Water Vapor and Cloud Feedbacks

Is anyone interested in starting a separate thread on water vapor and cloud feedbacks? While it is of some relevance to the hurricane/global warming topic, the relevance is indirect and this topic certainly has enough scientific meat for its own thread (provided people are sufficiently interested.

Comment by Judith Curry “¢’‚¬? 17 September 2006 @ 9:39 am | Edit This

Your wish is my command.

There are a number of topics in this area that intrigue me. In the mid-1990s, Ramanathan and others wrote about “anomalous atmospheric absorption” in which they argued that atmospheric absorption of near infrared solar radiation was up to 25 wm-2 greater than modeled (or in Trenberth’s standard figures). IPCC TAR used TRenberth, mentioning Ramanathan’s concerns only in passing without analyzing or explaining them. IPCC 4AR makes no reference to the issue at all as far as I could tell (I requested an explanation but probably won’t get one.)

A number of recent articles by specialists in water vapor have proposed significantly higher NIR absorption by water vapor than in climate models. Belmiloud and Tennison have written scathingly about this. There was an interesting clerical error in NIR absorption levels in HITRAN-96 used in all IPCC TAR models – the error in wm-2 was greater than 2x CO2. I’ve seen some discussion that observed NIR absorption can be (sort of) explained if cloud droplet sizes are assumed to be three times larger than generally believed. (I’ll check this reference).

I get the impression that recent work mostly attributes “anomalous” absorption to aerosols and, to some extent, I get the feeling that aerosols and higher parameterizations for NIR absorption by water vapor/liquid/clouds are competing mechanisms. Intuitively it seems to me that this would result in quite different feedback consequences – higher NIR absorption from water cycle would be a negative feedback whereas aerosols are not.

The recent discussion at realclimate [link] was also interesting – I was very struck by Held’s surprise that cloud feedbacks were all uniformly strongly positive – Held said that he expected the feedbacks to be evenly distributed between positive and negative. It’s a bit disquieting that such a leading expert should be surprised by this.

138 Comments

  1. Posted Sep 17, 2006 at 10:54 AM | Permalink

    As a start I’d like to submit this website from Physics World which shows the lack of understanding by the warmers of the role of water vapor:

    http://physicsweb.org/articles/world/16/5/7/1

  2. Steve McIntyre
    Posted Sep 17, 2006 at 11:25 AM | Permalink

    #1 that has a discussion of “anomalous” absorption. The topic is not discussed in IPCC 4AR. Water vapor feedback is a crucial issue. Arguably it should have had a large section on its own in IPCC 4AR in which each detail was discussed since these are the major issues in quantifying 2xCo2. Unfortunately IPCC WG1 tends to provide little discussion on such matters, making it much less useful as a guide than one would hope for. My comment here is not just backbiting – as I suggested such discussions both in IPCC First Draft reviews and even a couple of years ago in some correspondence with people connected with IPCC.

  3. TCO
    Posted Sep 17, 2006 at 1:20 PM | Permalink

    Where are the Held remarks? Can’t find them.

  4. Judith Curry
    Posted Sep 17, 2006 at 2:26 PM | Permalink

    re anomalous absorption. This issue had been hanging around at a low level for several decades

    and then in 1995, with the near simultaneous publication of the Cess and Ramanathan papers, the issue “exploded”. If climateaudit had been around in 1995, you would have had a field day with the statistics and analysis techniques, these studies make hockey stickers and hurricaners look like statistical geniuses. This issue was subject to intense debate and investigation in the late 1990’s, mainly by the DOE Atmospheric Radiation Measurement program (see http://www.arm.gov, search “anomalous absorption” and tons of papers will pop up).

    The papers that arguably put this issue to rest:

    Sengupta M, Ackerman TP, Investigating anomalous absorption using surface measurements JOURNAL OF GEOPHYSICAL RESEARCH-ATMOSPHERES 108 (D24): Art. No. 4761 DEC 17 2003 (note unpublished report version of this is found on the arm website).
    Ackerman TP, Flynn DM, Marchand RT, Quantifying the magnitude of anomalous solar absorption JOURNAL OF GEOPHYSICAL RESEARCH-ATMOSPHERES 109 (D9): Art. No. 4273 MAY 9 2003

    Basically, the field was sent on a merry goose chase over Cess and Ramanathan’s papers. The timing of the Cess and Ramanathan papers was such that this issue was barely mentioned in 2nd assessment report, and by the time of TAR, it was pretty much a dead issue.

    The refractive index for liquid water drops is well known. arguably we do not have a firm theoretical foundation for how we treat water drops with an embedded aerosol particle, but there is no apparent great sensitivity to this. The other issue is 3D cloud effects; we know how to treat these as a function of cloud geometry (but certainly not all climate models deal with this adequately). These small uncertainties pale in the face of inadequate simulation of cloud properties by climate models (if your clouds are far from reality, these small details like drop/aerosol mixed refractive index don’t make that much difference).

    In terms of water vapor, the big challenge is continuum absorption (on the doe arm site, search “continuum absorption”). the climate models mostly use empirical relationships to treat this; the theoretical foundation for this is still fuzzy. The most recent breakthrough in this area has been clarification of the “dirty window” (search “dirty window” on the arm site) in the water vapor rotation band, which is mainly of relevance for parts of the atmosphere where the humidity is low. Basically, the “clear sky” radiative transfer problem is regarded as pretty much solved in terms of radiative fluxes. This does not mean that all climate models adequatly treat this. I note here that the GISS GCM in particular has paid a substantial amount of attention to its radiation code and it is definitely state of the art.

    cloud and water vapor feedback are colossally difficult to evaluate and the processes that determine the cloud and water vapor characteristics are mostly “sub gridscale” and thus parameterized (search arm site “cloud feedback” to get a lot of free, up to date info).

    The best paper on cloud feedback IMO is Stephens GL. Cloud feedbacks in the climate system: A critical review JOURNAL OF CLIMATE 18 (2): 237-273 JAN 15 2005

    But separating cloud from water vapor feedback is rather artificial, they are both totally entertwined.

    (more later on all this)

  5. eduardo zorita
    Posted Sep 17, 2006 at 3:19 PM | Permalink

    Which would be the way to make progress in this issue? Climate models will for a long time lack the neccesary spatial resolution to represent all these microhysics processes, so that coarse parametrizations, with their related uncertainties, will be needed. If the spread in the IPCC projections within a single scenario are mainly due to cloud feedback, does this mean that the uncertainty in temperature projections is not going to be narrowed any time soon?

    From the empirical side, could long time series of observed cloudiness be of help? I am aware there exists some series over hundred years long in Spain and Austria

    eduardo

  6. Steve McIntyre
    Posted Sep 17, 2006 at 3:57 PM | Permalink

    Judith, thanks so much for the thoughtful summary. I thought about clouds flying back from Europe looking at the miles and miles of different types of clouds. The patterns are often so beautifully fractal. I guess Joni Mitchell’s anthem still stands for climate models.

    I’m intrigued by Ou’s maximum entropy approch to clouds which I posted up on here. It seems quite possible to me that some approach like this might end up being just as successful as trying to generate individual clouds, but who knows?

  7. Judith Curry
    Posted Sep 17, 2006 at 4:01 PM | Permalink

    NSF is spending alot of money on this issue, see

    http://welcome.colostate.edu/index.asp?url=media_nsf

    I am not hugely optimistic about that this project will “solve” the problem, but the colorado state group is exploring one possible strategy, whereby the imbed cloud resolving models in each climate model grid cell. This strategy may introduce more problems than it solves, and in another decade, climate models will be run at really high resolution anyways (say 10 km). Resolution isn’t the only problem tho.

    Here are some of the main challenges:

    Climate models are colosally complex dynamical systems, many orders of magnitude more complex than the systems that dynamical systems people (mostly mathematicians and physicists) deal with. So on one level, we really don’t know what we are doing with these models, but on another level we are building confidence in these models through comparison with observations. The atmospheric component of climate models have been evaluated against observations for decades as a part of numerical weather prediction/forecasts (no the models don’t do a great job of forecasting surface weather at a particular point, but they do get the large scale aspects of weather systems fairly well on time scales out to 10 days). And increasingly ocean models and coupled atm/ocean models are being evaluated in the context of seasonal forecasts. So in spite of their complexity, the models seem to “work” on some level.

    The problem with the sub grid parameterizations (this is how we deal with the unresolved degrees of freedom) is that the are typically based on a little bit of physics and a little bit of statistics, and tuned so that the model “works”. This is sort of OK for a weather model (i.e. it “works”), but in climate models used to make doubling CO2 projections, these parameterizations clearly aren’t “tuned” appropriately for the new climate.
    So the parameterization challenge is to develop parameterizations that
    1) capture in some stochastic way the necessary effects of the parameterized process
    2) are physically robust outside the current climate regime
    3) do not adversely affect the stability of the calculation
    4) work in tandem through appropriate linkages with other parameterizations

    Many of the things that need to be parameterized are very very difficult. the problems with parameterizing convective clouds and subtropical stratocumulus clouds bump up against what is arguably the outstanding unsolved problem of classical physics: turbulence. Clouds and water vapor and the associated changes of phase are one of the greatest challenges in modelling a nonlinear system: the presence of phase boundaries. So there are some fundamental physics challenges here, and climate modelers try to do the parameterizations in some sort of sensible way, but they almost certainly miss processes that would contribute to feedbacks in a perturbed system by neglecting what may be important degrees of freedom. So modelling paleoclimates is an important way to test climate models.

    But there is no easy solution to these problems. We need to develop more sophisticated ways to test the models and parameterize the unresolved degrees of freedom. It would also be great if the dynamical systems people could develop some theories about how such large models work, which could help us answer questions about how the resolved degrees of freedom can act to compensate (or not) for the missing degrees of freedom, how to really evaluate feedback in such models (the way we do it now assumes that the models are linear, which clearly can’t be correct). And so on.

  8. Judith Curry
    Posted Sep 17, 2006 at 4:07 PM | Permalink

    Steve, there must be some way to use entropy, it is a hugely underutilized piece of physics in climate science. My brief foray into this ended in frustration

    Duane, G. and J.A. Curry, 1997: Entropy of a convecting water-air system and the interpretation of cloud morphogenesis. Quart. J. Roy. Meteorol. Soc., 123, 605-629

    Again, help from the dynamical systems people is needed here in terms of integral constraints associated with entropy on dynamical systems

  9. Steve McIntyre
    Posted Sep 17, 2006 at 4:12 PM | Permalink

    One mathematical question that I’ve been turning over in my mind – and try to think about it as a problem in functional analysis rather than climate – if you have a model in which you have a few forcing series and with physics operating on a spherical shell in which your only output of interest is a time series of averages taken in equal time intervals (“years”) over the entire sphere (i.e. very ergodic averages), does there exist a 1-D atmosphere or alternatively a circular shell (2-D) atmosphere such that applying physics to the simpler structure approximates the ergodica average of interest to a desired accuracy?

    I talked to a Dutch mathematician about this and he said that he didn’t know the answer to this and it sounded like a very hard problem. It would be fun to formulate a mathematical question that was hard for mathematicians.

  10. Judith Curry
    Posted Sep 17, 2006 at 4:50 PM | Permalink

    People have tried to formulate low-order one and two dimensional climate models (Jerry North has done some of the more interesting work on this). Toy models are thought provoking and can help us learn, but until we do the problem “correctly” with a more sophisticated model, we probably wouldn’t have any confidence in the forecast capability of the toy model (if there was any).

    We really do need help from the mathematicians and physicists on this (note the big colorado state project does have some of them involved, will be interesting to see if they can contribute something to the cloud parameterization problem).

  11. Posted Sep 17, 2006 at 4:54 PM | Permalink

    Judith, you mentioned the refractive index for water drops, but as a question in line with the cloud feedback: what happens with IR wave absorption/reflection/transmission in clouds?

