What Causes the Large Swings in Global Satellite Temperatures?

March 16th, 2012 by Roy W. Spencer, Ph. D.

One of the most frequent questions I get pertains to the large amount of variability seen in the daily global-average temperature variations we make available on the Discover website. From Aqua AMSU ch. 5, these temperatures can undergo wide swings every few weeks, leading to e-mail questions like, Is the satellite broken?

Unusually good examples of these have occurred over the last couple months. In the following plot I took from today, we see January and February of 2012 experiencing two full cycles of temperature variations of about 0.5 deg. C (click for large version):

We have observed this behavior ever since John Christy and I started the satellite-based global temperature monitoring business over 20 years ago.

These temperature swings are mostly the result of variations in rainfall activity. Precipitation systems, which are constantly occurring around the world, release the latent heat of condensation of water vapor which was absorbed during the process of evaporation from the Earth’s surface.

While this process is continuously occurring, there are periods when such activity is somewhat more intense or widespread. These events, called Intra-Seasonal Oscillations (ISOs) are most evident over the tropical Pacific Ocean.

During the convectively active phase of the ISO, there are increased surface winds of up to 1 to 2 knots averaged over the tropical oceans, which causes faster surface evaporation, more water vapor in the troposphere, and more convective rainfall activity. This above-average release of latent heat exceeds the rate at which the atmosphere emits infrared radiation to space, and so the resulting energy imbalance causes a temperature increase (see the above plot).

During the convectively inactive phase, the opposite happens: a decrease in surface wind, evaporation, rainfall, and temperature, as the atmosphere radiatively cools more rapidly than latent heating can replenish the energy.

(As I keep emphasizing, a temperature change is caused by an imbalance between energy gain and energy loss. You can cause warming either by increasing the rate of energy gain, or by decreasing the rate of energy loss. This is how the “greenhouse effect” works, by reducing the rate of radiative energy loss by the surface).

There are observed cloud and radiation budget changes associated with ISOs, as described in our 2007 paper which analyzed the average behavior of the ISO over a 6 year period (see Figs. 1, 2, and 3 in that paper). The main mode of ISO activity is called the Madden-Julian Oscillation.

How Can Rainfall Cause Warming?

I will admit, even as a fresh PhD researcher, the idea that rainfall causes heating seemed counter-intuitive. This is because we are used to the cooling effect of rain when it falls to the surface, which is where we live. But high up in the atmosphere where the rain forms, huge amounts of heat are given up in the process. Just as water evaporating from your skin feels cool, that extra heat is given up when the water condenses back into liquid.

The most vivid example of this process is the warm core of a hurricane, which is heated by the rainfall occurring around the hurricane eye. What actually happens is that the latent heat release within clouds causes localized warming which is almost immediately “relieved” by ascending motion (convective updrafts, which frequent flyers are familiar with).

Because rising air cools, there is little net temperature change in the ascending parcels of air…but all of that rising air also forces air outside of the cloud systems to sink, causing “subsidence warming”. That is where most of the actual temperature increase takes place. The hurricane eye is the most extreme example of this process, where subsiding air gets concentrated in a relatively small region and the temperature can rise many degrees. More commonly though (such as with the ISO phenomenon), the subsidence warming gets spread over hundreds or thousands of km of atmosphere, and the temperature rise is only a fraction of a degree.

33 Responses to “What Causes the Large Swings in Global Satellite Temperatures?”

Toggle Trackbacks

  1. CoRev says:

    Thank you Dr. Spencer. I hope your daughter is doing better.

  2. Svend Ferdinandsen says:

    Good explanation Roy.
    Would you also comment the annual variation and the causes for that? It seems from the picture to be around 2K.

    • The annual variation is due to the fact that most of the land mass (which has lower heat capacity) is in the Northern Hemisphere.

      So, during N.H. summer, the land surface warms more than the ocean surface does in S.H. summer.

      What is interesting is that the sun is farthest from the Earth during N.H. summer, when global-average temperatures are highest.

      • Dave Springer says:

        The tropical ocean stores summer heat and releases it in the winter. This is what causes what was coined “continentality” 200 years ago. Continentality is the fact of continents experiencing a much larger seasonal temperature variation than oceans at the same latitude.

        The ability of the ocean mixed layer to store heat for 6 months is what confounds researchers who find that clouds exihibit a positive feedback. The effect of clouds over the tropical ocean in the summer plays out in the winter. I believe, Dr. Spencer, that in your dust-up with Dessler recently you re-analyzed the data with a longer time constant and found a signal between summer clouds and winter SST that Dessler’s shorter time constant entirely missed.

  3. matthu says:

    Thank you, Dr Spencer – I was one who asked you to provide some explanation and you have done so very comprehensively!

  4. Fred from Canuckistan says:

    Years ago, when I was still teaching geomorph stuff, I just explained what they already knew . .

    To turn liquid water into vapor – like water in a kettle boiling dry, energy has to put IN the water – vapor is the high energy state of water

    To turn the vapor back to liquid, energy has to be released.

    Energy released results in heat.

    They all seemed to get the concept.

