Dr. Roger Pielke Senior posted this today, since he has no comments on his blog, I felt it would be good to repost it here to allow discussion – Anthony
There is an excellent collection of interviews posted by Sam Patterson on April 23 2011 on the weblog Climatequotes.com titled
I have reposted his very informative set of interviews and commentary below.
_________________________________________________
From: Climatequotes.com
Note: I wrote this post many weeks ago and never posted it because I was waiting for some more feedback. However, Pielke Sr. has posted specifically on this issue recently and Watts ran it also, so I feel now is a good time to post it.
This post deals with the the question of whether or not climate sensitivity should be measured by global average surface temperature anomaly. I asked multiple climate scientists their opinion, and their responses are below. First, some background.
Over at The Blackboard there is an interesting guest post by Zeke. He attempts to find areas where agreement can take place by laying out his beliefs and putting a certain confidence level on them. This idea was commented upon by several blogs and scientists. Judith Curry, Anthony Watts, Jeff Id, and Pielke Sr. all contributed. I want to focus on Pielke’s response, because he challenges a core assumption of the exercise.
In Zeke’s post, he gives his position on climate sensitivity:
Climate sensitivity is somewhere between 1.5 C and 4.5 C for a doubling of carbon dioxide, due to feedbacks (primarily water vapor) in the climate system…
Here is Pielke’s response to this claim:
The use of the terminology “climate sensitivity” indicates an importance of the climate system to this temperature range that does not exist. The range of temperatures of “1.5 C and 4.5 C for a doubling of carbon dioxide” refers to a global annual average surface temperature anomaly that is not even directly measurable, and its interpretation is even unclear…
Pielke goes on to explain that he has dealt with this issue previously in the paper entitled “Unresolved issues with the assessment of multi-decadal global land surface temperature trends.” Here is the main thrust of his response:
This view of a surface temperature anomaly expressed by “climate sensitivity” is grossly misleading the public and policymakers as to what are the actual climate metrics that matter to society and the environment. A global annual average surface temperature anomaly is almost irrelevant for any climatic feature of importance.
So we know Pielke’s position. He is adamantly opposed to using surface temperature anomaly when discussing climate sensitivity, for various reasons, not the least of which is it ignores metrics which actually matter to people.
I haven’t heard this view expressed very often, so I decided to contact other climate scientists and find out their opinions on this issue. I asked the following questions and invited them to give their general impressions:
1. Do you believe that global annual average surface temperature anomaly is the best available metric to discuss climate sensitivity?
If yes to Question 1, then:
2. Could you briefly explain why you consider global annual average surface temperature anomaly the best available metric to discuss climate sensitivity?
If no to question 1, then:
2. What do you believe is the proper metric to discuss climate sensitivity, and could you briefly explain why?
John Christy
1. Do you believe that global annual average surface temperature anomaly is the best available metric to discuss climate sensitivity?
No. The surface temperature, especially the nighttime minimum, is affected by numerous factors unrelated to the global atmospheric sensitivity to enhanced greenhouse forcing (I have several papers on this.) The ultimate metric is the number of joules of energy in the system (are they increasing? at what rate?). The ocean is the main source for this repository of energy. A second source, better than the surface, but not as good as the ocean, is the bulk atmospheric temperature (as Roy Spencer uses for climate sensitivity and feedback studies.) The bulk atmosphere represents a lot of mass, and so tells us more about the number of joules that are accumulating.
Patrick Michaels
I think it is a reasonable metric in that it integrates the response of temperature where it is important–i.e. where most things on earth live. However, it needs to be measured in concert with ocean measurements at depth and with both tropospheric and stratospheric temperatures. For example, if there were no stratospheric decline in temperature, then lower tropospheric or surface rises would be hard to attribute to ghg changes. Because we don’t have any stratospheric proxy (that I know of) for the early 20th century, when surface temperature rose about as much as they rose in the late 20th, we really don’t know the ghg component of that (though I suspect it was little to none).
Having said that, I suspect that where we do have such data, it is indicative that the sensitivity is lower than generally assumed, but not as low as has been hypothesized by some.
Gavin Schmidt
Your questions are unfortunately rather ill-posed. This is probably not your fault, but it is indicative of the confusion on these points that exist.