    – (low level) clouds are reflecting ocean emitted IR in such quantities, that they reduce the overnight cooling of the surface (as discussed on RC). But as clouds are mostly fine water droplets, that means that the IR is mostly immediately reflected or absorbed/re-emitted (to all sides, including – in part – through the cloud to space), and that little (if any) is permanently absorbed. If substantial amounts of heat would be absorbed, this should warm the cloud droplets and ultimately the cloud would disappear.
    – If most of the cloud droplets re-emit (reflect) IR, shouldn’t that be the case for the skin of the ocean surface too. That means, immediate re-radiation (reflection) without much absorption, thus no/little heating of the ocean below the skin.
    – What about the different wavelength spectra: ocean surface emitted IR has a temperature dependent spectrum. Does this change for CO2 (or other air molecules) re-emitted heat (after absorption) and/or cloud droplets, due to the colder temperatures at the altitude of re-emission?

  12. David Smith
    Posted Sep 17, 2006 at 5:20 PM | Permalink

    Thanks for the links – I have a lot to read.

    I have wondered about the effects of near-saturation (high humidity) on the transmission of IR.

    Here in the high-humidity near-tropics we cool at night to near the dewpoint by about 2AM, and that’s it – the temperature drop slows substantially. That has always impressed me. The conventional explanation is that we are still radiating away lots of heat, but the energy removed is resulting in condensation (dew, fog).

    Condensation indeed happens, but I am not so confident that tells the whole story. I wonder if something puts “the brakes on” and slows radiation loss substantially during the high humidity of late night.

    That would be significant because, in effect, it means that Earth has some reserve of radiative capacity. If there is excess heat due to excess CO2, then the high-humidity-caused slowdown of radiation is delayed until, say, 4AM instead of 3AM, but the earth still radiates away the excess heat before sunrise.

    Anyway, just a thought for me to explore as I read.

  13. John Creighton
    Posted Sep 17, 2006 at 5:21 PM | Permalink

    I looked up functional analysis and it sounds a bit like boundary value problems. I remember something in electromagnetism about the spatial and time parts being separable so we new the form of the functions were decaying sinusoids. We then integrate along the boundaries to find the weight each function contributed to the solution of the partial differential equation. In the examples I did in school the math was easy the basis functions were orthogonal. However you don’t necessarily need orthogonal basis functions to approximate a solution.

    In terms of climate models, I think you could use such approaches to fill in missing information from your measurements but things like clouds do not have nice smooth functional forms so some thought needs to be given once you decide on the basis functions how you define the optimal combination of them. For clouds you suggest a fractal model. Some thought must be given about how we define closeness between a fractal approximation of clouds and the actual state of the clouds.

    Turbulence is another interesting one. How might we approximate the turbulence when we can’t possible measure every vortex. How are the vortices spatially distributed and how do there intensities vary. Maybe we shouldn’t just pick one set of basis functions but randomly select from a group of likely basis functions to get an idea of the uncertainty. Of course this only describes how we might initialize a global climate model and does not address if the simulation methods are appropriate. If clouds are fractal then no amount of resolution will every completely describe the structure of clouds and thus rather then shrinking the resolution maybe it is more appropriate to increase the number of states for each of the smallest blocks in the simulation.

  14. Judith Curry
    Posted Sep 17, 2006 at 5:45 PM | Permalink

    re #11, 12: the imaginary part of the refractive index determines whether a substance (e.g. water) can absorb/emit radiation at that wavelength. We know the refractive index for water, but radiative transfer in clouds is determined by a combination of emission and multiple scattering that determine the “bulk” radiative properties of clouds like reflectivity, transmissivity, absorptivity.

    The tropics are very interesting from a radiative point of view. There is so much water vapor in the lower atmosphere that adding CO2 or a cloud doesn’t make all that much difference to the local transfer of IR radiation. The so called “water vapor window” which is mostly transparend to IR (around 8-12 microns) is mostly closed up in the tropics so you don’t see much nightime surface cooling.

    cloud drops have pretty much the same temperature as the surrounding air. if you look at the heat budget of an individual drop, you have absorption and emission of IR, heat conduction from the atmosphere, and latent heat associated with any condensation or evaporation (which depends on the local air humidity). You don’t get the cool skin effect on a drop such as you do on the ocean surface (the drops are too small).

  15. Posted Sep 17, 2006 at 5:47 PM | Permalink

    Judith, I did send my remarks on cloud feedback (strongly negative) on the other thread, but a recent correction of the ERBS satellite data (see Wong ea.) does change the radiation budget in sign…
    Instead of the original 3.1/-2.4/-0.7 W/m2/decade for LW/SW/net radiation (TOA), after correction, this is changed to 0.7/-2.1/1.4 W/m2/decade for the 1990s vs. the 1980s and for the 20N-20S equatorial band.

    There still are a lot of questions left, like the difference of ~5 W/m2 between ERBS (1999) and CERES satellites which operate since 2000, with a gap of one year, which prevents intercalibration. The consequences for the radiation budget for the subtropics are not given. For the period 1990-2000, the radiation budget should equal the observed increase in ocean heat content (I wonder how that can be, as the satellites don’t cover the highest latitudes). Further, that still doesn’t explain the observed reduction in ocean heat content 1980-1990. And last but not least, the changes in radiation budget still are an order of magnitude higher than what can be expected from the increase in GHGs in the period of interest (be it now with a positive sign)… Thus a positive feedback for increasing SST (or the cause of higher SSTs?) in the tropics?

    This doesn’t change the findings of J. Norris, who found a global decrease of 1.3% in upper-level clouds since 1952 over the oceans. Which normally indicates more cooling (the results for low-level clouds were ambiguous). But he too used the previous ERBS radiation budget data to calculate the overall radiation budget change…

  16. Steve McIntyre
    Posted Sep 17, 2006 at 6:09 PM | Permalink

    #8. Entropy has just been used in proving tghe Poincare conjecture (Perelman, The entropy formula for the Ricci flow and its geometric applications). This is a huge accoplishment; Perelman’s proof has been studied for about 3 years and recently endorsed. Mathematicians don’t rush to accept something because of a press release or an abstract in NAture. There’s a profile of PErelman in NEw Yorker recently – he’s a strange guy. He just turned down a $1 million prize for solving the Poincare conjecture. His proof is online – he didn’t bother publishing it in a journal. I’ve browsed it but can’t even find a toehold to begin understanding anything about it, although I studied algebraic topology when I was young.

    As I posted once before, another Clay Institute prize is for making any substantial contribution towards understanding the Navier-Stokes equations, which are presently intractable. Navier-Stokes are the equations that govern climate, although the uncertainties of the mathematicians don;t seem to trouble modelers.

  17. Steve McIntyre
    Posted Sep 17, 2006 at 6:12 PM | Permalink

    If you look at the category http://www.climateaudit.org/index.php?cat=25 and scroll back, I did some posts last year on water vapor especially NIR absorption issues. Some of those topics might be worth reviving.

  18. Ross McKitrick
    Posted Sep 17, 2006 at 8:05 PM | Permalink

    Welcome back Steve. Re #9, I was looking up some info on Levy distributions in air pollution and came across a very readable paper by Schertzer and Lovejoy here that might pertain to your question. It’s mostly over my head, but there’s a discussion of the literature on the search for low-dimensional climate attractor models, as well as ergodicity in geophysical systems. From what I understand, initial hopes on each concept have not panned out, though many theoretical insights have been generated along the way.

  19. Steve McIntyre
    Posted Sep 17, 2006 at 8:21 PM | Permalink

    #18. Reference isn’t what I had in mind. I don’t see why chaos theory would help. Ergodic theory is essentially unchaotic i.e. averages exist. The question is more – if you have an evolving average global average, is there in some sense aqn average atmosphere that can yield a result arbitrarily close to the evolving average.

  20. KevinUK
    Posted Sep 18, 2006 at 12:47 AM | Permalink

    #1, Richard, great link which sums up the state of our somewhat limited knowledge of the radiative and absorptive physical properties of water vapour.

    #2, Steve,

    I think at times you are far to polite. That’s your choice and I’m sure that we all respect it but for some of us the explanation as to the lack of mention of water vapour in the IPCC TAR and it would now seem complete lack of any mention in the IPCC draft 4AR is obvious.

    If other (more plausible – based on sound physics) explanations for global warming were brought into the equation then the case for anthropogenic global warming (AGW) would be underminded. In that circumstance just how long do you think the funding for the IPCC would last if it turned out that currently observed global warming turned out to be due almost entirely to natural causes? I wager that the funding would not continue for very much longer. Instead they (the eco-theologian alarmists within the IPCC) choose to take advantage of our current lack of knowledge of the physics of this clearly dominant greenhouse gas i.e. water vapour and choose instead to overplay the contribution from CO2 through invented positive feedback effects which enable them to make alarmist predictions of rapid temperature increases over short periods to underpin their primary objective, namely to force us to reduce our dependence on fossil fuels.

    KevinUK

  21. maksimovich
    Posted Sep 18, 2006 at 1:59 AM | Permalink

    re 9 Erik Bollt uses some intersting techniques in a paper called targeting control of chaos.The use of OGY allows pole placement and stabalization of unstable orbits.The OGY is as he suggests with targeting is ergodicty plus time.

    You may find some serendipity with Perelman and the poincare solution although this deals with sectional solutions.

    The problematic problem that will never be overcome with models is the reliance on moving parameters to provide an homegenous state,thereby creating the initial state paradox.

    The solution is control of each set of variables within its own system.

  22. Dave Dardinger
    Posted Sep 18, 2006 at 11:21 AM | Permalink

    The title of this thread, Water Vapor and Cloud Feedbacks sounds nice and compact, but as I’m sure we all know, there are a multitude of various subjects under that umbrella topic. Does anyone know of a comprehensive listing of what all might fall under this rubric? Lacking that, is there something similar to the diagrams which are available for the Earth Radiation Budget or the Carbon cycle?

    I see the risk in a thread like this is that we’ll discuss a bit of this and a bit of that and no single topic will be explored in sufficient detail to allow anyone to reach a reasonable conclusion.

    Let me start to throw out fairly focused topics which might either be dispatched with a single citation or might evolve into a major discussion. Here’s the first one.

    1. As updrafts of air heated at the surface rise, they eventually cool to the dew point of the air and form clouds. Obviously this varies in terms of height of cloud formation, etc. in a fairly simple manner based on the absolute humidity of the air, barometric pressure, temperatures in the upper air, etc. I’d guess that this sort of thing is too small scale for a GCM to account for expect as a parameterization. Does anyone know just how detailed the parameters are which are maintained for each grid cell and how they are used? In particular, is the sort of typology (direction of prevailing winds, average elevation of land, etc.) of the cell taken into consideration or is is just something like the % of water surface in the cell?

  23. John Creighton
    Posted Sep 18, 2006 at 12:38 PM | Permalink

    It is not enough for them just to be cooled to the dew point. They need something to condense on. Thus if the atmosphere lacks aerosols and dust in an area, you could have super cooled water vapor.

  24. Dave Dardinger
    Posted Sep 18, 2006 at 12:44 PM | Permalink

    re: 23 Are there studies which actually looked for degree of super-cooling in an area? And there is a limit as to how much supercooling can occur, if I recall. Finally there’s the cloud-chamber effect with suggestions that this is a way that the connection between sunspot number and temperature could be amplified.

    Also, doesn’t a lot of sea salt get caught up in updrafts over oceans so that they carry their own seeds with them?

  25. Dano
    Posted Sep 18, 2006 at 1:41 PM | Permalink

    23:

    It’s a rare parcel indeed at or near the lifted condensation level that doesn’t have a cloud condensation nucleus around. The atmosphere without a lithometeor? Nah.

    22:

    Parameters start.

