RealClimate logo


Ce n’est pas qu’une histoire d’albédo de la glace…

Classé dans: — rasmus @ 22 décembre 2006 - (English)

Traduit par Valérie Masson Delmotte

A partir de nouvelles observations satellitales, un article récent de Francis & Hunterfournit une discussion intéressante sur les raisons de la diminution actuelle de l’extension de glace de mer (banquise) dans l’Arctique. On invoque souvent à propos de la glace de mer une rétroaction positive parce que la glace modifie l’albédo planétaire (la manière dont notre planète réfléchit le rayonnement solaire vers l’espace avant que cette énergie n’entre dans “le système climatique”). Pourtant, l’histoire est plus complexe, car la glace de mer agit également comme un couvercle isolant posé sur la surface de la mer.

Certains effets sont subils, ainsi, la planète perd davantage de chaleur depuis les mers libres que depuis les zones couvertes de banquise (une partie de cette chaleur est absorbée par l’atmosphère, mais les régions englacées ont un climat d’hiver plus continental : sec et froid). La perte de chaleur par l’océan dépend bien sur de la température de surface de la mer (SST, acronyme de “Sea surface temperature”). Les mers libres sont aussi une source d’humidité, par contraste avec les les zones de banquise (parce qu’elles sont froides, et non parce qu’elles sont sèches); mais l’humidité atmosphérique est aussi influencée par le transport d’humidité lié aux vents (advection de vapeur d’eau).

Francis & Hunter ont découvert une corrélation positive entre le manque de glace de mer et le rayonnement infra-rouge incident, un phénomène qu’ils attribuent principalement à l’effet de la nébulosité. Par conséquent, les nuages jouent un rôle important, à la fois en modifiant l’albédo et également en piégeant les pertes thermiques. Francis & Hunter suggèrent que les changements de rayonnement infra-rouge sont plus importants que la modulation du rayonnement solaire incident par les nuages.

Par le passé (IPCC TAR), les modèles de la glace de mer étaient connues pour leurs difficultés à fournir des simulations réalistes. Une partie du problème pourrait cependant provenir du couplage entre l’océan et la composante atmosphérique. Je ne sais pas si ces aspects ont été améliorés dans le prochain rapport IPCC; cependant, Francis & Hunter proposent plusieurs facteurs qui pourraient affecter la localisation du front de glace de mer, comme l’influence de l’océan, du débit des rivières, des effets non linéaires, l’effet des vents et du transport de chaleur, et trouvent qu’ils jouent des rôles différents en différents endroits.

Les régions polaires ont des nuits blanches en été avec un ensoleillement 24h sur 24 (mais avec un angle rasant) et l’obscurité totale 24h sur 24 en hiver. La rétroaction d’albédo de la glace de mer ne peut être active qu’en été, où il y a du rayonnement solaire à réfléchir. L’extension de glace de mer varie aussi selon les saisons, avec une zone d’eau libre plus étendue en été et une couverture de banquise plus large en hiver. Enfin, la glace de mer est déplacée par les vents de surface. Au final, l’effet de changements d’extension de glace de mer sera différent selon les saisons. Une diminution d’extension de la glace de mer en hiver entraînera une perte de chaleur plus grande depuis l’océan libre – cependant, à partir du moment où les températures chuteront suffisamment, il se formera à nouveau de la glace. L’océan libre fournira aussi des conditions plus favorables à la genèse de dépressions polaires (tempêtes). Le refroidissement de la couche de surface et le mélange par l’exposition aux vents pourrait ensuite modifier les courants océaniques et peut-être la formation des eaux profondes. En été, il est probable que c’est l’effet de l’augmentation de l’absorption de rayonnement de courte longueur d’onde qui va dominer.

