Tsonis and Teleconnections

Tsonis et al


  1. Posted Oct 20, 2007 at 8:47 AM | Permalink

    Equatorial Pacific warm water volume (courtesy TAO) continues to drop. This is consistent with expectations of an ongoing La Nina stretching into the NH winter.

    “Warm water volume” is the amount of near-equator water in the Pacific above the 20C isotherm. It’s a meaure of upper-ocean heat content.

    Somewhat related to this, I saw a reference which stated that cold-water upwelling in the eastern Pacific declined by 10 to 15 Sv from the 1970s to the 1990s. (An SV is a flow of a million cubic meters per second. For size comparison, the North Atlantic thermohaline flow is 15 to 20 Sv, so the decline in Pacific upwelling is not small potatoes.) So, the eastern Pacific upwelling decline, presumably driven by natural multidecadal variation, has reduced upper-ocean cooling by a significant amount since the 1970s. If the multidecadal upwelling variation is driven by PDO-type factors, and if the PDO is converting back towards its pre-1976 state, then we may see a trend into lower tropical Pacific SST, with some impact on global temperature.

  2. Posted Oct 20, 2007 at 9:33 AM | Permalink

    ref: 8 David Smith, I like that topic, it ties in with Tsonis et al and leads to answering John V.’s question, at least partially.

  3. Larry
    Posted Oct 20, 2007 at 11:35 AM | Permalink

    The conclusions seemed a bit nebulous. Am I missing something?

  4. Posted Oct 20, 2007 at 12:20 PM | Permalink

    No, I unfortunately can not stay online long enough to find and post the link to the enso, nao, PDO study Tsonis did that showed synchronization changes in 1910’s, 1940’s and 1970’s. That paper indicated that the 1970’s through 1990’s temperature increase was a function of synchronization of the major oscillations. That warming trend is natural, based on his results, though it may be superimposed on a CO2 trend.

    As the synchronization of the ENSO, PDO and NAO decrease, a cooling trend is expected. As ENSO affects the Pacific up welling, that paper ties in nicely with Davids. If someone can post that link, things will be clearer.

  5. Posted Oct 20, 2007 at 1:52 PM | Permalink

    Now that the internet connection is repaired, Try this link for more details.

  6. Posted Oct 20, 2007 at 1:59 PM | Permalink

    There is an overview paper on ocean multi-decadal cycles here:

    Under the heading “Natural” see Ocean Cycles (of course).

    There’s also an interesting paper on solar cycles where solar experts use methods other than satellite measurements to estimate solar activity.


  7. Posted Oct 20, 2007 at 2:56 PM | Permalink

    My conjecture on the mode shifts of the late 20’th Century is summarized here .

    I use mid-troposphere data because I view the troposphere as a heat exchanger, which can change temperature quickly, whereas the surface (ocean)is a heat reservoir that changes temperature relatively slowly. I use NOAA’s RATPAC because it stretches back to the 1950s.

    What I see is a global tropospheric temperature jump in 1976, associated with the “great climate shift”. (The surface temperature record lagged somewhat this tropospheric change.)

    Things were nearly flat (my red line may exaggerate the possible rise) from the early ’80s to the early ’90s. Then, in my estimation, a second slower shift began, associated with changes in the Atlantic. The upturn is somewhat masked by the big volcano in the early 1990s and then the big swings (La Nina / El Nino / La Nina) in the late 1990s.

    Since this last mode shift (completed around 2001) we are once again into a more-or-less flat period.

    At some points it’s likely that nature will swing things in the cooler direction and we’ll see steps downward.

    Now, in this conjecture there is room for CO2-driven AGW, which may be working “in the background”, making things warmer than they would otherwise be regardless of any natural shifts. I have no problem with that and would be surprised if it is absent. In my view increased CO2 is akin to “fouling” of a heat exchanger, which affects efficiency.

    One point to emphasize is that mode shifts may not only affect heat stored in the ocean but also may affect the efficiency of the troposphere’s heat removal. Mode shifts may affect the global troposphere’s efficiency by a few watts per square meter and in fact it would be surprising to me if the shifts did not affect heat removal.

    Tsonis’s article is fascinating and has the appearance to me of groundbreaking work. His writing bends over backwards to simplify and demystify the topic, which is often a sign of someone who is mastering a subject. I’ll be studying the plots at the back of the article later this evening (probably over a glass of good red wine).

  8. Posted Oct 20, 2007 at 4:26 PM | Permalink

    ref 7 David Smith, I had a Partial diffeques professor that described simplicity as elegance. I found Tsonis’ writing and approach very elegant. There is plenty more by Tsonis I have to read. Searching CA I found a few places where he was discussed, but not many. RC has no mention of his work at all. One thing I have not found yet is negative commentary on the quality of his work.

  9. Mhaze
    Posted Oct 20, 2007 at 6:43 PM | Permalink

    Tsonis has a home page with all his papers in downloadable pdf.


  10. Posted Oct 20, 2007 at 8:49 PM | Permalink

    Ref David Smith #7: David I’m no longer confident that “the surface (ocean)is a heat reservoir that changes temperature relatively slowly”. Steven Schwartz of the Brookhaven Labs has an elegant paper on the subject demonstrating that the time lag in reaching equilibrium is relatively short, in the order of 5 years.

    “The time constant pertinent to changes in GMST is determined from autocorrelation of that quantity over 1880-2004 to be 5 +/- 1 yr. The resultant equilibrium climate sensitivity, 0.30 +/- 0.14 K/(W m-2), corresponds to an equilibrium temperature increase for doubled CO2 of 1.1 +/- 0.5 K.”

    Heat capacity, time constant, and sensitivity of Earth’s climate system. Schwartz S. E. J. Geophys. Res. D, in press, 2007. URL: http://www.ecd.bnl.gov/steve/pubs.html#preprints

  11. Steve McIntyre
    Posted Oct 20, 2007 at 8:59 PM | Permalink

    #10. It argues the point. I doubt that it “demonstrates” the point but this is a detailed story.

  12. Posted Oct 20, 2007 at 10:19 PM | Permalink

    Here’s a list I made some months ago of parameters that noticeably changed circa 1976:

    Pacific Warm Pool SST
    Northern Indian Ocean tropical cyclones
    Temperatures in several Arctic regions (eg, Alaska)
    Lower stratosphere temperature
    Tropopause temperature
    Global geopotential heights
    Aspects of the Hadley-Walker cell outfow
    Southern Indian Ocean SST
    SST in the eastern South Pacific
    PDO Index

    I’m sure there are more – I just stopped making a list.

    Fred, surprisingly the sort of ocean response time frame I had in mind was on the order of 5 to 10 years, not very far from what Schwartz estimates.

  13. Tom Vonk
    Posted Oct 22, 2007 at 9:00 AM | Permalink

    This paper is not so much unlike the thesis of a girl (Byrnes ?) that we saw some 6-8 months ago .
    If I remember well she argued that the global temperature changes were only driven by multidecadal events , spcifically ENSO .
    Here we have a kind of generalisation to the interaction of 5 events .
    Personnaly I have always been convinced for fundamental dynamic reasons that the climate evolution was dominated by deterministic chaos as well as the weather is but with lower frequencies .
    As the deterministic chaos can also be seen like a non linear interaction of a great number of events with different frequencies , this paper goes in the right direction .
    If that thesis is right , follows that whatever will happen in the next decades regardless of
    some trace gas concentrations will NOT be a straight increasing line .
    So the best estimate is that the discrepancies between the AGW global models and the reality will become obvious even for the most obtuse Gavins of this world latest in 10 years .

  14. Posted Oct 22, 2007 at 10:23 AM | Permalink

    Ref: 13 Tom Vonk, Synchronized chaos does seem to have merit looking at past climate relationships. The strength of the tropical network influence on global climate is obvious in the papers which make a great deal of sense. A change in the trade winds due a strong La Nina synchronized with a solar minimum should cause a noticeable cooling trend. Will it be another thirty year trend?

  15. Posted Oct 22, 2007 at 5:26 PM | Permalink

    Is it Global Warming or changing weather patterns?

    Pronounced changes in the wintertime atmospheric circulation have occurred since the mid-1970s over the ocean basins of the Northern Hemisphere, and these changes have had a profound effect on surface temperatures. The variations over the North Atlantic are related to changes in the North Atlantic Oscillation (NAO), while the changes over the North Pacific are linked to the tropics and involve variations in the Aleutian low with teleconnections downstream over North America. Multivariate linear regression is used to show that nearly all of the cooling in the northwest Atlantic and the warming across Europe and downstream over Eurasia since the mid-1970s results from the changes in the NAO, and the NAO accounts for 31% of the hemispheric interannual variance over the past 60 winters. Over the Pacific basin and North America, the temperature anomalies result in part from tropical forcing associated with the El Niño-Southern Oscillation phenomenon but with important feedbacks in the extratropics. The changes in circulation over the past two decades have resulted in a surface temperature anomaly pattern of warmth over the continents and coolness over the oceans. This pattern of temperature change has amplified the observed hemispheric-averaged warming because of its interaction with land and ocean; temperature changes are larger over land compared to the oceans because of the small heat capacity of the former. James W. Hurrell, NCAR 1996

    The 1970 to present is generally used as definitive evidence that AGW has outpaced solar irradiance as the driving force of global temperatures. Tsonis et al noted that synchronization of decadal climate oscillations can be the causes of three roughly thirty year climate trends. Rising NH temperatures from 1910’s to 1940, stable temperatures from 1940 to 1970 and the rise in temperature from the 1970’s to 2000. Until the 1970s, total solar forcing somewhat matched the first two trends. From 1970 to 2000 the Total Solar Irradiation (TSI), well I noticed a few oddities that may be discussed in the soon to be announced solar thread.

    My question is since most of the warming from 1970 to 2000 was in the northern hemisphere, most of the warming per Tsonis appears to be natural in origin, how can it be proven that CO2 may contribute much less than that estimated by the team?

  16. Posted Oct 22, 2007 at 8:07 PM | Permalink

    Here’s Tsonis’ Powerpoint on ENSO and global warming ( link ).

    My read is that, as the globe warms, heat accumulates in the Pacific Warm Pool, which causes the trade winds to increase, which causes greater interannual variation in windspeed, which leads to more El Nino events.

  17. Posted Oct 22, 2007 at 9:02 PM | Permalink

    16 Until there is a shift, then things change. If the trades stay Easterly, the up-welling cools the Eastern Pacific, La Nina cool waters move to the west. A strong La Nina leads to cooling and a shift in weather patterns. If that is out of phase with the NAO and AMO the there is a slight reduction in global temperature rise. If the La Nina is in sync with a negative NAO and/or AMO thing chill out quickly. Each may feedback on the other, it is interesting what happens.

  18. Posted Oct 23, 2007 at 7:35 PM | Permalink

    captainD, the climate-related oscillations which I can recall (AMO, NAO, PDO, etc) are characterized in conventional ways, mostly surface pressure patterns. I wonder if there are yet-to-be-discovered oscillations involving outgoing longwave radiation (OLR). Perhaps there are multidecadal atmospheric modes characterized by persistent regional changes in OLR which, for the globe, alter the net radiation.

    Maybe for instance there’s an East Pacific Oscillation which involves greater marine low cloud coverage and less cirrus, thus increasing OLR enough to affect the global energy balance.

    It seems to me that there’s no reason to think that multidecadal oscillations are necessarily radiation-neutral: perhaps some increase the globe’s heat loss while others reduce it.

    have you come across any such conjecture in your reading?

  19. Posted Oct 24, 2007 at 4:55 AM | Permalink

    David, nothing yet. I have been looking at rainfall reconstructions that I believe would be a good indicator of oscillation patterns. The various monsoons would indicate changes in convection and OLR changes with convection. I haven’t had much time to research the past week though.

  20. JP
    Posted Oct 26, 2007 at 9:12 PM | Permalink

    The $100 question is how much do the major teleconnections affect short term synoptic scale weather patterns. Everyone knows about ENSO and the NOA, etc… But noone has yet crafted a medium range forcaste model that can tie either a plantary teleconnection to a weather forcast. WHen was the last time some organization has accuately forecasted the weather 24 months out?

    Tsonis is on the right track, but so much more needs to be discovered. GCMs that attempt to forecast major weather cycles 100 years from now are worthless.

  21. captdallas2
    Posted Oct 30, 2007 at 2:20 PM | Permalink

    20 jp, you are right. I am running my own models of Noel through my mind right now to determine if I need to move the boat tonight(I hate being on the edge of the cone). Tsonis does seem to be on the right track. I have been reading as much as I can, but haven’t found anything earth shattering other than he downplays AGW somewhat. He and Elsner have one or two peer reviews that address limits of statistical methods that were interesting.

    I haven’t found the paper yet, but there was one that implied a lower climate sensitivity to CO2. There are also implications ( by other authors referencing Tsonis) that the solar influence may be understated. It is tough finding sources that have unbiased ethos.

  22. Posted Nov 6, 2007 at 8:00 PM | Permalink

    Here is an unsurprising, but neat nevertheless, time series. The tropical troposphere temperature anomaly, as derived by RSS from satellite data, is plotted. Also plotted is the ENSO multivariate index (3-month average), which indicates whether the Pacific is in a warm El Nino (high value) or cool La Nina (low value) state.

    The r value for the two is 0.63, not bad.

    Clearly, ENSO behavior in the Pacific plays a major role in tropical atmospheric temperature. In turn, it also plays a noticeable role in global temperature.

    Over the last ten years both the ENSO index and tropical tropospheric temperature have trended sideways. I’ll extend the plots backwards in time to see if that pattern holds.

  23. Posted Nov 6, 2007 at 8:54 PM | Permalink

    For fun I plotted the differences between the tropical temperature and the ENSO index (see #22), to see if the relationship stayed about the same over time. The plot is here .