    20:

    the eco-theologian alarmists within the IPCC

    You may want to be careful. Many here harrumph and dudgeon when they perceive something to be ad hom. Just a thought.

    Best,

    D

  26. Dave Dardinger
    Posted Sep 18, 2006 at 2:20 PM | Permalink

    re: #25 Dano,

    Thanks for the link, which I’m bookmarking for future use. However, it would be nice if you’d actually look up what was asked for rather than (apparently) just assuming it’s there. Here’s the link for cloud formation (about halfway down the page) which says:

    Cloud Formation
    Stratiform cloud formation is based on the relative humidity diagnostic form of Slingo (1987) [26]. Clouds are of 3 height classes: high (sigma levels 0.189-0.336), middle (sigma levels 0.500-0.664), and low (sigma levels 0.811-0.926). The fractional amount of each type of cloud is determined from a quadratic function of the difference between the maximum relative humidity of the sigma layer and a threshold relative humidity that varies with sigma level; for high and low cloud the threshold is 60 percent humidity, while for middle cloud it is 50 percent. In addition, following Rikus (1991) [8], low cloud forms when the relative humidity at the lowest atmospheric level (sigma = 0.991) exceeds 60 percent, and is capped by strong static stability in the layer immediately above (i.e., a temperature inversion is present). In this case, the amount of low cloud increases with the strength of the inversion. (See Convection for the treatment of convective cloud and Radiation for cloud-radiative interactions.)

    Anyone have any comments on whether this seems a sensible way to parameterize cloud formation? And what atmospheric activities would not be taken into consideration by such a method?

    Obviously this description is pretty sketchy and practically every word could elicit a question like, “Why 60%?” or “What’s the definition of ‘strong’ static stability?” But I’m just trying to get an overview at this point so I’m not worried by the lack of detail.

  27. TCO
    Posted Sep 18, 2006 at 2:37 PM | Permalink

    Steve: #3.

    Judy (#7): Given the uncertainties and paramaterizations, how useful are reconstructions that use climate models for some element of validation? Note, I’m not saying people shouldn’t work on hard, intractable problems (maybe we’ll get fusion licked one day). Just wondering how much of a reed these things are to hang anything on. For me, the simple coincidence of temp rise and CO2 rise has always been the simplest and most relevant gut rationale for AGW as a phenomenon. The reconstructions and models seem very strained…(as a fundamental physical scientist…but one who has done complicated business analysis in uncertain situations.)

  28. eduardo zorita
    Posted Sep 18, 2006 at 3:23 PM | Permalink

    #9

    Steve,

    my guess is that such simplification would not work due to the non-linearities of the system. If you start with a three-dimensional equations discretized on the sphere and subsequently consider zonal, meridional and hight averages you will probably found terms that represent the averages of correlations of variables, e.g. T and,U,V. These averaged correlations would have to be parametrized in terms of the averages of the variables themselves. For instance, the meridional heat transport, caused in part by the correlations between T and V, would have to be represented as functions of the meridional T gradient, and these function is not known. I guess this is the same as the closure problem in turbulence.

    eduardo

  29. John Creighton
    Posted Sep 18, 2006 at 6:39 PM | Permalink

    Hi Dano,
    I was hoping the link you gave would give a direct disruption of the parameters used. Anyway, I think the observation of the correlation of low level clouds with sun spot number is an indication that aerosols effect the condensation rate. Perhaps there is sufficient CCN in the air to keep the water vapor from becoming super cooled but shouldn’t rain deplete the air of CCN. Does the sun replace CCN fast enough to make up for what is depleted by the rain. I think as rain falls the droplet size must increase. I am curious about when the rain starts to fall. It must depend on the air currents. I would think the droplet velocity should be different then the air velocity while before it starts to rain. I also think that there would be a distribution of droplet velocities and sizes.

  30. Jeff Weffer
    Posted Sep 18, 2006 at 6:59 PM | Permalink

    I like Willis’ proposition that water vapour and clouds provide a balancing effect to the climate. The Earth warms (or let’s say even a small local region warms) and water vapour and clouds serve to balance that heat out. The Earth cools, less water vapour and clouds, and the Earth warms.

    The best evidence that this occurs is (your own personal observation of a warm day of course) but mainly the fact that Earth’s climate has been remarkably stable over the last 1 billion years at least.

    There has primarily been warm periods where the Earth’s average temperature was about 20C. Then there have been cooler periods of ice ages which seem to be related primarily to the position of the land masses on Earth.

    The more the continents are weighted toward the poles, the cooler Earth becomes. There have been several Snowball Earth periods which are directly tied to the continents being locked up together over or close to the south pole. Glaciers build up at the poles, spread out over the land and more sunlight is reflected back into space. The glaciers get bigger and more sunlight is reflected and so on.

    At least twice, it appears the entire Earth froze over, even the oceans. Without significant land masses near the poles, the glaciers and polar ice can’t spread far from the poles. When continental drift moved enough of the land masses away from the poles, the Earth warmed back up again.

    When the land masses are weighted away from the poles, the climate of the Earth seems to settle into that 20C range.

    The ice ages of the last 3.5 million years seem to be directly related to the movement of North America and Eurasia just slightly north, just slightly north enough to give us periodic ice ages related to the Earth’s orbit and the sun’s cycles.

    The planet has had a remarkably stable climate despite there being significantly different levels of CO2 and O2 in the atmosphere over time. Other than the ocassional time period when we had Antarctica times 10 over the south pole or an asteroid impact, it has been quite pleasant for a hairless ape in the Earth’s climate history. There must be a very strong balancing factor and that would be water.

  31. Paul Linsay
    Posted Sep 18, 2006 at 7:15 PM | Permalink

    #29

    Anyway, I think the observation of the correlation of low level clouds with sun spot number is an indication that aerosols effect the condensation rate.

    Actually this is one of the prime pieces of evidence for the Svensmark effect, see Figure 4, that cosmic rays seed clouds and are part of the solar influence on climate. Why should aerosols, which are of terrestrial origin, be correlated with the solar cycle?

  32. Judith Curry
    Posted Sep 18, 2006 at 8:18 PM | Permalink

    will be back tues nite, will try to address some of these very interesting issues.

  33. TCO
    Posted Sep 18, 2006 at 8:21 PM | Permalink

    28. Maybe it won’t work in the end, but some good insights would be gathered to help the problem-solving. I’m concerned that the huge complexity of the problem, combined with all the fudge factors (I mean paramaters) and the politization of the field (as well as the less then Sherlock Holmes intelect of atmosphere phycisists) is a prescription for screw ups.

  34. Posted Sep 19, 2006 at 6:03 AM | Permalink

    Re #31:

    The link between GCR (galactic cosmic rays) is quite dubious in recent years at one side (but there is -again- discussion about the accuracy of satellite measurements…), see Fig. 1 in Kristjansson, but a recent article again links GCR and clouds, by looking at cloud formation in a part of the globe where there is a low anomaly of the earth´s magnetic field. Clouds increase from the outside to the inner part of the area. See Vieira and Alves da Silva

    Nevertheless, there is a significant (inverse) correlation between solar irradiation at the top of the atmosphere (TOA) and low cloud cover, see again the same Fig.1 of Kristjansson. How that works, is quite difficult to know. There are sun induced observed cycles in the stratosphere for ozone, temperature and jet stream position. The latter influences cloudiness (and rainpatterns) towards the poles. Or it could be that increased vis/NIR absorption by water vapor/drops reduces/vaporises drop counts and thus cloudines…

    More reflecting aerosols seems to increase cloudiness (more nuclei – smaller drops and increased lifetime of the clouds), while absorbing aerosols (heated by sunlight?) seems to decrease cloudiness. See Kaufman and Koren. Their calculation that aerosols cause a 5% increase in cloudiness is IMHO a little (?) overblown…

  35. Dave Dardinger
    Posted Sep 19, 2006 at 8:22 AM | Permalink

    I’ve been reading in a book I had sitting around on Weather and it’s interesting looking at all the things entering into cloud formation. For instance one common situation is warm humid air meeting a cold air mass. The warm air rides up over the cold air as it’s lighter and that makes clouds form in a line. But whether it’s the cold air moving toward the warm air or vice versa makes a difference in structure of the resulting clouds and precipitation. So I’m wondering how well models can track this sort of thing. I’m sure the average can be figured out an plugged in, but there may be structual reasons why an uncommon sort of storm / cloudiness situation would prevail in part of a grid cell and this could have long-term affects over an entire region. Just allowing, for instance, for the average elevation of a grid cell wouldn’t let you know if it consisted of one large valley area and mountains on either side, a plain and a plateau, a mountain ridge to the west and a basin, a jumble of terrain, etc.

    Continuing on my questions, 2. Are there parameters in GCMs which account for these things?

  36. Richard deSousa
    Posted Sep 19, 2006 at 10:16 AM | Permalink

    I have a question regarding the Physics World article I posted at the beginning of this blog. At the end of paragraph 7, page 3, the authors say “Crude calculations suggest that the two effects approximately balance each other, and that water vapour does not have a strong feedback mechanism in the Earth’s climate.” It seems to me that this is Lindzen’s Iris at work so there should be a negative feedback.

  37. Richard deSousa
    Posted Sep 19, 2006 at 10:19 AM | Permalink

    Oops! I should have also included this previous sentence “Water vapour in the atmosphere can change phase, which leads to more clouds, and greater cloud cover means that more sunlight is reflected straight out of the atmosphere.”

  38. Steve McIntyre
    Posted Sep 19, 2006 at 10:20 AM | Permalink

    #36. Lindzen’s iris effect is one particular effect which is controversial and quite a different issue from NIR absorption by water vapor.

  39. Richard deSousa
    Posted Sep 19, 2006 at 10:33 AM | Permalink

    I reckon you’re referring to inbound radiation being absorbed in the NIR by the atmosphere which would change the results of the GCMs.

  40. Paul Linsay
    Posted Sep 19, 2006 at 4:09 PM | Permalink

    #34: With regard to Kristjansson, it would be nice to see the cloud cover data before all the corrections. I also can’t understand why they say that there is strong low cloud correlation with solar irradiance but not cosmic rays when it is quite clear just by looking at Figure 1 that CRs and SI are very strongly correlated. Flip one over and they are nearly identical. Do you have an explanation?

  41. Murray Duffin
    Posted Sep 19, 2006 at 5:09 PM | Permalink

    #34, #40, There has been a lot of discussion since Kristjansson 2002 thAT STILL LEFT Marsh and Svensmark alive and well in early 2005 when I stopped following it. Has anyone continued to follow in 2005/6? Murray

  42. Joel
    Posted Sep 19, 2006 at 7:30 PM | Permalink

    Re #30: Jeff, your conclusions from looking at past climates differ rather drastically from the conclusions of those who work in that field…and who have looked to see how the climate models deal with simulations of past climates. See, in particular, this paper: D.P. Schrag and R.B. Alley, Science 306, 821-822 (Oct. 29, 2004) [http://www.sciencemag.org/cgi/content/summary/306/5697/821]. Here is the summary of the paper: “Climate models and efforts to explain global temperature changes over the past century suggest that the average global temperature will rise by between 1.5º and 4.5ºC if the atmospheric CO2 concentration doubles. In their Perspective, Schrag and Alley look at records of past climate change, from the last ice age to millions of years ago, to determine whether this climate sensitivity is realistic. They conclude that the climate system is very sensitive to small perturbations and that the climate sensitivity may be even higher than suggested by models.”

    Also, while admittedly there are still considerable uncertainties regarding clouds, I think that there is now some decent evidence that the climate models are handling water vapor at least roughly correctly. See, for instance, B.J. Soden et al., Science 310, 841-844 (October 6, 2005) [http://www.sciencemag.org/cgi/content/abstract/310/5749/841] and B.J. Soden et al., Science 296, 727-730 (April 26, 2002) [http://www.sciencemag.org/cgi/content/abstract/296/5568/727].

