Ocean Heat Storage: Implications for Climate Change Research

[sylas]

Please excuse a confused layman's question.
I found this graph: http://omlc.ogi.edu/spectra/water/gif/segelstein81.gif

and similar here: http://www.lsbu.ac.uk/water/vibrat.html (scroll down)

Nice diagrams. Thanks for the links...

It is possible (probable) I am misinterpreting this data. It appears to me that infrared radiation above 4 microns has no heating effect on water but possibly an increase in water vapour release.

The energy of sunlight all gets absorbed (less what is reflected). In shallow water a few meters deep, you still get plenty of visible light penetrating to the bottom, where it heats the floor; and then the water gets heated from there by convection. In deep water nearly all radiant energy is absorbed within a couple of hundred meters.

From your diagrams, there's not much radiant energy getting below 10m, and below 100m it's going to be quite dark, as nearly all the light gets absorbed by then.

Water vapour depends on the temperature of the surface, mostly, but any energy taken into those upper layers will contribute to a warmer surface. The two go together. You get water vapour by heating water.

This indicates to me that the oceans are heated by sunlight only (below 3 microns) and greenhouse gases play no part in heating over 70% of this planet. This would also apply to all rivers and inland waters.

This is a non-sequitur. Greenhouse effects apply in precisely the same way over land and over ocean. They are determined by the atmosphere, not the surface, and they result in a greater flux of infrared radiation down to the surface from the atmosphere at around 10 μm, just where water absorbs energy most effectively. This is no accident... water is the strongest greenhouse gas in the atmosphere.

Greenhouse gases are not a direct source of energy in their own right. They are like a blanket in this respect. All the energy, ultimately, is from the Sun. What greenhouse gases do is absorb some of the thermal radiation coming up from the surface, and this makes it harder for energy to get out to space. So the surface heats up more, until it is hot enough to shed the energy it absorbs back to space.

Another completely equivalent way of considering this is "backradiation". By Kirchhoff's law, gases emit strongly in the same wavelengths that they absorb; and they radiate in all directions. The end result is that there is a large flux of energy coming down from the atmosphere... though by the second law it is not as much as is going up from the surface. The atmosphere is heated from the surface, and this is just another way of saying that the atmosphere makes it harder for the surface thermal radiation to get out into space.

The additional flux of thermal radiation coming down to the surface peaks around 10 microns.

Increased water vapour = increased cloud cover = reduced solar radiation at sea level = eventual reduction in ocean heat. All in all, a self regulating negative feedback system.

You have correctly identified an important negative feedback process.

However, there's a lot more to cloud than this effect. It's one of the hardest parts of the whole climate issue to model physically. Clouds are very good at absorbing thermal radiation, as well as reflecting sunlight. This means that they add considerably to the greenhouse effect themselves. Think about night time -- which is just as important as the day for averaging temperatures. A clear night is much colder than a cloudy one, and that is because clouds have such a strong interaction with thermal radiation. The effect of cloud depends very strongly on their altitude and composition. In general, cloud feedback is thought to be a net positive feedback; although with the largest uncertainty of any climate feedback process, to the extent that we can't be sure it is a net positive.

But the most important factor for water vapour, by far, is its direct greenhouse effect by infrared absorption in the atmosphere generally. If you look at your diagrams, to the right of the visible light band, you see water absorbing 10 μm very effectively, and that is right where Earth's main thermal emissions lie. This positive feedback is the strongest part of the effect of water vapour.

Reference: Bony, S. et. al. How Well Do We Understand and Evaluate Climate Change Feedback Processes?, in J. of Climate, Vol 19, 1 Aug 2006, pp 3445-3482.

Bony et al give the water related feedback estimates (mean and standard deviation) as follows (measured in W/m²/K)

• Water vapour direct feedback: 1.8 +/- 0.18
• Lapse rate feedback: -0.84 +/- 0.26
• Cloud feedback: 0.69 +/- 0.38

As you can see, the cloud feedback estimate is less than two standard deviations away from zero, so the sign is actually uncertain. Water vapour direct effects are strongly anticorrelated to lapse rate effects, so when combined their uncertainty is actually reduced. All told, the water vapour plus lapse rate feedback is dominant, but there's a lot of ongoing work to try and constrain cloud feedbacks better, since they are the major source of uncertainity.

Cheers -- sylas

 

[chriscolose]

I think that Richard's point was more so that shortwave radiation "penetrates more" into the water layer than does longwave (infrared) radiation. It does, but this doesn't mean infrared doesn't heat water, or that it isn't as important a term. The top 100 m or so of ocean is well stirred by turbulence, so energy dumped into the top skin layer gets mixed downward rapidly. It isn't as though the greenhouse effect suddenly shuts off as you move from land to water.

Furthermore, the enhanced greenhouse effect from anthropogenic activities results not only in enhanced downward infrared emission, but all the heat fluxes which couple the air to the surface. The increased downward emission is more so due to higher temperatures than it is directly due to higher CO2 levels, since the whole troposphere more or less warms up together.

As far as clouds, many models and limited observation suggests decreases of low cloud cover in a warmer climate, or at least a greater increase in high cover relative to low cloud cover changes (i.e., a positive feedback). As sylas noted, this is a very active topic of research and not limited to "more water vapor = more clouds," particularly since cloud formation is not determined by absolute amounts of vapor, but rather by relative humidity, which changes little on a global scale. The sign of cloud impact depends on height, latitude, optical thickness, and other things. Hartmann's FAT hypothesis (probably the leading competitor to the IRIS proposal by Lindzen et al) has high cloud top temperature changing little, which decouples the emission temperature from the surface temperature and becomes a positive feedback.

 

[joelupchurch]

I'm not comfortable with how they mix in the lower quality pre argo data. There seems to be a sharp change in the trendline around 2004. Is that real or just because they are mixing apples and oranges?

Excuse me for referencing whatsupwiththat, but it looks like I wasn't the only one who thinks the transition from preargo to argo data didn't look right. It looks like there is a problem stitching the two datasets together. If you look at the 3 month line, the change is almost all in one quarter. I was a year off reading the scale, since the post references 2002-3, not 2004.

http://wattsupwiththat.com/2009/06/02/anomalous-spike-in-ocean-heat-content/#more-8132"

 

[Richard111]

Thank you sylas #36 and thank you chriscolose #37, that was indeed my point. When I read the http://www.climates.com/SPECIAL%20TOPICS/GW/greenhouse.htm I imagined the top skin of water would vaporise. (I know it doesn't as I live by the sea.)

For heat energy reaching the surface of the Earth from all sources, the budget is as follows:

1. 34.1% comes from the Sun as direct and diffuse solar radiation (insolation). Of this amount, 20.1% is as direct solar radiation, and the remaining 14.0% as diffuse solar radiation.

2. 65.9% comes from the “greenhouse” gases of the atmosphere. Of this amount, 40.0% is from water in its various phases, 14.1% from carbon dioxide, and the remaining 7.8% from ozone and oxygen.

I must go and read up on surface turbulance and heat distribution in water.

Regards

Richard

 

[Richard111]

Well my ignorance was politely pointed out by email that liquid water responds to IR like any other substance that is opaque to IR, in fact better because of the conductive ability of water. IR, in time, will warm a brick, which is opaque to IR. (sigh) Its a long hard learning curve.

 

[Patriot Vet]

relative humidity, which changes little on a global scale.

It has been changing a lot for the past 50 years, globally. Dropping 21% in fact.

http://www.cdc.noaa.gov/cgi-bin/data/timeseries/timeseries1.pl"

Also documented by:
‘http://www.theclimatescam.se/wp-content/uploads/2009/03/paltridgearkingpook.pdf" ’ by Garth Paltridge, Albert Arking and Michael Pook

ABSTRACT:

Water vapor feedback in climate models is positive mainly because of their roughly constant relative humidity (i.e., increasing q) in the mid-to-upper troposphere as the planet warms. Negative trends in q as found in the NCEP data would imply that long-term water vapor feedback is negative—that it would reduce rather than amplify the response of the climate system to external forcing such as that from increasing atmospheric CO2.

Discussion:

First, the observations of relative humidity RH do not support the proposition that emerges from the behavior of general circulation climate models that the value of RH at any given height in the troposphere remains fairly constant under the influence of global warming (e.g., Soden and Held 2006; Pierrehumbert et al. 2007).

 

[sylas]

It has been changing a lot for the past 50 years, globally. Dropping 21% in fact.

This is looking at the wrong thing. You need to look at specific humidity, not relative humidity. The drop in relative humidity is mainly because of the rise in temperature. But feedbacks depend on the amount of water...

You've actually put your finger on one of the reasons the feedback from clouds is not as strong a negative feedback as suggested above. Cloud depends largely on relative humidity.

But the much stronger water vapour feedback depends on specific humidity, which has been increasing -- though the year to year variation is very high so measurements come with substantial uncertainty.

There are some major problems with your data source. You are using the NCEP/NCAR reanalysis data... which is not actually measurement. It is a climate model, which attempts to give a detailed climatology constrained by measured data, but the actual data given is the model, and it has known issues.

It comes with many caveats, and it's particularly dubious as a substitute for measurement of one particular aspect of the climate system.

The actual science on this indicates increasing moisture in the upper troposphere, where the feedback effect is strongest. Note that these results are consistent with reducing relativity humidity. Remember... it is specific humidity that gives the feedback, and as temperatures rise you get more water and more feedback even as relativity humidity may be dropping.

Refs:

• Soden, B.J. et al. http://www.sciencemag.org/cgi/content/abstract/310/5749/841, in Science, Vol. 310. no. 5749, pp. 841 - 844, (4 Nov 2004).
• Dessler, A.E. and Sherwood, S.C., http://geotest.tamu.edu/userfiles/216/dessler09.pdf in Science, Vol 323, no. 5917, pp. 1020 - 1021 (20 Feb 2009)

 

[chriscolose]

Patriot Vet,

I don't agree with sylas that RH is not something important to look at, mainly because while the actual water vapor feedback depends on specific humidity, the constancy of relative humidity tells you how well the change saturation vapor pressure follows expected changes from Clausius-Clapeyron. The relative humidity is roughly balanced by increases in temperature and increases in moisture. A decrease in relative humidity however does not a priori mean a negative feedback. But he is correct that this particular data is not very useful, and in fact probably worthless in determining the long-term humidity trends (see Trenberth, Fasullo and Smith 2005 and the Soden et al paper referenced above for discussion). I have blogged about the Paltridge paper and the corresponding discussion on other secondary sources, but it's probably not worth going into in more detail because it is not something that is going to change our picture of the subject and is not a useful contribution to further understanding of water vapor feedback or humidity trends. Sylas' two references are good ones to look at.

 

[sylas]

... I don't agree with sylas that RH is not something important to look at, ...

That was not my intended implication; and I am glad to accept the correction. Thanks.

My point is that it is the wrong thing to look at for water vapour feedback, which depends on the total amount of moisture in the air: specific humidity.

However, looking over the thread I see that patriot vet was actually quoting chris colose on relativity humidity and cloud feedback... and of course relative humidity is important for cloud feedback.

My understanding of the data is that there is indeed a small overall trend of reducing relative humidity, but considerable regional variation in trend. This is based on direct measurement, rather than the model based NCEP/NCAR reanalysis product, which is a poor guide.

So where are we. Richard raised the matter of feedback processes associated with water vapour. This is complex, as there are multiple effects involved.

For a useful short background and summary of the current state of knowledge, I recommend the paper cited above: http://geotest.tamu.edu/userfiles/216/dessler09.pdf . This is a short and clear review article specifically about empirical constraints on feedback processes associated with water vapour. The conclusion is that theory and observation imply a strong positive feedback; and the magnitude is most likely about 1.5 to 2.0 W/m²/K. This refers to the main water vapour feedback.

This positive feedback might be offset by a weaker negative feedback from increased cloud cover. This feedback is particularly hard to constrain, but estimates are that it is actually an additional positive feedback; and a part of the reason for this is that cloud depends in relative humidity rather than specific humidity. Relative humidity shows considerable regional variation (Dressler et al 2009 again) so a single trend gives the wrong idea. But over all, the is probably a net fall where it matters, and this contributes to cloud being an additional net positive feedback on top of the strong positive water vapour feedback.

Cheers -- sylas

 

[joelupchurch]

I think Patriot Vet has a point. If RH is dropping then the positive forcing is probably at the low end of the range that the IPCC is using. That would give a warming of 0.2 degrees centigrade per decade or even lower.

That doesn't change the ultimate problem, but gives you a lot more options for solutions.

What is really interesting is the back story on the Paltridge paper. You can read about it here:

http://www.climateaudit.org/?p=5416"

 

[chriscolose]

I think Patriot Vet has a point. If RH is dropping then the positive forcing is probably at the low end of the range that the IPCC is using. That would give a warming of 0.2 degrees centigrade per decade or even lower.

That doesn't change the ultimate problem, but gives you a lot more options for solutions.

What is really interesting is the back story on the Paltridge paper. You can read about it here:

http://www.climateaudit.org/?p=5416"

It would be very nice to stick with refereed sources or something from ".edu" and similar-quality sites. I understand that it is very "sexy" to blog about how all those AGW people are ignoring their paper, silencing dissenting views, or what have you. The fact is that if this paper is brushed to the side then the experts who actually study these things have reasons to brush it to the side, perhaps reasons which neither Paltridge or McIntyre are aware of. It may surprise some but many papers (either good or bad) in the literature are just not very important. In some cases it may be best to just tell them why their paper is not very important, or why it does not offer robust conclusions, or why other evidence is better and leave it at that.

//"--------------------------------------------------------------------------------
I think Patriot Vet has a point. If RH is dropping then the positive forcing is probably at the low end of the range that the IPCC is using"//

This really has nothing to do with the forcing, but rather the strength of the water vapor feedback, which has been the subject of study for many decades. The general picture has not changed very much.

Estimates of climate sensitivity come from the paleoclimate record and suggest that on the timescales relevant to anthropogenic perturbations, the climate is dominated by positive feedback and the equilibrium response to a doubling of CO2 is roughly 2 to 4.5 C. Evidence accumulated over the past few decades has done a good job putting constraints on various feedbacks, clouds still remaining at a very unacceptable status of understanding. Modelling and limited observations suggest decreases in low cloud cover in a warming climate. Some of the "feedback" is not actually a feedback in the traditional definition of being a response to the temperature anomaly, but rather is a rapid response to the increased forcing from CO2 independent of deltaT. If the overall cloud feedback is negative, it does not appear to be very strong.

 

[sylas]

Chris beat me to it... but since I had started writing this I'll post anyway.

I think Patriot Vet has a point. If RH is dropping then the positive forcing is probably at the low end of the range that the IPCC is using. That would give a warming of 0.2 degrees centigrade per decade or even lower.

Humidity makes no difference to forcing. I presume you meant "feedback" there, which in turn determines the climate sensitivity.

The relative humidity does not have much effect on feedback. The feedback is mainly from specific humidity (Dressler 2009). I don't believe that the humidity data is accurate enough to constrain sensitivity measurements beyond the range given in the literature and IPCC reports. If anything, the work coming out on humidity is making the low end of the range LESS likely... but with the kind of uncertainties that are reflected in having a large range of possible sensitivities.

Some references:

• Soden, B.J. et. al. (2006) An Assessment of Climate Feedbacks in Coupled Ocean–Atmosphere Models, in J. of Climate, Vol 19, July 2006, pp 3354-3360.
• Bony, S. et. al. (2006) How Well Do We Understand and Evaluate Climate Change Feedback Processes?, in J. of Climate, Vol 19, 1 Aug 2006, pp 3445-3482.
• Dessler, A.E. el. al. (2008) http://geotest.tamu.edu/userfiles/216/Dessler2008b.pdf , in Geophys Res Lett, Vol 35, L20704, doi:10.1029/2008GL035333, (2008).
• Dessler, A.E. and Sherwood, S.C., (2009) http://geotest.tamu.edu/userfiles/216/dessler09.pdf in Science, Vol 323, no. 5917, pp. 1020 - 1021 (20 Feb 2009)

IPCC sensitivity values are 2.0 to 4.5 K per CO2 doubling. Sometimes 1.5 is given as the low end of the range. Using W/m² as the forcing units, and 3.7 W/m² for a doubling, that is around 0.5 to 1.2 K per W/m².

The relationship between sensitivity and feedbacks is explained in Bony et al (2006). Basically, it works like this. Let F be a forcing, and T be a temperature. The feedback basically like a forcing you get from temperature, and is given as λ. There's a basic Planck no-feedback response (λ_P, or sometimes λ₀). The equilibrium response to forcing is given by the equation

$$F + (\lambda_P + \lambda) T = 0$$​

The Planck response is negative (higher temperature result in more emitted energy) with a value of -3.2 W/m²/K. The sensitivity is simply T/F. If we let s be the sensitivity in K per W/m², then

$$\begin{align*} s = T/F &= \frac{-1}{-3.2 + \lambda} \\ \frac{1}{s} &= 3.2 - \lambda \\ \intertext{Hence} \lambda &= 3.2 - \frac{1}{s} \end{align*}$$​

Low sensitivity means low feedback parameter. Sometimes rather than the sensitivity s in K per W/m², a paper focusing on feedbacks might just give the total feedback which is the negative inverse of s. This is done in Soden (2006), for example.

Here are the values for feedback and sensitivity, at the high and low ends of the range.

$$\begin{array}{r|llll} & \mbox{sensitivity} & \mbox{sensitivity} & \mbox{feedback and base response} & \mbox{feedback} \\ & & & \lambda+\lambda_P & \lambda \\ \mbox{units:} & K/\mbox{2xCO2} & K W^{-1} m^2} & Wm^{-2}K^{-1} & Wm^{-2}K^{-1} \\ \hline \mbox{IPCC v. low}&1.5&0.41&-2.47&0.73\\ \mbox{IPCC low}&2&0.54&-1.85&1.35\\ \mbox{Soden low}&2.31&0.63&-1.6&1.6\\ \mbox{Soden high}&4.11&1.11&-0.9&2.3\\ \mbox{IPCC high}&4.5&1.22&-0.82&2.38 \end{array}$$​

The numbers from Soden (2006) are theoretical (model based) expected values. The largest feedback is the water vapour feedback, from about 1.5 to 2.1 W/m²/K (table 1, page 3355).

When compared with empirical measures of humidity, estimates of water vapour feedback are at the high end, rather than low end. Dressler et. al. (2008) estimates 2.04 from recent empirical data, but with substantial variation, from 0.94 to 2.69.

What is really interesting is the back story on the Paltridge paper. You can read about it [...snip...]

I strongly recommend we do NOT go into secondary speculations from blogs about "back stories". I'm very tempted to go down this side track, but let me just say that this blog link is a perfectly dreadful way to get the so-called "back story". What you have there is mostly a bad case of sour grapes over a paper with little relevance or value to the realities of climate, given its ... odd ... choice of data and methodology. It would be easy to see OTHER perspectives on this alleged "back story", but let's avoid that whole quagmire and stick to the science please... and the forum guidelines.

Paltridge et al is NOT using empirical measurement, but a model based reanalysis product. The NCAR/NCEP reanalysis is useful for various applications when used with due regard for its limitations. Unfortunately it has also been used for a lot of poorly founded inference that conflicts when more credible science properly based on empirical measurement directly. The paper itself tends to acknowledge the limits of the approach rather better than the pundits and bloggers who have picked it up and inflated its significance.

Cheers -- sylas

 

[joelupchurch]

Normally, I wouldn't reference a blog posting, but since this was an actual quote from the author of the paper, I decided to mention it. If it had just been a comment by Steve McIntyre, I wouldn't have brought it up.

When I see a paper mentioned, I usually check Google Scholar and see if there is a place I can download it for free and see what other papers reference it and what the other papers say. That didn't work, so I did a regular search and found the statement by the author. The statement by the reviewer was so biased that I found it remarkable that the editor didn't get another scientist to review the paper.

I looked at IPCC AR4 and section "3.4.2.2 Upper-Tropospheric Water Vapor" and there are multiple places where they say the observations are consistent with a constant Relative Humidity. That implies to me that if the RH is decreasing, that the Specific Humidity is increasing by less than expected.

As to the comment about NCAR/NCEP being a "model based reanalysis product", I should point out that applies to most of the data we are dealing with. When dealing with data collected over a period of decades, I don't see how that can be avoided. I'm not clear about why radiosonde data has more issues than the buoy data this thread started with.

I can accept that the data, especially the older data, has a high uncertainty. I made comments about pre-1970 buoy data being almost useless for this reason. But when the results show decreases in RH at every pressure over a period of decades, then explanations about this sort of bias could have creep in are in order instead of just saying the data is suspect. I looked at Wang et al., 2002a, but it isn't clear why issues with one model of radiosonde would justify discarding 50 years of data.

 

[sylas]

Normally, I wouldn't reference a blog posting, but since this was an actual quote from the author of the paper, I decided to mention it. If it had just been a comment by Steve McIntyre, I wouldn't have brought it up.

Understood -- and if the focus was on the actual science then I would be more interested. But as you said: it's all about the "back story", which means he's bellyaching about reviewers and editors. Pretty poor, IMO. There's two sides to that, and neither of them are about the science.

In any case, Paltridge DID get his paper published, albeit in a less prestigious journal. The reference is:

• Paltridge, G. et. al. (2009) Trends in middle- and upper-level tropospheric humidity from NCEP reanalysis data, in Theoretical and Applied Climatology, online Feb 2009, doi 10.1007/s00704-009-0117-x.

The major defect with this work stands out immediately in the abstract: his results are largely inconsistent with satellite data, and rely on a reanalysis which (for humidity) is based on balloon data that he identifies right from the start as "notoriously unreliable". But rather than focus identifying inaccuracies, he seems to focus on what it would mean if they were correct. What the...? My reaction would probably be pretty much the same as the reviewer he called "hysterial".

When I see a paper mentioned, I usually check Google Scholar and see if there is a place I can download it for free and see what other papers reference it and what the other papers say. That didn't work, so I did a regular search and found the statement by the author. The statement by the reviewer was so biased that I found it remarkable that the editor didn't get another scientist to review the paper.

There were other scientists reviewing the paper; and although Paltridge objects, I'm inclined to think the anonymous reviewer quoted may have nailed it. The approach of the paper is very odd indeed, and there's nothing improper about a reviewer making such judgments. It is part of the job. Reviews are MEANT to weed out most submissions to a journal. I take with a grain of salt Paltridge's claim that the editor should have ignored such a review. I've been involved in journal review (maths and computer science), and in my experience there's no problem with strongly negative reviews... only with wildly inconsistent reviews. We're not in a position to judge that; and speculations by all kinds of bloggers who have picked up on Paltridge's objections have jumped to conclusions about reasons for rejecting the paper that they are in no position to make. We can now judge the paper itself on its merits, and it's no surprise to me at all that it was rejected by Journal of Climate. Good for them.

I looked at IPCC AR4 and section "3.4.2.2 Upper-Tropospheric Water Vapor" and there are multiple places where they say the observations are consistent with a constant Relative Humidity. That implies to me that if the RH is decreasing, that the Specific Humidity is increasing by less than expected.

Two issues here. First and most importantly, there's no good reason to think there's a strong decrease in relative humidity. Paltridge is using data known to be inaccurate, which makes the whole exercise something of a waste of time.

Second: there's no particular "expectation" for constant relative humidity. The expectation is rather than the change in RH is small. That fits observations. There IS some evidence of small decreases in relativity humidity; though with sufficient uncertainty that observations are consistent with constant relative humidity. Some models involve increases, and some involve decreases, but none involve large changes.

What section 3.4.2.2 actually does (page 273) is start out by noting explicitly that humidity is uncertain, and then give a quick list of various different results. It cites Wang for evidence of increases in relative humidity, and Minschwaner and Dessler for decreases in relative humidity; then small trends both increasing and decreasing in relative humidity from Bates and Jackson, and others. The major implication is that changes in relative humidity are comparatively small... and that means specific humidity must be increasing, given the temperature trends.

My understanding is that ongoing work tends to be along the lines of Minschwaner and Dessler, for a small decrease in relative humidity.

The top of page 273, right column, is particularly relevant, since it describes the errors in radiosonde data, and the difficulty in developing accurate corrections. Satellite data is generally more reliable, though any measurements over an entire atmosphere have difficulties. Basically this is the reason for Paltridge's results... the data he's using is known to be inaccurate and lacks good error correction approaches. The trends in the graph provided in this thread don't actually mean anything much.

There is also substantial regional variation in humidity trends.

As to the comment about NCAR/NCEP being a "model based reanalysis product", I should point out that applies to most of the data we are dealing with. When dealing with data collected over a period of decades, I don't see how that can be avoided. I'm not clear about why radiosonde data has more issues than the buoy data this thread started with.

Flatly wrong. There's a drastic difference between empirical measurement and model based reanalysis. Most of the data we are dealing with is NOT a model based reanalysis, but is based on empirical data, with clearly identified differences between different instruments.

I don't know why or how you are comparing atmospheric radiosondes with bouys in the ocean. Every measurement has its own sources of error. The IPCC report clearly tells you the basis of problems with radiosondes for humidity measurement. From page 273:

In general, the radiosonde trends are highly suspect owing to the poor quality of, and changes over time in, the humidity sensors.​

The proper comparison here is radiosondes and satellites; and it looks pretty clear that for humidity, the satellites are more reliable.

I can accept that the data, especially the older data, has a high uncertainty. I made comments about pre-1970 buoy data being almost useless for this reason. But when the results show decreases in RH at every pressure over a period of decades, then explanations about this sort of bias could have creep in are in order instead of just saying the data is suspect. I looked at Wang et al., 2002a, but it isn't clear why issues with one model of radiosonde would justify discarding 50 years of data.

I've had a look also, now. Here's the reference for others:

• Wang, J.H. et al. (2002) Corrections of humidity measurement errors from the Vaisala RS80 radiosonde – Application to TOGA COARE data, in J. Atmos. Ocean. Technol., vol 19, pp 981–1002.

I don't know what you mean by "discarding". This is data that is used as best they can, and people are continuing to try and improve the error corrections. This is hard; and it means quite simply that we don't have a good picture of relative humidity trends before satellite records became available. It doesn't mean that data is simply discarded. The thing is that you have to know the problems with what you are using and give due caution to the inferences you might draw. And for times when you've got better data, then you prefer conclusions based on better data.

Note that the papers trying to measure feedback magnitudes are not trying to project back into the past without valid data. They are simply using the period when there IS good data available sufficient to help constrain sensitivity. There's no good reason at all to think that sensitivity was drastically different before we had satellite humidity measurements; only that we don't have the kind of detail in that time that would allow useful constraints to be inferred.

As for corrections: note that they work better in the lower troposphere than the upper troposphere. For more detail... and for the comparisons indicating the superior reliability of satellite data, see:

• Soden B.J. (2004) An analysis of satellite, radiosonde, and lidar observations of upper tropospheric water vapor from the Atmospheric Radiation Measurement Program in J. Geophys. Res. Vol 109, D04105, doi:10.1029/2003JD003828.

Cheers -- sylas

 

[sylas]

There is a new paper just out on heat content of the world ocean. I'll summarize the results as I understand them, in the context of numbers from earlier in the thread.

Since writing that, I have asked Gavin Schmidt about it, at his blog. Gavin was a co-author of Hansen et al 2005. In the responses I find that Hansen has indeed reduced his estimate in recent lectures, and draws a plain distinction between measurements and models. See, for instance, a talk given by Dr Hansen earlier this year:

Chart 14:

Modeled Imbalance: +0.75 +/- 0.25 W/m²
Ocean Data Suggest: +0.5 +/- 0.25 W/m²​

Now, the ultimate question: can we stabilize climate? We would need to restore the planet’s energy balance. The underlying imbalance (averaging over short-term fluctuations) is probably close to 0.5 W/m².

—Air Pollutant Climate Forcings within the Big Climate Picture, Talk given by J. Hansen at the Climate Change Congress, “Global Risks, Challenges & Decisions”, Copenhagen, Denmark, March 11, 2009​

Note that the error bounds here are quite large. He is quoting model numbers here that range from 0.5 to 1.0, whereas in Hansen 2005 the range was 0.7 to 1.0. Personally, I expect that as the dust settles we'll end up with a number a bit less than 0.5. That's not an expert opinion, but an amateur guess based on reading quite a number of different papers working towards nailing this down. I think nearly everyone agrees that the major part of this imbalance corresponds to heating within the upper 700m of the ocean; and the data on that is consistently hovering around 0.25. So it will be more than 0.25, when you consider contributions to stored energy on land and in the deep ocean, but I doubt it can be as much as 0.5.

This is by no means settled business, but it is converging towards a solution. Direct measurements of the heat content have been plagued by instrument problems, and that's still not cleared up.

The new paper is:

• von Schuckmann, K., F. Gaillard, and P.-Y. Le Traon (2009), http://www.agu.org/pubs/crossref/2009/2008JC005237.shtml, J. Geophys. Res., 114, C09007, doi:10.1029/2008JC005237

Abstract:

Monthly gridded global temperature and salinity fields from the near-surface layer down to 2000 m depth based on Argo measurements are used to analyze large-scale variability patterns on annual to interannual time scales during the years 2003–2008. Previous estimates of global hydrographic fluctuations have been derived using different data sets, partly on the basis of scarce sampling. The substantial advantage of this study includes a detailed summary of global variability patterns based on a single and more uniform database. In the upper 400 m, regions of strong seasonal salinity changes differ from regions of strong seasonal temperature changes, and large amplitudes of seasonal salinity are observed in the upper tropical and subpolar global ocean. Strong interannual and decadal changes superimpose long-term changes at northern midlatitudes. In the subtropical and tropical basin, interannual fluctuations dominate the upper 500 m depth. At southern midlatitudes, hydrographic changes occur on interannual and decadal time scales, while long-term changes are predominantly observed in the salinity field. Global mean heat content and steric height changes are clearly associated with a positive trend during the 6 years of measurements. The global 6-year trend of steric height deduced from in situ measurements explains 40% of the satellite-derived quantities. The global freshwater content does not show a significant trend and is dominated by interannual variability.​

What this means is that they've looked at data from the Argo floats, which we've mentioned in the thread. The paper looks at seasonal variations, and also finds a strong increase in heat content over the 6 years of the study period.

In the paper, the heat content is given in units per square meter, and I think this is for the world ocean; not the whole surface of the Earth. The study is "gridded", which means that they are looking closely at different areas; and also different depths, and finding patterns of change. They also obtain totals by integrating over all regions.

Over the six years 2003-2008 inclusive, there is a trend in total heat content for the upper 2000m of ocean corresponding to 0.77 W/m² +/- 0.11. Since this is over the world ocean, to compare with the numbers used previously in the thread we need to multiply by about 0.7; the proportion of the Earth's surface than is ocean. This gives 0.54 W/m² +/- 0.08.

These figures mean that global warming has been continuing over this period; and that when the ocean heats up enough to shut off the energy flow, there will be another 0.54 W/m² of forcing expressed as a surface temperature, just based on what changes have already taken place up to this point.

In the conclusion of the paper, they compare with Levitus et al (2009) which was introduced in [post=2186603]msg #2[/post] of the thread, and discussed in following posts. Joel, in [post=2188093]msg #6[/post], mentions the fact that the ocean heat content numbers from Levitus et al (2009) have a sharp rise around 2003 and then flatten out a bit. This is seen in figure 1 of that paper, also displayed in [post=2188248]msg #10[/post], but described simply as ocean heat content, which was not strictly accurate. It is the heat content of the upper 700m.

The same plateau effect is mentioned here, in the conclusion of the new paper:

During the 6 years of in situ measurements, an oceanic warming of 0.77 ± 0.11 W m−2 occurred in the upper 2000 m depth of the water column. This number is roughly consistent with the 10-year heat content time series published by Willis et al. [2004]. Although this represents a significant increase in the rate of warming, the updated long-term study of Levitus et al. [2009] shows that the upper ocean (0–700 m) heat content increases to a plateau during 2004–2007, i.e., in the time domain of our study.​

Cheers -- sylas