Une autre question intéressante est la suivante : comment la perte de glace de mer peut-elle affecter la “machine thermique planétaire”, quand davantage de chaleur est perdue aux hautes latitudes? Si la perte de chaleur dans l’Arctique est amplifiée quand la glace de mer se réduit et découvre une surface de la mer avec une température plus élevée? Ceci pourrait être totalement contre-balancé par une diminution de l’albédo local et par une augmentation du transport de chaleur par l’océan et l’atmosphère, ou par une intensification de l’AGW. Il est probable qu’un “effet de serre amplifié par la nébulosité” dans l’Arctique pourrait supprimer une partie de la perte de chaleur par l’océan, et cet effet est cohérent avec les conclusions de Francis & Hunter. Ainsi, l’augmentation de la nébulosité et l’augmentation du rayonnement infra-rouge incident qui en résulte pourraient agir comme une rétroaction positive supplémentaire pour la surface de l’Arctique, en plus de l’effet d’albédo, alors que la perte supplémentaire de chaleur agit comme une rétro-action négative. Quoique subtiles, les résultats de Francis & Hunter ont des implications importantes pour le bilan de chaleur de notre planète.

Sea-ice area from NSIDCLes rétroactions positives qui ont lieu à l’échelle locale pourraient donner lieu à des variations locales plus fortes. En plus de l’effet de la dérive de la banquise sous l’effet des vents, de telles rétroactions pourraient expliquer les grandes anomalies locales de température et l’amplitude de la variabilité naturelle dans l’Arctique. Les régions polaires représentent des régions relativement “petites” en terme de degrés de liberté, une facette qui est souvent négligée. De plus, les fluctuations de température montrent une bonne cohérence sur de larges régions (il y a cependant une anti-corrélation dans la structure en dent de scie associée à la NAO, Oscillation Nord Atlantique). Les profils de moyennes zonales (la valeur moyenne de mesures effectuées par bandes de latitude autour du globe) intègrent la comparaison de zones très étendues autour de l’équateur à des zones plus resserées près des pôles (en tenant compte de la surface des différents “anneaux”). A cause de la corrélation spatiale, cependant, les bandes équatoriales includent beaucoup plus de degrés de liberté que celles près des pôles. De tels profils latitudinaux de moyennes zonales reviennent donc à comparer des valeurs moyennes d’échantillons de tailles différentes (un peu comme comparer des valeurs journalières à des moyennes mensuelles ou annuelles). L’implication de cette analyse est que des évènements comme le réchauffement du début du siècle (dans l’Arctique) ne sont pas aussi significatifs qu’un réchauffement de même amplitude apparent dans un profil zonal aux latitudes plus basses.

72 Responses to “Ce n’est pas qu’une histoire d’albédo de la glace…”

  1. 51
    Jim Cross says:

    Re #34, #50 Amending my previous posts

    My view is GCR theory is more related to explaining thousand year changes; however, I recognize others have extended it to try to account for most of recent warming.

    On the hand, I haven’t received any response to my #34 post about the reason the Ewing-Donn model fell into disfavor, aside from the sea depth issue.

    The thing that has somewhat mystified me about the glacial periods is: “Where did most of the snow/ice come from to form the glaciers?” Did it come from the warm tropics and temperate oceans as global temperatures fell? Ewing-Donn suggests it came from a relatively ice-free Arctic.

    Is there a model somebody has run with the Earth axial tilt changes cycling the Arctic from relatively warm, lesser ice periods leading to heavy snow in the higher latitudes with more ice, frozen, dry periods until the glacial maximum is reached? This seems to be the core part of the Ewing-Donn model.

    Or, is this simply not how it happened?

  2. 52
    gary says:

    The thing that has somewhat mystified me about the glacial periods is: “Where did most of the snow/ice come from to form the glaciers?”

    What is the possibility that an ice-free Arctic Ocean will bring about ocean effect snows, capable of spawning very intense bands of precipitation, which deposit at a rate of many inches of snow per hour?

    http://en.wikipedia.org/wiki/Lake_effect

    [Response: The oceans. But snowfall over the Laurentide region probably came from the Pacific sub-tropics/mid-latitudes with some transport from the Gulf of Mexico, rather than the Arctic. – gavin]

  3. 53
    shargash says:

    If GCRs were responsible for recent warming, wouldn’t atmospheric temperatures by altitude have a different profile than for AGW? And if the atmosphere doesn’t exhibit that profile, isn’t that a big problem for GCR-induced warming?