  43. beng
    Posted Sep 19, 2006 at 7:33 PM | Permalink

    RE 12:
    Condensation indeed happens, but I am not so confident that tells the whole story. I wonder if something puts “the brakes on” and slows radiation loss substantially during the high humidity of late night.

    Something does if radiational ground fog forms — depending on density, it optically reduces/stops IR to space from solid or liquid surfaces beneath it. Same effect as low clouds. I note that, at least in summer here, dropping to the dewpoint isn’t necessary to produce an increasing low-level “haze” during the nite, producing the same, if lesser, effect. I can gauge that from the visual “transparency” of the nite-sky (my area is very dark) under different temps & dew points.

    The formation of radiational-cooling ground fog or haze is a good example of a negative feedback effect on (surface) temperatures.

  44. Judith Curry
    Posted Sep 21, 2006 at 4:50 AM | Permalink

    Re #15 Ferdinand, I have a hard time drawing any conclusions about what is going on in the climate system from the broadband radiative flux measurements at the top of the atmosphere (TOA). The error bars seem to be the same size as the effect we are trying to measure. Even if the measurements are accurate, it is difficult to know how to interpret them in terms of what is going on in the climate system. The linking of TOA radiation fluxes to things like heat storage in the ocean seems to adopt a 1D radiative convective model of the climate system, which is WAY overly simplistic. There is a proposal on the table by Richard Goody and Jim Anderson to start measuring spectral radiative fluxes at TOA; looking at the spectral signal would help ALOT in terms of actually what is going on below the TOA.

  45. Judith Curry
    Posted Sep 21, 2006 at 5:00 AM | Permalink

    Re #26 Dave, you describe accurately the “voodoo” of how the subgridscale aspect of cloud parameterization is handled in global climate models. Sophisicated microphysical parameterizations like nucleation of cloud and ice particles from aerosols, how this forms rain, etc. are hampered in model with low resolution (vertical and/or horizontal); we can’t assume that clouds form at relative humidity around 100%, since the model grid cell relative humidity (say in a box 200 km in horizontal dimension) would never come close to that value. So people who parameterize cloud formation then need to come up with some sort of statistical relationship between the cloud properties and the resolved variables (this “tuning” needs to be redone every time the model resolution is changed). Some more sophisticated methods are starting to be used whereby the subgrid cloud parameterizations are tied to a subgrid distribution of stochastical fluctutions (which are then related to grid resolved variables, one needs to decide whether a gaussian or skewed distribution is appropriate, etc.)

    Does all this mean that climate models are unreliable, given the obvious importanice of clouds to the radiation balance and to climate? Maybe not (see my next post, i have a week internet connection, making the posts short).

  46. Judith Curry
    Posted Sep 21, 2006 at 5:38 AM | Permalink

    Re #27 and #30: TCO and Jeff,

    Jeff (and the post by willis that i couldn’t find?) raises a key issue about water and the terrestrial water cycle:they act to regulate the Earth’s climate. The Earth’s climate sits pretty close to the triple point of water (273K), so with its abundant water there is lots of opportunity for the water to change phase and to so change the planetary radiation balance. I think that clouds do act as a regulator/stabilizer overall on the Earth’s climate (some sort of negative feedbacks almost certainly must be associated with clouds).

    Does the fact that our treatment of cloud processes in climate models (which are inherently subgridscale) is rather crude imply that the climate model simulations are not reliable? Maybe not. Going back to my very ad hoc dynamical systems analysis that i started discussing in my original post to this thread, if you have a dynamical system with a very large number of degrees of freedom, it seems likely that some internal adjustments will take place if one of the degrees of freedom is not handled accurately. The substantial variations in the time/space distribution of clouds simulated by climate models reflects slight differences in the simulated dynamical thermodynamical fields. The dynamical and thermodynamical fields then adjust slightly to account for any changes to surface conditions or atmospheric thermodynamics caused by clouds

    The bottom line is that i don’t think cloud feedbacks are evaluated correctly in climate models. People only have the baseline (current) and perturbed (double CO2) runs to evaluate. You can’t really do the job of evaluating feedbacks just from these simulations.

    So I think that clouds act as regulators. In my opinion the key issue is to have the appropriate numbers and types of degrees of freedom in your cloud parameterization, rather than to have a perfect parameterization for a small number of degrees of freedom. For example, adding the degrees of freedom whereby cloud condensation and cloud freezing interact with atmospheric aerosols is essential, even if the treatment isn’t perfect. This brings us back to not understanding the role of phase boundaries in chaotic dynamical systems (or how to model them, or whether we should worry about how we model them).

    Then if you look at the broader issues of the hydrological cyclne (like surface snow and ice), you also see some regulatory mechanisms at work. In a warmer climate you get more snowfall which adds to the glacier mass and then increases the surface albedo

    Does this mean the hydrological cycle will keep the earth’s climate stable even in the face of doubling or tripling CO2? It will certainly help keep the climate relatively stable (i.e. more stable than if we didn’t have water on the planet). The earth’s climate has been much more stable than mars and certainly venus over the planetary history. But “stable” is a relative thing; you can consider it from the perspective of planetary history, or from the perspective of human history. Events or regimes that may seem relatively stable from a planetary history perspective would seem catastrophic from a human history perspective. The effects of doubling or tripling CO2 (with the most likely projection 3C warming for the doubling according to the models), would arguably lie in the catastrophic category in terms of human history (if not in terms of planetary history).

  47. Paul Linsay
    Posted Sep 21, 2006 at 6:58 AM | Permalink

    #46

    The bottom line is that i don’t think cloud feedbacks are evaluated correctly in climate models. People only have the baseline (current) and perturbed (double CO2) runs to evaluate.

    I take it from this statement that the models don’t ramp up CO2 gradually as is actually happening? If that’s so, it would be a major flaw since the atmospheric adjustment would likely be very different between incremental CO2 additions versus a CO2 bomb.

  48. TCO
    Posted Sep 21, 2006 at 7:50 AM | Permalink

    Judy: I have no problem with people working on the models even if there are all kinds of things wrong with them. I just wonder if, given all their issues, whether they are really any kind of independant support for AGW. (Note, I think AGW is likely happening.) I also wonder about various people who are doing other derivative studies in statistics, reconstructions, etc. and are using the models for smaller issue inferences. I wonder whether that is a week reed to tie in to experimental work.

  49. Posted Sep 21, 2006 at 11:39 AM | Permalink

    Thanks Judith for your thoughts in #44. It seems to me that the adjustment of the LW (and SW) radiation measurements at the TOA will have an impact on the conclusions of a lot of papers published in past years. The proposal to measure TOA fluxes over the (full?) spectrum is very interesting. I have asked something like that about downgoing LW radiation, as reaction on the paper by Philipona, to know what was really responsible for the “dimming” (water vapor, aerosols, GHGs,…)

  50. Steve Sadlov
    Posted Sep 21, 2006 at 1:35 PM | Permalink

    Piekle Sr. dropped the following bomb today:

    http://climatesci.atmos.colostate.edu/2006/09/21/a-presentation-by-v-ramanathan-at-the-sorce-meeting-entitled/

    This is amazing.

    This probably warrants its own thread.

  51. TCO
    Posted Sep 21, 2006 at 1:45 PM | Permalink

    Am I missing something? Don’t see a peer-reviewed paper. Don’t see a bomb. Not sure what specifically on that site that I should look at.

  52. Steve McIntyre
    Posted Sep 21, 2006 at 2:33 PM | Permalink

    I’ll probably put up a thread. Ramanathan is an interesting author. I once got intrigued a couple of years ago at where the figure 4 wm-2 came from. It’s not easy to find a primary calculation of this figure. IPCC uses the figure but their sources are all dead ends. I tracked it back to some publications by Ramanathan in the late 1970s by stumbling across a publication in 1983 which linked back. In my opinion, Ramanathan’s presentations are about the only presentations of the AGW effect from first principles in anything other than trivial terms. If anyone can provide more recent articles of similar scope, I’d love to see them. If you googl;e Ramanathan at the online Am Met Soc site, you can locate most of the articles.

  53. David Smith
    Posted Sep 24, 2006 at 4:53 PM | Permalink

    Here’s a question for anyone familiar with global warming theory and the climate models:

    I’ve been looking at historical upper-air charts, covering the globe, stretching back to about 1950. (These are mostly balloon-derived, I believe.)

    What the charts show is that, above about a kilometer (higher than 925mb), the air has been drying for 50 years. This is especially evident at higher altitudes.

    Now, below a kilometer, it has gotten more humid. On balance, the overall atmosphere has tended to dry (as measured by precipitable water) over the last 50 years, regardless of ground temperature trends.

    That surprises me. I thought that water vapor levels would rise throughout the atmosphere as the earth warms.

    It also makes me wonder about changes in cloud cover. If the atmosphere is drier, then are we having fewer clouds?

  54. David Smith
    Posted Sep 25, 2006 at 2:57 PM | Permalink

    Re: trends in upper atmosphere humidity

    Linked below (I hope) is a chart of the water content of the worldwide atmosphere at roughly 9 kilometers up. This would be considered the middle troposphere. The chart covers about the last 50 years.

    The trend in water content (specific humidity) is clearly downward.

    Similar decreases in water content are seen at other altitudes above two kilometers while, near the surface, the water content of the air has increased.

    To me, this seems significant. It seems like it affects the ability of the mid and upper troposphere to radiate away IR, it seems like it affects cloud cover, it seems like it says something about changes in atmospheric mixing.

    Judith, Peter H or others, have you seen anything that explains why this water content decrease is happening, and whether it is a feature of global warming?

    Thanks

    link

  55. Dave Dardinger
    Posted Sep 25, 2006 at 3:29 PM | Permalink

    re: #54

    Let’s go through this with an eye toward causes.
    1. Humans produce CO2 by burning fossil fuels
    2. This produces warming of lower atmosphere; about 1 deg C for a doubling of CO2.
    3. This warmer atmosphere must be offset by a colder upper atmosphere; otherwise more heat would go out than comes in, an impossiblity.
    4. This cooler stratisphere / upper troposphere propagates down until we have balance.
    5. Because the upper troposphere is cooler, it can hold less water vapor and this causes a lower specific humidity over time.
    6. This lower humidity causes the slightly lower parts of the troposphere to radiate a bit more to space. Net result, only the lowest parts of the troposphere will warm while upper parts will stay the same or cool.
    7. Over oceans (i.e. the SH), the warmer atmosphere will cause more evaporation and thus a larger amount of convective action, resuting in not much net temperature rise.
    8. Over land (i.e. the NH), there won’t be much extra water to evaporate meaning that the warmer troposphere will result in a warmer surface as well.
    9. Final conclusions, what we see is what we get. Temperatures over land will continue to rise a bit while over water there won’t be much rise. Eventually I’d say the extra evaporation over water will stop the surface from rising any more and a new equilibrium not much higher than the present will occur.

  56. David Archibald
    Posted Sep 25, 2006 at 3:34 PM | Permalink

    Re 54, that is a very interesting graph. The peaks correlate with the solar cycle peaks, and the 1998 el nino. The 80 ppm fall in water content is roughly equivalent to the 80 ppm rise in carbon dioxide level over the same period.

  57. Douglas Hoyt
    Posted Sep 25, 2006 at 3:59 PM | Permalink

    According to the climate models with their assumption of constant relative humidity, an 80 ppm increase in carbon dioxide will be accompanied by about an 160 ppm increase in water vapor, globally averaged. I deduced this number years ago from looking at some plots of GCM outputs expressed in ppm increase in water vapor, so it is only approximate. It would be nice if the modelers would explicitly state the number so people could get an idea of what is going on in the models.

    In any case, a decrease in specific humidity is inconsistent with GCMs.

  58. bender
    Posted Sep 25, 2006 at 4:01 PM | Permalink

    Re #53

    are we having fewer clouds?