    It looks like there was a drift in the relationship, 1997 to around 2005. Then there’s what appears to be a step and then flatline behavior for about the last two years.

    Or, it could be an illusion.

    The drift downwards is consistent with the tropical troposphere warming faster than what would be expected due to ENSO alone. Then, the drift stops.

    I’ll plot this back to 1979. We’ll see if the apparent pattern holds.

  24. Posted Nov 7, 2007 at 10:02 PM | Permalink

    Here are several more plots on tropical temperature anomalies and ENSO.

    The temperature anomaly is from RSS’ estimate of the lower troposphere (LT) value ( link ) . It covers 20N to 20S.

    The ENSO multivariate index is derived from pressure, windspeed, sea surface temperature, air temperature and % cloud cover values ( link ) . High index values are associated with the warm “El Nino” while low index values are an indication of the cool “La Nina”.

    Q: Has the ENSO index values increased (= tendency towards warm El Ninos) during the satellite era (1979-present)?

    A: A plot of the index is here . It shows variation but no rise or fall of ENSO over the last 27 years. So, the contribution of changes in ENSO to the tropical warming trend during the satellite era is little or none.

    Q: How has the tropical LT temperature trended in the satellite era?

    A: I’ve plotted the temperature anomaly as well as the ENSO index here . It indicates a relationship of ENSO to tropical temperature, moreso directionally than in absolute values. Interestingly, the two curves line up rather well from 1979-1991 but then begin a divergence around 1992.

    Q: What about that divergence?

    A: A plot of the difference between tropical temperature and the ENSO index is here . (Note that the absolute value of this difference is without meaning – what counts are trends in the relationship.) The plot shows a sideways pattern from 1979-1991 then a switch to a ten-year (more or less) period in which a trend developed (downwards is consistent with warming). In the last year or two it seems like the downward trend has ended, with perhaps sideways movement since 2005.
    To me, this means that the tropics saw little or no net increase in temperature from 1979 to about 1992, then a progressive non-ENSO driven warming until about 2005. This 1992-2005 warming is consistent with Atlantic (AMO) activity, I believe.

    Q: Around 1976 there was a shift from a climate regime in which cool La Ninas dominated to one where warm El Ninos dominated. Over the same period the tropical temperature increased. How much of the circa 1976 temperature rise might be attributable to changes in ENSO, rather than due to CO2/aerosols/solar etc?

    A: Step one is to make a scatterplot of ENSO versus temperature, which I’ve done here . The temperature values are 3-month smoothed and are lagged by 3 months, which seems to about how long it takes for the tropical temperature to fully respond to changes in the ENSO index factors. The period covered is 1979-1991, an era in which the ENSO/temperature relationship was steady (no warming trend).

    It’s a respectable-looking scatterplot.

    Step two is to calculate the average ENSO index value for, say, 1965-1975. Step three is to apply the relationship factor (looks like about 0.03C per unit change in the ENSO index).

    I have not done two and three yet, but suspect we’ll see a noticeable piece of the post-1976 warmup is associated with ENSO activity. Anyone care to venture a guess (10%? 30? 70%?).

    I’ll do the math tomorrow.

  25. Posted Nov 7, 2007 at 10:13 PM | Permalink

    As happens, one nanosecond after I I hit “transmit” I realized that I used the ONI index, not the multivariate index. The results are the same but I’ll cange my labeling, tomorrow.

  26. Posted Nov 8, 2007 at 8:26 AM | Permalink

    #24 RE:

    “Q: Around 1976 there was a shift from a climate regime in which cool La Ninas dominated to one where warm El Ninos dominated. Over the same period the tropical temperature increased. How much of the circa 1976 temperature rise might be attributable to changes in ENSO, rather than due to CO2/aerosols/solar etc?”

    Theodor Landscheidt has demonstrated pretty convincingly that ENSO is solar driven. He is the only person to have accurately predicted El Ninos consistently.

    Email me at facingthefuture@shaw.ca and I will send you the relevant papers.

  27. Earle Williams
    Posted Nov 8, 2007 at 12:06 PM | Permalink

    Re #25

    David Smith,

    A bit off-topic, but I’ve noticed an almost super-human increase in my proofreading skills in that brief window of time after clicking to submit but before the web page has refreshed. It seems there should be an appropriate term to describe this phenomenon. I’m not too familiar with the terminology of behavioral psychology, but something that produces the acronym ‘DOH!’ seems appropriate. 😉

  28. Posted Nov 10, 2007 at 3:41 PM | Permalink

    Sometimes one plays with data and the result goes nowhere in particular. This is one of those occasions. 🙂

    But, before I bury the plots, I thought I’d share a couple of plots. Nothing profound.

    My goal was to answer the question, “What does the tropical temperature trend look like if ENSO (El Nino and La Nina) are removed?” I used RSS lower troposphere anomaly data for the temperature and ONI values to represent ENSO. I used five-month smoothing.

    The method was to estimate the ENSO effect on temperature and back that out of the actual temperature. From earlier work I’d found that a change in ONI of 0.1 unit affects tropical temperature by about 0.03C with a lag of about three months. So, I calculated this ONI effect and subtracted that from the actual temperature value with a three month lag.

    The plot is here , with the blue line representing the “de-ENSO’d” anomaly. The red dashed line is the “un-deENSO’d” temperature anomaly.

    What I see is reduced variation, which means the method may have achieved something. I also see evidence of the El Chicon and especially Pinatubo volcanic eruptions. The Pinatubo looks to have had about 0.3C impact on tropical temperatures, which is broadly consistent with the 0.5C global impact stated in the literature.

    What I also see/imagine is indicated by the green line in this modified plot . It looks to me like tropical temperatures went sideways until the early to mid 90s, then underwent a substantial rise until the early part of the 2000s, then has been sideways since. The timing is somewhat consistent with natural Atlantic AMM/AMO activity (switch to the active phase of the AMO). What I don’t see is a monotonic rise.

  29. DeWitt Payne
    Posted Nov 10, 2007 at 4:00 PM | Permalink


    Interesting. All sorts of things shifted in the mid 1990’s. The Atlantic dipole index went positive. North Atlantic hurricane numbers may have increased. The UAH NoPol anomaly shows a significant increase in slope around 1994 as well. I should look at the RSS data too. The NoPol data seems to have peaked and may be heading down. Graphs and discussion are here.

  30. Posted Nov 10, 2007 at 4:36 PM | Permalink

    Speaking of the Arctic, DeWitt, have you noticed the recent behavior of the Arctic ice report ? No doubt it’s thin ice but, thin or not, the coverage growth has been extraordinary in recent days.

  31. DeWitt Payne
    Posted Nov 11, 2007 at 12:54 AM | Permalink

    If anything, the RSS Arctic temperature anomaly data is trending down faster than UAH. I need to get a copy of a paper on constructing EWMAST charts. The EWMA part is easy, calculating control limits to see if a change is significant is the tricky bit. It’s one of the few downsides to being retired that I don’t have easy access to the literature any more.

    That is pretty rapid coverage growth. If it continues just a little longer, the anomaly will go positive. If it does, don’t expect to read about it on the front page of the NYTimes.

  32. captdallas2
    Posted Nov 11, 2007 at 9:55 AM | Permalink

    Tsonis mentioned that the NAO had a third influence from the Mongolia region I believe. As a tripole, it would be more complicated to determine solar influences I would suspect.

    Solar may be subtle, but its influence shouldn’t be blown off the way real climate likes to insist.

  33. Posted Nov 11, 2007 at 10:20 AM | Permalink

    Re #29 The reference mentions the Latif et al paper in the footnotes. Good paper on multidecadal oscillations and interesting thoughts on the relationship between the NAO and Atlantic SST behavior.

    Re #32 I wonder if there are atmospheric patterns and modes yet to be recognized, perhaps ones which are more important than those currently named. I that that is particularly true when it comes to stratospheric behavior, and stratospheric/tropospheric interactions, and parameters for which multidecadal data is limited, like OLR.

    Solar changes may have a greater atmospheric impact at higher levels so maybe we’re missing recognition of some solar-influenced high-atmosphere modes.

  34. DeWitt Payne
    Posted Nov 12, 2007 at 1:02 PM | Permalink


    What everyone either forgets or doesn’t know is that it is possible for the global average temperature to change even if there is no change in forcing. The simplistic calculation of greenhouse warming compared to an isothermal body underestimates the true situation. The earth isn’t isothermal. A non-isothermal, non-conducting body with infinite heat capacity has a temperature gradient from the equator to the poles. The equator is warmer than an isothermal sphere, but the poles are much colder. That lowers the average temperature a couple of degrees. Infinite heat capacity gives constant temperature with longitude or a zero diurnal temperature range. That is obviously incorrect as well. Lower the heat capacity, the DTR increasses and the average temperature drops some more. It doesn’t drop much for realistic DTR’s and heat capacities, though. For an extreme example, look at the Moon. The DTR is on the order of 300 K and the average temperature is very low. Now lets add some heat transport from the equator to the poles by air and ocean currents. That makes the planet more isothermal and raises the average temperature. Vary the heat transport and the average temperature changes even though nothing else has changed. That’s one factor that makes teasing out the effect of CO2 quite difficult. I’m also pretty sure the GCM’s don’t model any cyclic circulation changes like ENSO either.

    Steve: There’s a good point here that I’ve been meaning to post on for some time. A way of simplifying one’s understanding of this is to view the earth as having two boxes – tropical and extratropical – and you increase the heat flow from the warm box to the cold box, while keeping the sum( T^4) constant, you increase the average temperature. The percentage increase in heat flow needed to cause a 0.4 deg C increase in average temperature (formed from sum (T^1) ) is not as large as one would think. Calculations using integrals are fancier but the point is the same.

  35. mzed
    Posted Nov 12, 2007 at 1:26 PM | Permalink

    #34, I have a hard time visualizing your claim. I see no reason why it would be true. Surely one way to look at the “average temperature” is to sum all the temperature measurements, then divide by the number of measurements (assuming a 2D system and a perfectly regular grid of cells, each with its own measurement–or imagine integrating in a perfect world) This sum does not change with heat transport–it remains the same. And so do the number of cells.

  36. mzed
    Posted Nov 12, 2007 at 1:30 PM | Permalink

    And by “perfectly regular grid”, I mean a grid of equal-area cells, naturally…

  37. captdallas2
    Posted Nov 12, 2007 at 3:02 PM | Permalink

    #35 mzed, I don’t grasp it all, but it makes sense. The tropics have much greater energy than the poles that can be dissipated through local convection or through energy transferred via moving storms, currents and decadal oscillations to higher latitudes. The interaction of the various oscillations have a huge influence on average temperature.

    Even with the PDO, warm low pressure over land versus over ocean would increase average temperature more than the reverse in the northern hemisphere. It is an interesting puzzle.

  38. mzed
    Posted Nov 12, 2007 at 3:14 PM | Permalink

    Well, alright, if we are just talking about atmospheric temperatures, I can see the point. (Though arguably a change in any of the parameters you mentioned could be called a “change in forcing”, but this may just be a quibble.)

  39. DeWitt Payne
    Posted Nov 12, 2007 at 3:19 PM | Permalink


    It’s easy enough to model in Excel. Insolation at any latitude is a function of cos(latitude). Using a sphere with axis of rotation perpendicular to the ecliptic, insolation at the poles will be zero so, in the absence of an atmosphere and assuming an emissivity/absorptivity of one, the temperature at the pole will be 2.7 K (Cosmic Microwave Background Temperature) and increase to 296 K at the equator. The isothermal blackbody temperature is 278.7 K [(1368/(4*sigma))^0.25] and the non-isothermal latitudinal, isothermal longitudal (infinite heat capacity) average on the basis of area is 275.2 K. Finite heat capacity is only slightly more complex. Zero heat capacity gives a maximum temperature at noon at the equator of 394.1 K and a minimum of 2.7 K and an average over the whole sphere of 156 K or -117 C. My simplistic model oscillates at low heat capacity, but for finite heat capacity, the minimum temperature goes up rapidly with increasing heat capacity, the maximum comes down somewhat less rapidly and so the average goes up and approaches 275.2 fairly rapidly. Maximum temperature moves to later in the day and minimum temperature is just after sunrise. With a heat capacity that gives a DTR of 70 C at the equator, maximum temperature (at the equator) is 331 K at 16:50 local, minimum temperature is 261 K at 6:40 and the average is 293 K. Note that the (max + min)/2 average is higher, 296 K, because there is less than twelve hours between max and min. I could put the spreadsheet on esnips, I guess, if you are interested. Here’s a graph of equatorial temperatures for the two non-isothermal cases mentioned.

  40. captdallas2
    Posted Nov 12, 2007 at 3:52 PM | Permalink

    REf 39, Well thank you so much Dewitt Payne! Tomorrow I will be head scratching all day during my charter! I kinda of sorta see your basic idea but I was so looking to simplify, since a few tenths of a degree C may be the ENSO sensitivity. Crap! Back to the books.

  41. DeWitt Payne
    Posted Nov 12, 2007 at 5:06 PM | Permalink

    Steve McIntyre

    I haven’t tried to put latitudinal heat transfer in my simple models yet. It doesn’t look to be all that difficult, I just haven’t done it. It would be nice to have longitudinally averaged global temperatures for each 5 degrees of latitude for comparison. It’s pretty obvious that there is substantial heat transfer, as a chart in Petty’s book on Atmospheric Radiation shows an annual excess of outgoing radiation compared to incoming on the order of 100 W/m2 at the poles and a corresponding deficit of outgoing radiation of 50 W/m2 at the equator. The crossover point from deficit to excess is about 40 degrees latitude N and S. For the isothermal (perfect heat transfer) blackbody model, the deficit at the equator would be 93 W/m2 and excess at the poles would be 342 W/m2 (smaller for a gray body with an albedo of 0.31 like the Earth). The non-isothermal models, of course, have no deficit or excess. The axial tilt means that the poles do receive insolation, unlike my simple model, but I don’t think it changes the basic conclusion.

