2011-09-17 05:59:30Response to Pielke Q3, Q4, Q5...
Rob Honeycutt

robhon@mac...
98.207.62.223

Starting up this thread and reposting the response to Q3...

3. What is your preferred diagnostic to monitor global warming?

All of the above.  This is how science operates.  It's extremely important to look at a problem from as many angles as reasonably possible.  Each of the methods for monitoring global warming has it's own challenges.  We know you prefer ocean heat content, and specifically the upper 700m data.  OHC content, I think you would agree, is very challenging and nacent research.  Ocean circulation and sparse coverage of the ARGO data leaves uncertainties regarding whether all the data is indeed being fully captured.  Satellite data is relatively easier to collect and process, it has much broader coverage and can be measured in greater detail.  We know we can capture good data but that data is also challenged in that it's only measuring a small portion of the heat energy in the climate system being that most of the actual heat energy is in the oceans.  TOA radiation imbalance is merely another tool, or dial on the control panel, that we should be using to evaluate and monitor global warming.  All of these have to make sense within the broader context of the physics of the atmosphere.

Hanging your hat on only one best diagnostic would be a mistake.  A pilot does not fly an aircraft by looking at only one dial on the panel, nor should we rely on only one diagnostic to monitor global warming.

3b. What is your best estimate of the observed trends in each of these metrics over the last 10 years and the last 20 years?

For the best estimate of the observed trends we would defer to the published research for each data set but we would always be wary of shorter periods that do not rise to the 95% confidence level.  We also be very wary of anyone defining trends in exact terms without estimations of uncertainties.

The job here is not to pick the exact time for the winning dog in the race, it's to try to comprehend what the many lines of evidence tell us about global warming.

2011-09-17 06:06:43
dana1981
Dana Nuccitelli
dana1981@yahoo...
64.129.227.4

I'd drop "as we understand it".  Makes us sound amateurish, and that is how science operates.

3b should start out by noting that 10-year trends are not statistically significant, at least for surface and atmosphere temps (I presume the same is true for ocean heat content).

2011-09-17 06:41:10Re the statisitcal signifigance of 10-year trends..
John Hartz
John Hartz
john.hartz@hotmail...
98.122.98.161

Check out the exchange between Dikran Marsupial and Pielke on the comment thread to Dana's article.

2011-09-17 07:04:41
nealjking

nealjking@gmail...
91.33.125.111

Rob,

Who do we talk to to really understand all these different measures? Do we have someone who really knows the ins and outs of this stuff, or are we basing this on reading various articles in no specific order? Do WE know which papers/authors are really authoritative, or are we going by a general sense of who is prominent?

2011-09-17 07:21:07
Rob Honeycutt

robhon@mac...
98.207.62.223

Neal...  My personal sense is that I try to listen to all sides.  Honestly I even try to keep Lindzen and Spencer in the realm of potential reality as data points rather than definitive answers.  I am more personally compelled to believe AGW is real because of the overwhelming evidence of the entire scientific community rather than the positions of any given scientist.

2011-09-17 07:27:23
nealjking

nealjking@gmail...
91.33.125.111

I think we need to be careful. We need to make it clear that we are relying on what appears to be generally accepted by the working body of climate scientists, that is not being actively shredded in analysis by experts. We are not relying solely on our own expertise.

Related question: How do you decide/come to the conclusion that a particular article/proposal/conclusion is just crap? What's your methodology?

2011-09-17 07:51:43
Riccardo

riccardoreitano@tiscali...
188.152.84.214

There's no need to mention his preference on the metric. I like the way you started, saying that each has its own problems and limitations,but we may look at the problem from several view angles.

Use the OHC as an example and maybe you could add imbalance at TOA, which we still cannot measure reliably.

The second part becomes irrelevant. Different datasets will give different results, but hardly of any significance over just 10 years. Anyway, for this there's the scientific littrature.

2011-09-17 08:18:32Comment
Robert Way

robert_way19@hotmail...
142.162.205.182

"All of the above"

Instead of that we should have a list of the different measures with a link to graphs from each. Evidence comes from Surface thermometer readings (citations for the 3 major ones), sea surface temperatures, radiance measurements of oxygen particles (Christy, spencer, mears etc) of the atmosphere, OHC (0-700 m, references), Deep ocean heat content (links), TOA measurements (santer et al... trenberth...), LW radiation measurements (Chen et al, harries et al,), LW down radiation measurements (forget the paper), Ecosystem indicators (90% point to warming) and spectrometry done on the LW down showing our W/m2 measurements are appropriate (Evans, 2006).Sea level rise too..