    I think I’ve seen this question asked before in the literature, with the answer that we don’t have enough good quality data to say.

  59. bender
    Posted Sep 25, 2006 at 4:05 PM | Permalink

    Would be interesting, therefore to know if the graph in #54 is correlated with cloud over time periods where we do have good data.

  60. David Smith
    Posted Sep 25, 2006 at 4:37 PM | Permalink

    Here’s a plot of global precipitable water (PW). PW indicates the water content of the atmosphere, g water per kg atmosphere. Not much of a pattern over the years.

    Humidity near the ground has increased while the upper 80% of the atmosphere has become drier over the last 50 years. Interesting.

  61. David Smith
    Posted Sep 25, 2006 at 4:38 PM | Permalink

    Oops, here’s the link

    link

  62. David Smith
    Posted Sep 25, 2006 at 5:03 PM | Permalink

    One final plot, for fun.

    This one is the temperature of the tropopause (the top of the troposphere).

    Note the dramatic change circa 1976. (This shows up in plots of nearby atmospheric levels, too, so it’s “real” and not due to some definition change.)

    Note that 1976 was also the time that the earth’s atmosphere began its recent “unprecedented” temperature rise, especially in the Arctic regions.

    It’s as if something very big happened to the atmosphere. I’ve heard it attributed to one of the Pacific oscillations, but that explanation was never clear to me.

    It also makes me wonder whether the cooling of the lower stratosphere, attributed to AGW and CFCs, may in part just be a recovery from the Big Event of 1976.

    I don’t know. I haven’t a clue. My only point is that, until we fully understand things like this, it’s gotta be hard to reliably model the atmosphere.

    tropopause temperature trend

  63. David Smith
    Posted Sep 25, 2006 at 5:18 PM | Permalink

    Uh, I guess I told a fib – here’s the final plot I’ll post.

    This one is of the temperature of the upper stratosphere (10mb level), where 99% of the atmosphere lies below it.

    Notice the dramatic shock in temperature in 1976, which coincides with the change of direction from global cooling to global warming.

    Seems reasonable that, at that height, the main thing that affects the atmosphere’s temperature is the sun and particle-physics stuff. That is my guess.

    I wonder how this ties in with aerosol theory.

    I have no notions on this, other than that nature is something to behold.

  64. David Smith
    Posted Sep 25, 2006 at 5:20 PM | Permalink

    link for #63

  65. bender
    Posted Sep 25, 2006 at 5:54 PM | Permalink

    But this is not inconsistent with Warmer thinking on aerosols – that the HS blade would have started up in the 50s if it hadn’t been for the unexpected cooling effect of aerosols in the 60s and 70s. Right?

  66. Paul Linsay
    Posted Sep 25, 2006 at 6:26 PM | Permalink

    #65, No need for aerosols in the 50s. If you look at solar activity and temperature for the last 150 years they correlate nicely including falling from the 40s to the 70s. link

  67. Steve Sadlov
    Posted Sep 26, 2006 at 12:03 PM | Permalink

    RE: # 60 – “Humidity near the ground has increased while the upper 80% of the atmosphere has become drier over the last 50 years.”

    Juxtapose this against the explosion on irrigation (both the agriculture and landscape varieties). Contemporaneously, there has been an explosion in air conditioning, especially in the developing world. Meanwhile, with forests and other wild lands having less extent, and paving going from being an urban luxury to being more or less the norm, I’d have to imagine that overall evapotranpiration on the global scale has gone down. Then the question must be asked, what of bulk water temperatures of all natural waters? Have these perhaps actually gone down overall over the past 50 years? And the final question as already asked above, more or less vertically developed clouds?

  68. David Archibald
    Posted Oct 1, 2006 at 8:19 AM | Permalink

    Re #54, the upper atmosphere tends to fractionate according to molecular mass. CO2 is 1.5 times heavier than air. If the upper atmosphere is fully saturated with water all the time, then the addition of another heavy molecule might displace water. This might explain why the fall in humidity is offset by a rise in CO2 level by the same amount. If true, then there is no positive feedback mechanism from rising CO2 levels for a large part of the atmosphere.

  69. TCO
    Posted Oct 1, 2006 at 9:14 AM | Permalink

    If the upper atmosphere is fully saturated with water all the time, then the addition of another heavy molecule might displace water.

    This hypothesis should be proveable by by basic thermo. I can’t remember what happens with partial pressures in a gravity field, but my intuition is that there is some logical flaw. That the C02 and H20 as both being trace gasses at low pressure will have no effect on each other.

  70. Willis Eschenbach
    Posted Dec 16, 2006 at 12:09 AM | Permalink

    There’s a fascinating new paper at arXiv that is of great interest:

    The Antarctic climate anomaly and galactic cosmic rays

    Henrik Svensmark

    Center for Sun Climate Research, Danish National Space Center,
    Juliane Marie Vej 30, 2100 Copenhagen àƒËœ, Denmark

    ABSTRACT

    It has been proposed that galactic cosmic rays may influence the Earth’s climate by affecting
    cloud formation. If changes in cloudiness play a part in climate change, their effect changes sign
    in Antarctica. Satellite data from the Earth Radiation Budget Experiment (ERBE) are here used
    to calculate the changes in surface temperatures at all latitudes, due to small percentage changes
    in cloudiness. The results match the observed contrasts in temperature changes, globally and in
    Antarctica. Evidently clouds do not just respond passively to climate changes but take an active
    part in the forcing, in accordance with changes in the solar magnetic field that vary the cosmic-ray
    flux.

  71. maksimovich
    Posted Dec 16, 2006 at 3:10 AM | Permalink

    re 70

    “‚⠠ High energy particles, such as galactic cosmic rays and solar protons are modulated by the 11-year solar cycle and by short-term (hours to days) solar changes. In some cases, high energy particles penetrate to the surface, which presents the possibility of atmospheric radiative changes linked to the particlesàƒ’€š’ passage through the stratosphere and troposphere. Changes in atmospheric ionisation at the surface can be caused by substantial solar changes.

    “‚⠠ Aerosol microphysics, such as particle nucleation, coagulation, and scavenging can be expected to be modified through changes in cosmic ray ion production.

    “‚⠠ There is detailed theoretical support for ultrafine aerosol production in the atmosphere from ions, in the regions where there is condensable vapours with no substantial competing vapour sinks. Large cluster ions, which are expected in the initial phase of particle formation, have also been detected in the free troposphere. The combination of theoretical prediction and supportive (but preliminary) experimental atmospheric results now gives reasonable confidence that particle formation occurs from ions in the troposphere, and therefore from cosmic rays. The geographical distribution of particle formation and its frequency is not known with any confidence.

    “‚⠠ Charging of aerosol particles and droplets occurs on the particle and cloud boundaries, related to global circuit properties. Theory supports the enhanced removal of charged particles to cloud droplets, even at low particle charge levels. There is some basis for an effect of charged aerosol on freezing of super-cooled clouds through electro-scavenging, but this has not been experimentally demonstrated in the atmosphere.

    “‚⠠ Currents flow in the global atmospheric electrical system as a result of the partial electrical conductivity of air, caused mainly by cosmic ray ionisation. The global atmospheric electrical circuit drives vertical electric currents in fair weather regions.It provides a global teleconnection, communicating atmospheric electrical changes throughout the stratosphere and troposphere, down to the surface.

    “‚⠠ Cloud-retrievals from the ISCCP satellite program show a strong correlation between low liquid water clouds with galactic cosmic rays from July 1983 to September 1994. Without detailed model calculations combining the aerosol and cloud microphysics and estimating the effects, it is not possible to be sure whether the correlation results from a direct cosmic ray effect on low clouds, or has another origin. If a calibration correction to the data is accepted, on the basis of the absence of a polar orbiting satellite, the correlation continues until September 2001; there is some recent evidence that the correlation is stronger at high latitudes than low latitudes.

    “‚⠠ Cosmic rays and Total Solar Irradiance variations are often closely correlated. This relationship has been used to infer past TSI changes. However it also means that cosmic ray and TSI signals may be ambiguous on some timescales: which has implications for the use of cosmogenic isotopes as proxies. The proxies actually measure the cosmic ray variations. Consequently if cosmic-ray induced aerosol microphysics changes couples strongly through to clouds, climate signals attributed through proxies to solar TSI changes could also be explained by a direct atmospheric effect of cosmic rays.

    A Review of The Influence of Solar Changes
    on the Earthàƒ’€š’s Climate
    L. J. Gray, J. D. Haigh, R. G. Harrison

  72. bender
    Posted Dec 16, 2006 at 3:25 AM | Permalink

    Any warmers out there want to revise their prior estimate of A in AGW in light of this new information? Every Good Bayesian Devours Fudge.

  73. Steve Bloom
    Posted Dec 16, 2006 at 2:57 PM | Permalink

    Re #72: bender, a strict Bayesian, or anyone even faintly connected to reality, would bear in mind that much of Svensmark’s past work on this subject has been, um, open to question.

    At a quick read I see that Svensmark is a fan of strong water vapor feedback. I thought you guys didn’t like that. Also, if I recall correctly Svensmark’s work is extremely dependent on the 10Be proxy record, which I believe has been thrown into question by a recent paper.

    Nobody questions, I think, that there is some kind of effect here. The question is whether it is a significant driver of large-scale clinate trends. The fundamental problem I have with it is that a glance at the climate history of the last 50,000,000 years or so makes it hard to imagine that cosmic rays are the main thing going on.

  74. bender
    Posted Dec 16, 2006 at 3:15 PM | Permalink

    Re #72/73 Ok, Bloom’s a “no”. That’s, um, surprising.

  75. Willis Eschenbach
    Posted Dec 16, 2006 at 9:22 PM | Permalink

    Re #73, Steve, I don’t know many people who don’t think that there isn’t any positive water vapor feedback due to increased IR. The unanswered question is not that. It is, what is the strength and the sign of the overall climate feedback to increasing temperature.

    AGW adherents think the sign is positive. Me, I doubt that very much, because if it were, the earth would exhibit much larger temperature swings than it has.

    In this connection, it is helpful to visualize what would happen to a watery planet if the sun’s power started at zero, and then began to rise. Initially, there would be no liquid water, and very little water vapor in the atmosphere. At some point the ice would melt, and water vapor would rise. At some point, clouds would form, and the amount of sunlight reaching the surface would begin to decline, cooling the surface. But increased IR absorption by increaced water vapor would tend to heat the surface. On the other hand, the various parasitic water vapor related losses (evaporation, transpiration, vertical transport in thunderstorms, hydrometeors, etc.) would also rise, cooling the surface. It is important to distinguish between these two processes, the cloud feedback and parasitic losses on the one hand cooling the earth, and the increased water vapor IR warming the earth.

    At any given stage in the sun’s warming, the earth would have an equilibrium temperature. This temperature would be less than the theoretical temperature, and would be determined by the relative balance of the increasing parasitic losses and increasing cloud reflection cooling the earth, and the increasing greenhouse effect of the water vapor warming the earth. For any given sun strength, the earth will not warm beyond a certain point, the point at which the parasitic losses (which are driven by ‘ˆ’€ T) and the cloud albedo reflections (driven by evaporation) matches the greenhouse gains. In other words, the heat engine of the climate is always running as hot as it can. If the earth heats a bit, parasitic and albedo losses increase and reduce the temperature. If the earth cools a bit, losses decrease, and the earth warms up.

    The fact that the earth’s temperature is limited, not by the strength of the sun, but by the ‘ˆ’€ T dependent losses, is a crucial point which is often not accounted for.

    Now in the midst of all of this, what will be the overall effect of a slight increase in CO2 forcing? About the only thing we can say for sure is that it will be less than it would be if there were no parasitic losses. This is because the net balance between water vapor feedback and parasitic losses increase with increasing temperature. It has to, or the temperature would increase until it does.