  42. bender
    Posted Nov 12, 2007 at 6:17 PM | Permalink

    Re #34

    I’m also pretty sure the GCM’s don’t model any cyclic circulation changes like ENSO either.

    Sure they do. At least they try. See the Smith et al. (2007) paper being discussed in “unthreaded #24”.

  43. Curt
    Posted Nov 13, 2007 at 1:26 AM | Permalink

    #34 DeWitt & Steve:

    I’ve been playing with some of these ideas myself for a while. I long ago realized from straight mathematics that in a system that radiates its heat away as a function of T^4, the less the temperature variation in the system, the higher the mean temperature must be. However, I didn’t crank the numbers, assuming that this effect would be tiny.

    However, a few weeks ago, Anthony Watts posted some interesting data on his blog from Jim Goodridge, California’s former climatologist.


    Part of the post (I’ll explain below) got me thinking about this again, so I made a few calculations.

    The first calculation was almost exactly like Steve’s. Divide the world into two zones of equal area: tropical (-30 to +30 latitude), and extra-tropical (+/-30 to +/-90). For a start, assume a temperature of 300K (~27C) in the tropical zone, and 280K (~7C) in the non-tropical zone. Compute 0.5*Trop^4 + 0.5*ExTrop^4 as the normalized radiative power. Next lower the tropical temp to 299K and raise the extra-tropical temp to 281K to keep the same mean temp. The radiate power output is reduced a bit. To get the same radiative output with an 18K difference as the original case with a 20K difference, the mean would have to increase a full 0.1K.

    I next tried a continuous function, with a linearly varying temperature ranging from 268K (~-5C) at the pole to 308K (~35C) at the equator. (Note: not varying linearly with latitude but with equal surface area along the function. The mean would be 288K (~15C). Now reduce the slope of this line so that the range is 270K to 306K, keeping the same mean of 288K. The radiative power output that is proportional to T^4 is reduced, and the mean temperature with this same spread would have to increase almost 0.14K to get back to the old power output.

    Well, so what? What got me thinking about this was some data on “atmospheric angular momentum” that Jim plotted. He claimed that the AAM index he used was related to the ratio of north/south winds to east/west winds. The first thing I noticed about the plot was the dramatic change since the 1970s, when these measurements started — with the trend of steadily increasing north/south winds. In the same period, extra-tropical temperatures have increased far more than tropical, according to virtually all accounts. If there is any kind of cause/effect relationship here, I would think it is in the direction of increased north/south winds reducing the temperature spread, not the other way around.

    Changes in AAM are counterbalanced by the angular momentum of the solid earth, because total angular momentum is conserved. So changes in AAM cause a small but measurable change in the length of a day (LOD), with almost perfect correlation between the two numbers. Because LOD has been measured far longer than AAM, older values of LOD can pretty safely be used as a proxy (dreaded word here, I know) for the AAM index in the pre-satellite era.

    Here’s the kicker (to me, at least). The plot of 20th century LOD looks very much like the 20th century US temperatures. Anthony plotted LOD against a global mean temperature set and showed a half-decent relationship, but the US values eyeball as much closer. (I have not had time to run the numbers.)

    Does this amount to anything? I have no idea. But it sure is intriguing. I haven’t had much time to look into this, and my inquiry to Goodridge has not been answered yet. Anyway, it’s probably time to throw out the idea to see if it gets shot down.

  44. DeWitt Payne
    Posted Nov 13, 2007 at 10:07 AM | Permalink


    And look when the AAM index peaked and headed strongly south, the mid-90’s. Why am I not surprised.

  45. Steve Sadlov
    Posted Nov 13, 2007 at 10:12 AM | Permalink

    If we keep going on this path, I’ll be forced to break out my old Physics books and relearn how to do Hamiltonians. LOL!

  46. Posted Nov 14, 2007 at 10:33 PM | Permalink

    Here’s a time series showing global precipitation anomaly based on NOAA estimates. The heavy blue line represents 1-3-4-3-1 smoothing.

    There’s a hint of decadal-scale variability and maybe something of a lower frequency too.

    One day someone bright will decipher the behavior of this curve as well as that of things like Holgate’s 20’th century sea level plot and the Atlantic decadal-scale variability in SST and atmospheric pressures.

  47. Posted Nov 22, 2007 at 9:23 AM | Permalink

    Here’s an update on the most famous teleconnection, ENSO.

    The cool anomaly in the tropical Pacific ENSO region is now the greatest since the 1980s. There is usually a lag of three or so months before the full effect is felt globally.

    Interestingly, the winds (as indicated by the SOI) have not been particularly favorable to the current La Nina. Favorable winds produce a “pumping action” that helps bring cool subsurface water to the surface along the equator. Recently, though, the winds have become more favorable, so maybe this La Nina has not maxed out yet.

  48. Posted Nov 22, 2007 at 6:29 PM | Permalink


    The reason the SOI has not been playing ball with the developing La Nina as indicated by SST’s in the Nino regions until now is because another less well known teleconnection, the Indian Ocean Dipole (IOD), has been stubbornly refusing to cooperate with La Nina by remaining steadfastly +ve in spite of the urging of the La Nina to do otherwise.

    A +ve IOD means cool waters around northern Australia and warm off Africa, while a -ve IOD means warm waters off northern Australia and cool off Africa – you can find out more about the IOD here:


    The IOD has only recently (i.e. the last few weeks) began to cooperate with La Nina by showing signs of switching into -ve mode, with the oceans warming off northern Australia and the cool anomoly formerly entrenched there moving NE past Indonesia towards Africa – and there have been two remarkably early cyclones (TC Lee, TC Guba) in the Australian region since the IOD started shifting towards the -ve.

    I speculate that the IOD will complete it’s switch into -ve mode over the next few months, and the current La Nina will deepen considerably as the now more favorable winds from the IOD shift are already starting to reinforce it – I also predict a very active cyclone season in the Australian region as a result of the more -ve IOD reinforcing La Nina.

    BTW, I find myself wondering what effect dueling teleconnections from various climate oscillations have on factors said to contain teleconnected climate signals, such as tree rings, as in many areas there will be periods of time where these oscillations will confound each other and times where they reinforce each other.

    For example, in spite of a strong La Nina SST signal in the E Pacific for many months now, drought breaking winter rains that often accompany a developing La Nina failed to materialise in south-eastern Australia due to the stubborn +ve IOD preventing moisture plumes coming over the continent from the Indian Ocean to the NW.

    I speculate that the drought breaking rains for SE Australia will materialise in our summer due to an increase in NW moisture plumes and landfalling cyclones moving S through the inland areas, and continue through autumn and winter with the more favorable -ve IOD wind regime firmly holding the La Nina in place.

  49. Posted Nov 22, 2007 at 7:12 PM | Permalink

    Ref 47 & 48 Interesting. I just glanced at the North Atlantic Oscillation. It seems to be trending downward.

  50. Posted Nov 22, 2007 at 7:17 PM | Permalink

    Carl, thank you and wow! I have never heard of the IO dipole and now have some reading to do on this intriguing topic.

    The 1976 climate shift included some very pronounced changes in the Indian Ocean region and I’m now wondering how this newly-identified teleconnection might have been involved.

    This is the energizing part of CA – learning stuff.

  51. Jimmy
    Posted Nov 22, 2007 at 10:42 PM | Permalink

    David – here’s a good jumping off point for that newly-identified teleconnection:
    Webster PJ, Moore AM, Loschnigg JP, Leben RR, NATURE Volume: 401 Issue: 6751 Pages: 356-360 Published: SEP 23 1999

  52. John M
    Posted Nov 23, 2007 at 6:44 AM | Permalink

    Now THIS is a teleconnection.

    Sounds like we need another UN committee to investigate.

  53. Posted Nov 23, 2007 at 7:06 AM | Permalink

    Re #51 Thanks, Jimmy. Looks like it was identified in 1999 by Yamagata.

    One of the things I wonder is whether long-term changes in this affects the Arabian Sea bulloides proxy – dunno, it’ll have to wait until I finish my weekend travel. I did see a short reference to a relationship to coral records.

  54. Posted Nov 23, 2007 at 7:16 AM | Permalink

    Re #52 Wow, now we’re even destroying the universe via the ultimate spooky action at a distance. I look forward to Lubos’ take on this.

  55. kim
    Posted Nov 23, 2007 at 7:54 AM | Permalink

    At a small enough scale, it is all garbage.

  56. Posted Nov 24, 2007 at 10:55 PM | Permalink

    Here’s a plot of the Indian Ocean (IO) dipole (June-Nov, the northern monsoon season). Basically, positive values mean anomalously warm water near Africa and cool water near Australia and negative values are vice-versa.

    What catches my eye is the apparent shift in the early- to mid-1990s. This corresponds to a shift in the Atlantic Ocean dipole ( see Latif’s graph here , where the black line is the dipole value).

    And, of course, this 90s IO shift corresponds with the sudden mid-90s shift towards greater Atlantic hurricane activity.

    Regardsing global impact, one thought which has crossed my mind is that there was a global warmup from 1995-2000 which was mostly masked by ENSO activity. See an older graphic here . It was a stretch to think that Atlantic activity alone (stronger thermohaline) could account for that global rise but perhaps there was more of the globe involved in the 90s warmup, as indicated by the IO dipole shift.

    Also of interest is whether this is an indication that the Indian Ocean is involved with shifts in thermohaline activity.

  57. Posted Nov 24, 2007 at 11:13 PM | Permalink

    bender, I noticed an apparent 5-year periodicity ( see the green dots here ) in the IO dipole data. This seems to have popped up in some of the hurricane data, too.

  58. Posted Nov 24, 2007 at 11:18 PM | Permalink

    I forgot to mention earlier the apparent change in year-to-year IO dipole variability starting in the mid-90s.

  59. Posted Nov 25, 2007 at 6:14 AM | Permalink

    Ref David,

    The shift in the IOD lags the shift in ’87-’90 temperature shift in Northwestern Europe three to four years. I wonder if that is due to the third NAO influence mentioned by Tsonis?

    The IOD and NAO generally show an inverse relationship except for the ’50s to ’70’s where both treaded downward. Further evidence I would think of another teleconnection whether it is de-synchronization of global oscillations or one or more poorly understood forcings.

    Interesting stuff.

  60. Erl Happ
    Posted Nov 26, 2007 at 2:55 AM | Permalink

    I have been following David Smiths comments with interest. I would like to point out a solar driver for SOI/ENSO and low latitude temperatures in general. The SOI responds to jerks in the aa index of geomagnetic activity. The aa index reflects the solar wind, plasma ejected from the suns corona that has the capacity to alter the density of that part of the neutral atmosphere that is coextensive with the ionosphere. This runs from the Tropopause through to about 100 km in elevation. I conceptualise an atmospheric window, largely cloud free, lying between 30° north and 30° south latitude and moving with the seasons. However, the extent of the cloud free status depends upon the energy driving the Hadley cell and the absolute area of the subtropical high pressure cells. These cells depend in turn upon convection at the intertropical convergence. A small change in energy received can be amplified by other processes including the height of the tropopause and the resultant humidity of the upper atmosphere below the tropopause.

    Up to 1976 a gradually increasing level of geomagnetic activity both at solar minimum and within the period of the solar cycles is associated with spurts of temperature increase in the northern hemisphere (El Nino events). This was especially the case after 1976. In times where the SOI index is cumulatively positive over the solar cycle or two or three cycles together, especially from 1945 through to 1976, when one might expect cooling, the base strength of aa activity and the solar wind was increasing and temperatures tended to stability.

    We are about to enter a period where the SOI index will be cumulatively positive again but this time the aa index is in decline. I don’t know what is responsible for this swing in the cumulative SOI but the cycle is very clear, has been evident since 1880 and is unlikely to disappear now. The current fall in the aa index seems to relate to a decay in the strength of the suns poloidal magnetic field that began with solar cycle 20 and this is reflected in the subsequent toroidal field and sunspot activity that is an expression of the destruction of that toroidal field.

    From 2007 we should see northern hemisphere cooling following a strong fall in water temperatures in the tropical oceans and the mid Pacific in particular. This will feed through to the northern hemisphere in half a year or less. If you look at the way the waters of the major oceans circulate you will see that most of the heat that is gained by the tropical waters is transferred to the northern hemisphere. The nearest equivalent to the Gulf Stream that we have in the Southern Hemisphere is off Madagascar.

    Autumn is currently warm in North America as a direct result of the collapse of convection in the tropics. This is directly linked to the strength of the cooling downdraught at the poles. Compare sea surface anomalies January with October at http://www.osdpd.noaa.gov/PSB/EPS/SST/climo.html. and the effect will become plain. However, I note the current rapid growth of ice in the Arctic. Thanks for pointing that out. Northern Hemisphere Spring will be delayed. This will affect the grain belt. Planting will be late and the frost free season short.

    Nine of the last 12 solar cycles exhibited La Nina activity at close of cycle followed by strong El Nino activity as soon as sunspot activity became apparent. It looks like solar minimum is still 18 months off. This developing La Nina could be a whopper. Thanks for pointing out that sea surface temperature in mid Pacific is now much lower than for many years.

    I hope that this makes sense. I am happy to provide a paper on this if you email erlathappsdotcomdotau Non peer reviewed of course and very unlikely to survive the exercise given the determination of the bulk of the community to blame man for the warming of the planet.

    By the way, there is no correlation between sunspot activity and temperature that I can see. At solar maximum the aa index frequently falls initiating a fall in the temperature of the tropical oceans. About half of the cycles exhibit El Nino’s at solar maximum and half the reverse.

  61. bender
    Posted Nov 26, 2007 at 9:35 AM | Permalink

    Interesting. Maybe there is a similar driver somewhere in the Atlantic, yet to be discovered, that explains the 5y hurricane PAC? Looking at the time-series behavior, one wonders about the applicability of the phrase “deterministic nonperiodic flow”. Questions.

