22 February 2010

Dare to ask a question!

I sometimes browse the referrals to my site, and was surprised to see someone commenting that they 'dare not' post here.  Most peculiar.  Unless the author of that comment is incapable of writing without profanity, or can't staying on topic (and his comment there showed no such signs), they certainly may 'dare' to post here.  At least as far as I'm concerned.  If there are employer-based reasons, well, that's out of my hands.

It's a while since I've put up a 'question place' note, so here's one for current questions.  Be daring, ask a question!


carrot eater said...

An open-ended opinion question:

If you were directing research, what aspect of climate science would you mark as a top priority? Meaning, in what area would some progress be most beneficial to the overall understanding?

No cheating by copying their answers, but the question is inspired by this feature:

Unknown said...

A vastly narrower question:

If the atmosphere of a planet did not interact with radiation (no CO2, water vapor, ozone - nothing but Argon, for example) what kind of temperature variation with altitude would you expect?

Michael Hauber said...

At Weatherzone in Australia we've been discussing an apparent large jump in SST's, and the WZ chart appears to be sourced from http://polar.ncep.noaa.gov/sst/ and this site has your name on it.

Our discussion is at: http://forum.weatherzone.com.au/ubbthreads.php?ubb=showflat&Number=811862&page=21

Looking at polar.ncep data, the half degree data looks roughly what I'm used to, whereas the 12th degree data looks a lot warmer and the same as the Weatherzone chart we think is too warm.

Would you care to comment?

Anonymous said...

Grrrr! Why on earth would Nature put such a topical news piece be behind a firewall? Do some of our leading science journals want to limit access to thoughtful journalism?

Robert Grumbine said...

I'll take up your question in its own post. Should be fun. I have that issue of Nature sitting on the table, so could cheat, but won't.

I'll guess that Nature wants the folks paying $199/year to feel like they are getting something for their money.

From your list of gases you want to exclude, I'll take you to mean that there would be no interaction with earthly radiation -- no greenhouse effect. But I'll leave interaction with solar radiation turned on. The structure of the variation I'd expect is pretty much what we have now. Start from the top of the atmosphere. The thermosphere and ionosphere form because they absorb very short wavelength radiation from the sun, and have little greenhouse gas content to help them cool, so that wouldn't change. The stratosphere and mesosphere form because oxygen and ozone absorb solar ultraviolet. If those layers couldn't cool by radiating in the infrared -- from ozone, I'd expect the layers to be quite a bit warmer than they are now. Not sure they'd get to thermospheric levels (1000 K and up), as mixing processes are better down here. In the troposphere, I'd expect the temperatures to cool with elevation from the surface, along a moist pseudo-adiabatic -- as they currently do. The change would be that they'd be cooling from a much colder temperature. Without the greenhouse effect of H2O, CO2, etc., the starting temperature would be more like the 255 K we expect from the simplest climate model. There'd also be some change to how fast the temperatures cooled, as there'd be less water vapor in the atmosphere to work with. Now, if pseudo-adiabatic means nothing to you, don't worry. I'll be taking a question closely related to yours in its own post soon. What sets the basic structure of the atmosphere?

More detailed questions about the stuff at polar.ncep.noaa.gov/sst are best sent to the email addresses on that web page. That's work stuff, and the blog here is non-work. Still, I can mention a couple of things. One is, starting February 16th, there should have been changes, almost strictly in the near-shore zones (less than 35 km from coast could have seen large changes, 35-200 km should see progressively smaller, and farther than that should have been almost identical). Two reasons. One is, there's a new land mask involved, one that is more permissive of water. Second is, the prior method for retrieving temperatures near land had some bad biases, so very near land (the 35 km) a different method is used. 35-200 km is a blending zone between the two, and more than 200 km it still uses the same method as before (it's better once land issues are resolved). So that's one side of things. The other side is that it looks like there was an error in the update process, and verification statistics (you can see this on the site itself) have gotten much worse. (Keeping in mind that 'much' is 0.5 K rms.)