2011-09-17 08:28:10
nealjking

nealjking@gmail...
91.33.125.111

Rob,

I am still interested in my question: How do you detect crap? I think this could be important, because Pielke knows as well as we that you can't rely on one measurement technique alone. I guess that he's looking for us to say something like "We rely on X," so he can say "But X is not reliable in THESE circumstances." So if we say we rely on everything, he could ask, "If different techniques give different answers, what do you do?"

Personally, I don't deal with recent research, so I don't have experience with this issue. But if I had to make some kind of evaluation:

- I would consider the reception of any research by the climate-science community: If there are no eruptions, and it seemed to be non-contradictory to other uncriticized work, I would admit it as acceptable and to be taken into account.

- If there was some kind of objection to it, I would study the paper more carefully, to see if I understood the general logic, and possibly the essential technique. If there was serious criticism, I would try to understand the criticism; likewise the support, if any. Opinions based (apparently) on politics alone would be discounted.

I think this is important: When you say you are evaluating the scientific techniques, that can be said a few different ways depending on how "hands on" you can be with the evaluation. Speaking for myself: If I'm talking about the quality of research that applies general relativity, I will rely heavily on the opinion of big names in the field, because I'm only familiar with the basic concepts. If I'm talking about a paper that depends on classical mechanics or basic quantum mechanics, I will trust my own understanding quite a bit. So in the first case, I might say, "This research seems well accepted/rejected by the experts."; but in the second case, I would say, "Their analysis seemed correct/completely ridiculous." It's important, in dealing with Pielke, that we don't give the impression of assuming that we have more expertise than we actually do. THAT would be the fatal error.

2011-09-17 08:35:14
Albatross
Julian Brimelow
stomatalaperture@gmail...
199.126.232.206

"3b. What is your best estimate of the observed trends in each of these metrics over the last 10 years and the last 20 years?"

He has done the maths on this, it is a trap-- his son Roger was prattling on to someone recently about X number of Joules etc. recently-- sorry but I cannot recall where, sorry!  Pielke should know the answer, why is he asking us?  One should be looking at about 20 years worth of data from multiple metrics.

2011-09-17 08:55:01
nealjking

nealjking@gmail...
91.33.125.111

Yes, we have to address the question, but we don't have to answer it the way he is asking it. We're not taking an examination.

The real question is NOT the numbers, the real question is how we go about approaching it. The right answer is that we depend on the general scientific consensus; and we give some general sense of how we determine that, without claiming to be up-to-the-minute. We can cite some current numbers; but we shouldn't claim to be au courant (unless somebody really is).

2011-09-17 08:56:35
Albatross
Julian Brimelow
stomatalaperture@gmail...
199.126.232.206

"Yes, we have to address the question, but we don't have to answer it the way he is asking it. We're not taking an examination."

Agreed Neal!

2011-09-17 09:13:54
Rob Honeycutt

robhon@mac...
98.207.62.223

Neal...  I would say that I rely on what is consistent with other research.  One reason I don't trust Lindzen and Spencer's estimations of low climate sensitivity is that it doesn't square with vast amounts of other research that show sensitivity much higher.  I think everyone is vulnerable to being wrong, sometimes very wrong, even if they have the best of intentions.  The reason science works so well is it's okay to be wrong.  Science allows people to put out science questioning established assumptions.  BUT it has to fit.  It has to answer all the questions that were answered with those previous assumptions.  That is where I think Lindzen and Spencer fail so profoundly.  They publish what they say is an answer to the question of climate sensitivity without ever addressing the fact that their answers do not square with the entire paleoclimate record.  They pull out the cornerstone and completely ignore that the entire structure collapses when they do.

I don't know if that is answering your question or if it's addressing what should be said in response to Pielke's question.  Is there something you're edging in on here that I'm not comprehending?