  62. bender
    Posted Nov 26, 2007 at 10:08 AM | Permalink

    David Smith, recall Mann & Emanuel (2006) EOS 87: 233-244. Check Fig 1d. Which frequency range dominates?

  63. John M
    Posted Nov 26, 2007 at 12:52 PM | Permalink

    David Smith, No. 12

    Here’s something to add to your list of things that changed around 1976. It deals with Great Lakes water levels and the balance between precipitation and evaporation (and lots of other stuff too).

    Evaporation made a dramatic switch in 1977 from less than –1 millimeter per year (mm/yr) to almost 4 mm/yr, according to the team’s modeling. “We see some patterns but not enough [to be] consistent with climate change,” Sellinger says, referring to both increased evaporation and decreased precipitation.


  64. Posted Nov 26, 2007 at 4:24 PM | Permalink

    This new paper on solar proxies may help explain some of the oscillation shifts. If solar is now a valid climate driver once again it could get very interesting.

  65. David Smith
    Posted Nov 26, 2007 at 5:16 PM | Permalink

    Re #64 Captain is the link OK? The link seems to circle back to the front page of CA.

  66. Posted Nov 26, 2007 at 8:02 PM | Permalink

    Sorry David I posted the discussion not the paper. This should be the link. It is an odd but interesting way of comparing proxy reconstructions to determine solar influence on global temperature change.

  67. Posted Dec 11, 2007 at 9:41 PM | Permalink

    Here’s some conjecture regarding tropical temperature and ENSO, in this time series . It’s an unusual time series, but interesting.

    In this I took the difference between the tropical temperature anomaly (RSS lower troposphere for 20N to 20S) and the ENSO indicator (ONI SST anomaly, which indicates the SST anomaly in a key part of the tropical Pacific). I call this difference an “index”, one which no strict physical meaning other than a temperature contrast between two entities.

    Generally speaking, when the ENSO indicator (Pacific SST) rises the global tropical temperature rises. Conversely, when the ENSO indicator drops the tropical temperature drops.

    What the chart shows is that for roughly 1980-1995 the global tropical temperature had no net change. The tropical temperature fluctuated year-to-year due to ENSO changes but there was no net temperature change.

    Then from about 1995 to sometime in the mid-2000s, the tropics progressively warmed independently of ENSO. ENSO still produced interannual fluctuations but something else drove a trend of rising temperature.

    Then, over the last year or so, the trend was broken by a small but distinct cooling.

    Now in this exercise I’m reading a lot into small apparent changes which I realize may or may not real. Also, the breakpoints may be off considerably.

    Perhaps the 1995-2005 warming is related to events in the Atlantic.

    What I expected to see was a more-or-less continuous background warming across the full 25+ years. Instead, what I actually see are long periods of flatness punctuated with sudden changes and a decadal period of steady background warming. In other words, what I actually see is a tropospheric temperature anomaly pattern that looks like it involves natural phenonema.

  68. Posted Dec 12, 2007 at 9:29 PM | Permalink

    Further to #67, the thing that puzzles me the most about the global warming of the last 35 years is the “chunkiness” of it. There are some sizeable jumps in temperature (“1976 regime shift”, for instance) over the period which strike me as natural. That’s not to say that CO2-forced warming has not also been happening but rather to suggest that nature may have played a sizeable role.

    I cleaned the graph a bit from #67 link . What strikes me as most interesting is the 15-year period (1979 to mid-90s). The tropics saw El Ninos and La Ninas and volcanoes but pretty much returned to the same lower-troposphere temperature after each event.

    Then in the mid-90s a warmup began, lasting around a decade. What drove this? My suspicion is that it is related to Atlantic behavior, specifically the AMO/AMM phenomena. To look at this, here is the AMM index. It shows a dramatic shift in the mid-90s. This shift is associated with the start of a tropical Atlantic warmup that, in the last couple of years, has leveled or slightly declined.

    The thought occurs to me that, if the tropical warmup is largely natural in origin, then these naturally-warming tropics can also naturally cool.

  69. bender
    Posted Dec 12, 2007 at 10:09 PM | Permalink

    You suggest these patterns are natural, and I don’t disagree. But I can’t agree either. Here’s the problem. The warmer modelers are, post-hoc, developing ever-more inclusive and complex ideas of the nature of the AGW “fingerprint”. It gets so their hypothesis is virtually untestable. When every dip, bump, shift, and anomaly is assumed to carry the sign of the devil, then you’re never going to find evidence to the contrary. As the models are always going to be in a state of flux, I don’t see any end to this witch hunt. There is not a single discordant observation that can not or will not be made to fit “the model” by a re-tuning or re-parameterization.

  70. Mark T
    Posted Dec 12, 2007 at 11:27 PM | Permalink

    It almost seems like trying to find the Taylor series that represents (fits) the function you want. It will work perfectly over the very narrow range for which it is defined, but diverges rapidly outside of that…


  71. Philip_B
    Posted Dec 12, 2007 at 11:41 PM | Permalink

    You can add cloud cover into the mix. While I realize different kinds of clouds have different GHG and albedo effects, no one would dispute that in temperate and warmer regions more clouds means less daytime warming at the surface. Here is the cloud cover trend for Australia. Note the 1975 shift to decreasing cloudiness.

  72. Posted Jan 6, 2008 at 9:41 AM | Permalink

    Here are a few notes on ENSO and tropical temperatures.

    First, the ENSO index I use is the ONI ( Oceanic Nino Index ) which is a measure of SST in a key equatorial Pacific region.

    Side note: a time series of ONI over the satellite era (1979-present) shows lots of variation but there is no upward (warming) trend. That is sort of interesting to me – I expected to see some rise over time in this SST.

    Anyway, first I wanted to see how ONI relates to tropical lower-troposphere temperature anomalies . I chose the period Jan 1979-Dec 1990 because it looks trendless and avoids the problems caused by Mt Pinatubo’s eruption (though it includes the smaller El Chicon eruption).

    A plot of the relationship is here . I was surprised at the robustness of the fit – I’ve rarely encountered r-squares of 0.71 on climate-related data. (Side note one: I marked the El Chicon period points with mauve (my wife says it’s purple, whatever) as those were affected by volcanic cooling.)

    Now to the meat of the matter – I used this fit to “remove” ENSO from the atmospheric temperature over 1979-2007 and see what tropical lower-troposphere trends were present sans ENSO. This is not a sophisticated approach but for my purposes I think it gets me close-enough.

    This “de-ENSO’d” plot is here . I’ve marked the volcanic periods, where temperatures were suppressed for a known cause. What I think I see is trendless behavior from 1979 into the early to mid 1990s. Then, there’s a sizeable (0.4C) rise over a fairly short 5 years or so. Then, a decline begins. I’ve marked these impressions on this plot as “A”, “B” and “C”.

    Assuming these are real, as opposed to random variation, then I wonder

    * why did the tropics stay trendless for 12 years in a period of rising greenhouse effect?
    * why the sharp rise circa 1995-2002?
    * why the decline in recent years, despite rising greenhouse effects?

    Side note: tropical lower troposphere temperature anomalies are correlated with global temperatures, as indicated here where dark blue is tropical and mauve (whatever) is global.

    Part of the evidence offered for AGW is that the post-1975 global temperature rise reportedly can’t be explained by natural factors. To me that implies that A,B and C above are therefore attributable to greenhouse effects. That being the case, what are the mechanisms and do the GCMs demonstrate this kind of behavior?

  73. Erl Happ
    Posted Jan 13, 2008 at 3:03 AM | Permalink

    I have made a couple of posts on Svalgaard 2 (296, 297) that are relevant. I think it would be difficult to remove ENSO effects via statistical techniques. There is a base load of geomagnetic activity and fluctuation on top of that. Base load is concealed. Check out geomagnetic activity for solar cycles 21-23 late in the cycle.

    One can not assume that ENSO effects are climate neutral. The opposite is the case. These are patently not ‘internal oscillations of the climate system’!

  74. Posted Mar 7, 2008 at 5:17 PM | Permalink

    A mildly interesting graph is here . This shows correlations of various climate indices with global mean temperature (GISS version). I did a monthly lead/lag so as to get a feel for whether they lead global temperature on a short-term basis.

    * Indo-Pacific Warm Pool index relates closely with global temperature, moreso than ONI (ENSO)

    * Indo-Pacific Warm Pool index and PDO do not seem to lead or lag global temperature, at least on a short-term basis

    * ONI (ENSO) leads global temperature

    * Solar flux seems to have little short-term relation to global temperature

    If anyone would like to see other indices plotted, let me know.

  75. Posted Mar 7, 2008 at 7:34 PM | Permalink

    This plot is a bit interesting:

    It correlates global temperature anomaly with the Indo-Pacific Warm Pool Index. R-squared =0.78, which is pretty good for climatology.

    I’m unsure which calls the dance tune, but my conjecture is that the ocean rules, especially it has such a large heat content and can release that via evaporation.

  76. John M
    Posted Mar 8, 2008 at 6:45 PM | Permalink

    David Smith

    Your comments above have nudged me enough to finish up something I’ve been dabbling with for a while. I’m sure I’m re-inventing the wheel here, but it’s such a nice wheel. Below, I’ve taken a figure from here. Actually, I captured it a couple of months ago, so it looks like the website has added about a year to the PDO figure. Under it, sized to the same time scale, I’ve pasted a plot of the HadCRUT3 temperature data (again, downloaded a few weeks ago).

    The slopes shown in the bottom figure are calculated for the years 1925-1946, 1947-1976, and 1977-2007. The first two ranges are those assigned to cool and warm PDO phases by the jisao site.

    Note that the very slight warming in mid-century was at a time when major news organs were blasting away about a coming ice age. That period of “cooling” has now been adjusted to show minor warming.

    If one were to estimate an honest warming, I would guess that one should start in 1947, not 1977, since the rate of warming late in the century is about the same as in the previous PDO warm phase. Note, of course, that the slopes can vary a few tenths of a degree/decade, depending on when one chooses to start and stop the PDO phases. I’ve used the timing used by the jisao site to avoid any bias I might have.

  77. Posted Mar 9, 2008 at 8:22 AM | Permalink

    Re #76 Interesting comparison, John M. I can think of several possible hypotheses for the temperature trend change:

    1. Perhaps changes in PDO drive changes in ENSO behavior. These ENSO changes have longer-term, cumulative effects. Repeated warming(cooling) episodes results in heat accumulation (loss) elsewhere in the globe.

    2. Perhaps PDO and ENSO are parts of a larger climate “mechanism” which affects global temperature in ways we don’t yet grasp.

    3. Perhaps the data quality is too poor for use in a comparison.

    4. Perhaps it’s chance

    5. Perhaps it’s some combo of the above.

    I imagine I’ve missed a few.

    My inclination is to suspect it’s a combo of #1 and #2, with a dash of #3. I wonder if a cool-phase PDO increases the trade winds, which increases the likelihood of La Ninas, which reduces the typical amount of sensible heat and water vapor in the troposphere, which increasingly cools portions of the globe away from the ENSO regions.

    If that is correct, and if we are switching to a cool-phase PDO era, then we’ll see a cooling element added to whatever else is happening (AGW, solar, AMO, etc).

    By the way, a very nice website for climatic indices is this one . It gives numeric values for just about any index. It also allows a quick-and-dirty correlation estimation.

  78. Posted Mar 9, 2008 at 8:50 PM | Permalink

    Here’s today’s Teleconnection Selection. It may take a few minutes to focus on the plot:

    ( link in case the image fails to appear. )

    There are three variables. One is the Indo-Pacific Warm Pool (“IPWP”), which is defined as the SST of the region 15S to 15N and 60E to 165E. The second is the global lower troposphere temperature anomaly from RSS (“Global Temp”). The third is an ENSO indicator, the “ONI”, which is a measure of the SST in a key part of the tropical Pacific separate from the IPWP.

    All values are monthly except for ONI, which is a three-month average. The period used is the last ten years (January 1998 thru December 2007).

    I compared these variables two-at-a-time on a lead/lag plot. In this kind of plot one variable is advanced or retarded versus the other, to see how a time lead/lag might affect the relationship. I calculated the r-squared value for each lead/lag comparison.

    For example, take the pink line at a +2 value. What that point means is that, if Global temperature for a month is compared with the ENSO value two months earlier, then the r-squared for that relationship is about 0.44, which is pretty good for unsmoothed data in the world of climate.

    What does all this indicate?

    * ENSO activity (El Nino and La Nina) lead global temperature by perhaps three to six months (see pink line). If a La Nina drops the temperature this month then the full global cooling effect is perhaps three to six months from now.
    Conversely, there’s little indication of global temperature changes causing a change in ENSO (ONI).

    * The Indo-Pacific Warm Pool and global temperature (blue line) appear to have a “quickly-coupled” relationship, suggesting that changes in one affect the other on a very short timeframe. (I would not put much weight on the small difference between the lead and lag sides.) That close relationship is pretty much as I expect, with IPWP being the 500-lb gorilla in this case.

    * What surprises me is the green curve. It appears that changes in ENSO lead to changes in the IPWP in a positively-correlated way. An El Nino begins (say this month) and begins to affect IPWP SST, with the full effect perhaps six months from now. I had expected the IPWP, due to its huge size and hob-gobs of 28C+ water, to be the 500-lb gorilla that pretty much does whatever it wants. Maybe ENSO activity affects the IPWP’s ability to shed its heat, thus warming and thus adding to the ENSO effect on global temperature.

    Maybe that’s how ENSO works – ENSO undergoes a change, which over several months affects the IPWP which then affects the global temperature through its tighly-coupled relationship.

    Some thing to ponder. As always, I’m quite open to critiques of methodology or interpretation – this is an exploration and learning experience for me, not a statement of position.