    [Response:If it were due to an increased level of solar activity, one might expect a heating of the upper stratosphere (the ozone depletion also plays a role, but closer to the tropopause). quite the opposite is oberved. There is also the point that any albedo-based amplification ought to be more effective on the day-side of the Earth, where there is sunlight. It’s would therefore be difficult to explain why, on global average, tendency is that night-time increases faster than the day-time temperature. But the real emporer’s cloths is the lack of trend in the GCR, and as far as I know, lack of trends in the cloudiness… It really should be op to the GCR-antagonists to present evidence for trends in these. -rasmus]

  4. 54
    Jim Cross says:

    #52 Gavin Comment

    Are you referring to some of the studies showing El Nino like conditions during the last Ice Age?

    http://www.earthinstitute.columbia.edu/news/story07_11_02.html

    [Response: No. It’s just a statement about where rain/snow falling in North America was generally evaporated. ENSO conditions may make some difference but I would expect my statement to be pretty close to true regardless. – gavin]

  5. 55
    Jim Cross says:

    #54 Gavin Comment

    Reading the qualifications in your remarks, I’m guessing that the topic of the source of ice age glacial ice hasn’t been studied much either because everyone thinks they already know the answer or maybe because it is hard to study.

    I have always thought the Arctic Ocean to have been mostly covered with ice through most of the last Ice Age, only to begin breaking up with the Holocene. At the same time, I found it hard to understand how the temperate and tropical zones with reduced atmospheric and ocean temperatures during the Ice Age could evaporate enough water to drive enough snow far to the North to form the glacial ice. Hence, the mystery for me about glacial ice.

  6. 56
    Joseph O'Sullivan says:

    #39 David Iles

    I understand what you are commenting about. Scientific writing can be very technical and the writing style, apart from the technicalities, can be dense and convoluted. Legal writers have the same habits. Some in the legal community favor the “plain english movement”.

    Its where the writing is only technical as it needs to be and the writing is in a clearer and simpler writing style. The goal is to make legal literature more accessible to the general public.

  7. 57

    Liked Rasmus explanation ln #39, because it brings up a clear fact that the Polar long nights in darkness have a huge amount of lower latitude advection, primarily because vast areas of heat sources overwhelm the smaller much colder Polar zones. I must add that since the Arctic has been largely warmer now a days, it does create a lower latitudinal warming feedback of sorts. Heat from advection lasts longer in the Arctic, reducing the size of cold air mass formations considerably, making in effect the South of the Arctic largely warmer for much longer periods of time, reducing once upon a time long bitterly cold winters, to occasional irregular “weekend” winters largely dominated by fall like conditions. One can see quite often much smaller cold air masses struggling to linger, this is a sure sign that warmer air is taking over the Northern hemisphere at quite a remarkable pace.

  8. 58
    Adrianne M says:

    Very interesting! I recommend reading http://www.1ocean-1climate.com, where you can find more data on the Arctic-sea ice extent.

  9. 59
    Thomas M says:

    I see very little information regarding the water cycle in the discussions. It seems that as global ocean temperatures rise and the water cycle increases the additional temperature transfer from surface to atmosphere would be very great – not to mention the extra convection. Again, it seems this would be a lagging but extremely powerful negative forcing. If there are specific links regarding this I would be very interested in learning more.

  10. 60

    Re “I see very little information regarding the water cycle in the discussions. It seems that as global ocean temperatures rise and the water cycle increases the additional temperature transfer from surface to atmosphere would be very great – not to mention the extra convection. Again, it seems this would be a lagging but extremely powerful negative forcing. If there are specific links regarding this I would be very interested in learning more.”

    I think any gain in cooling from evaporation would be countered by the increased greenhouse effect from having more water vapor in the air.

  11. 61
    david iles says:

    Thank you Rasmus for your clarification, it helps me understand. I have been reading here for a year or so and just try to get what I can out the posts and comments. I appreciate anything you can do to help me understand what you are saying. And I also see why you would want to use the language you have developed to communicate with others in the field because by the length of your explanation it is obvious that posts could become incredibly long if you had to describe every detail in more public terms.
    Thanks for your effort.

  12. 62
    Hank Roberts says:

    Re “Very interesting! I recommend reading …” — I don’t think I’d recommend that site for anything beyond the owner’s belief that war caused the climate to change. It’s an opinion.