2011-09-17 09:18:31
Rob Honeycutt

robhon@mac...
98.207.62.223

Neal... "but we shouldn't claim to be au courant"  I completely agree.  He's trying to tie us down with his numbers and I'm trying to say that all the numbers are important because anyone might be wrong.  The numbers that are more likely to be correct are those that agree best with the broader science.  In essence, yes, we rely on the broader scientific body of research than any specific "diagnostic".

2011-09-17 09:28:57
Dikran Marsupial
Gavin Cawley
gcc@cmp.uea.ac...
88.108.224.131

Statistical power is the problem with the 10 to 20 year trends, and Prof. Pielke's evasion suggests he doesn't have a good grasp of the statistical issues.

In standard hypothesis "works" by working out the probability of observing a trend at least as extreme by chance and rejecting the null hypothesis (that the observed trend was due to chance) if this probability is too low.  However if you don't have much data, almost anything can be explained by chance, and hence it is unlikely that a significance test will reject the null hypothesis when it is false.  The probability of rejecting the null hypothesis when it is false is known as the "statistical power" of the test.  Very few scientists pay much attention to statistical power (though they should), but if a test has very little statistical power then it is essentially meaningless.  Phil Jones clearly understands the issue well (c.. that BBC interview) possibly intuitively rather than theoretically.

Wanting to talk about short term trends is ideal for someone who wants to play up uncertainty, but if there is data over  longer period that makes the argument with better power, it is not science to ignore it.

The RankExploits test Prof. Pielke linked to is also bogus.  Rather than testing the observed trend with the error bars of the ensemble mean trend, the proper test is to see if the observed trend lies within the distribution of trends from individual model runs.  The reason for this is that the ensemble mean is an estimate of the forced trend, whereas the observed trend is a combination of the forced and unforced trends.  There is no reason to expect them to be the same.

2011-09-17 09:31:47
Rob Honeycutt

robhon@mac...
98.207.62.223

I have to add that I think Pielke is on the right track in saying that OHC is a better measurement of global warming, but I also think he is WAY too quick to make sweeping claims based on the quantity and quality of the data that is available.  

2011-09-17 09:37:27
nealjking

nealjking@gmail...
91.33.125.111

Dikran,

It may be that Pielke understands the point but prefers not to acknowledge it.

It's probably enough for us to say that it is our understanding that 10-year trends are not statistically meaningful with regard to climate issues.

2011-09-17 09:47:47Comment
Robert Way

robert_way19@hotmail...
142.162.200.7

Dikran Marsupial,
The test is bogus anyways because the other issues which are relevant to the models have not be addressed i.e. increased aerosols that were not included in the models from volcanism and southeast asia. The models will undoubtedly be running as an ensemble mean too hot because those negative forcings aren't included...

2011-09-17 10:00:52Attempt at Re-write
nealjking

nealjking@gmail...
91.33.125.111

In light of the above comments, I'm going to try a re-write of Rob's draft. I am going to avoid some of the specific details that could be used to trip us up:

 

3. What is your preferred diagnostic to monitor global warming?

We don't have a "preferred" diagnostic. We are not actively engaged in research, so we depend upon the general consensus of the research community: We think that it is most appropriate to rely upon the experts with hands-on experience and involvement to judge what appears to be most reliable at the present time, and to sort out any discrepancies amongst them. When we are put in the position of evaluating a new approach, we certainly study the technique; but it bothers us if there seems to be a general outcry against the competence demonstrated in a paper.

In general, we are aware that many techniques, such as you have enumerated, are used. That makes sense to us: Just as a pilot cannot depend on just one instrument, climate studies cannot rely on just one diagnostic.

3b. What is your best estimate of the observed trends in each of these metrics over the last 10 years and the last 20 years?

For the best estimates of the observed trends, we would defer to the published literature. However, we don't take much interest in 10-year trends for climate, as they are statistically not meaningful. Our interest is not in picking the exact temperature trajectory, but to comprehend what the many lines of evidence are telling us about global warming.