  79. Kenneth Fritsch
    Posted Mar 10, 2008 at 9:33 AM | Permalink

    What happens to the Global Temp vs ENSO and ENSO vs IPWP correlations beyond the six month lead/lag times?

  80. David Smith
    Posted Mar 10, 2008 at 10:57 AM | Permalink

    Re #79 Good question – they drop. I’ll post an extended chart, out to 12 months, this evening.

  81. Posted Mar 10, 2008 at 8:35 PM | Permalink

    Re #78-79 Kenneth here is the extended chart for the two variables:

    I’m glad you asked because it shows something I had not expected. Also, I found one point in error (green line at -1) which I corrected here.

    ENSO seems to drive both global temperature and Warm Pool temperature with about a six-month delay. I was stunned to see the second set of (notably weaker) humps in the 24- to 30-month period. What’s that about??

    I dunno.

    Does ENSO (El Nino phase) put extra heat and water vapor into the upper troposphere which warms the global atmosphere which makes it difficult for the Warm Pool to shed heat? Are ENSO and Warm Pool behavior coupled (with a delay) and then the warming Warm Pool drives up the global temperature? What’s the connection?

    Above all, I’d like to understand those second humps. Are they an analytical artifact or chance or an illusion created by a “valley” at 15 months? Are they real? If they’re real, what’s the verbal explanation? Things to ponder.

  82. Kenneth Fritsch
    Posted Mar 11, 2008 at 10:05 AM | Permalink

    I was assuming I could say that the suspense was over after you extended the lags, but with those second set of humps I think the plot thickens.

  83. MarkW
    Posted Mar 11, 2008 at 10:57 AM | Permalink

    How can you say the plot is thickening. Those lines are the same width all the way through.

  84. Posted Mar 11, 2008 at 8:16 PM | Permalink

    Mark, I bit on your bait for about 20 minutes before realizing the joke. I hope it’s due just to tiredness and not a sign of neural thinning…

    Kenneth I wonder if the apparent 6-month and 27-month impacts of ENSO in #81, if they are real, also affect our favorite Atlantic hurricane seasons. Every so often a season is expected to see a strong impact from ENSO but maybe that doesn’t work if the 6-month impact goes one way and the 27-month impact goes the other.

    I’m really fascinated by this topic at the moment. If anyone knows of a literature reference that explains the ENSO /Warm Pool connection and the apparent 6-month and 27-month impacts, please post.

  85. Posted Mar 11, 2008 at 8:49 PM | Permalink

    I think I’ll ask the namesake of this thread for references and/or an opinion, as he’s one of the ENSO experts.

  86. Posted Mar 14, 2008 at 8:50 PM | Permalink

    Here are several assorted plots:

    The above shows the RSS lower-troposphere temperature anomaly minus the estimated effect of ENSO. (There’s nothing that I can do to remove the cooling effect of the two major volcanoes but I did circle the periods.) Note how the impacts of the 1998 El Nino and adjacent La Ninas are muted by this adjustment method.

    The visual impression, sans ENSO, is a general rise from 1979 to perhaps 2000-2001, then flatness.

    If I overlay the Indo-Pacific Warm Pool SST on this plot I get:

    Looks like some degree of correlation between global temperature and the Warm Pool temperature. But, which drives which? And, if global temperature somehow drives the Warm Pool temperature, how doitdodat?

    And here are some background plots:

    Indo-Pacific Warm Pool map

    Global correlation between SST and temperature with Warm Pool marked

    Monthly anomaly ranking for Warm Pool (60=warmest in last 60 years, and so forth). This is simply an alternate way of looking at trends in anomalies.

  87. Posted Mar 14, 2008 at 9:10 PM | Permalink

    A better (annual) plot of global SST correlation with global temperature is here .

  88. Posted Mar 15, 2008 at 8:55 AM | Permalink

    If the Warm Pool has a big influence on global temperature, which seems reasonable given its heat content, then an important question is, what drives the Warm Pool temperature?

    Indeed the increasing atmospheric carbon dioxide could play the key role, directly or indirectly. Another possible factor is ocean behavior and net upwelling of cold subsurface water. There is no quick answer to this, to my knowledge, especially in the Indian Ocean region where ocean behavior still has its mysteries.

    Anyway, it’s worth taking a glance at current subsurface ocean temperatures, to see if there’s anything worth watching. The current global subsurface anomaly map is below, courtesy of the Australian BOM:

    Note that, in the tropical eastern and central Pacific, La Nina (blue colors)is alive and well. There is also a hint of the cool-phase PDO pattern in the central North Pacific. The North Atlantic remains in its AMO warm phase, more or less.

    Of interest is the strong blue region in the Indian Ocean. The Indian Ocean subsurface is dynamic with blues and reds coming and going but this patch of upwelling is large and appears to be strong and may influence surface temperature in the coming months.

    Does this blue patch mean anything in and of itself? No. But, we need to watch for evidence of changes in blue-patch extent in the Warm Pool in the coming months and years.

  89. John M
    Posted Mar 15, 2008 at 9:22 AM | Permalink

    David Smith,

    I’ve been dropping your name (positively) at some of the other blogs.

    Hope you don’t mind.

  90. Francois Ouellette
    Posted Mar 15, 2008 at 10:34 AM | Permalink

    #88 David,

    It’s not just the surface temperature that’s important. It’s how deep the thermocline. The ocean’s heat content is huge compared to the atmosphere, and a small change in the stratification of temperatures can mean a lot of heat stored (or removed).

    Lately I’ve become interested in how the oceans affect the CO2 in the atmosphere. Fascinating subject. There is a lot that is only beginning to be discovered, particularly about the influence of biological activity. The so-called “biological pump” removes a lot of CO2. It is thought that without the biological activity of phytoplankton, there would be almost twice as much CO2 in the atmosphere (estimates vary). Yet it also turns out that the biological activity (or productivity) is highly temperature-dependent, sometimes by a factor of 20 for a few degrees. There is also a good correlation between global temperature and productivity that has been found in recent years. All this points to a much more dynamic regime for CO2, something that is in accord with my own little finding.

    I’m currently modeling these oceanic pumps to try and see if I can find parameters that give a good fit with the observed variations in CO2 uptake. I first hypothesized that the oceans can absorb all the CO2 we emit, but I’m not so sure now. Yet you’ve got to reconcile all the facts, not just the convenient ones. And it is an undeniable fact that the CO2 uptake is much more related to temperature variations than it is to human emissions. I think most researchers in the field are pretty much aware of that, and it is tacitly admitted as something that has yet to be resolved.

  91. Posted Mar 15, 2008 at 11:47 AM | Permalink

    John M I’m honored and will tell the boss (wife).

    Francois, one of the off-the-wall thoughts I’ve had is whether phytoplankton change color on centurial timescales, or dominant types or distributions change, affecting ocean albedo. which in turn affects climate. Maybe solar activity plays a role. Far-fetched, I know, but it’s an excuse to read and learn about phytoplankton.

    Pielke Sr has many good articles on oceanic heat content and believes (correctly, in my opinion) that oceanic heat content, if it can be determined accurately, is the better indicator of planetary warming.

  92. Francois Ouellette
    Posted Mar 15, 2008 at 11:55 AM | Permalink

    #90 I should have read the thread more carefully before commenting. Hadn’t noticed the 150 m depth. Also, Shwartz’s paper referred to above is very interesting. Now I’m wondering if I couldn’t correlate CO2 and ocean heat content even better. This would give even more plausibility to the biological pump hypothesis.

    I’m trying to get a hold of Levitus’ data on ocean temperature. I have access to a lot of journals, but unfortunately not the AGU journals. If anyone has his papers, please let me know.

  93. Francois Ouellette
    Posted Mar 15, 2008 at 12:20 PM | Permalink

    #92 Oops, found everything at NODC.

  94. John Lang
    Posted Mar 15, 2008 at 4:40 PM | Permalink

    In the last 6 weeks, there has been a rather interesting large reduction in sea surface temperatures in the Indian Ocean.

    That doesn’t match up with the more recent temperature changes which show rather large increases in temps in the majority of the world, especially Europe, Asia and Eastern Antarctica.


  95. Francois Ouellette
    Posted Mar 15, 2008 at 4:54 PM | Permalink

    #91 David,

    There’s an interesting connection between phytoplankton and aerosols, found by two australian researchers. Don’t have time now to give you the reference (hockey game starts in two minutes), just do a Google search and you’ll find it.

    All in all, the living matter is much more important for climate than one might think.

  96. maksimovich
    Posted Mar 15, 2008 at 6:38 PM | Permalink

    re 91

    one of the off-the-wall thoughts I’ve had is whether phytoplankton change color on centurial timescales, or dominant types or distributions change, affecting ocean albedo. which in turn affects climate. Maybe solar activity plays a role. Far-fetched, I know, but it’s an excuse to read and learn about phytoplankton.

    Dickie and Falkowski use the following example

    One of the clearest examples of biology affecting physical processes is the modulation of upper ocean heating rates by variability in phytoplankton and their associated pigment concentrations and related optical characteristics. Because phytoplankton absorb visible radiation in spectral regions that are relatively transparent for water itself,these photosynthetic organisms are potentially capable of altering the upper ocean heat budget. The extent to which this occurs depends on the concentration and vertical distribution of pigments within the water column, as well as the incident spectral irradiance. Intuitively, one can understand the effect by considering two bodies of water, lying side by side—swimming pools, for example. If one adds black ink to one pool while keeping the second clear, the darker pool will absorb virtually all of the incident solar radiation and become warmer faster. This effect is used to heat water in rooftop solar systems for homes. Similarly, the addition of phytoplankton to the upper ocean can have measurable effects on the rate of heating of the euphotic zone, with consequences for the depth of the upper mixed layer and vertical eddy diffusivity (e.g., Lewis et al., 1983, 1988).

    … This work is supportive of the previous assertions concerning the importance of the penetrative component of solar radiation and more generally biogeochemical processes. For example, it was determined that common values of the penetrative solar flux are about 23 W m−2 at 30 m(the climatological mean mixed layer depth), and thus a large fraction of the climatological mean net air–sea flux of about 40 W m−2. Synoptic scale forcing (e.g., wind bursts) were found to lead to tripling of phytoplankton pigment concentrations and a reduction in penetrative heat flux of 5.6 W m−2 at
    30 m,or a biogeochemically mediated increase in the radiant heating rate of
    0.138C/ month.In-depth analysis of the radiant heating and parameterizations of light attenuation for this experiment are given in Ohlmann et al. (1998)

    Shell et al 2003 (co author Richard Sommerville) draw similar conclusions.

    Phytoplankton alter the absorption of solar radiation, affecting upper ocean temperature and circulation. These changes, in turn, influence the atmosphere through modification of the sea surface temperature (SST). To investigate the effects of the present-day phytoplankton concentration on the atmosphere, an atmospheric general circulation model was forced by SST changes due to phytoplankton. The modified SST was obtained from ocean general circulation model runs with space- and time-varying phytoplankton abundances from Coastal Zone Color Scanner data. The atmospheric simulations indicate that phytoplankton amplify the seasonal cycle of the lowest atmospheric layer temperature. This amplification has an average magnitude of 0.3_K but may reach over 1_K locally. The surface warming in the summer is marginally larger than the cooling in the winter, so that on average annually and globally, phytoplankton warm the lowest layer by about 0.05_K. Over the ocean the surface air temperature changes closely follow the SST changes. Significant, often amplified, temperature changes also occur over land. The climatic effect of phytoplankton extends throughout the troposphere, especially in middle latitudes where increased subsidence during summer
    traps heat.The amplification of the seasonal cycle of air temperature strengthens tropical convection in the summer hemisphere. In the eastern tropical Pacific Ocean a decreased SST strengthens the Walker circulation and weakens the Hadley circulation. These significant atmospheric changes indicate that the radiative effects of phytoplankton should not be overlooked in studies of climate change.

    Changes to absorption and emission of nutrients are also responsive to changes in both the type and spectra of radiation, these inhibit some populations and enhance others.

    Indeed what we can see is the ecological communities of microflora, changing rapidly to meet their changing levels of nutrients and energy, is a Belousov-Zhabotinsky reaction diffusion mechanism operating far from equilibrium in a perpetual state of self-organization (changes to the population as responders. ie adaptive and defensive strategies for say fluctuations in frequency of radiative spectra are present always and everywhere)

    For example, some chemical reactions are provoked only by light of frequency higher than a certain threshold; light of frequency lower than the threshold, no matter how intense, does not initiate the reaction.

  97. steven mosher
    Posted Mar 15, 2008 at 7:01 PM | Permalink

    Re 88:

    “If the Warm Pool has a big influence on global temperature, which seems reasonable given its heat content, then an important question is, what drives the Warm Pool temperature?”


  98. Posted Mar 15, 2008 at 7:10 PM | Permalink

    It would be a hoot if changes in solar irradiance affect phytoplankton which affect net albedo which drive climate. What a thought: phytoplankton, the Hoi Polloi of the Seas, are found to be the missing link in the solar/climate puzzle and thus a master of our future.

    John, something I noticed during our cold January was the large amount of very dry upper-troposphere air in the tropics. That air moistened in February and into March but appears to be drying again, coincident with apparent global cooling in the last week. This is anecdotal but worth watching.

  99. steven mosher
    Posted Mar 15, 2008 at 7:45 PM | Permalink

    re 98. Sponge bob would not be happy if plankton ruled.

  100. Francois Ouellette
    Posted Mar 15, 2008 at 7:47 PM | Permalink

    #96 Not only that. Marine life also drives oceanic currents: Dewar et al. Journal of Marine research 2006

    Ocean mixing is thought to control the climatically important oceanic overturning circulation. Here we argue the marine biosphere, by a mechanism like the bioturbation occurring in marine sediments, mixes the oceans as effectively as the winds and tides. This statement is derived ultimately from an estimated 62.7 TeraWatts of chemical power provided to the marine environment in net primary production. Various approaches argue something like 1% (.63 TeraWatts) of this power is invested in aphotic ocean mechanical energy, a rate comparable to wind and tidal inputs.