  13. 63
    William Astley says:

    In reply to:
    1) “bottom line is: there is no systematic trend in GCR or other solar indices that can explain the warming over the past ~50 years. -rasmus”

    2) “everything in the GCR-idea and the recent warming hinges on whether there have (has been?) a trend in GCR or not. So far, the GCR-theory proponents have not offered any explanation for how GCR can cause a warming if there has been no trend in the GCR since 1952. -rasmus”

    In reply to 1): See figure 6 in the attached paper that shows there is close correlation between observed the global temperature anomalies, including the last 50 years, to the level and changes in the solar index “ak”.

    From that paper: “It has been noted that in the last century the correlation between sunspot number and geomagnetic activity has been steadily decreasing from – 0.76 in the period 1868-1890 to 0.35 in the period 1960-1982, … According to Echer et al (2004), the probable cause seems to be related to the double peak structure of geomagnetic activity. The second peak, related to high speed solar wind from coronal holes (my comment: For example coronal hole 254 that produced the Dec 16,2006 peak in solar wind, during a sun spot minimum, see attached link to Solar Observation Data), seems to have increased relative to the first one, related to sunspots (CMEs) but, as already mentioned, this type of solar activity is not accounted for by sunspot number. In figure 6 long-term variations in global temperature are compared to the long-term variations in geomagnetic activity as expressed by the ak-index (Nevanlinna and Kataga 2003). The correlation between the two quantities is 0.85 with p< 0.01." (For the years 1856 to 2000) 2005 paper by Georgieva, Bianchi, & Kirov â��Once again about global warming and solar activityâ�� http://sait.oat.ts.astro.it/MSAIt760405/PDF/2005MmSAI..76..969G.pdf

    Solar Terrestrial Activity Report
    http://www.dxlc.com/solar/

    In reply to 2): From Fangqun Yu’s 2002 paper “Altitude variations of cosmic ray induced production of aerosols: Implications for global cloudiness and climate”

    “…Lockwood et al. (1999) have shown that from 1964 to 1996 the strength of the solar magnetic flux, shielding the earth from GCR, has increased by appox. 41% while GCR has decreased by approx 3.7%. The ion chamber measurements (Ahluwalia, 1997) also indicate that the sea level GCR intensity has decreased by 2% from 1979 to 1994. Then GCR intensity decrease is expected to be larger at higher altitudes in the troposphere. From the data available, we estimate that the decrease in GCR fluxes during the past two decades (1979-1999) is 1/3 to 1/2 of maximum variations during the last solar cycle.”

    If the above data is accepted, there is smoking gun evidence that GCR flux changes AND solar changes could be responsible for a PORTION of the recent observed global temperature change (1950 to 2005) in addition to a portion of the temperature changes observed for the period 1900 to 1950.

    The question as to what are the likely climatic mechanisms which could have enable GCR and solar activity changes over the last 100 years, to have likely caused a PORTION of the observed temperature changes in the last 100 years, will be addressed in the next comment.

    [Response:Thanks for your comment. I’m sorry, but I didn’t get the link to the paper to work. However, it’s still problematic that the direct measurements of the GCR do not show any trend, and if the solar index ‘ak’ shows a trend but not the others, then that may suggest that there are some inconsistencies between your indeces. Other indicators, such as 10.7cm flux do nto indicate any trend either. For more discussion on this, please see: Benestad, R.E. (2005) A review of the solar cycle length estimates GRL 32 L15714, doi:10.1029/2005GL023621, August 13. -rasmus]

  14. 64

    Re “From that paper: “It has been noted that in the last century the correlation between sunspot number and geomagnetic activity has been steadily decreasing from – 0.76 in the period 1868-1890 to 0.35 in the period 1960-1982, … ”

    This is a lot less impressive than it sounds. The coefficient of determination is the square of the correlation coefficient, so r = 0.35 means only 12% of the variance is accounted for.

  15. 65
    gary says:

    http://science.nasa.gov/headlines/y2006/21dec_cycle24.htm?list88679

    “Astronomers have been counting sunspots since the days of Galileo, watching solar activity rise and fall every 11 years. Curiously, four of the five biggest cycles on record have come in the past 50 years. “Cycle 24 should fit right into that pattern,” says Hathaway.”