  5. Kasuha says:

    Recently there was a paper which pointed out one more form in which energy is released to atmosphere during rain. Aside of releasing heat during condensation, there’s also certain amount of energy lost to drag. Droplets in clouds have certain amount of potential energy given by their height and when they fall, this energy is released in the form of heat. However this energy depends on the cloud height while the energy released directly on the ground is only proportional to their terminal velocity, the rest is lost to drag (and, I assume, also evaporation). I did a crude calculation and it came out that a 10 mm rain (fairly heavy rain) may warm up the air column above it by almost 0.1°C if there was no evaporation.
    I’d like to know what is your opinion on this.
    Thank you.

    • It is true falling rain exerts drag on air, when means downdrafts (true of falling snow, too). Since all kinetic energy from air motions gets dissipated as heat, then it would cause this effect, but I have never heard a quantitative estimate.

      On this subject, it is estimated somewhat over 1% of absorbed sunlight gets converted to the kinetic energy of wind, and (of course) the same amount is continually being dissipated through friction.

    • Dave Springer says:

      Rain doesn’t necessarily dissipate gravitational potential energy in sensible frictional heating. The friction can cause evaporation at the surface of the raindrop and thus become latent heat. Evaporation is the dominant mode of removing heat from the earth’s surface on average. Conduction (sensible) is the least mode and can be ignored in a first approximation because air is a very poor conductor of heat. The ratio of latent to radiative heat loss is pretty much determined by how much water is available at the the surface. Where there’s an infinite supply then radiative loss plays a small role and where there’s little then radiative loss plays a large role. Given there’s a lot more ocean than desert on the surface the overall winner is evaporation.

  6. “Fred from Canuckistan says:
    March 16, 2012 at 2:27 PM”

    Or in the form of equations which can go both ways, see below. The top three are obvious. The bottom three have to follow from the law of conservation of energy.

    ice + heat –> liquid
    liquid + heat –> gas
    ice + heat –> gas (sublimation)
    liquid –> ice + heat
    gas –> liquid + heat
    gas –> ice + heat

  7. Kasuha says:

    Regarding drag, this is the paper that I was referring to:

  8. I thought that if this is correct, then there should be many cases where the lower troposphere warms but the sea surface cools and vice versa. It is difficult to assess this since many trends last a week or two, but we only have monthly values for UAH and hadsst2. However what I plotted was very revealing. I plotted UAH and hadsst2 from 2009 to 2012. I scaled the UAH by 0.5 and offset UAH by 0.2 for easier comparison. I was surprised at the number of times the two went in opposite directions. I counted 16 out of 36 or 44% of the time. I am not a statistics expert, and I could be wrong here, but it seems to strongly corroborate what Dr. Spencer said. See:

    • Steve Hempell says:


      I have noticed and wondered about this too. There seemed to be a inverse relationship between the TLT temperature and the AQUA sea surface temperature. That was until the instrument for sea surface temperature went kaput a while back.

      according to Bob Tisdale the SST took a nosedive recently.

      So Roy, is there a casual relationship here? Ocean releasing heat to the troposhere or is this just my imagination at work?

  9. coturinx says:

    Strengthening MJO they say is a sign of progressing global “warming’… but more likely climate change, for something has to change as co2 rises; if not temperature than what? Could be (and should be) circulation patterns… i seen a paper on da internets that said that qualitative changes to global circulation should already be expected as co2 approaches and crosses 420 ppm.. pretty soon it is. Too bad they were talking about oceanic circulation. But who knows… all we know is that modern climate, the one we are used to, is pretty unusual and transient on geologic timescale… not only climate changes all the time, it also prefers to change elsewhere, not around current ‘equilibrium’… it would probably be good if climate changed not towards ice age, although i, the cool temperate grown person, can’t stand subtropical summers; can’t imagine how you people can live in alabama during summers.

  10. Yet Dr. Spencer states the amount of IR emitted to earth into space has decreased by 1% ,which infact is not the case. Infact ,if anything the amount of IR going into space from earth has been on the increase,despite co2 increasing, since year 2003.

    The icecap website, has a good article on this subject.

    Dr. Spencer shoud know given the many problems with data ,that you can’t make blanket statments like he did recently, about the 1% decrease in IR from earth over the past 40 years.

    Also 1% is to low to show a trend, and if the bands were not already totally saturated ,and the wings as Dr. Spencer maintains not, then we would have had a clear concise down trend in the amount of IR being emitted by earth into space. We don’t, and the data is conflicted if nothing else.

    In addition co2 follows temp. does not lead it.

    Last and most important none of what Dr. Spencer ,ever presents adresses the issue of abrupt climate change. Until then, what he has to say about earth’s climatic system is weak in my opinion.

    I have addressed this issue, many times on these boards.

  11. JSlycke says:

    The strong, cyclic, variation during the first months of this year is indeed intriguing. Although earlier year’s data sometimes also show strong variation, they usually look more random in nature. I do not question Roy’s explanation, but there are likely to be many different drivers for both short term and long term variations. I tested a wild idea to find out if the phases of the moon (http://aa.usno.navy.mil/data/docs/MoonFraction.php), in part, could account for the strong cyclic variation seen during the first months of this year. You find my result here: http://tinypic.com/r/347wkus/5. The visible correlation between AQUA Ch05 and the Lunar phase is strong. I have not checked for correlation in previous years, but believe that there is likely to be some degree of influence also there. It is therefore likely that the moon in some way, besides other influences, contributes to the tropospheric temperature variation.