“Climate sensitivity” is *defined* as being the equilibrium response of the global mean surface temperature to a change in radiative forcing while holding a number of things constant (aerosols, ice sheets, vegetation, ozone) (c.f. Charney 1979, Hansen et al, 1984 and thousands of publications since). There is no ambiguity here, no choice of metrics to examine, and no room for any element of belief or non-belief. It is a definition. There are of course different estimates of the surface temperature anomaly, but that isn’t relevant for your question.
There are of course many different metrics that might be sensitive to radiative forcings that one might be interested in: Rainfall patterns, sea ice extent, ocean heat content, winds, cloudiness, ice sheets, ecosystems, tropospheric temperature etc. Since they are part of the climate, they will be sensitive to climate change to some extent. But the specific terminology of “climate sensitivity” or the slightly expanded concept of “Earth System Sensitivity” (i.e Lunt et al, 2010) (that includes the impact on the surface temperature of the variations in the elements held constant in the Charney definition), are very specific and tied directly to surface temperature.
People can certainly hold opinions about which, if any, of these metrics are of interest to them or are important in some way, and I wouldn’t want to prevent anyone from making their views known on this. But people don’t get to redefine commonly-understood and widely-used terms on that basis.
I sent a response to Gavin clarifying my questions, and including Pielke Sr’s comments. Here is his response to Pielke’ comments:
I disagree. Prof. Pielke might not find the global temperature anomaly interesting, but lots of other people do, and as an indicator for other impacts, it is actually pretty good. Large-scale changes in rainfall patterns, sea ice amount, etc. all scale more or less with SAT. (They can vary independently of course, and so ‘one number’ does not provide a comprehensive description of what’s happening).
Kevin Trenberth
1. Do you believe that global annual average surface temperature anomaly is the best available metric to discuss climate sensitivity?
This is not a well posed question. This relates to definition: the sensitivity is defined that way. It is not the best metric for climate change necessarily
If yes to Question 1, then:
2. Could you briefly explain why you consider global annual average surface temperature anomaly the best available metric to discuss climate sensitivity?
I think the best metric overall is probably global sea level as it cuts down on weather and related noise. But global mean temperature can be carried back in time more reliably and it is reasonably good as long as decadal values are used.
If no to question 1, then:
2. What do you believe is the proper metric to discuss climate sensitivity, and could you briefly explain why?
However, it is all variables collectively that make a sound case
Pielke Sr.
We have already discussed Pielke’s position, but I contacted him to find out what metrics he would prefer to use. Here is his response:
1. Do you believe that global annual average surface temperature anomaly
is the best available metric to discuss climate sensitivity?
NO
If yes to Question 1, then:
2. Could you briefly explain why you consider global annual average
surface temperature anomaly the best available metric to discuss
climate sensitivity?
If no to question 1, then:
2. What do you believe is the proper metric to discuss climate
sensitivity, and could you briefly explain why?
The term “climate sensitivity” is not an accurate term to define how the climate system responds to forcing, when it is used to state a response in just the global average surface temperature. This is more than a semantic issue, as the global average surface temperature trend has been the primary metric used to communicate climate effects of human activities to policymakers. The shortcoming of this metric (the global average surface temperature trend) was discussed in depth in
“National Research Council, 2005: Radiative forcing of climate change: Expanding the concept and addressing uncertainties. Committee on Radiative Forcing Effects on Climate Change, Climate Research Committee, Board on Atmospheric Sciences and Climate, Division on Earth and Life Studies, The National Academies Press, Washington, D.C., 208 pp. http://www.nap.edu/openbook/0309095069/html/“
but has been mostly ignored in assessments such as the 2007 IPCC WG1 report.
A more appropriate metric to assess the sensitivity of the climate system heat content to forcing is the response in Joules of the oceans, particularly where most the heat changes occur. I discuss this metric in
Pielke Sr., R.A., 2008: A broader view of the role of humans in the climate system. Physics Today, 61, Vol. 11, 54-55.