2011-09-17 10:37:21
dana1981
Dana Nuccitelli
dana1981@yahoo...
69.230.106.125
I don't really like that first paragraph after the first sentence. I'd just say we don't have a preferred diagnostic because we have to take all data into account. Nice and simple. For 3b we could provide some numbers for 20 year trends. I can pull up the figures for surface and TLT data easily. We just need to find a paper with the ocean trends, like von Schuckmann [I probably mangled that spelling]. I'd also reference Santer et al. in noting that models predict decades with dampened warming trends, and 17 years of data are necessary to discern the human influence on TLT temps, and link to my post on the paper.
2011-09-17 10:48:44
nealjking

nealjking@gmail...
91.33.125.111

dana,

My concern is that if he starts probing our expertise in the interpretation, comparative reliability and possible discrepancies among the diagnostics, how deeply are we prepared to go?

To put it bluntly, if he starts pressing on these points, are you prepared to go head-to-head with him?

'cause I'm not.

2011-09-17 10:51:38
dana1981
Dana Nuccitelli
dana1981@yahoo...
69.230.106.125
There's not much probing he can do if we say we look at all the data in totality. He thinks OHC is the most important metric. If we pick another, like surface temperature, he can needle us on it. If we say we'd look at them all, that includes OHC and he's got nowhere to go.
2011-09-17 11:10:54
nealjking

nealjking@gmail...
91.33.125.111

Why not propose specific changes to the text?

Or go back to Rob's text, if you prefer.

2011-09-17 14:19:21
dana1981
Dana Nuccitelli
dana1981@yahoo...
69.230.106.125

3. What is your preferred diagnostic to monitor global warming?

We don't have a preferred diagnostic.  It's important to look at all the data in totality to monitor global warming (surface temperature, ocean heat content, atmospheric temperature, TOA energy imbalance, sea level rise, receding ice, etc.).

3b. What is your best estimate of the observed trends in each of these metrics over the last 10 years and the last 20 years?

10-year trends are not statistically meaningful (see Santer et al. 2011, for example).  The best estimate observed trends for some of these metrics over the last ~20 years are as follows.  TLT: 0.18°C per decade.  Surface temperature: 0.20°C per decade.  OHC upper 700 meters: 6.3 x 1022 J per decade.  Sea level rise: 32 mm per decade.  Arctic sea ice volume: -2900 km3 per decade.

2011-09-17 14:33:21
Rob Honeycutt

robhon@mac...
98.207.62.223

I think there is probably a very good combination of all these versions.  I suggest we blend them into one.

2011-09-17 15:04:07
dana1981
Dana Nuccitelli
dana1981@yahoo...
69.230.106.125

I intentionally kept my answer short.  K.I.S.S.  The less we say, the less he has to nitpick.  This is a straightforward one - we look at all the data, 10-year trends are not statistically significant, here are the 20-year trends.  I really don't think it requires anything more, and the more we say, the more material we give him to work with and play his silly games.

2011-09-17 15:14:27
Albatross
Julian Brimelow
stomatalaperture@gmail...
199.126.232.206

Dana,

Agreed, good strategy IMHO.  The RSS TLT that you link to is 0.19 :)  Yes, 20 years is a very reasonable number and is supported by the literature.  There is alo a paper out by Peterson et al. (2011) that looks at the increase in moist-static energy from surface data.  But Pielke has strong opinions on that, well same goes for everyhting els too....  ;)  

DelSole et al. (2011) say at least 16 years IIRC when looking at SSTs.  W ehave to support ourt choices with references, that will make him look like an outlier if he chooses to cherry pick a single metric.  OHC has issues, so it is far from perfect, at least in the real world, and that is what we are dealing with, and that is why we need to look at the body of evidence b/c all the metric have their limitations and do not epxlain the whole picture-- I wonder if anyone says that in a published paper?

Whatever we say he will nit pick and then reference some of his own papers (notice how ;skeptics" like to reference their own work and believe it to be of more importnace than that of others?) to try and call it into question, so the less we give him the better.

2011-09-17 18:53:34
nealjking

nealjking@gmail...
91.33.99.16

I like the direction this is heading in. I like the brevity of dana's version. However, I wonder if we can toss in a reference to the fact that we're trying to keep up with consensus of the research community. The point: I don't ever want to get into a situation where he says, "I think the trend is X, and I've got years of professional experience over you."; because he does, and for him to point that out is a little embarassing for us. I want to warn him off, by making it clear that our next move would be, "The articles by A, B, C and D show that the trend is between 2X and 3X." In other words, I want him to realize that we're not going to let it become a contest between our years of climate-science professionalism and his, but between the collective YPE of the entire climate-science community and his: He may be a battleship and outrank us, but we're a lighthouse and we're sitting on an island.