    Thus, the general assumption that the water vapor feedback will increase the effect of a CO2 change cannot be true. It is true that the direct water vapor feedback is positive, but the overall effect must be negative.

    Now, how do cosmic rays fit into all of this? In general, it seems that they increase cloudiness, which should cool the earth. But as Svensmark’s latest paper seems to indicate, this may cause different effects at different latitudes. At present, I don’t think that we have enough information to make any definitive statements about that, other than the usual mantra … there’s a lot about the climate that we don’t understand.

    w.

  76. Dave Dardinger
    Posted Dec 17, 2006 at 9:06 AM | Permalink

    re: #75 Willis,

    You say,

    In other words, the heat engine of the climate is always running as hot as it can.

    But isn’t it actually that the climate is always running as cool as it can? That would seem to me to be the meaning of the maximum entropy theory, which has been discussed here before. We know by definition delta S = delta q/T. So to maximize S you need to minimize T. That is, the heat energy must be produced at as low a temperature as possible. Of course in the case of radiation it’s a bit tricky since heat lost to space isn’t part of the delta q in the equation. So if something like CO2 blocks part of this q(r), then other things being equal T would increase, decreasing delta S which nature doesn’t like. And, of course, the actual total delta q can’t change since it’s fixed by solar radiation incoming.

  77. jae
    Posted Dec 17, 2006 at 10:34 AM | Permalink

    70: Yeah, that all backs up Veizer’s paper.

  78. bender
    Posted Dec 17, 2006 at 10:52 AM | Permalink

    much of Svensmark’s past work on this subject has been, um, open to question

    1. That’s, um, poisoning the well. What precisely is, um, wrong with the paper Wilis has cited?
    2. Are the, um, multiproxy reconstructions and GCMs NOT open to question? Thought so.

  79. Brooks Hurd
    Posted Dec 17, 2006 at 1:25 PM | Permalink

    Re: 73,

    Steve B,

    Everyone’s work is open to question. That is part of the scientific method.

    Several articles have thrown Mann’s work into question. Does this mean that you reject everything that he has done?

  80. Willis Eschenbach
    Posted Dec 17, 2006 at 7:11 PM | Permalink

    Dave D, you raise a fascinating point when you say:

    re: #75 Willis,

    You say,

    In other words, the heat engine of the climate is always running as hot as it can.

    But isn’t it actually that the climate is always running as cool as it can? That would seem to me to be the meaning of the maximum entropy theory, which has been discussed here before. We know by definition delta S = delta q/T. So to maximize S you need to minimize T. That is, the heat energy must be produced at as low a temperature as possible. Of course in the case of radiation it’s a bit tricky since heat lost to space isn’t part of the delta q in the equation. So if something like CO2 blocks part of this q( r ), then other things being equal T would increase, decreasing delta S which nature doesn’t like. And, of course, the actual total delta q can’t change since it’s fixed by solar radiation incoming.

    Actually, both are true, although “minimizing entropy” is not the fullest description of the process. A better description is the constructal law. One description of this says that a flow system acts to maximize the interface between the parts of the system. The author’s description of constructal theory is here. The constructal law shows how the system physically restructures to minimize entropy.

    The system, then, is running at the coolest temperature at which it can dissipate all of the energy entering the system. This is also as hot as it can run, given the constraints of the physical laws governing the system.

    w.

  81. Willis Eschenbach
    Posted Dec 17, 2006 at 7:20 PM | Permalink

    Well, looks like you can’t use a blockquote inside a blockquote, it gets confused. Here’s what it should have said:

    Dave D, you raise a fascinating point when you say:

    re: #75 Willis,

    You say, “In other words, the heat engine of the climate is always running as hot as it can.”

    But isn’t it actually that the climate is always running as cool as it can? That would seem to me to be the meaning of the maximum entropy theory, which has been discussed here before. We know by definition delta S = delta q/T. So to maximize S you need to minimize T. That is, the heat energy must be produced at as low a temperature as possible. Of course in the case of radiation it’s a bit tricky since heat lost to space isn’t part of the delta q in the equation. So if something like CO2 blocks part of this q( r ), then other things being equal T would increase, decreasing delta S which nature doesn’t like. And, of course, the actual total delta q can’t change since it’s fixed by solar radiation incoming.

    Actually, both are true, although “minimizing entropy” is not the fullest description of the process. A better description is the constructal law. One description of this says that a flow system acts to maximize the interface between the parts of the system. The author’s description of constructal theory is here. The constructal law shows how the system physically restructures to minimize entropy.

    The system, then, is running at the coolest temperature at which it can dissipate all of the energy entering the system. This is also as hot as it can run, given the constraints of the physical laws governing the system.

    w.

  82. Steve Bloom
    Posted Dec 18, 2006 at 1:59 PM | Permalink

    Re #78: Gee, bender, have you become a solarphile? I had thought you knew better.

  83. EW
    Posted Dec 18, 2006 at 2:47 PM | Permalink

    #70
    At RealClimate someone said that Svensmark’s cosmical ray theories should be tested in Antarctica. Seems that it just happened.

  84. Lee
    Posted Dec 18, 2006 at 2:57 PM | Permalink

    re 77 – in Veizer’s paper, look at figure 14 on page 10. What happens post 1985?

  85. jae
    Posted Dec 18, 2006 at 3:33 PM | Permalink

    #70: It seems that the Danes are doing a lot of really good research on solar effects. An interesting quote from the paper you linked:

    Physics history comes full circle. More than 100 years
    ago, C.T.R. Wilson developed the cloud chamber to try
    to understand natural clouds but he was diverted by his
    detection of ionizing particles. In 1937 the first known
    muons turned up in a cloud chamber[24]. Now Wilson’s
    initial purpose is fulfilled in a fresh understanding of the
    physics of natural low-level clouds, and in the evidence
    presented here for its relevance in the real world.

  86. jae
    Posted Dec 18, 2006 at 3:40 PM | Permalink

    84: Lee, I don’t know for sure, but I don’t think Veizer used the best independent variable. He should have used Solar cycle length. See this paper.

  87. jae
    Posted Dec 18, 2006 at 3:42 PM | Permalink

    We will have a nice out-of-sample test for the effects of solar cycle length soon (a couple of years, perhaps), since Solar Cycle 24 is late. I’ll bet that we will be seeing a cooling over the next few years.

  88. Paul Dennis
    Posted Dec 18, 2006 at 5:51 PM | Permalink

    re 84

    Lee, you might care to read what Jan Veizer says about the post 1985 data illustrated in Figure 14. It’s in his article under ‘The decadal to annual record of the last century’ on pages 19 and 20. He doesn’t hide from any ‘divergence problem’.

  89. Lee
    Posted Dec 18, 2006 at 8:30 PM | Permalink

    88 –

    Actualy, reading that it quite interesting.

    Among other things, he argues that solar is “the primary climate driver,’ as if there can be only one unchanging climate driver. Yes, past temperature changes before significant anthropogenic carbon inputs were not anthropogenic. Duh.

    He mentions the breakdown in correlation after 1985 as follows: “…except for the last two decades of the 20th century that may or may not be an exception to this pattern.”

    Uhhh – the two traces diverge completely after 1985. It may be an open question as to whether that ‘exception’ will continue It is NOT an open question that it is an “exception.” He even admits that the difference …”has to be attributed, therefore, to greenhouse gases, specifically CO2.”

    He is admitting an apparent shift in ‘the primary climate driver’ but then argues that since solar accounts for more of the variance it must be the primary climate driver. This argument ignores his earlier admission , on the same page, that it is NOT the primary driver later in the record. In essense, he is claiming that since he can correlate temp changes to his solar measures for most of the record, it is simplest to attribute solar for all of the temp record, event hough we can see that the correlation breaks down in the later record

    His unstated assumption is that if he can derive a single ‘primary climate driver’ it must be the only important one – and that he can continue to attribute it as the primary driver even when the record shows that the correlation is broken. He is ASSUMING that if the primary driver is solar, then it must be primary for all time in that trace. This is absurd.

  90. Steve McIntyre
    Posted Dec 18, 2006 at 10:09 PM | Permalink

    #88. Paul, did you get any answer from Science or Thompson to your inquiry about Guliya?

  91. Paul Dennis
    Posted Dec 19, 2006 at 1:29 AM | Permalink

    re 89 Lee, I think if you read the section in full you will come to a deeper understanding of what Jan Veizer is saying and not the distortion you presented.

    He does not admit that the difference has to be attributed to greenhouse gases and specifically CO2. He says that others have attributed the apparent fact that direct estimates of the TSI flux could not apparently explain the entire observed magnitude of the temperature rise to greenhouse gases. He even cites the papers (Ramaswamy et al., 2001, Solanki, 2002, Solanki et al, 2004 and Foukal et al, 2004). Nowhere does Veizer argue that solar effects are not the primary driver as you assert.

    He then goes on to make the case that the GCM’s are essentially water cycle models that generally do not incorporate an active carbon cycle. Carbon dioxide is prescribed and inputted in the form of an energy term. He quotes 4 W. per metre squared for CO2 doubling. As he points out these models would yield outcomes in the same direction regardless of the source of the additional energy.

    He then goes on to show some other possible sources of sdditional energy including the general decline in earth albedo (figure 15). His graph from 1985 to 1997 shows a decrease in albedo equivalent to a 7 W per metre squared change in the energy loading.

    Thus he concludes that the celestial cause as a primary driver appears to be a more consistent explanation.

    Reading the whole paper rather than picking sections out of context will lead to abetter understanding of what Veizer is saying.

  92. Paul Dennis
    Posted Dec 19, 2006 at 1:30 AM | Permalink

    Steve, I haven’t had any replies yet. As soon as I do I will let everyone know.

  93. Paul Linsay
    Posted Dec 19, 2006 at 7:42 AM | Permalink

    #89, Lee

    the two traces diverge completely after 1985

    It’s quite likely that there is no divergence problem because the the ground based temperature measurements are wrong, i.e., much too high. A temperature record after 1980 that used the satellite temperatures would show a much smaller, possibly no, divergence. Just to get an idea of how poorly the ground based measurements cover the earth, especially after 1990, take a look at the station coverage of the earth. No coverage of the cooler 70% of the earth, the oceans, and the remaining ground stations heavily contaminated by UHI. There are any number of pictures at Roger Pielke Sr.’s website of the ground station thermometers next to parking lots, on building roofs, and other completely inappropriate places for an unbiased measurement of atmospheric temperature.

  94. Steve Sadlov
    Posted Dec 19, 2006 at 10:45 AM | Permalink

    RE: #93 – Since the first installation of power grids there has been a logarithmic increase in anthropogenic power / thermal dissipation especially in the industrialized Western countries and Japan. Since the 1960s it has also encompassed the parts of the developing world in Asia, Latin America and Eastern Europe. The only place this has not occurred is Africa (which, on the other hand, has probably experienced radical albedo and vegetation type modifications due to slash and burn agriculture and overgrazing). And while cities are interesting in terms of their large increases in net anthropogenic energy flux within the first 1000 feet above local ground level, even more interesting has been the drastic change in rural development levels which turned rural areas from extremely energy poor areas of subsistance and small farming into energy rich areas of intensive industrial acriculture and husbandry. Most of this latter element happened even in the developed countries mostly after World War 2 and almost completely during the 20th century. Combine that with the vast expansion of area under cultivation made possible by large water projects and modern arid climate farming methods aka the Green Revolution. Pielke Sr seems to have a really good handle on the potential implications of all this vis a vis climate models and global energy flows.

  95. jae
    Posted Dec 19, 2006 at 10:46 AM | Permalink

    89: And you cannot even look at the data and methodology that is used to estimate “global average surface temperature.” The climatologists should demand that this information be released.