    Here’s the phytoplankton-aerosols connection, from R.A. Cropp and A.J. Gabric ‘Evidence for global coupling of phytoplankton and atmospheric aerosols’

    Biological coupling between the ocean and atmosphere may have profound implications for global climate change. Atmospheric aerosols such as dust can directly influence the radiative balance of the earth, but evidence is accumulating to suggest it may also have more subtle effects. Biologically available iron in dust, for example, may stimulate phytoplankton blooms, which then draw down carbon dioxide and emit dimethylsulphide (DMS) to the atmosphere. The latter reacts to form methanesulphonate aerosols (MSA) that may also influence the radiation budget. Such coupling processes have been demonstrated in mesoscale experiments in two HNLC regions of ocean. We present a global analysis of remote sensed data that suggests coupling between phytoplankton and atmospheric aerosols may be more widespread in the worlds oceans than previously thought. Collocation of clusters of covarying phytoplankton and aerosols with physical processes is compelling evidence for close biological coupling of ocean and atmosphere.

  101. jae
    Posted Mar 15, 2008 at 7:56 PM | Permalink

    Fascinating. Maybe that’s also related to subtle changes in Sun, i.e., the postulated solar amplification.

  102. Francois Ouellette
    Posted Mar 15, 2008 at 8:11 PM | Permalink

    #101 What it all means in particular is that all GCM’s that don’t take into account these various interactions (i.e. the great majority, since adding the biosphere is still in its infancy) are likely to be completely useless.

  103. Steve McIntyre
    Posted Mar 15, 2008 at 8:33 PM | Permalink

    Just so you don’t think that this topic’s been ignored – here’s an old post on an interesting article about phytoplankton http://www.climateaudit.org/?p=213

  104. Posted Mar 16, 2008 at 12:18 AM | Permalink

    Re 90

    Thank goodness someone is looking at the powerful biology of the oceans WRT AGW. Does your thinking extend as far as an explanation of the C12/C13 ratio? A biological response to warming oceans (warmer water causes stratification and less nutrient flow to the upper levels — http://www.agu.org/pubs/crossref/2008/2007GL031745.shtm –) could well be a switch by certain phytoplankton from C3 to C4 metabolism, which would draw down an anomalously large amount of C13 and leave a C12 atmospheric signal.

    One of the more persuasive arguments for the increase in atmospheric CO2 being anthropogenic is the light C signal. However, I am unconvinced that this ratio is a good proxy for the amount of ACO2 remaining in the atmosphere. Has this relationship been studied? I am unable to find graphs of varying isotopic ratios which makes me puzzled — surely this most important evidence for the A in AGW must be laid out in great detail somewhere easily accessible.

    I’m rather pleased about Polivna’s result — I predicted it a year ago.

  105. MarkR
    Posted Mar 16, 2008 at 5:15 AM | Permalink

    Francois Ouellette. I’m amazed more attention isn’t being paid to your paper

    Click to access 0802.3130.pdf

    a brilliantly simple, but statistically significant analysis of what drives CO2 levels.

  106. MarkR
    Posted Mar 16, 2008 at 5:21 AM | Permalink

    David Smith. Your correlation of SST and Temp seems to show a pattern of high correlation along a line latitudewise in the southern hemisphere. Perhaps the trade winds are the mechanism for the dispersal of the heating or cooling? Likewise the Monsoon Winds for the Indian Ocean area?

  107. John F. Pittman
    Posted Mar 16, 2008 at 5:39 AM | Permalink

    Francois Ouellette, I agree with MarkR #105. I am wondering if you could use the model as proposed and substitute the anticipated cooling trend that is expected to last till 2011 or 2013. If your model is correct, or close, it would be nice to see a prediction that would meaningfully beat the IPCC and other’s models.

  108. Francois Ouellette
    Posted Mar 16, 2008 at 9:59 AM | Permalink

    #105, 107 John, Mark,

    Thanks for the kind words. Since I wrote that (2-3 weeks ago), I have read on the topic almost day and night, and cooked up my own model of the oceanic pumps, based on and analogy with interconnected capacitors, which I’m still working on. I realize I was quite naive in my speculations, but that’s not always a bad thing.

    One thing is for sure, anthropogenic CO2 emissions do exist, and that CO2 is going somewhere, so any model has to take that into account. But the one thing that has been very frustrating for me is that, just like the derivation CO2 sensitivity, the modeling of the CO2 buildup is nowhere to be found in the literature. I know it must be there somewhere, but I’ve done extensive searches for that one paper, without success. There were early conjectures, by Bolin, and Keeling, but at that time, so little was known about the oceanic pumps that their results can only be meaningless. And some of the early modeling papers can’t be found on the net. So every paper seems to take it for granted that anthropogenic CO2 just accumulates and that’s it. Yet nobody has a decent explanation for why half of it is taken up by the oceans, and not more nor less. And nobody ever mentions the simple correlation between uptake and temperature that I have been using.

    Yet one major enigma in that field is why CO2 concentration fell so low during the glaciation. There are tons of paper attempting to explain that, but no definitive theory. Just goes to show that we don’t know everything about the carbon cycle. What I suspect is that those icecore reconstructions are just false in the end. I know it’s a big chunk, but if you rather take the stomatal index evidence (and obviously both techniques cannot be right so it’s one or the other), then you find that CO2 could have been both higher and lower during interglacial, from, say, 240 to 340 ppm , as opposed to the “stable” typical 280 ppm that is always cited. So the difference between glacial and interglacial would then be much smaller (240 to 180 instead of 280 to 180). At the same time, it is clear that interglacial temperatures are tightly bound. Holocene temperatures fluctuate mostly within 1 deg.C, maybe even less. So this means that somehow, the biological pump (this is phytoplankton using CO2 and then sinking to the bottom with it when they’re dead) is much more sensitive to temperature, and causes large changes in CO2 concentration. It also means that CO2 sensitivity is not very large, and that there are feedback mechanisms constraining temperatures within that range. So far, there is ample evidence in support of such a viewpoint, and not much to contradict it. The icecores are the only sticking point. That’s when you realize that much of AGW relies on them. Maybe that’s why the IPCC completely ignored the stomatal index results in their chapter 7 (they are not mentioned, and are not in the references. That alone is enough to discredit the entire report for me.).

    So what I hope to be able to do with the simple model, is to find a set of parameters that reconciles an ensemble of observations in a coherent way. At least that would form a testable basis. I don’t pretend to overthrow everything, I’m just having fun and learning. But if I do get results that make some sense, I might give it a try for a peer reviewed paper.

  109. John F. Pittman
    Posted Mar 16, 2008 at 4:49 PM | Permalink

    #108 Ouellette: You said “I realize I was quite naive in my speculations, but that’s not always a bad thing. ”

    No it is not! As I was told as I pursued my engineering degree: “with 11 parameters you can describe an elephant”. The last time I saw (read) such scientific disagreement was over PCB’s, DDT, and eggshell thickness! {{{For the auditors who understand that our history is one shot, one chance:::the explanation was DDT, but subsequently PCB’s were not used and in areas PCB’s were used, the eggshell thickness was negative to PCB’s, not DDT(as best one shot wonders could tell): nevermind all the 3rd world “darkies” who died by this (DDT=dead birds) assumption. Note in areas of DDT the shell thickness was the same.}}}http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6V78-4G4MMDX-5&_user=10&_rdoc=1&_fmt=&_orig=search&_sort=d&view=c&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=e69f558fcff476e95ea5fbb4956cfd28http://ambio.allenpress.com/perlserv/?request=get-document&doi=10.1639%2F0044-7447(2004)033%5B0495%3AROORIG%5D2.0.CO%3B2&ct=1

    Hope these come through correctly. There are many more!

    Time to reread “Silent Spring” and understand what ASSUMPTIONS really mean.

  110. John F. Pittman
    Posted Mar 16, 2008 at 5:45 PM | Permalink

    #108 Francois Ouellette says “Yet one major enigma in that field is why CO2 concentration fell so low during the glaciation. There are tons of paper attempting to explain that, but no definitive theory.”

    This is only because as posted on another thread (with 4 college bound kids and 3 nephews wanting to play on my computer, I can’t take time to find it:) ) there are two physical relationships that (think that if you and I are wrong, ALL Global Climates Models are wrong): THE NUMBER ONE ASBORBTION (for our earth) of CO2 is calcium, then water; both with inverse relationships to temperature. Note there is an inverse relationship for CO2 absorbtion (not ADsorbtion as some in IPCC confuse).

    There is a virtual dearth of half-step expositions on global warming. Let’s do a simple series.!. Using the physical characteristics of just CO2, that we will use, CO2 and all gasses are inverse to temperature (and don’t discout pressure)!!.

    We will discount pressure, and get rid of half-steps, at first!! LOL. Just to make it understandable (eat your heart out, you Climate scientists that refuse to seek wisdom from statisticians!!)

    As the world goes into glaciation, more and more CO2 will be absorbed (not adsorbed) into ALL water forms. Water in the air, water running in and over land, water under the land, water in the oceans….ALL water will adsorb more CO2. So #108, ask yourself what is the nature of this absorption? It is >>>>>tah dah>>>> a SELF LIMITING FUNCTION!!! It has boundaries! Obviously (despite what Mann and computer modelers postulate), all functions are self limiting, according to the limits of the control volume and the physics of the assumptions.

    What does this mean? Now we will do some half-step estimates!! As temperature goes DOWN, more and more CO2 is ABsorbed into all water forms. At the same time, more and more calcium is soluable as the temperature goes down. Calcium does not like to be ALONE

  111. John F. Pittman
    Posted Mar 16, 2008 at 6:46 PM | Permalink

    #108 Francois Ouellette CONTINUED

    Now we will examine calcium. Calcium is “NEAT”. The defintion of CRUD is clear river unexplained deposits, before wiki and online defintions. Unfortunately the net has little on this, shows my age, LOL.

    Calcium gets more soluable with decreasing temperture! Google calcium scale and hardness, but note IPCC are not engineers or boiler people!! Hardness is by defintion (talk to your pool or boiler persons) Ca!!! equivalents. Calcium LOVES to be joined with CO2. AND get this, Calcium and CO2 are inverse with respect to temperature.

    So if you ADD them together, negative to temperature as you posted, it is not 1 to 1, but a 2.5 exponential (what is the log decay when all the world is an iceball?) to 1 relationship. I leave the math to tamino since he seems to be mathematically correct, but somewhat wisdom challenged.(Hint 1+1+(1 influences the other 1, assume 50%)=2.5).

    Please note, all RC addicts need not wonder how this will impact GCMs since the science is done. OF course it means BY DEFINITION, increased heat releases more CO2 past a tipping point, because by THE IPCC DEFINITION, when man first made a fire he condemned the world to catastrophic warming.http://itsgettinghotinhere.org/2007/08/17/climate-tipping-points-get-scarier/http://www.independent.co.uk/environment/climate-change/too-late-to-avoid-global-warming-say-scientists-402800.html http://www.celsias.com/2007/10/10/has-the-tipping-point-been-reached-already/

    Too bad nature has limiting functions such as “water vapor”!!! YOU can’t make it more than 100%!!???!!!!

    “Damn those 100% limits”. That heat actually has to “HEAT” something.

    The above rant is to project the fact that as far as we ALL know, functions are self limiting. For the philosophically and physically challenged, the universe did NOT enter into an ever increasing tipping point with respect to the “Great Bang”. Do a mass and energy balance on this using accepted IPCC assumptions on how CO2 and other GHG’s (and all other mass and energy assumptions) are positive feedbacks.

  112. John F. Pittman
    Posted Mar 16, 2008 at 8:09 PM | Permalink

    Francois Ouellette says:#108 “Yet one major enigma in that field is why CO2 concentration fell so low during the glaciation. There are tons of paper attempting to explain that, but no definitive theory.”

    This is a strawman as far as I can tell.

    As temperature decreases more CO2 is absorbed, not adsorbed, into all forms of water, within reason(you cannot put 150% into a 100% boundary layer, nor -110% of 100% boundary layer). The major “sink” of CO2 is calcium. CO2 and Ca++ have an inverse relationship to temperature. As the temperature drops, more CO2 is absorbed and more Ca++ is available to absorb CO2. As they interact, the most likely asssumption is a 2.5 exponential function ((1+1+(50% of 1+1 at .5 decay)=2.5)

    Model temperature vs CO2 at a 2.5 exp decay. Note that thus the physics of CO2 versus temperature indicates that the global sensitivity should be the 2.5 root of the Arrhenius not the 2.5 power that IPCC assumes.

  113. Posted Mar 19, 2008 at 6:39 PM | Permalink

    Here’s some conjecture, and a prediction with which to test it, for fun:

    Satellites monitor the water vapor content of the upper part of the troposphere. A current example of the water vapor imagery is here . This shows the air high above the tropical Atlantic. The browns indicating very dry air while the whites and blues (mostly in the non-tropics) indicate moister air.

    This image is “browner” than normal in the tropics, having increased its “brownness” (dryness) over the last week. The last time this happened was in December, when the global temperature began to cool (in anomaly terms).

    By January there was considerable brownness (dryness) around the tropical NH, coinciding with the quite cool month. Then the tropical dryness lessened into February and the first part of March, coinciding with NH warming.

    If the upper tropical drying has resumed, and if it indeed affects temperature, then we should begin to see its fingerprint on global temperature on this Hovmoeller temperature/latitude/time plot .

    What I currently see on this plot is evidence that March to-date has been warmer (in anomaly terms) than January and probably February. But I also see a hint that cooling has resumed in recent days.

    The length of time, and thus the evidence, is modest. I’ll watch, and report back here in about two weeks, as to whether the dryness continued and global conditions indeed cooled. I think they will. We’ll see.