  12. Dave Springer says:

    Rainfall causes atmospheric warming. It causes surface cooling. The top millimeter of the ocean surface is 0.5C or more cooler than the what’s immediately below it. It’s removing heat faster than the cooler, denser water can sink!

    So we need to be careful about whether we are talking about air temperature or surface temperature. Given that in the big picture the sun heats the ocean and the ocean heats the atmosphere when I say surface I’m talking about the surface of the ocean not the lower troposphere. There’s so little heat capacity in the atmosphere it is like the tail on the dog and the tail of course does not wag the dog.

  13. Dave Springer says:

    “As I keep emphasizing, a temperature change is caused by an imbalance between energy gain and energy loss.”

    In the long term. The reality is we humans live in the short term and in the short term the bulk of the global ocean (90%) lies below the thermocline at a temperature of 3C. If the exchange rate between the surface mixed layer and the frigid bulk of the ocean were to increase for some reason there’s going to be some atmospheric cooling totally disconnected from TOA energy balance.

  14. Vlasta says:

    I explained these swings similar way on a forum with no response .
    Adding to it also falling and then melting snow .
    Thats why temperatures changes to current warming dont support it .
    Werner hit it on the spot .
    The ammount of snow is way too small compare to slight cooling of the oceans .
    What a time AMSR-E packed up .
    Then we could start talking about lagg also …..

  15. Eli Rabett says:

    It might help if you guys areleased all of your secret sauce code for the TLT product. Crowd sourcing software is a useful think as Nick Barnes has shown GISS

    • Richard LH says:

      The rascally Rabbit just needs any answer that shows that Dr Spencer’s figures are wrong somehow. Any answer (right or wrong) will do as he cannot accept that these figures just might be more accurate than other sources which he prefers.

      Unfortunately for him the IPCC seems to agree that Dr. Spencer is closer to the truth than others (when cross compared to other sources).

      Mind you I doubt that any facts will change his mind, that is already made up.

  16. Philip Bradley says:

    These intraseasonal effects on global troposphere temperatures from changes in the hydrological cycle, begs the question, are longer term effects on the hydrological cycle similarly affecting temperatures?

    I’m thinking of aerosols which are known to have large effects locally on cloud formation and precipitation.

    Dr Spencer, what is your view on global trends in aerosols influencing global temperatures through the hydrological cycle, say over decadal timescales?

    I realize this is a complex subject with several inter-related processes at work (latent heat transport, solar insolation, LWR), but I would be interested in a brief opinion.

  17. BenAW says:

    Dr. Spencer

    I’m curious how you compensate for the effects of wind speed over the oceans surface.
    According to this paper:
    the difference between calm winds and higher wind speeds can be as high as 5K for the same position within very short time periods.
    On the other hand the daily (day/night)variation is below 1K if some wind is blowing.


    Ben Wouters

  18. Doug Cotton says:

    It’s good to see you sticking to weather topics, Roy.

    Climate is all too hard, isn’t it.

  19. Bob Tisdale says:

    Steve Hempell, sorry for the delay in finding your March 17, 2012 at 2:54 PM comment.

    It’s not your imagination. The following animation compares SST and TLT anomalies during the 1997/98 El Nino and 1998/99/00/01 La Nina. The graph infills and can be used really as a reference for timing. (It compares East Indian-West Pacific (60S-65N, 80E-180) SST anomalies to scaled NINO3.4 SST anomalies.
    It’s from this post:

  20. Andrew McRae says:

    This might seem really out of left field, but I noticed a new paper in machine learning which shows a new technique for determining which of two free variables is the cause of the other:

    I believe you’ve argued in the past that mixing up cause and effect between clouds and warming has been a major blunder in the global warming issue. When you get half a moment perhaps you can check that paper to see if their technique is in any way applicable to diagnosing forcing vs feedback?
    They assume the causal function is invertible, but I think the radiative transfer fits that requirement.

    Like I said, it may not be useful, but… you never know if you don’t ask.

  21. hr says:

    If the code gets released the rabbit can spend the rest of his life proving Spencer is wrong.
    And the IPCC can think Spencer is so great they can give him part share in their Noble prize.
    But neither of these seem to preclude the sauce code being released. That matter seems to be important for far less trivial reasons.

  22. Katlyn Schwanebeck says:

    I learned something newnow now I’m set for now. Best!

  23. Frank says:

    Very informative.

    Do climate models show this type of behavior in the tropics? (AR4 has a statement that about the absence of the MJO in climate models.)

  24. Bill Yarber says:

    Then explain why power plants use evaporative coolers and A/C systems use evaporative chillers?


  25. Bill Yarber says:

    Dr Spencer

    In your graph, shouldn’t your (cooling) note say (latent cooling…), not (latent heating …) as in the first note?