Different sort of thing I'll mention. MMAB (the branch that produces the RTG SSTs you're looking at) is looking for people to keep an eye on the products and let them know when things look wrong. This is mostly aimed to the coastal and inland waters, but any error anywhere is something to hear about. So, again via the email addresses on the web site, you and the weatherzone folks are welcome to sign up to be providers of feedback.

Unknown said...

Thanks for the response. Actually, I was trying to start even simpler, with an atmosphere that didn't interact with solar radiation either. It seems to me that such an atmosphere should (eventually) be isothermal, with a temperature matching that of the ground. My understanding is that the adiabat represents roughly the maximum allowable temperature decline with altitude, with convection preventing the lapse rate from being any steeper - but that smaller lapse rates would be stable, giving rise to no convection. In the absence of radiative interaction (by assumption) and convection, this would leave conduction to bring the whole atmosphere to the same temperature as the ground, however slowly (ignoring questions of latitude, day/night, etc.) Does this sound reasonable?

I'll be looking forward to your post.

carrot eater said...

I'll try a sceptical question.

The various coupled ocean-atmosphere models featured in the IPCC report tend to agree with each other when hindcasting over the last 100 years. Yet they diverge from each other somewhat when projecting the next 100 years, for a given scenario of forcings.

Why is this? Are they cheating in the hindcast? Are they tuning some parameter to fit the past record, even though they're not supposed to?

Anonymous said...

What would be the forcing resulting from the change in albedo if summer ice were gone from the Arctic?

I have pushed the limits of my Google-fu, but all I can find are qualitative statements (it will be warmer; other ice will melt faster) when what I want is a quantitative estimate: Arctic sun falling on open water all summer = x W/m^2 forcing.

Willing to share my embarrassingly amateur efforts to calculate on request.

Robert Grumbine said...

carrot: I'll come back to your question. I'd like to provide some cites, which will take longer.

anonymous: a rigorous number of the sort you seem to be looking for will be hard to provide. I can do a fair non-rigorous number. But, since you mention having done some preliminary calculations, I'm intrigued. Please do show how you were going about the question. We'll start from there and head for more elaboration if it seems needed.

EliRabett said...

Dave, the first thing is that the surface temperature would be set by the equilibrium between absorbed solar and radiated thermal, that means a lot cooler than when the ghgs are present.

The thermal structure of the atmosphere would be set pretty much by the dry adiabatic lapse rate, but remember that in your model, you are starting from a colder base temperature. That is if Ts is the surface temperature, T(z) =To -Gz where G is the scale height.

Eli is also assuming no condensible gases and no ionization way up.

Unknown said...

Eli - I'm happy with the notion that the surface would be cooler in the zero-GHG situation I'm imagining (an atmosphere with no radiatively active constituents.) But I don't yet see why temperature would still decrease with altitude.

My (tentative) understanding is that the adiabatic lapse rate is basically a limit on temperature gradient imposed by convection. In an atmosphere with a gradient steeper than the adiabatic lapse rate, a packet of air which happened to rise a small distance and cool adiabatically would be warmer and more buoyant than the surrounding air, and would therefore keep moving - transferring heat aloft by convection, and bringing the temperature gradient back to the adiabatic lapse rate.

But that only prevents the gradient from being too steep. If the gradient is shallower than the adiabatic lapse rate, that rising packet of air would find itself cooler and less buoyant that its surroundings, so the atmosphere would be stable and convection would not occur.

Since this atmosphere (by assumption) can't lose heat to space by radiation, it seems to me that even if we assumed an initial temperature profile that was steeper than the adiabiatic lapse rate, the subsequent convection would be temporary and eventually (through conduction) the entire atmosphere would reach the same temperature as the surface.

Does this sound correct, or am I missing something about the connection between lapse rates and convection?

Anonymous said...

I've got one related to Dave W's question, but it applies to the world as it is, rather than a hypothetical.

Why is the deep ocean so cold? If the average ocean surface temperature is something like 15C (this page says 17 - http://www.windows.ucar.edu/tour/link=/earth/Water/temp.html ) then without large scale circulation, the deep ocean should average about the same, but instead the deep ocean is around 3C. The only way that makes sense is if the deep ocean water at any latitude has much better thermal contact (through circulation) with high latitude surface waters than with any other surface waters.

EliRabett said...