  114. Posted Mar 22, 2008 at 12:34 PM | Permalink

    Here, for variety, are several plots related to US precipitation.

    I’ve been exploring the topic in extreme precipitation (basically heavy rain) and whether it is increasing as the GCMs expect. At first I thought this would be a straightforward exercise, as what could be simpler than putting a cylinder outside to catch rain? But, there are many wrinkles and the deeper I get into it the more I encounter adjustments, interpolations and assumptions.

    There are several studies ( here for example ) in recent years which report increases in extreme precipitation in the US. But since they mention the scarcity of early records (especially in the western US and prior to the 1930s-1950s drought decades) I decided to do a simple but limited check using data at-hand.

    The data at hand are the lists of top-10 rainiest days of various US sites. An example of a list is Saint Louis, Missouri . I chose 60 sites across the US which spanned at least 100 years and I tried to space them geographically ( map ). These sites are predominantly cities which have been around a while (by US standards). As expected, data in the western US is much scarcer than in the older eastern US which creates a few gaps like eastern Idaho.

    Anyway, I tallied the top-10 rainiest days for the 60 locations and plotted them by the year in which they occurred:

    What I see are hints of a heavy-rain era in the early part of the 20’th Century then a dry period then perhaps a return to a heavy-rain period in the 1990s. That’s important because some of the studies start in the dry 1930s-1950s and because even those which extend back 100 years have to address large gaps in geographical coverage (esp the western US), changes in measurement devices and locations, etc in that early heavy-rain era. There’s a lot of best-guesswork.

    And, any trend in heavy-rain record events is hard to detect.

    I also did a plot of US temperature and rainfall (NCDC data) alongside my 60-city review:

    What I see is are the familiar temperature pattern, a trend of increasing precipitation and little if any trend in the 60-city heavy-rain plot. I also ran the basic correlation checks and found a lack of meaningful correlation between any pair.

    Now this is a crude and limited exercise, with warts like the use of only 60 locations and the presence of some 19’th century values, but what it does is to indicate that any report of increasingly extreme rain events needs to say clearly how the authors accounted for the historical gaps and discontinuities.

    (Side note: in reading about the topic I learned that windspeed plays a nontrivial role in how much precipitation falls into a collection device. I figured that is true for snow but not so much for rain, but apparently even for rain it can be in the single-digit percent range, which sounds small but can be important in climatological studies.

    Anthony Watts’ study shows that the co-op thermometers have moved towards human habitation in recent years – I wonder if the rain gauges have made a similar move. If they have, then there may be increasing numbers located in less-windy sites, near buildings and vegetation, which might increase the amount of precipitation they collect. That increase would have little effect on heavy-rain events but might introduce an unexpected variable into long-term precipitation studies.)

  115. Chris Knight
    Posted Apr 4, 2008 at 5:39 PM | Permalink

    Walter Munk’s paper on Tidal Friction puts the dissipation of energy for the Earth-moon system at estimates of 1.1 x10^19 ergs/sec or 2.4X10^19 ergs/sec, or 1.1 TeraWatts or 2.4 TeraWatts, depending on the calculation method. This is a very small forcing, but does not happen at a uniform rate (viz. delta LOD), nor is split in a constant ratio between land tides or ocean tides, since neither the land nor the ocean topologies are constant or uniform (cryosphere, atmospheric pressure systems, sea level change, plate spreading, volcanism etc.). The energy of ocean tides is dissipated in the shallow coastal water boundary with the land (or ice), and will cause localised warming in response to changing currents, lunar orbital eccentricity (apsidal cycle) and Lunar orbital tilt precession (nodal cycle, earth’s polar nutation), together with reinforcement of the solar tide with the annual perihelion and aphelion events.

    Perhaps some models incorporating well known tidal data could be run to determine where and when extreme ocean tidal friction events may be expected in regard to generation of Ocean Oscillation events. May find that Tidal forcing is a force to be reckoned with.

  116. Chris Knight
    Posted Apr 5, 2008 at 5:24 AM | Permalink

    115 (me): Correction, that should be 1.3 and 2.7 in the first sentence.

  117. Andrew
    Posted Apr 24, 2008 at 3:37 PM | Permalink

    Chase, T.N., R.A. Pielke Sr., and R. Avissar, 2007: Teleconnections in the Earth system. Encyclopedia of Hydrological Sciences, M. Anderson, Editor-in-Chief, John Wiley and Sons, United Kingdom, 2849-2862.

    Click to access CB-48.pdf

    Seems relevant. I’m not misunderstanding the vocabulary here, am I?

  118. Posted Sep 24, 2008 at 6:15 AM | Permalink

    I noted Gavin posting this comment at RC in a thread related to SLR,”NAO and PDO have little or no correlation to the global mean temperature, and your implication that you have 20 year predictability in either is laughable.”

    This would seem a bit over the top given that Tsonis make a very good case that Gavin’s et al take on the role of natural climate variation on GMT is at least under estimated.

    Since the thread is closed at RC, I am hoping that Eric or Gavin will respond here. While I doubt Tsonis reads CA, a discussion on this topic by all three parties would be interesting.

    • bender
      Posted Sep 24, 2008 at 9:59 AM | Permalink

      Re: captdallas2 (#118),
      The idea of 20-year predictability IS laughable. What’s the problem?

  119. Posted Sep 24, 2008 at 12:16 PM | Permalink

    It may not be as laughable as thought. Tsonis’ Dynamical paper indicated that natural cycles or oscillations do impact global temperature and may have predictable trends. At least that is what I took from the paper. Predicting whether the next trend will be up, down or sideways I would say is laughable at this time. But once a NAO/PDO/ etc. synchronization/de-synchronization trend starts, the trend duration appears to be reasonably predictable. The problem is determining when a trend starts. We cannot say we are currently in a downward trend, basically because we can’t determine a trend until we are nearly out of the trend. However, there are indications we may be in a sideways or downward trend, since roughly 2000.

  120. Posted Oct 14, 2008 at 6:53 PM | Permalink

    The cold “bubble” continues to rise towards the surface in the tropical pacific (near 140W, 10N). Here’s the temperature profile, with the anomalous bubble circled:

    The anomaly plot is here:

    As indicated, water 5C below normal is just 10 meteres below the surface and 10C below normal is 20 meters down. This anomalous region has been rising towards the surface for several months. Will it make it? Dunno.

    This is simply a curiosity so far as I know but I do wonder what causes such heavy cool water to rise so strongly. I also wonder if these cool pockets become more frequent and widespread during a PDO cool phase.

    • John M
      Posted Nov 3, 2008 at 6:25 PM | Permalink

      Re: David Smith (#121),

      Looks like the bubble may be about to break the surfcae.

      In addition, note the intense cool anomaly right off the Central America coast.


  121. Geoff Sherrington
    Posted Oct 15, 2008 at 3:58 AM | Permalink

    I asked our locals about a similar cold upwelling off New South Wales a couple of years ago and was told it was from eddy formation but that it involved complicated maths like Coriolis and Navier-Stokes, inferring they did not want to discuss it. Lasted several months at least, maybe more, I do not recall the start/finish dates. Was perhaps 200 km diam depending how you measured it.

  122. David Smith
    Posted Oct 15, 2008 at 6:58 AM | Permalink

    Re #122 I wonder about the effects of a 200km cold upwell in the tropical Pacific, especially if the cooler water rises far enough to begin wind-driven mixing with the surface water. A 5 C or 10C upwell which approaches the surface could mix with surface water and lower the mix by 1C or 2C. That doesn’t sound like much but, in the case of the tropics, it could make a noticeable difference in water vapor pressure and evaporation. In the specific case of this upwell or bubble a cooling might affect the extent of low cloudiness, which in turn could affect the radiation balance and wind.

    My impression, which is based in observing the SST maps for the eastern Pacific in recent years, is that there are occasional upwells or cool bubbles which do reach and affect the surface and lower the SST by 5C or so over a small region, but they are spaced out in location and frequency. Perhaps they become more frequent and clustered in a cool-phase PDO. It’s something to occasionally ponder and watch as this cool-phase PDO progresses.

  123. Posted Oct 21, 2008 at 7:34 PM | Permalink

    Here’s a simple question which puzzles me. If anyone has the answer, please post. Thanks in advance:

    The question concerns ENSO and the sun. Below, I have plotted the SST change in ENSO region 3.4 (5N to 5S; 120W to 170W). Each point is the temperature of that month minus the temperature of the prior month. (A footnote is that I used 3-month averages centered on a month rather than the single month value, for ease of viewing, but that has no effect on my puzzle.)

    As the plot shows, there is a peak in warming when the sun crosses the equator in September/October, every 12 months, with only a few exceptions.

    If I plot the monthly averages then I get this:

    What it confirms is the clear peak in ocean surface warming circa the September equinox, which seems sensible due to the strong sunlight. My puzzle is that there is not a similar clear peak with the March Equinox, as the sun passes over the equator headed south. Why not?

    Perhaps it is related to asymmetric ocean currents or wind strength. Or, something else. I think it’s a somewhat important question, because the mechanism might vary over time, affecting ENSO and thus global temperature. The effect would be minor, but minor counts when we’re looking at tenths of a degree C.

    • jc-at-play
      Posted Oct 22, 2008 at 2:46 PM | Permalink

      Re: David Smith (#124),

      Wild guess – could it have something to do with Earth’s varying distance from the sun?

      [Paradoxically, perihelion occurs almost simultaneously with winter solstice, just a couple of weeks later.]

  124. Bob Koss
    Posted Oct 21, 2008 at 11:49 PM | Permalink

    Just speculation here.

    The north pacific ocean is almost landlocked at around 60N and the south pacific hits land at around 75S. More cold water in the south to be heated and moved toward the equator by the gyre.

    September to March the sun is south of the equator trying to heat a greater volume of water.

  125. Posted Oct 22, 2008 at 4:23 AM | Permalink

    I found an labeling error in my second plot. It doesn’t affect the core question (asymmetry) but it does switch which equinox is muted:

    Here’s the corrected image:

    • Posted Oct 22, 2008 at 11:41 AM | Permalink

      Re: David Smith (#126),

      I’d think of deep water upwelling. Maybe estearn winds are sesonaly stronger during the NH summer.

    • Erl Happ
      Posted Oct 23, 2008 at 10:16 AM | Permalink

      Re: David Smith (#126),

      Many people over a very long period of time have maintained that the solar wind as reflected in the aa index of geomagnetic activity affects terrestrial temperature. I hypothesize that it does so by thinning the atmosphere over the tropics. There is a strong coupling of the solar wind with the Earths magnetosphere about the time of the equinoxes. Because of the orbital factor which yields a 7% increase in the general strength of the suns irradiance as seen at Earth on January 3d the temperature peak due to the coupling effect in both September and March is displaced towards January. The effect seems to be greater in September so we see maxima in October or even later. The effect can be seen in temperatures at the top of the stratosphere at 1hPa here: http://www.cpc.ncep.noaa.gov/products/stratosphere/strat-trop/

      It is also reflected in 10hPa temperatures over the warmest part of the globe. That data is downloadable from here:http://www.cdc.noaa.gov/cgi-in/Timeseries/timeseries1.pl.
      I show 10hPa temperature data with sea surface temperature data here: http://i249.photobucket.com/albums/gg220/erlandlong/SSTand10hPatemp0-10N.jpg

      So, you can see that the effect is present at the surface.

      I have been rabbiting on about the mechanism to Leif Svalgaard for some considerable time, but without any other result other than to offend his sense of what is good science. Briefly, ultraviolet radiation excites ozone in the upper troposphere where dry air lies over cold ocean, warming the air, lowering relative humidity and reducing cloud cover. The penetration of ultraviolet radiation into the upper troposphere depends upon the depth and density of material that it must traverse. That’s where the solar wind has an impact. The secondary effect of a change in the solar wind is UV penetration and that is what changes the temperature of the air.

      If one looks at geomagnetic data there is very strong variability in March and September. If one looks at temperature data for all levels of the atmosphere down to 500hPa (I have not gone further) that same strong variability occurs in March and September. You would expect strongest temperature variability in January perhaps when the sun is closer, but no, its in March and September.

      So, the sun changes the temperature of the globe by changing the cloud cover.

      At the moment we have virtually no sunspots so less very short wave radiation, particularly in the very short X ray spectrum. We have very little geomagnetic activity. All upper layers of the atmosphere are very cool as can be see here:http://discover.itsc.uah.edu/amsutemps/execute.csh?amsutemps The SOI is rising as can be seen here:http://www.eldersweather.com.au/climimage.jsp?i=soi

      Finally, according to Livingston and Penn there will be no sunspots by 2015, which at the current rate of progress should be just past the peak of Solar cycle 24, if it ever appears at all.

  126. Pat Frank
    Posted Oct 22, 2008 at 8:06 PM | Permalink

    Take a look at the yearly track of the sun across the globe here, David. It’s not symmetrical about the equator. The southern ocean receives its heating across a much wider area than the northern ocean.

    In the north, the sun’s path actually traverses itself. I’d suspect this radiant asymmetry has something to do with the effect you’ve noticed. Certainly, a combination of more focused insolation plus a very different basin geometry and volume might well cause a different upwelling and incurrent dynamic, and so a difference in equatorial heat content.

    All hand-waving, of course, but still, the north-south solar asymmetry is really there.

  127. Posted Oct 23, 2008 at 4:42 AM | Permalink

    Thanks for the posts. I suspect, based on what you’ve mentioned, that solar asymmetry plays a key role and likely accounts for March exceeding September. It makes sense.

    I made similar plots for the Atlantic. March is stronger than September but not by nearly as much as in the Pacific 3.4 region. That leaves me wondering what accounts for the apparent dampening of the Pacific September behavior.

  128. Posted Oct 23, 2008 at 9:20 PM | Permalink

    Re #130 This is a minor and inconsequential matter, I know, but it’s fun to wander down these alleys sometimes.

