [ocean heat uptake] pattern, and that equilibrium climate sensitivity cannot be reliably estimated from transient observations.”
“We demonstrate that a single realization of the internal variability can result in a sizable discrepancy between the best [climate sensitivity] estimate and the truth. Specifically, the average discrepancy is 0.84°C, with the feasible range up to several °C. The results open the possibility that recent climate sensitivity estimates from global observations and [intermediate complexity models] are systematically considerably lower or higher than the truth, since they are typically based on the same realization of climate variability.”
Trenberth and Fasullo (2013) also note that ocean heat content (OHC) variability can strongly impact the 'instrumental' climate sensitivity estimates (emphasis added).
"Climate sensitivity estimates are greatly impacted by such variability especially when the observed record is used to try to place limits on equilibrium climate sensitivity [Otto et al., 2013], and simply using the ORAS-4 estimates of OHC changes in the 2000s instead of those used by Otto ... changes their computed equilibrium climate sensitivity from 2.0°C to 2.5°C, for instance. Using short records with uncertain forcings of the Earth system that is not in equilibrium does not (yet) produce reliable estimates of climate sensitivity."
None of these papers or concerns with ‘instrumental’ climate sensitivity estimation methods are mentioned in the GWPF report. Instead, the report argues that this approach provides the only reliable method for estimating climate sensitivity, and that all other methods that produce higher estimates (e.g. paleoclimate and GCMs) are wrong.
However, an important new paper just published by Drew Shindell at NASA GISS reconciles the difference between the climate sensitivity estimates in these varying approaches, but not in the direction advocated by the GWPF report. Shindell notes that the ‘instrumental’ approach studies preferred by the GWPF report assume that the global mean temperature response to all forcings is equal. His study investigates this assumption by comparing GCM temperature responses to greenhouse gases with their responses to aerosols and ozone.
Shindell, who was a co-author on Otto et al. (2013), notes that “forcing in the NH extratropics [above 30° latitude] causes a greater global mean temperature response than forcing in the tropics”; a result noted by Hansen et al. (1997):
“A forcing at high latitudes yields a larger response than a forcing at low latitudes. This is expected because of the sea ice feedback at high latitudes and the more stable lapse rate at high latitudes”
The forcing from aerosols and ozone isn’t globally uniform, but instead focused more in the northern hemisphere extratropics. Hence it results in a relatively larger temperature response than an equivalent forcing from greenhouse gases, which are well mixed throughout the atmosphere.
When assuming equal sensitivity to all forcings, Shindell estimates the transient climate response (TCR) at 1.0–2.1°C, most likely 1.4°C, which is almost identical to the Lewis GWPF report estimate (1–2°C, most likely 1.35°C) and also similar to the estimate in Otto et al. (2013). However, when Shindell accounts for the higher sensitivity to the aerosol and ozone forcings, the estimated TCR range rises to 1.3–3.2°C, most likely 1.7°C. Compared to the IPCC estimated TCR range of 1–2.5°C, and the range in climate models of 1.1–2.6°C, Shindell's results give a low probability for the low end of the range and higher probability for the high end; the opposite of the GWPF report. Given the strong correlation between TCR and equilibrium climate sensitivity, Shindell’s results also suggest that the lower climate sensitivity estimates are unlikely to be accurate.
Accounting for Cloud and Water Vapor Observations
The GWPF report also notes that changes in cloud cover in a warming world are a key to determining climate sensitivity. On this topic the report merely claims “Observational evidence for a positive cloud feedback is weak, at best.” However, there have been several studies comparing observed changes in cloud cover to cloud simulations in climate models. For example, Fasullo and Trenberth (2012) used satellite data from the NASA Atmospheric Infrared Sounder (AIRS) and Clouds and the Earth's Radiant Energy System (CERES) to examine the relationship between seasonal changes in relative humidity (RH) in the dry subtropics and the Earth's albedo via cloud cover.
"...the results strongly suggest that the more sensitive models perform better, and indeed the less sensitive models are not adequate in replicating vital aspects of today’s climate. The correct simulation of the vertical structure of RH and clouds should be a prerequisite for developing confidence in projections for the future."
Sherwood et al. (2014) built on the work of Fasullo and Trenberth by looking at the way that various climate models handle the cloud feedback. They found GCMs with a low climate sensitivity were inconsistent with observations. It turns out that these models were incorrectly simulating water vapor being drawn up to higher levels of the atmosphere to form clouds in a warmer world. In reality (based on observations), warming of the lower atmosphere pulls water vapor away from those higher cloud-forming levels of the atmosphere, and the amount of cloud formation there actually decreases. The diminished cloud cover leads to greater warming (a positive feedback), and is better reproduced in the GCMs with higher climate sensitivities.
These studies were omitted from the GWPF report, but they provide yet another line of evidence for high and against low climate sensitivity.
Climate Policies are Insufficient in Any Case
The GWPF report concludes by complaining that by not emphasizing the lower climate sensitivity estimates, the IPCC has “inadequately informed” policymakers about the state of the science. However, from a policy standpoint, we're not doing nearly enough to reduce emissions even in the best case scenario. As Myles Allen, co-author on Otto et al. (2013) noted of the GWPF report,
“Their prediction of 1.35 degrees C [TCR] is, even if correct, only 25% lower than the average of the general circulation models used in the IPCC 5th Assessment. A 25% reduction in TCR means the warming we might have expected by 2050 might take until the early 2060s instead.”
Moreover, as detailed above, the full body of scientific evidence suggests that climate sensitivity is relatively high. Even if you believe the GWPF report is right, there's a good chance it's not. Proper risk management therefore mandates that we must take action to mitigate the threat of dangerous climate change.
But in any case, the full body of evidence is firmly against the conclusions of the report. The authors merely dismiss or ignore the research that doesn’t support their desired conclusion, and overlook the shortcomings of the research that does.
Note: the discussion of the Shindell results has been incorporated into the Advanced rebuttal to the myth 'climate sensitivity is low'.