  16. 66
  17. 67
    Steve Sadlov says:

    RE: # 57 – How do you account for what has been happening during the winters, in the mid latitude Western US, since, more or less, the late 1990s? That does not square with the mechanism you have described. Are Yukon and Alaskan cold air masses somehow exempted?

  18. 68
    Steve Sadlov says:

    RE: #66 – The data are from 22 Dec. There have been some (minor) changes in the sea ice pattern since then. Most notably, the Bering Sea ice edge has swiftly moved to the south southwest, and Hudson’s Bay has essentially completed its ice closure and will now remain ice bound until this summer’s melt ensues. One item of note – the necking of ice between Greenland and Iceland has been coming and going (changing with winds) – we might see closure there this season, especially if the pattern that was prevalent in October and November returns.

  19. 69
    Adrianne M says:

    RE #62: The thesis on climate change by naval war is
    based on the assumption that the term climate should
    be defined: as the continuation of the oceans by other
    means, as defined by the author in a letter to NATURE
    in 1992 (Vol. 360, p. 292), see:
    http://www.oceanclimate.de , right column.

  20. 70
    Urs Neu says:

    Re 63

    Trends in solar activity and Cosmic rays in the 1950-present interval depend on the time period chosen. While there is a positive trend from 1960 to 1990 there is a negetive trend since 1980 and no trend at all for the whole period since 1950. This is the case for any geomagnetic activity index, also for
    the Ap index which corresponds to the ak index used in the cited paper. The Ap data can be found on
    ftp://ftp.ngdc.noaa.gov/STP/GEOMAGNETIC_DATA/INDICES/KP_AP/MONTHLY.DAT
    I was not able to reproduce the corresponding part of figure 6 in the cited paper.
    Even if there is a trend for some part of the time period, there should be an overall trend since the 1950ies to explain the temperature increase since that time. The 1977-2006 mean (last 30 years) of the Ap index is the same (in fact a touch lower) as the 1935-1965 mean.

  21. 71
    Hank Roberts says:

    Urs, you wrote “there should be an overall trend since the 1950ies to explain the temperature increase”

    Doesn’t that presume the mechanism, whatever it may be, acts with little or no time lag? There are so many possible correlations here — like the well established one between cosmic ray counts and murder and homicide rates — that it’s hard to figure out what’s connected to what and how.

    Suppose (speculating in the absence of any knowledge) that, say, the sun’s varying magnetic field exerts, say, a varying drag on the rotation of Earth’s core by interacting with Earth’s magnetic field — that might warm the planet, but the lag time for the heat to reach the surface would be considerable.

  22. 72
    Urs Neu says:

    Hank, of course you are right that there could be a time lag. But if there is one, it is highly probable that this time lag always exists because it is a charachteristic of the mechanisms. Thus you should find it in the data. This is the case e.g. for El Nino and global temperature, where there is a strong correlation and a time lag of about 5 months (that’s why we will see the effect of the current El Nino on global temperature mainly this year). However, I haven’t seen any evidence for a coherent time lag concerning any solar parameter, neither sun spots nor cosmic rays nor magnetic field etc, and neither on the short term nor on the long term.

    If there is a time lag of global temperature to changes in solar activity or the magnetic field of more than a few decades, this should be detectable in the time series since the Little Ice Age. However, there isn’t any. The proponents of the solar influence always present the ‘perfect’ match of solar parameters and temperature since the Little Ice Age, with supposed time lags of a few years.

    If there is no trend since 1950, and solar parameters should explain the recent temperature rise, there would have to be a time lag of at least 50 years. But if there is a time lag in the order of 50 years, we should see it in the data of the last few centuries somewhere, but we cannot see it. Of course theoretically there could be a time lag of a century or more so that we would see now the effects of the rise after 1700 or 1800. But you will have a hard time to find corresponding physical mechanisms and to find a forcing-reaction model which reproduces the temperature evolution of the last few centuries in that way. Maybe it is possible but there isn t any evidence, which makes it quite unlikely.