http://pielkeclimatesci.files.wordpress.com/2009/10/r-334.pdf
Pielke Sr., R.A., 2003: Heat storage within the Earth system. Bull. Amer. Meteor. Soc., 84, 331-335. http://pielkeclimatesci.files.wordpress.com/2009/10/r-247.pdf
More generally, in terms of true climate sensitivity, more metrics are needed as we discussed in the 2005 NRC report. The Executive summary includes the text [http://www.nap.edu/openbook.php?record_id=11175&page=4]
“Despite all these advantages, the traditional global mean TOA radiative forcing concept has some important limitations, which have come increasingly to light over the past decade. The concept is inadequate for some forcing agents, such as absorbing aerosols and land-use changes, that may have regional climate impacts much greater than would be predicted from TOA radiative forcing. Also, it diagnoses only one measure of climate change “global mean surface temperature response” while offering little information on regional climate change or precipitation. These limitations can be addressed by expanding the radiative forcing concept and through the introduction of additional forcing metrics. In particular, the concept needs to be extended to account for (1) the vertical structure of radiative forcing, (2) regional variability in radiative forcing, and (3) nonradiative forcing. A new metric to account for the vertical structure of radiative forcing is recommended below. Understanding of regional and nonradiative forcings is too premature to recommend specific metrics at this time. Instead, the committee identifies specific research needs to improve quantification and understanding of these forcings.”
It is, therefore, time to move beyond the use of the global annual average surface temperature trend as the metric to define “climate sensitivity”.
Differing views
There are clearly differing views on this subject.
John Christy does not support the metric. He points out that the surface temperature is affected by numerous things other than greenhouse forcing, and then gives two metrics which he prefers. The first is the change in joules in the system, with particular emphasis on the oceans. The second is bulk atmospheric temperature.
Patrick Michaels supports using the metric. He points out that the metric is important because it addresses the area where people live. However, he emphasizes that the surface temperature must be taken in concert with measurements such as ocean temperature at depth, and tropospheric and stratospheric temperatures. Without these other measurements, it would be difficult to assess the impact of GHGs on surface temperature.
Gavin Schmidt supports the metric unreservedly. He and Trenberth rightly point out that climate sensitivity is defined by global average surface temperature anomaly. Of course, the point of my question is challenging whether or not this is the best definition. Gavin seems to think so, and points out that the metric is “commonly-understood and widely-used”. He states that other metrics such as rainfall patterns and sea ice amount track very well with surface air temperature.
Trenberth is very brief, but states that global average surface temperature anomaly is not necessarily the best metric to use for climate change. He considers that global sea level is a better metric because it cuts down on weather related noise. However, he also points out that global average surface temperature anomaly is useful because it can be applied to the past more reliably. He also states that all variables taken together make a sound case.
Pielke Sr. is adamantly opposed to using this metric. We’ve already discussed his reasons. He also proposes a different metric for assessing climate sensitivity, “A more appropriate metric to assess the sensitivity of the climate system heat content to forcing is the response in Joules of the oceans”. He supports these claims with several of his own papers as well as a NRC report.
Conclusion
Pielke and Christy want to stop assessing climate sensitivity by using global average surface temperature anomaly, and both recommend using a change of joules (particularly in the ocean) as a better metric.
Michaels and Trenberth support the metric while emphasizing that other metrics must also be taken into account. Schmidt does not mention any drawbacks and emphasizes that the metric is already widely used and it works well with other metrics.
It seems to me the main problem here isn’t the metric itself, but the emphasis placed on it. I don’t believe that Pielke or Christy believe the metric has no value at all, only that it is a poor choice to use as the main metric when discussing CO2′s impact on climate. In Pielke’s case, the emphasis on CO2 itself is a problem, as he believes that other human impacts are far more important.
Climate science so frequently focuses on CO2 and temperature that it seems natural climate sensitivity would be measured by global average surface temperature anomaly. A shift away from this metric seems unlikely. However, if it can be shown in the future that a change in joules in the ocean directly contradicts other metrics then I’m sure this discussion will come up again. Pielke’s paper mentions an apparent contradiction found by Joshua Willis of JPL, although the measurements are only taken over a four year period. Only time will tell which metric is most valuable.
cba says:
April 30, 2011 at 2:40 pm
———
You have to think of it as, the majority of the solar radiation is still getting through the clouds. So it gets through the clouds and then still reflects off the surface.
What does cloud cover do over the Greenland Glacier. Nothing really, because the Ice is reflecting up to 83% (in some places) of whatever solar radiation (including UV which goes straight through most clouds) back out to space.
These amounts are higher than any cloud has so technically, cloud cover contributes 0 to Albedo over the Greenland Glaciers. Take the clouds away and Albedo is exactly the same.
In fact, take all the clouds away across the whole planet and what is the average surface Albedo – 13 to 14 percentage points (actually it is more like 16 but then there is re-reflection back from the clouds again which drops the numbers a little).
Bill,
70% of the surface is water. That’s less than 0.04 albedo for most of the surface where the majority of incoming power comes in. Reflectivity for non thin clouds tends to be given in the range of 40% to 80%, depending on type and other conditions. What gets through the cloud after the reflections ( scattering) and the internal absorption, a few %, is going to be more like half or less of the total striking the clouds. While greenland will have little cloud effect, 80% surface albedo only happens with fresh snow. What fraction of the average incoming power strikes greenland? What fraction of the incoming solar power occurs over snow or ice?
Of the less than half of the solar power coming through the clouds, the very small average surface albedo reflecting the power back through the clouds is going to have a good portion reflected back down and so only a small part gets scattered outward through the cloud. In any case, the atmospheric scattering back to space exceeds the average surface albedo. Even if snow and ice albedo were 100%, there is just not enough ice and snow to compensate for all that ultra low albedo ocean area.
cba says:
and this is with 70% of the surface covered in water, which runs less than 0.04 albedo
Henry@cba
I remember often standing at the beach and getting the sun’s reflection from the water in my face – where the water acts as a mirror, causing me to squint or to turn my eyes away. Counting also the waves that can get come into a right angle, I doubt if the albedo from “sea shine” is only 0.04. Where did you get that value and how was it arrived at ?
I think what you also forget is that a considerable % of that 70% is covered with ice and snow again.
I had a look at the CERES data for March.
I picked out two 1.4 degree by 1.4 degree blocks at -1.4S, (effectively, the Sun was directly overhead) that were 1) the most cloud-covered and had the highest Albedo (100% – Indian Ocean) and 2) the least cloud-covered (18% – near the International Dateline in the Pacific).
1) The 100% cloud-covered block (and had the highest Albedo/solar radiation reflected of several other blocks which were also at 100% – ie, it was thick high cloud) had a net Albedo of 52%.
2) The 18% cloud-covered block had an Albedo of 13%.
Just doing the math and using the 52% Albedo block as a max since it must have had the highest thickest cloud possible, the Ocean Albedo couldn’t be lower than 10% when the Sun is directly overhead in the real Ocean.
Remember there is waves on the Ocean. The amount of Solar energy which is reflected off water increases exponentially as the incidence angle increases. In practise, it only becomes important at about 70 degrees and higher (so this is important for open water at the poles for example and for waves on the Ocean).
Henry P, Bill
the wiki article on albedo has pretty good information.
http://en.wikipedia.org/wiki/Albedo
You’ll note that even at 60 degrees from the vertical, one has under 0.1 albedo for water and that is the point where the atmospheric thickness is double that of straight up. At the poles one sees that the water reflections are much greater due to the higher angles.
Bill, your numbers look reasonable but remember too that the reflectivity of clouds varies with the droplet size of the cloud particles. One can have significant variation in the reflectivity and in the albedo both by cloud fractional coverage and by the actual cloud reflectivity – the origins of Lindzen’s Iris theory. for 100% cloud cover 0.52 sounds reasonable or typical. That would be for a cloud average reflectivity of 52% and that is pretty much the middle of the road for typical cloud reflectivity.
Using the same value for 18% cloud cover and simply doing a weighted average with the surface being 0.04 reflectivity, one comes up with an average of 12.6% albedo, very close to your value of 13%.
If you take the idea that something significant portion of the clouds reflectivity is coming through the cloud from the ground, say for the 100% coverage, you’ve got about 50% of the light reflected from the cloud, leaving more like 45% striking the surface after absorptions. Since only 4% reflectivity is occurring, that provides about 1.8% reflecting back up to the clouds and assuming a 50% transmission through the cloud outbound, that reduces to under 1% the total of the incoming.
Just to comment here that the wiki article does not mention that the albedo “seashine” is only 0.04. It is clearly indicated at almost double that. But I still wonder how they measured it. Perhaps by looking at pictures like the one from WUWT here above? (if you look carefully at the picture you could say that the yellow/orange area is around 10%?)
Henry,
Check out the section on water in wikipedia’s albedo article, especially the graph and the part about waviness.