  96. jae
    Posted Dec 19, 2006 at 10:57 AM | Permalink

    Veizer’s paper also has a very interesting discussion of the interrelationships between moisture and temperature. He demonstrates that plants are generally limited by moisture in higher latitudes. Thus, higher temperatures would exacerbate this problem, due to both increased evapotranspiration and evaporation of soil moisture. IMO, this is why tree rings are primarily a moisture proxy. And I’ll bet this relationship also goes a long way in explaining the divergence phenomenon.

  97. Brooks Hurd
    Posted Dec 19, 2006 at 5:43 PM | Permalink

    Re: 93

    Paul,

    The other problem is that many rural stations have closed over the past 30 years.

    If you look at the station graphs above, the middle graph shows that compared to 1970, we have only about 40% of the stations still operating. I have looked through many GISS station data and it appears that many of those which have been closed are rural.

    I find it odd that even though the number of stations has decreased significantly since the 1970s, the coverage graph does not show the same sort of decline. I believe that this has to do with how Jim Hansen’s team defines the grid cells upon which coverage is based.

    The graphs are from the GISS site.

  98. jae
    Posted Dec 19, 2006 at 6:32 PM | Permalink

    Hmm, fewer and fewer rural stations, higher and higher surface temperatures, little change in tropospheric temperatures, many individual rural stations showing temperature declines (Idsos’ stuff). Sure looks like UHI effects could explain a lot of this unprecidented rise in temperature.

  99. Brooks Hurd
    Posted Dec 19, 2006 at 7:38 PM | Permalink

    And Phil Jones refuses to divulge his station data.

    Whatever the reasons, it does cause one to wonder.

  100. Brooks Hurd
    Posted Dec 19, 2006 at 7:43 PM | Permalink

    I expect that Steve Bloom will bring up the point that David Parker proved that the UHI did not exist in two papers over the past couple of years. I corresponded with Parker and I dissagree with his conclusions from the standpoint of heat transfer – both radiative and convective.

  101. Paul Linsay
    Posted Dec 19, 2006 at 8:40 PM | Permalink

    #97

    Brooks,

    The rightmost graph you provide looks nonsensical to me. If you look at the station data link I provided in 93 above, at their most extensive, all the stations covered about 15% of the earth’s surface. There’s essentially no ocean coverage so 70% is off the board immediately. The land is heavily covered in temperate North America and part of Western Europe. Everywhere else the coverage is spotty or nonexistent. Africa, S. America and large parts of Asia seem to be particularly poorly covered except for a short period in the 1960s.

    If you fiddle a bit with the mpg you’ll notice that during 1990-1991 there is a dramatic drop in coverage everywhere (apparent in the middle plot), followed by a second large drop in N. America coverage in January 1999. From 1990 on I’d say the station data is worthless in terms of coverage. It seems to be entirely restricted to heavily populated areas. The satellite coverage began in 1979 when the number of stations were starting to fall drastically.

    I’m sure that with care one can recover interesting local historical information from individual ground stations, but the plot makes it very clear that they do not form anything like a decent global network to measure the temperature of the entire planet.

  102. James Erlandson
    Posted Dec 20, 2006 at 3:48 AM | Permalink

    Re 101: Percent Coverage
    They use the definition “the percent of hemispheric area located within 1200km of a reporting station.” If you click most anywhere on the map at the bottom of the page it will show you a reporting station within 1200 km. For example, a spot in the South Pacific shows the reporting station at Mururora to be just 95km distant and a total of nine stations within 1200km.

    Combining the world’s temperature data into a single “global average temperature” is like condensing all the world’s literature in all languages into the letter “e.”

  103. Posted Dec 20, 2006 at 6:01 AM | Permalink

    #102

    They also provide a list of the stations actually used:

    http://data.giss.nasa.gov/gistemp/station_data/station_list.txt

    3D plot here:

    http://www.geocities.com/uc_edit/giss.html

  104. Paul Linsay
    Posted Dec 20, 2006 at 9:59 AM | Permalink

    ;)#102

    Ah yes, the famous 1200 km criterion. By my calculation it only takes 113 stations uniformly spaced over the earth’s surface to measure global temperature if that sampling radius is sufficient. All the continents would be covered by just 34 stations and the rest would cover the oceans. The division for the continents would be

    Asia: 10
    Africa: 7
    N. America: 5
    S. America: 4
    Antarctica: 3
    Europe: 2
    Australia: 2

    The oceans would need

    Pacific: 37
    Atlantic: 19
    Indian: 16
    Southern: 4
    Arctic: 3

    If some brave soul wants to find the appropriately spaced stations and compute a global temperature, that would be an interesting exercise.

  105. Nordic
    Posted Dec 20, 2006 at 10:41 AM | Permalink

    On the issue of Urban Heat Island effects. Does anyone have a number for the percentage of the earth’s surface and/or landmass covered by UHI effects? Where I live it would be very little (though our longest temperature records come from the Provo-Salt Lake-Ogden urban corridor), but I would think that significant portions of SoCal, the NE Corridor, Europe, coastal China, southeast Brazil, etc would either be considered urban areas, or close enough to large urban areas to be affected.

    I know that people have tried to adjust for the heat island effect in the past by adjusting the records of afffected stations downwards. Might it not be easier to simply figure out what percentage of each continent’s surface is a) an urban area, and B) within X km of a large urban area, then build your list of stations including only as many urban and near-urban stations as is appropriate given the spatial extent of urban areas. Mabe this has been done already – I’m just asking is all.

  106. Brooks Hurd
    Posted Dec 20, 2006 at 12:48 PM | Permalink

    Re: 105

    Nordic,

    The raw percentage is not difficult to calculate because census data includes urban area sizes. The difficulty is that census data varies form country to country, thus the basis for defining urban and rural is not necessarily consistent. As I said above, my perusal of the GISS data indicates that many of the closed stations are rural. This means that the ground station data is becoming more and more urban over time. If you combine this with the growth of urban area over the past years, the character of the ground station data is rapidly becoming urban.

    Hansen, who is in charge of the GISS data, has adjusted the records for UHI among other factors. How effective the adjustment can be in light of the change in character of the ground stations to an increasingly urban character is a question which needs to be addressed.

  107. jae
    Posted Dec 20, 2006 at 1:04 PM | Permalink

    Many of the “rural” stations may also have a problem, since they are located at airports and the thermometers are subjected to IR from the tarmack and buildings. I don’t think we have any idea what the “global average temperature” is. We don’t even know, for sure, whether it is even warming (though I think there has been some warming), let alone what is causing it. All this hype built on so much uncertainty. Amazing.

  108. Paul Linsay
    Posted Dec 20, 2006 at 1:38 PM | Permalink

    #105

    If you go to John-Daly.com and scroll down to the bottom where he has a US temp record you’ll see that the 1930s were the warmest of the 20th century. Read his comments.

  109. Nordic
    Posted Dec 20, 2006 at 3:43 PM | Permalink

    I went to the GISS data linked in the left-hand sidebar and looked quickly at the plots for my area (Utah). A couple of things jumped out. Most of the records run from the late 1800s and show an increasing trend until the 1930s or 40s, followed by a dip, then an increasing trend until the present. Unsuprisingly, virtually all the stations are located in communities. Most of these are rural communities, which in Utah means that they are surrounded by irrigated agricultural fields. The only sites not associated with towns of one size or another are a couple of National Parks. So the first question that comes up is how representative are these stations of a state that is overwhelminly composed of uninhabited mountains and deserts.
    The second is the question of heat islands. I wonder what the overal effect is likely to be in a fast-growing state where irrigated fields of alfalfa, sugar beets, pasture and small grains is giving way to buildings, pavement, and lawns. Salt Lake City is an obvious example. It has grown from 150,000 in 1940 to just 180,000 today. Not much of an increase, but when you look at the county data: 211,000 to 1,200,000 in the same period you can see how an urban center surrounded by agriculture has been transformed into a center in a larger sea of urban/suburban mix. Lehi is another example. It is listed on the GISS site as a rural station (which, incedentally only goes until 2003). That may have been true as recently as 20 years ago, but is not true now. Lehi has grown from 8500 people in the 1990 census to over 30,000 today and is part of a larger metropolitan area. Interestingly enough, the dropped Lehi station still had not attained the peak temperatures of the ’30s and ’40s as of 2003.
    One other observation: of 12 stations whose records do not extend to the present only one (Cedar City FAA) is not listed as a rural station. This list of dropped stations includes two stations not surrounded by irrigated lands: the Bryce Canyon airport, and Hiawatha, a Forest Service Guard Station.
    These are just observations. So if one were to try and construct a decent history of Utah (84,800 square miles), how many stations would you need? Should this be based on area, or on physiographic regions? (both toghether I wqould assume) Right now the record is dominated by the Wasatch front with few long-term records from the mountains. More records are available from desert regions, but they are almost all from towns founded where water was available to irrigate crops.

  110. Steve Sadlov
    Posted Dec 20, 2006 at 4:16 PM | Permalink

    Even in “rural areas” there has been an immense change as follows. Rural electrification and development have “urbanized” the rural areas. Dirt roads have been paved, task lighting put in place for 24 X 7 X 365 operations, etc.

    The “distance from urban area X” notion ought to be modified to some sort of integral. The integral would be the integration of some sort of function which, for each coordinate location, would look at the distance of that coordinate from each “population unit” – in other words, the cumulative impact on that coordinate of the overall global population density. Integrate this density function across all the coordinate points, and that would yield a function as follows:
    F(x,y,z) = G (Xn,Ym,Zp) where function G is itself the integral of all heat flows from the population density distribution. Bottom line is, anthropogenic energy dissipation impact upon any measurement point is the integral of all the individual anthropogenic “finite elements'” individual impacts.

    This is a much more sophisticated (and difficult) approach than the vastly oversimplified notion of “islands” of heat only over connurbations.

  111. jae
    Posted Dec 20, 2006 at 5:19 PM | Permalink

    Roger Pielke, Sr. put up some interesting comments on the problems associated with characterizing climate change by “average global temperatures.” He also notes that there is really only one data source for this average–Jones.

  112. Paul Linsay
    Posted Dec 20, 2006 at 7:47 PM | Permalink

    110, Steve,

    Nice idea, but the problem with it and any other correction scheme is that you have nothing to calibrate against. How do you know that you’ve made the right correction? There’s no physical theory that can tell you what to do, it’s purely ad hoc. That’s the problem with the GISS and HadCRU corrections.

    The ground stations also have a problem in that they’re just that, 2 m off the ground and in the boundary layer. The weather extends up about 5 km. Again, how do you correct for the poor vertical sampling of the atmosphere?

    It would be better to just end all reliance on the ground stations for anything but some very careful history.

  113. Posted Dec 20, 2006 at 11:58 PM | Permalink

    re #109, Nordic,

    FYI there are a couple of summary charts of Utah maxima and minima (all stations) from the USHCN/Daily
    data set at

    These charts aren’t much use for any one station, but do give a good bird’s eye view of the long term sweeps.

  114. Cliff Huston
    Posted Dec 21, 2006 at 2:09 AM | Permalink

    RE: #113
    Somebody should call Utah and tell them that they have their TMAX Hockeystick backwards.

  115. Posted Dec 21, 2006 at 5:44 AM | Permalink

    re #114

    A beguiling thought, but I should point out that the early part of the simple average series is dragged upwards
    by series 38, Zion National Park, which has the first displayed point at 77F, and to a lesser extent by series 3,
    Blanding, which has the first displayed point at 64.7F. The average is more reliable from 1928 onwards, when the
    majority of the stations come on line. The first ten years of each series are suppressed from display as the
    moving average kicks in.

  116. Steve Sadlov
    Posted Dec 21, 2006 at 9:48 AM | Permalink

    RE: #112 – My goal is not a correction. It is simply to prove or disprove the thesis that anthropgenic energy flux makes the surface record useless.