My original wording above made that point, but probably comes across as a bit too weak. I would appreciate some help along these llines.

When we're through with this, I also think we can handle Q4 & Q5 along very similar lines.

2011-09-18 02:01:19
Rob Honeycutt

robhon@mac...
98.207.62.223

He's going to have all kinds of reasons why he believes OHC is the right measure and he's continually stated that it should replace the surface temperature sets as the proper metric of global warming.  But I think his intent is incredibly disingenuous.  I was just reading Eli Rabbett's responses in the thread about this and he says that OHC will be ADDED to the other data sets rather than replace them.

I think those guys talk about OHC mostly because measure of joules removes the connection that people have to temperature.  It makes global warming more impersonal, more of an abstract.  It's a good way to measure, especially if they can really get the kinks worked out, but it should be additive and not supplantive. 

The key point that needs to hit home is that we don't rely on one diagnostic.  Redundancy is a good thing.  It's also a good chance to drive home John's oft repeated line:  "...many lines of evidence."  So maybe we stick in there the statement that "We prefer to rely on many lines of evidence."

2011-09-18 02:09:55
nealjking

nealjking@gmail...
91.33.99.16

Rob,

An important point is that we're under no obligation to either agree with or disprove his favorite answer. We just need to give an answer that most people will find reasonable.

"Lines of evidence": This doesn't fit so well here, because the issue now is monitoring, not proving a story. Dana already mentions all the different diagnostics.

2011-09-18 02:18:01
Rob Honeycutt

robhon@mac...
98.207.62.223

Neal...  I'm just offering that the phrase is a useful ("sticky") rhetorical technique that echoes what's been said here many times before.  This is a very important aspect of good messaging.  I'm not saying anything different that Dana.  Pielke is framing with HIS own sticky rhetorical phrasing.  Don't let him do it.  We should respond and repeat with OUR framing.

2011-09-18 02:19:40
Rob Honeycutt

robhon@mac...
98.207.62.223

For example...

3. What is your preferred diagnostic to monitor global warming?

We prefer to rely on many lines of evidence.  It's important to look at all the data in totality to monitor global warming (surface temperature, ocean heat content, atmospheric temperature, TOA energy imbalance, sea level rise, receding ice, etc.).

2011-09-18 02:21:12
Rob Honeycutt

robhon@mac...
98.207.62.223

See what what I mean?

Don't repeat HIS words.  Reframe and use OUR words.

Pielke's gig is trying to reframe the entire concept of global warming by turning it into a single, non-temperature related, "diagnostic."  Don't let him do it.  Never repeat his framing.

2011-09-18 02:31:51
nealjking

nealjking@gmail...
91.33.99.16

Rob,

OK, the last full picture we had was Dana's. Why don't you produce a revision of that?

2011-09-18 02:39:19
Rob Honeycutt

robhon@mac...
98.207.62.223

3. What is your preferred diagnostic to monitor global warming?

We prefer to rely on many lines of evidence.  It's important to look at all the data in totality to monitor global warming (surface temperature, ocean heat content, atmospheric temperature, TOA energy imbalance, sea level rise, receding ice, etc.).

3b. What is your best estimate of the observed trends in each of these metrics over the last 10 years and the last 20 years?

10-year trends are not statistically meaningful (see Santer et al. 2011, for example).  The best estimate observed trends for some of these metrics over the last ~20 years are as follows.  TLT: 0.18°C per decade.  Surface temperature: 0.20°C per decade.  OHC upper 700 meters: 6.3 x 1022 J per decade.  Sea level rise: 32 mm per decade.  Arctic sea ice volume: -2900 km3 per decade.

2011-09-18 02:47:23
nealjking

nealjking@gmail...
91.33.99.16

For the reason mentioned before, I don't like the use of "lines of evidence" here:

"We prefer to rely on many lines of evidence."

=>

"We prefer to rely on many types of evidence."

"We prefer to look at all available diagnostics."

Something like that. I believe that the term "line of evidence" finds its home in the framework of an argument leading to a logical conclusion. But here we're not talking about a conclusion, we're talking about a degree of progress.

2011-09-18 03:04:46
Rob Honeycutt

robhon@mac...
98.207.62.223

Neal...  I believe the rhetorical impact of repeating John's oft used line is more powerful here.  Repetition has high impact.  Altering the phrasing lessens the impact.  The very next line discusses monitoring.  

I think you're right that technically these two don't exactly match but, personally, I don't think that's nearly as important as driving home a consistent message.  Framing is extremely important and doesn't get used by our side nearly enough.  

2011-09-18 03:12:18
Rob Honeycutt

robhon@mac...
98.207.62.223

Please look at that first line as the "frame."  Then look at the frame in relation to each element laid out after that.  All the statements after the frame fit the frame.

Most of all it's important to avoid using the word "diagnostics" because that is Pielke's framing.  If you use that word then you are agreeing to his framing and he wins.  If we get him to use our framing of "many lines of evidence" then we win.

This is really important.  John's framing of "many lines of evidence" is probably the most powerful rhetorical tool used on this site.  It's clear.  It's easy to understand.  It's reinforced in how every article on this site is written and cited with published papers.  It's a perfect form of framing that everyone here should be repeating ad nauseum.

2011-09-18 03:13:50
nealjking

nealjking@gmail...
91.33.99.16

I guess it can pass.

It seems to me that we should be able to cook up Q4 & Q5 along similar lines.

2011-09-18 03:33:03
Rob Honeycutt

robhon@mac...
98.207.62.223

If we can find places to reinforce this I think that would be extremely powerful.  

2011-09-18 04:11:53Q4 & Q5
nealjking

nealjking@gmail...
91.33.99.16

OK, let's move onward: For the next one, couldn't we basically re-use the answer to Q3?

Q5: What are your preferred diagnostics to monitor climate change?

 

And for this one: Dana, any ideas?

Q4: What do the models predict should be the current value of these metrics?

Again, I'm a bit concerned this could get us into some tricky process situations.

2011-09-18 05:35:35
dana1981
Dana Nuccitelli
dana1981@yahoo...
69.230.106.125
I don't like the question. How is 'climate change' defined here? That's a pretty damn broad swath. It might be better to use John's "we look at all the lines of evidence" in this one and just leave it vague. For Q4, we can say the surface temp trend is consistent with model predictions [got plenty of links to support this]. TLT is slightly lower than models, but within the margin of error [see my Santer et al post]. I know sea level rise is above model projections, and Arctic sea ice decline too. Not sure about OHC. I can write some specific text later when I'm on my computer rather than iPad.
2011-09-18 05:38:18
dana1981
Dana Nuccitelli
dana1981@yahoo...
69.230.106.125
Oh and Alby, the RSS trend is about 0.16C per decade, and UAH is 0.19C. I just took the average of the two - that's the 0.18C.
2011-09-18 05:54:46
nealjking

nealjking@gmail...
91.33.99.16

Dana,

I think we could just repeat the Q3 answer for Q4: We're not obliged to accept Pielke's wording (or framing, as Rob keeps insisting).

Is there anything useful to be said about the models? I'm not quite sure why he's asking: Does he want to find out if we know which models are considered respectable? If we're current with the latest versions? or what? Got any clue?

2011-09-18 06:30:09
dana1981
Dana Nuccitelli
dana1981@yahoo...
69.230.106.125
I assume Q4 is meant as a follow up to Q3 [don't have the Qs in front of me right now], so he's trying to get a model-data discrepancies on the relevant global warming metrics. I would guess his motive is to argue that the recent slowdown in OHC trend is inconsistent with models. But that relies on interpreting short-term data, which we won't do. But like I said, most of the data we talked about in Q3 is consistent with model projections. Just need to find some model projections of OHC.
2011-09-18 10:51:00
dana1981
Dana Nuccitelli
dana1981@yahoo...
69.230.106.125

new thread with suggested answers to all Qs

2011-09-18 15:11:20
Agnostic

mikepope_9@hotmail...
118.208.81.161

Sorry - wrong comment, wrong place.

2011-09-18 18:53:32END OF THREAD
nealjking

nealjking@gmail...
84.151.56.122

OK, let's terminate this thread!