Dave, the lapse rate is a change in temperature per unit height imposed by the change in density (that's where the g comes in)

This does not change if the surface temperature varies, but the temperature profile shifts to a lower temperature. Take a look at the fourth figure from the top at this link

Unknown said...

I don't think the temperature drop is necessarily required by the change in density.

Suppose the surface temperature, pressure, and density are T0, p0, and ρ0 which satisfy the ideal gas law p0 = ρ0 R T0. Then we could have a uniform temperature profile
T(z) = T0
and exponential pressure and density profiles
p(z) = p0 exp(-z/h)
ρ(z) = ρ0 exp(-z/h)
where the scale height h is R T0/g. These profiles satisfy the ideal gas law everywhere, as well as the hydrostatic equation dp/dz = -ρ g.

Most derivations of the adiabatic lapse rate that I've seen appear to assume the presence of convection, which implies heating from below and cooling from above. I'm suggesting that if the atmosphere weren't radiatively active, it couldn't cool from above, so in the long run there wouldn't be convection and it would match the surface temperature throughout.

MikeCoombes said...

Several questions:

I keep reading Anti-AGWers mention global temperature "rebound" from the Little Ice Age or the last ice age. What are they talking about? Are they saying that there is energy coming from outside the earth warming it up such as a brighter sun or orbital variation. Or are they saying heat was stored deep in the oceans and is now coming to the surface? Is there any science to the concept? Shouldn't a rebound have a cause?

2. Similarly there is talk of "natural variation" to account for long term global temperatures changes. I assume most don't mean Milankovich Cycles since they are so slow. What are they talking about?

Perhaps I am too wedded to Cause and Effect and Conservation of Energy to understand the arguments.


Robert Grumbine said...

Eli, Dave W.:
I look at it a little differently. The lapse rate is how fast temperatures would change if you lift a parcel through the atmosphere, keeping it in an insulated, but flexible bag. For a dry atmosphere, that's about 10 C per km. That value has nothing to do with the radiative properties of the atmosphere, being strictly a function of the nature of the gas (predominantly diatomic molecules -- O2, N2) specific heat, and gravity. An atmosphere with such a profile would be neutrally stable. Any parcel could be lifted or lowered to any other location and it would stay there. That is a constant potential temperature. This is more or less your (Dave) convective atmosphere.

Radiation has been turned off, so there's only one more method for moving energy around -- conduction (diffusion). Diffusion tends to make temperatures (local temperatures, not potential temperatures) the same. Given a long time, and no horizontal motions, diffusion in Dave's non-radiating atmosphere should move the atmosphere towards constant temperature. This will mean increasing potential temperature with height, and will be stable.

The nearest example we have on the earth is the lower stratosphere, which was known in the early 1900s as the 'constant temperature' layer. Balloons couldn't get high enough to see the ozone layer, and its corresponding temperature increase.

Mike Coombes:
Always best to ask the person making a claim what they mean by it. Maybe you can check with one of your recent examples and see what they mean. For here, I'll go with what I've seen people mean when making the claim.

One error they regularly make is to cherry pick a starting date and conclude that there is now a cooling trend.

What their claim is, is that the people they disagree with are making the same error when they conclude from the temperature records that there is now a warming trend. The world is cold during an ice age, so of course we're warmer now. The 'little ice age' was cold, so of course we're warmer now. To this minimal degree, it is indeed correct to say 'of course we're warmer now'.

What is missed is that we quit warming up from the last ice age about 6000 years ago (the 'Holocene Optimum') The long term trend since then has been a cooling. In other words, the current warming cannot be a matter of 'coming out of the ice age'. We finished that a long time ago -- much farther back than the 100-150 years we're using to discuss current warming trend. So we're not making the mistake of cherry-picking our start date with respect to ice ages. I'll note, not all cherry picking is of malicious intent. It's one of the easier mistakes to make accidentally -- you start where your data starts, but it turns out that your data start in an unusual period. This is not what's happening when people take 1998 (or 2002, ...) as their starting point for talking about recent climate trends. But it can happen.