  117. Nordic
    Posted Dec 21, 2006 at 11:02 AM | Permalink

    DaleC: That is some interesting stuff. Does anyone know where to find information on where the data for each station was collected? – I know that many of these records are for stations that have moved over time.

    For instance, I live in Richfield, and our station is at the local radio station, a few blocks from my house. That building only looks to be about 20 years old – I wonder where the record was taken before then. This probably wouldn’t make much difference in a small town like Richfield, but I am exceedingly curious about the Salt Lake temperature record. That is the big one that is always reported on the news – and our local NWS page uses the Salt Lake KSLC record in their chart showing how each days temps. compare to the “normal” (I am still trying to track down what time period they use for normal) and record temps.

  118. Posted Dec 21, 2006 at 9:53 PM | Permalink

    re# 117, Nordic,

    Glad you found the charts interesting. Smoothed daily data seems to me to give better insight into behaviour over
    time than annual or decadal averages. Another advantage of working at the daily level is that wild outliers
    (which are often errors, even if flagged as valid) are not hidden.

    The data set and associated meta-data are described at

    http://cdiac.ornl.gov/epubs/ndp/ushcn/ndp070.html#back

    There is a file history.txt about a third of the way down the page which covers station moves.

    “This file documents station moves and instrument changes, lists station observers and observation times, and
    identifies suspect fields. For each station in the file there is an identification record followed by multiple
    data records describing station observing details over its period of record.”

  119. DaleC
    Posted Dec 21, 2006 at 11:20 PM | Permalink

    re# 117, Nordic

    FYI a comparative plot of Richfield max vs min is here. They diverge noticably about 20 years ago.

  120. beng
    Posted Dec 22, 2006 at 7:54 AM | Permalink

    I don’t even pay attention to the “surface” temps — they’re almost hopelessly inaccurate. Sat temps are the only thing I consider cause they provide almost complete horizontal & depth coverage . Here’s a link to a grid-display of sat temps:

    http://wwwghcc.msfc.nasa.gov/temperature/

    You can click on a grid-point (your own location for ex) & get a trend from 1979 on. Unfortuately, they haven’t updated the data since 2003. Ashamed, cause this is a neat tool.

    The better way even than sat temps is Pielke Sr’ ocean-heat content metric.

  121. Steve Sadlov
    Posted Dec 22, 2006 at 11:54 AM | Permalink

    RE: #120 – I agree 100% with your post.

  122. David Smith
    Posted Feb 26, 2007 at 9:20 AM | Permalink

    My impression, as a layman, is that observational data on atmospheric water vapor is not quite behaving as the models predict. This is a very important issue, because much of the AGW impact comes from water vapor feedback mechanisms.

    I plan to comment on this issue in small bites (= a number of shorter posts) rather than try to capture it all at once. This old water vapor thread, which never really took off as hoped, looks like a good place to park these posts.

    So, an important question is, how have the middle-to-upper levels of the tropics behaved in recent decades?
    First up is a slide #25 from Dr Held . My interpretation is that this slide (the middle map, of IR) indicates the impact that water vapor, at various altitudes and latitudes, has on global warming. This is per the global climate models (GCMs).

    The take-away points are:

    * the most important levels are the 300mb to 600mb

    * the most important region is the tropics (remember that most of the earth’s surface area lies between 30S and 30N, so the 30 to 90 portion is not as big as it looks)

  123. Gerald Browning
    Posted Feb 26, 2007 at 8:47 PM | Permalink

    David (#122):

    It is well known that the models do poorly in the equatorial regions for several reasons. One is that there is very little observational data in that region (see Roger Daley’s book). The second is that in that region the vertical velocity is proportional to the total heating (see references in my ITCZ post). Therefore, if there are any errors in the model forcings (parameterizations) near the equator, those errors will translate immediately into unrealistic results.

    Jerry

  124. David Smith
    Posted Feb 26, 2007 at 9:10 PM | Permalink

    Re #123 Jerry thanks for the info. I did not realize the GCMs struggle so much with the tropics, though I’m aware of the lack of data and clear understanding of things like tropical precipitation and cloud behavior.

    If the models miss the mark on this fundamental region then it’s hard to have much confidence in their projections on anything.

  125. David Smith
    Posted Feb 26, 2007 at 9:29 PM | Permalink

    Re #123 Something else about the models that surprised me is that apparently they struggle to simulate the Madden-Julian oscillation (MJO). The MJO plays an important role in tropical precipitation.

  126. Posted Feb 27, 2007 at 9:17 AM | Permalink

    How would the vertical temperature profiles of the atmosphere change if there was a decrease in lower cloud cover (as modeled by the solar-cosmic ray theory to explain recent warming). Would surface warming be more prevalent than warming in other parts of the troposphere as predicted by CO2-forced warming?

  127. Gerald Browning
    Posted Feb 27, 2007 at 12:22 PM | Permalink

    David (#124):

    I am in complete agreement with your last sentence.

    David (#125):

    There is also the double ITCZ problem.

    Recall my earlier mathematical argument that if the dissipation is incorrect (too large), then the forcings are necessarily incorrect (tuned).

    Jerry

  128. Gerald Browning
    Posted Feb 27, 2007 at 6:46 PM | Permalink

    Jason (#126):

    Near the equator, the vertical component of the velocity can be excited in a number of ways, e.g. warming of the ocean surface by the sun in a clear sky. Any combination of individual components of the total heating that leads to the same total heating in the troposphere will lead to the same vertical component of velocity of a slowly evolving solution in time. Thus one must know the size and type of increments in the components of the total heating due to any perturbations that are causing a change. If there is a dominant perturbation causing the change, then that is the likely cause. But it is also possible that two large perturbations cancel and a third smaller perturbation is the cause. Thus it is crucial that all the forcing components of the total real atmospheric heating be completely understood and that any errors in the formulation of the components not overwhelm the perturbations that cause the change.

    Jerry

  129. David Smith
    Posted Mar 4, 2007 at 4:06 PM | Permalink

    Richard Lindzen’s May, 2006 presentation includes a section on tropical cloud and precipitation behavior. This is connected to his Iris Hypothesis, which is a complicated and controversial topic.

    What I find interesting are his two plots (slides 41 and 42). They are evidence that, as sea temperatures warm, tropical thunderstorms tend to form proportionally more raindrops and proportionally fewer high-altitude clouds. This is important because high-altitude clouds are believed to be net trappers of radiation and, if these clouds spread out and precipitate, put water vapor into levels of the troposphere that would otherwise have low humidity.

    Lindzen’s hypothesis is that CO2 leads to warmer tropical SST, which lead to fewer high-altitude tropical clouds, allowing Earth to radiate away more IR. This is negative feedback, mitigating the impact of CO2. A lot of things have to fall into place for this hypothesis to be true and the jury is out as to whether those additional factors support the hypothesis. However, it’s intriguing that the first step (changes in thunderstorm behavior as SST rise) now has evidence to support it.

  130. cytochrome_sea
    Posted Mar 4, 2007 at 6:50 PM | Permalink

    Re: 129,

    David, there’s a bit more here.

  131. David Smith
    Posted Mar 4, 2007 at 8:21 PM | Permalink

    Re 3130 Excellent, thanks!

    I have no opinion (yet) as to whether Lindzen is right or wrong. But, I strongly believe that climate science would be farther along if it had more Lindzens.

  132. Posted Mar 5, 2007 at 4:17 AM | Permalink

    #129 David: please, tell me if you agree: “Lindzen’s hypothesis is that warmer tropical sst, regardless the source of this extra heat,….”

  133. David Smith
    Posted Mar 5, 2007 at 8:48 AM | Permalink

    Re #132 Paulo, interesting quesion. If the issue is strictly ocean temperature, and nothing else changes except the ocean temperature, then the source of the extra heat probably doesn’t matter.

    But, I imagine the question actually has to do with whether Lindzen’s experiment, which measured seasonal heating and cooling of the ocean, is really the same as increased trapping of IR by CO2.

    Increased IR trapping would warm the troposphere and affect the lapse rate, which to some degree would affect the behavior of the rising air in thunderstorms, but I think that impact would be relatively small. My guess is that thunderstorms would continue to make “more raindrops /less ice” with warmth from higher CO2, just as they seem to do today due to seasonal changes.

    Another aspect, as I see it, is whether the air in the thunderstorms will rise to the same height as today. I think this is an important question. If the air rises to higher, cooler temperatures, then it becomes drier. If the air doesn’t rise as far, then it is more humid. Ultimately the water vapor content of the upper tropical troposphere depends on the height to which the thunderstorms rise.

    My guess is that, for a doubling of CO2, the thunderstorm height probably doesn’t change much, and to the extent that it changes, it’s higher (cooler and drier) due to more overshoots. But, that’s something that can be plausibly argued both ways.

  134. Ken Fritsch
    Posted Mar 5, 2007 at 11:41 AM | Permalink

    Re: #122

    the most important region is the tropics (remember that most of the earth’s surface area lies between 30S and 30N, so the 30 to 90 portion is not as big as it looks)

    One half is the number I see most often as linked here.

    One-half of the Earth’s surface area is contained between 30S and 30N.

    Dave S, just to show I’m paying attention. This area apportionment came up in the Lorenz et al. (2006) thread and was an important factor there.

  135. jae
    Posted Mar 5, 2007 at 3:41 PM | Permalink

    129. Lindzen says he is part of the “consensus.” I guess most of us are also part of the consensus the way he defines it.

  136. David Smith
    Posted Mar 5, 2007 at 4:37 PM | Permalink

    Another interesting thing about 30 degrees latitude is that areas poleward of 30 degrees are net importers of heat. In other words, if it was not for heat transported from the tropics, San Francisco, Chicago, Toronto, London, Tokyo, Adelaide and so forth would be noticeably colder than they are today.

    In a sense, the tropics (the Hadley-Walker flows) are like a “furnace in the basement” that heats the global atmosphere. If the warm tropical waters expand (like in El Nino) then the heat output grows and the mid-latitudes warm. If the tropical waters contract (like in the upcoming La Nina) then the globe chills a bit. The impact of the tropics is real.

    In the 1970s, when the planet began its latest warming, something rather suddenly shifted in our basement furnace, the tropics. Its distribution of heat changed (see here ) . There seems to be more flow into the Northern Hemisphere since 1976, something like changing the ductwork settings in a house or apartment furnace.

    Our “furnace in the basement”, even though it seems mundane, is very important.

  137. Posted Mar 6, 2007 at 3:46 AM | Permalink

    David, you can find a nice picture of the latitudinal energy budget here.

  138. Alex Harvey
    Posted Feb 4, 2008 at 8:09 AM | Permalink

    Hi,

    I hope this is the best place for a question that’s been bothering me for a while.

    I’ve read most of the posts on Richard Lindzen now & it seems to me that his entire skepticism of AGW theory — at least as far as his scientific skepticism goes — follows from his “Iris hypothesis”. That is to say, if it were proved to his satisfaction that the Iris hypothesis was wrong, and that water vapour provides the positive feedback everyone else seems to say it does, then he would concede that AGW is probably true.

    Is this a fair comment?

    My other question is, does anyone happen to know when he is likely to publish again? His homepage at MIT suggests he has had the following two works in progress for some time:

    224. Lindzen, R.S. and R. Rondanelli (2006) On the need for normalizing satellite cloud data when applying results to climate. In preparation

    225. Rondanelli, R. and R.S. Lindzen (2006) Reexamination of Iris Effect using TRMM and Kwajalein Ground Radar. In Preparation.

    Best,
    Alex

One Trackback

  1. […] Comment on Water Vapor and Cloud Feedbacks by Alex Harvey Hi,. I hope this is the best place for a question that’s been bothering me for a while. I’ve read most of the posts on Richard Lindzen now & it seems to me that his entire skepticism of AGW theory — at least as far as his scientific … […]

Follow

Get every new post delivered to your Inbox.

Join 3,413 other followers

%d bloggers like this: