Wednesday, October 6, 2010

Can downwelling infrared warm the ocean?

Science of Doom has raised this question, which does crop up more frequently than it ought. Doug Hoyt, who should know better, has given it impetus.

It goes like this. Water is opaque to thermal IR, and not a good conductor. IR penetrates only a few microns. Therefore heat from incident IR from above must ...

be all converted to latent heat

or, be reflected, or ...

Land is also opaque to IR, but the idea that IR can't warm it doesn't seem to have the same attraction

Some thoughts below the jump.

Can the ocean emit thermal IR?

This is an equally good question. But it certainly does. In fact, that's how satellites measure SST (skin temperature).

A large part of the insolation that penetrates the sea, to a depth of several meters, is later radiated in this way. It's a big part of the thermal balance. Somehow, that heat is conveyed to the surface, and is emitted by the surface layer.

That pathway is a mix of turbulent advection, conduction in the top few mm, and radiation in the top microns. All those mechanisms are bi-directional. If heat can be emitted, it can be absorbed.

Heat balance

On average, the surface loses heat, by evaporation and radiation (and some conduction to air). Incoming IR does not generally need to be absorbed. It simply offsets some of the emission.

There is some diurnal effect, so that there depth intervals where the heat flow during the day is downward.

Thermal gradients

I've said that heat can be transported to and across the interface. But of course there is some resistance, and this produces a temperature gradient. The temperature differences are noticeable, but not huge. Dr Piotr Flatau has posted some results on Wikipedia:

Note the nearly logarithmic depth scale. The top 10 microns, in which there is nett daytime emission, shows a drop of a bit less than 1°C. There is another drop of about half a degree in the conductive layer. Then the gradient changes, reflecting presumably the upflux of heat from the more rapidly absorbed solar wavelengths. The gradient then diminishes as the source fades, and the turbulent eddies grow in size, increasing turbulent advection.

At night, the gradient becomes monotonic, with the advected flux uniformly upward. But again the temperature difference is less than a degree:


In a very technical sense, the sea is heated by sunlight rather than downwelling IR, as is the land. That's just trivial arithmetic - the sun is the heat source. And the upward IR flux generally exceeds the downward. But downwelling IR does add joules to the sea just as effectively as SW - there is no blocking thermal resistance.


  1. Nick,
    What you are trying to say? Your offering just adds to common confusion of climatists from both sides of AGW. In a very technical sense, you guys conflate the temperature of an object as a measure of its internal energy, and a PROCESS (heat, heat flux) that changes this internal energy. In a very technical sense the surface layer is heated by sunlight during daytime, and may be heated by warmer air or deep layers during nighttime while always emitting IR radiation. The trivial arithmetic is the overall balance of heat fluxes that result in instant temperature dynamics of each individual surface parcel. If the sum of all fluxes is positive at any instant, they "warm" the surface layer, such that it's temperature increases. If the sum is negative, the layer cools off and its temperature drops. It is an inseparable continuous process.

    In a very technical sense you should be asking "can downwelling radiation affect sea surface temperature"? The technical answer is "yes". While the upward IR flux does generally exceed the downdwelling IR, but the only thing that matters in radiative exchange is the flux DIFFERENCE, total flux. The opposite fluxes are not separable. Therefore bigger downwelling IR means less outgoing loss of energy, and therefore the surface parcel will lose the energy that was accumulated during daytime more slowly. On average the layer's temperature will be higher. If you like to conflate higher temperature as being "warmer", then yes, downwelling IR does warm sea surface. This all is so trivial, so I am confused, what is the point of your confusion in this whole issue?

  2. Al,
    I'm responding to SoD's post, in which he refers to a claim that downwelling IR can't "warm the ocean" because it doesn't penetrate. I've cited Hoyt's essay as an example.

    I don't conflate internal energy and flux. On the contrary, I think understanding the difference is vital.

  3. Nick,
    I don't think that Hoyt's arguments are correct. In particular, he argues that excess IR must be "mirrored" because "conduction is too slow". Too slow relative to what? In absolute terms, 10um layer of water has the thermal time constant of 0.7 milliseconds. It is how long does it take to "absorb" an instant temperature fluctuation of water surface. Same goes for the layer of microconvective (1mm) cells under the water surface.

    However, your compilation of subsurface temperature distribution might be quite educational to SST trend statisticians. If you combine your under-surface gradients with over-the-surface boundary layer of air, and mark the difference of temperatures at meteorological standard of "2m above surface" definition for lands, you might find a substantial inconsistency in readings that go into statistics. This bias would likely undermine the entire concept of fine temperature averaging, not speaking that the whole global average is not a proxy for direction of energy fluxes.

  4. Al, I think Flatau's diagram was drawn to illustrate the issues in SST measurement. His "skin SST" is what satellites pick up from IR, and subskin SST is from microwave. Foundation SST is what conventional measurement (engine intahe, buckets etc) gives. And this diagram is intended to illustrate possible biases in comparing different SST measurements.

    As with other temp measurements, though, it's not the absolute value that matters, but the anomaly - and generally, the trend.

  5. No, anomalies matter no more than absolute values; the "anomaly" is a linear operation of subtracting an arbitrary constant. Therefore all averaging operations and derived trends are absolutely equivalent, mathematically. More, if you still want at least a little bit of physical meaning out of global mean and want to associate changes with "warming", one degree of anomaly in equatorial area are twice more "energetically important" than one degree of anomaly in polar areas.

  6. Not an arbitrary constant - a number representing some prescribed local average for each site. It takes out the local variation effect, so it doesn't matter whether the measuring point is "2m above the surface" (hard to achieve at sea) etc.

  7. Sorry Nick, you are exercizing a logic of wishful thinking. Here is some simple mathematics of your "global anomaly":

    Let say we have n stations, 1...n. Each station has a time series record Tn(i), where i is some time index (daily, monthly, annual, whatever). Let's Bn is a somehow defined "baseline" for each station n. The baseline is a constant from arbitary selected subset of timeline i.

    So, we can construct an "absolute" global mean {T} of all station for each time interval i:

    [1] {T(i)} = SUM[Tn(i)]/n, where the SUM goes from 1 to n, global.

    Now, you like "anomalies", The "anomaly" for each record n at time instant i is:

    [2] An(i) = Tn(i) - Bn;

    You like to construct global average of anomalies:

    [3] {A(i)} = SUM[An(i)]/n, correct?

    Now, please follow some primitive equivalent algebraical transformations of this:

    [4] SUM[An(i)]/n = SUM[Tn(i) - Bn]/n = SUM[Tn(i)]/n - SUM[Bn]/n;

    But the first term in [4] is no more and no less than the absloute global mean {T(i)} defined in [1], and the second term is an average of all individual "local" site baselines, so I write SUM[Bn]/n = global_B; Therefore, we have:

    [5] {A(i)} = {T(i)} - global_B;

    As you see, the trend in global mean of "anomalies" differs from absolute data by a single arbitrary constant global_B, "global_mean_baseline".

    So, tell me please what kind of "better" information are you going to obtain from [5] that is not already contained in plain raw data [1]?

  8. Al,
    No, in your example, global_B is not a single arbitrary constant. It is inseparable from the choice of stations. If you changed the mix of stations, global_B would change. It isn't meaningfully global (or single).

    This was relevant to arguments a while ago about the changing mix of stations in GHCN. There are less high altitude stations now. That would raise T(i) in [1], and looks like artificial warming. But it also raises global_B. So the mean in [4] doesn't change - at least because of the altitude issue.

    In more concrete terms, imagine a small country with 4 sea-level stations and one mountain top, 10 C cooler on average. Obviously A(i) is less because of the mountain-top. If your using T(i), you'd have to worry about whether the mountain was representative of 20% of the country etc. But with A(i), you don't, at least re that 10C differential.

    If you replaced the mountain station with a sea-level one, T(i) would drop 2C. The countries climate didn't change, and neither did A(i).

  9. Nick: "Choice of stations...", "If you replaced the mountain station..."

    Why do you think that you have any "choice of stations"? From some latest discussions in blogosphere you should know that the overall sampling density is vastly insufficient to represent climatic average. Even Steve Mosher was inspired (although without any attribution) and made a PDF of cooling stations.

    How you cannot worry how the real mountain area contributes to radiative part of energy balance if you are rooting to detect a signal from global well-mixed and well-distributed "radiative forcing" from CO2 emissions?

    The small country's climate did not change, true, but you have no idea what it is/was in reality, so by manipulating your selection you do not gain any knowledge, because the initial data set was insufficient in first place. By believing in this tricks (in linear functions of averaging) and fudging your data set to get a better stable representation you simply fool yourself.

  10. Al: The small country's climate did not change, true, but you have no idea what it is/was in reality, so by manipulating your selection you do not gain any knowledge, because the initial data set was insufficient in first place. By believing in this tricks (in linear functions of averaging) and fudging your data set to get a better stable representation you simply fool yourself

    It is true that we don't know the temperature as precisely in the small country's climate as we do in countries where there is more sampling.

    But it's simply not true that we "have no idea what it is/was in reality." We're pretty sure for example it wasn't 0K or 5000 K. There is a bigger range of uncertainty introduced because of the sparser sampling, but that can be addressed directly here.

    No data is perfect, but we don't have to stop doing science until it is.

  11. Carrick,
    If your idea that 0K to 5000K range is a "good idea", then our ideas of what is good scientific accuracy with regard to particular applied problem of climate variability are very different.

    You don't have to stop doing science since you did not even started it in the area of data gathering. If you would start doing science by following established scientific methods for data acquisition, it would be a good start.

  12. "That pathway is a mix of turbulent advection, conduction in the top few mm, and radiation in the top microns. All those mechanisms are bi-directional. If heat can be emitted, it can be absorbed."

    Actually, it's mostly night-time convection that takes heat out of the ocean into the atmosphere, through the cooling of the surface via conduction to the cooler atmosphere, the subesequent sinking of the cooled surface waters, and their replacement by warmer water from below which was warmed by the sun during the day. Despite all attempts at redefining the word convection by scientistofdoom, convection doesn't work upside down...

  13. "subesequent sinking.."
    TB, that makes it sound as if the water is otherwise still. It isn't, and turbulent eddy motions dominate. Thermal buoyancy differentials are tiny in comparison with forced motions. Just look at the sea!

    Night-time conduction would have to remove a lot of heat - and then how does the air take it away?

    IR from the ocean is an easily measured (by satellite) fact. It's also readily computed by S-B. It's the main loss.

  14. Nick,

    I read through your comments at SoD, and I just have to ask. Do you accept or reject Minnett's theory for why GHGs heat the ocean?

  15. JCH
    I don't know. It would help if you gave a link, quote or something. I'm aware of Minnett as one of the MODIS people, but I don't know what you mean by his theory.

  16. This is the only place I've read about it:

  17. JCH,
    Thanks, that's an interesting article. And I agree with it. He says
    "Thus, if the absorption of the infrared emission from atmospheric greenhouse gases reduces the gradient through the skin layer, the flow of heat from the ocean beneath will be reduced, leaving more of the heat introduced into the bulk of the upper oceanic layer by the absorption of sunlight to remain there to increase water temperature."
    That's what I'm saying with
    "Incoming IR does not generally need to be absorbed. It simply offsets some of the emission."

  18. Yes, to me one the most interesting article in the RC archives. That's why I thought of the article when I read your comments, which were also interesting. A lot of people do not believe Minnett's theory - too wondrous in its accomplishments, this skin layer. As a lay person (I'm a science bonehead,) it sounds to me like the ratio between SW and LW retained in land and ocean is different - more sun heat in the ocean, and more GHG heat in land. Anyway, thanks for looking at it.

  19. I agree with the fact that where temp measurements are concerned it's the anomaly and trend that matters not the absolute value