Little ice age is much closer to our time, so might be more of a concern for accidental cherry picking. If the coldest part of the little ice age were around 1850 to 1880 -- when we start the instrumental temperature records -- then there would indeed be a natural recovery to warmer temperatures following that exceptional cold. The thing is, mid-1800s is when the Little Ice Age is considered to have ended -- temperatures and other climate features had returned to the baseline. Starting your trend computations at the time a climate feature had returned to baseline is the right thing to do, not a cherry-pick.

Robert Grumbine said...

You hit on the central feature of the global ocean. While most of the surface is warm, the overwhelming majority of the volume of the ocean is cold -- below 4 C.

You are also correct on why it is this way -- the deep ocean circulation is better-connected to the cold surface waters than to warm waters. If we look to a map of the ocean surface, we find very little is cold enough to be supplying the cold water to the deep ocean. So if you understood a rather small part of the ocean (turns out to be largely the northernmost Atlantic, north Pacific, and Antarctic), you understand what's driving the overwhelming majority of the volume of the ocean.

You have also hit on my dissertation area :-) So I'll take the question up as a full post in a few days.

GFW said...

(I was the anonymous who asked about deep ocean temps)

Thanks for the response, and I'll look for the full size posting. I have a followup question, but it can wait to be in better context on said upcoming post.

Unknown said...

Thanks, Robert. The reason I asked the question in the first place is that the answers I've seen to the question "why does temperature drop with altitude" have always left me dissatisfied. They tend to jump immediately to descriptions of air packets moving up and becoming cooler through adiabatic expansion, but this seems to confuse a mechanism for heat transfer – convection – with the fundamental cause, which is that the atmosphere is heated from the bottom and cooled from the top via radiation to space.

If a small amount of GHGs were added to the otherwise inert atmosphere I’m imagining, I think we could get a situation in which a small temperature gradient was compatible with the net upward transfer of energy via radiation and diffusion, without any convection. Increasing the GHG concentration would increase the gradient, until it reached the adiabatic lapse rate – at which point convection would occur, and the mechanism in the standard explanation would limit further gradient increases. Presumably we're well beyond this point now, so a moderate decrease in GHGs wouldn’t be enough to shut down convection – it would just shift the surface temperature, as in the figure Eli pointed to above.

Anonymous said...

I've a question about the lag rate between CO2 readings and temperature at Mauna Loa.

I've been 'debating' some denialists at a hobby site I frequent, and one pushed Atmospheric CO2 and Global Warming by Jaworowski as a definitive 'work'.

It reads like one to the layman - though the thought 'and why do the experts need a radiation scientist to point out the obvious to them?' occurs.

One part has me puzzled. Jaworowski refers to Kuo et al's "Coherence established between atmospheric carbon dioxide and global temperature," stating:

Kuo et al. (1990) noted that changing temperatures lead to changes in the amount of CO2 outgassing or dissolution in the ocean, and to variation in biological activity, and thus to CO2 level changes in the atmosphere. The five month lag time of CO2 changes behind temperature changes, indicates that the causality is: temperature-to-CO2.

Now, as the paper abstract itself says:

"The hypothesis that the increase in atmospheric CO2 is related to observable changes in the climate is tested using modern methods of time-series analysis. The results confirm that average global temperature is increasing and that temperature and atmospheric CO2 are significantly correlated over the past 30 years. Changes in CO2 content lag those in temperature by five months.

I know he's talking bullshit, however - do you have any idea of the physical reason for the lag?

Robert Grumbine said...

I think this is yet another example of not understanding elementary time series analysis. I can't get Kuo et al directly, but have checked references to it, one by Thomson (coauthor on Kuo et al.).

The thing is, for looking at time series, the first thing done is to remove the long term trends. Both CO2 and temperature have long term trends, and most of our concern about climate change rests on there being a connection between the trends. This trend removal amounts to a high pass filter (explained in the linked note).

What the authors are actually looking at is not the fundamental 'does CO2 cause warming', but 'do squiggles in CO2 cause warming or do squiggles in temperature cause CO2 to change'. To judge by the references to it, the authors were clear about this, looking at interannual variations in both CO2 and temperature. The lag they find, then, applies to the short term (few years) variations, not the long term -- climate -- connections.

There are several candidates for CO2 to lag the temperature for annual to short interannual basis. Most straightforwardly, it takes time for CO2 to cook out of the ocean or soil. And it takes time for forests to grow more (or less) in response to temperatures. These are the 5 months mentioned.

Anonymous said...

Cheers for that Robert, I shall study the link when I get some free time - then all I have to do is come up with a good analogy that might make sense to the denialists I'm currently 'debating'...


GFW said...

I have another question, but it's not a cold deep water question. Quite the opposite.

An El Nino event accomplishes a redistribution of heat stored in the ocean to the atmosphere. How "old" is that heat?

I realize one can't track individual joules around, but I think the intent of my question is clear. It's about how deep the heat goes before it's brought up again. If it's a shallow phenomena, then the answer might be "approximately the same time as the period between El Nino events". The deeper that "surface" heat can go, the more likely the answer is longer. Anyway, I thought I'd ask an ocean current expert. (A little searching while asking this question suggests that the question may be poorly formed. If I'm understanding equatorial divergence/upwelling, then El Nino puts very recent ocean heat into the air instead of allowing it to flow north or south in surface currents. It's not so much a release of long stored heat as it is an interruption in the upwelling of deeper colder water. Ok, I'll stop guessing now.)

carrot eater said...

Mainly, I just want to complain that 'deltoid' appears twice on the blog list.

I know you don't like personal questions, but I wonder if you are more careful when you write emails now, than in bygone days.

Robert Grumbine said...

Your guesses were pretty much on target, as far as I understand what you want. In the cold phase, warm water is pooled deep on the western side of the Pacific. In the warm phase (El Nino part of the Southern Oscillation), the warm water spreads to the east. Since it's mostly just a matter of warm upper ocean water sloshing around, the time scale of the storage/release is indeed also the time scale of ENSO.

Thanks for the reminder. I was having some problems with blogroll pointers not updating (still am -- Jules' blog has been updated much more recently than a year ago!).

On the emails ... the swifthack business doesn't really change anything for me. It was just another demonstration that 'seek and ye shall find' is still a valid principle. If a hacker goes looking for things to take out of context (the released mail is far less than 10% of what Jones alone wrote, and I understand it's not only his mail that was hacked) they can find things that lend themselves to such extraction and selective quoting. On top of that, given an audience of people seeking to find things to attack, the audience will find those things -- whether they're really present or not.

For my own email (which is something like 3000 per year, or 40,000 in the period the hacker was trawling through at CRU), for instance, there are many in which I say something like 'there is a problem with X', where X is some data source or model. The hacker could collect quite a few such messages and make it look -- to the folks on fault-finding missions -- like I think every model or data set I've ever looked at is bad (and then the fault-finders would trumpet how I agree with them, which would be far from true). In doing that, all he or the fault-finders would have to do is ignore my later messages in which I say 'The problem with X has been fixed.', or 'I was wrong about the problem with X because ...'.

It occurs to me, though, that if I thought I'd found an error in one of the series the fault-finders are concerned about, as a result of the swifthacking, I would probably be far less likely to discuss the issue with the authors by email. Phone call or paper mail instead. I think it's likely that the main result on science from swifthack is to impair the error-correcting mechanism.

Hank Roberts said...

I'd welcome a correction to my speculation here:

Anonymous said...

I'd greatly appreciate any explanations or references to help me with this question I have. Is there any isotopic difference between volcanic carbon dioxide and industrial sources of carbon dioxide?

I found C12/C13 ratios for voclanic, and C13/C14 for industrial. I'd like for comparisons sake to have them in the same isotope fraction. Is there anyway to reconcile these fractions?

Robert Grumbine said...

Your comment here is what prompted my Arctic sea ice update. I hope that addressed the question.

The best source I know for nonprofessionals is Jan Schloerer's CO2rise FAQ. He'll have values and, even better, citations that you have a chance to follow up. At least with library requests.

Both volcanic and industrial carbon is fully depleted in 14C. Volcanic carbon is close to reference (undepleted) in 13C. The carbon released by making cement (from limestone) is also close to undepleted. But fossil fuel carbon is markedly depleted in 13C.