    Here’s a plot of the Nino 3.4 SST change along with insolation (scaled) at the equator:

    Insolation is stronger in boreal spring than in boreal fall. In spring there’s a noticeable rise in SST as insolation increases. In fall the effect on SST is dampened. Why?

    • Erl Happ
      Posted Oct 24, 2008 at 4:05 AM | Permalink

      Re: David Smith (#132),
      Nino 3.4 SST change is the result of heating south of the Equator and that is strongest in January-March.

      Waters north of the equator are warmer. Extra energy results in evaporation not surface temperature increase. Waters south of equator are cooler. The extra energy goes less into evaporation and more into surface warming.

      I think its actually a very important matter to discuss. It relates to where the energy gets in that is responsible for the annual cycle of temperature change across the equator and the exaggerated departure from normal that we call an El Nino event.

      So, the question to ask is: What is responsible for the loss of high altitude cloud that lets the sun in to warm the cold waters south of the Equator.

  129. Posted Oct 24, 2008 at 4:52 AM | Permalink

    If I searched an answer to the differential response in sst changes at the Equinoxs, I’d look firstly at air-ocean circulation changes. It’s not easy for me to accept that the Sun radiation has all that direct influence, or that its energy is partitioned in a differential way between sensible and latent heat without other evidences.

    Earth is not a simmetric system and a different circulation, six months apart, is not to be ruled out. Of course a different pattern in air-ocean currents can have a remarkable impact also on cloud cover distribution.

    I’ve not been searching for the answer, It’s just my supposition.

  130. Posted Oct 24, 2008 at 12:36 PM | Permalink

    Here’s a bit more data.

    First up is a plot of the SST and insolation around 5N, the northern edge of the Nino 3.4 box:

    What I see is boreal Spring behavior in which SST peaks several months after the peak in insolation. That seems reasonable. But then, in boreal Fall, the SST declines despite the rising insolation. The visual impression is that the Fall insolation slows the rate of deline of SST but that’s about it.

    The second plot is SST and insolation around 5S:

    Here, too, the boreal Spring looks reasonable while the Fall shows relatively little response to the increased insolation.

    I’m beginning to wonder about trade wind strength and humidity, as perhaps the controlling factor in boreal Fall is evaporation rather than insolation.

    On a side note, I seem to recall that predictions of ENSO made in January have poor accuracy while predictions made in May have pretty good accuracy. I wonder if, and how, this March/April peak in SST relates to that change in predictability.

    • Erl Happ
      Posted Oct 24, 2008 at 5:55 PM | Permalink

      Re: David Smith (#136),
      What is the origin of the insolation data? Is this insolation as measured at the surface and at the Nino 3.4 longitudes? Is that data available for anywhere else in the Pacific? How is it obtained?

      In June the global land surface temperatures peak and cloud cover is down, the major factor in that warming. I find it hard to believe that insolation falls at 5°N unless this is a local event (east west movement of cloud) affecting this particular longitude. The ITCZ tends to run pretty consistently north of the Nino3.4 region all year. But the volume of moisture and cloud changes a lot depending upon the temperature of the waters.

      Your sea surface temperature data shows a shifting peak between March and June. If we look at the tropics as a whole (all longitudes) peak sea surface temperature is achieved in March (10-20S) April (0-10S) May (0-10N) and June 10-20N) and there is another smaller peak in the northern hemisphere in September (10-20°N) and October (0-10°N). The size of the peak to be achieved relates to the strength of the change from the previous month.

      So, the riddle needs to be looked at in terms of basic causation. Why is there a May-June peak in Sea surface temperature in the northern tropics 0-20°N that is unrelated to the passage of the sun. Could it be related to flow factors? Could the waters be acquiring their heat elsewhere rather than in the Nino 3.4 region?

      Then, there is the question of what local meteorological factors are affecting the temperature in the Nino 3.4 region and of this I have no knowledge. But if the cloud systematically increases convection draws in warm air and that heats the surface water.

  131. Posted Oct 24, 2008 at 2:17 PM | Permalink

    Tradewind strength looks like a factor, perhaps the main factor, in the SST asymmetry. That, I presume, is due its effect on east/west water motion and on evaporation.

    Here is a plot of zonal wind (expressed as an easterly wind) in Nino 3.4:

    The weakest wind, and thus weakest evaporative cooling and weakest movement of surface water from the cooler east, is in the boreal Spring, corresponding to the highest SST. And vice-versa for the boreal fall and early winter.

    Here, for fun, is the same plot but with the famous 1997 El Nino year added:

    The drop in easterly wind in 1997 is quite apparent.

    The next question is – why the seasonal change in zonal wind? This is all elementary to those knowledgeable in ENSO but it’s fun, and a learning opportunity, to explore the basic data.

    • Posted Oct 24, 2008 at 3:43 PM | Permalink

      Re: David Smith (#137),
      as I suggested in 127, you have showed that stronger eastern winds are likely the cause of the Equinox different behaviour.
      But you suggest that the stronger winds directly advect colder surface water from the Est and increase the evaporative cooling. Though this is true I insist on my supposition in 127. Since what keeps cold the equatorial sea water in the Pacific is primarily the deep water upwelling, which is controlled by the intensity of the easterly wind, I’m convinced that upwelling is tha main reason.

      Regarding your second question, i.e. why the seasonal change in zonal wind, I’d check the position of the ITCZ, whose movement through the seasons is not simmetric with respect to the Equator.

  132. David Smith
    Posted Oct 24, 2008 at 4:31 PM | Permalink

    Re #138 Paolo, you are right – upwelling in Nino 3.4 is the dominant source of the cooling.

  133. Erl Happ
    Posted Oct 24, 2008 at 7:59 PM | Permalink


    On a side note, I seem to recall that predictions of ENSO made in January have poor accuracy while predictions made in May have pretty good accuracy. I wonder if, and how, this March/April peak in SST relates to that change in predictability.

    Here is my interpretation for what it is worth.

    Blue line is average of monthly values of SOI since 1884. Pink line is difference between highest and lowest value for each individual month. You will know that blue line descending is El Nino.

    Blue line shows that the constant tendency is for warming in March- May. Between 10°N and 10°S the warm waters have their maximum extension eastwards between March and June as can be seen on a hovmoller. This is no doubt associated with slackening of the Easterlies as you show above. The strong variability at this time I put down to the solar wind influence and its effect on upper troposphere cirrus. If cirrus disappears in the east and south east Pacific as the Northern Hemisphere warms it will create convection in the east and slow the easterlies. Peak coupling with the magnetosphere is March and the orbital factor enhances the strength of radiation based on a January maximum. There is another period of lesser but significant variability in the SOI in August based upon the relative strength of northern hemisphere warming in June but it is muted because irradiance is weakest then.

    Given that SOI variability tends to be weaker Dec-March and there is a La Nina tendency at that time my guess is that colder waters coming in from the south, in part due to Antarctic melt at that time, are driving up air pressure in Tahiti. So, we have two drivers of surface winds between March and June. If the cirrus in the east stays in place there will be cooling in the Nino 3.4 because the easterlies stay strong. If the cirrus evaporates there will be warming in the Nino 3.4 zone as the Easterlies slacken. The northward movement of cold waters from Antarctica assists the disappearance of cirrus by reducing humidity. Cold waters tend to have less cloud above them.

    The external variable that affects this internal seasonal gyration is the influence of the sun on the upper troposphere cirrus in the rain shadow of the Andes.

    Ocean circulation and upwelling work together. The ocean circulation is there and it should not be ignored.

  134. Erl Happ
    Posted Oct 24, 2008 at 8:27 PM | Permalink

    Whats your take on Ed Berry’s analysis at http://weatherclimatelink.blogspot.com/2008/10/update-bear-atmosphere-stimulus.html

    I note this statement: SSTs have cooled across the equatorial Indian Ocean as a response to intense rainfall. Significant cooling has also been occurring in the region of the equatorial Dateline to Indonesia, as a result of a strong trade wind surge. In fact, the 29C isotherm has shifted back to 160E. The latter may be the initiation of a coupled ocean-atmosphere response toward La-Nina, which has been the character of the global circulation for at least the past several weeks.

    The analysis look to me like standing over a pot of marmalade jam boiling on the stove and examining the rate of circulation of the peel.

    I wonder what is happening to the cirrus in the south east Pacific?

    • Posted Oct 24, 2008 at 10:40 PM | Permalink

      Re: Erl Happ (#142), Very interesting website you linked. I, too, wonder whether La Nina, or at least La Nina-like effects, will expand globally this boreal winter. Earth’s boiler, the tropical Pacific and Indian Ocean, continues to sputter.

  135. Posted Oct 24, 2008 at 9:57 PM | Permalink

    Re #140 Hello, Erl. I plotted the top-of-atmosphere insolation using NASA’s calculator . Obviously that does not address what actually makes it to the surface.

    Speaking of data sources, do you know of a source that provides mean Pacific ITCZ locations for various times of the year?

  136. Posted Oct 25, 2008 at 12:12 PM | Permalink

    Here’s a comparison of Nino 3.4 easterly windspeed for 1950-1979 and 1980-2007:

    The plot is derived from NCEP reanalysis data.

    A look at annual data indicates that the shift occurred relatively quickly in the 1970s, with little change since then.

    Higher windspeed is associated with cooler SST, as noted in earlier posts. There is also some evidence that higher mean windspeeds are associated with a tendency towards more-frequent La Ninas.

    • Erl Happ
      Posted Oct 26, 2008 at 2:45 AM | Permalink

      Re: David Smith (#146),
      David, Thanks for this discussion of ENSO. My interpretation of your graph: Lower Easterly wind speed is associated with convection in the East of the Pacific, and a warmer tropical ocean overall, gradual warming of the world oceans (because the Pacific is just one of them and there are similar rain shadow zones elsewhere) and strong warming at high latitudes in winter time.

      I know that lower wind speed in the Nino 3.4 is associated with a dramatic increase in temperature at 200hPa on a global basis around the tropics and will also (I imagine) be associated with a change in wind velocity at 200hPa particularly in the rain shadow of the Andes plus a widening of the cloud free zone in that area.

      There seems to be a widely held view that cirrus cloud traps OLR (even Roy Spencer). I think its more important role is to reflect solar radiation. Anyway, rising temperatures at 200hPa mean less cirrus. Falling temperatures are associated with more cirrus so I don’t think that the warming idea has legs at all. More cirrus is the result of upper troposphere cooling.

      If one studies particular instances of major El Nino events like that of 1982-3 one notices an association with a rise in the aa index of geomagnetic activity or the early stages of a rise in sunspot activity at the start of a cycle. Below is a set of graphs that relate to the ‘Great Pacific Climate Shift’ of 1978. Notice the strong geomagnetic activity in the early part of the year in 1979 and 1982-83.The aa index has a seasonal fluctuation related to extent of coupling with the magnetosphere that is greatest at the equinoxes. A strong solar wind effect early in the year has a big impact because irradiance is 7% greater in January and there is a lot of ocean in the south to pick up the energy. This seasonal fluctuation in aa is shown as a dotted blue line in the first diagram on the left.

      The aa index commonly peaks one to three times in the decline phase of a solar cycle and that is when the big El Ninos happen. Some variability must be expected because the UV radiation from th sun also varies, particularly in the heavily ionising X ray wave lengths. Both short wave radiation and the solar wind are baskets of variables in themselves so don’t be too demanding with the statistical tests. Perhaps use some ‘fuzzy logic’.

      • Posted Oct 26, 2008 at 3:54 AM | Permalink

        Re: Erl Happ (#147),
        you make a lot of assumptions and it’s hard to me to follow your reasoning.
        I can’t understand why you think that cirrus clouds have a bigger role in reflecting SW than in scattering LW radiation. I think that an easy look at satellite imagery provide an answer of the role of a thin cirrus cloud layer.
        So what’s the problem with Spencer and Lindzen?
        You claim that surface wind speed in Nino 3.4 is associated with a dramatic warming in the upper troposphere. Assuming that you are right, your following conjectures of a change in 200 hPa wind over the Andes shadow region and cloud decrease cannot be used for other supposition without a proper demonstration.

        • Erl Happ
          Posted Oct 26, 2008 at 6:27 AM | Permalink

          Re: Paolo M. (#148),

          Thanks for the query. Let’s take one at time.

          I can’t understand why you think that cirrus clouds have a bigger role in reflecting SW than in scattering LW radiation.

          Not sure what the relevance of ‘scattering long wave radiation’ is. Do you mean ‘absorb’ causing a rise in temperature?

          Cirrus (frozen droplets) are between 550hPa and the stratosphere. But relative humidity is low and specific humidity is rather invariable above 700hPa. So, depending upon temperature movements the cloud comes and goes. The range of temperature fluctuation at 200hPa is more than twice that at the surface.

          Short wave comes in, long wave goes out. Specific humidity decreases dramatically with elevation so water vapour is a rapidly diminishing absorber of OLR and the greenhouse gas concentration is minute. OLR has really no significant impediment from the middle troposphere upwards because density falls so fast. It finds an absorber when it hits the tropopause and ozone producing a strong temperature peak in mid year between 100hPa and 20hPa. If there were significant absorbers of OLR in the upper tropopause there would be a mid year temperature peak there as well.

          Cirrus are multi branching ice crystals that are highly reflective of light.They give tremendous reflective bang for the specific humidity buck.

          About 50% of solar radiation doesn’t make it to the surface and just as well it doesn’t.

          Its funny that infrared can be used to take a photograph of the surface from a satellite through cloud and yet people imagine that if its going in the other direction the cloud is opaque.

        • Posted Oct 26, 2008 at 8:21 AM | Permalink

          Re: Erl Happ (#149),
          you need some clarifications about radiation in the atmosphere.
          GHGs (first of all H2O) abosorb radiation and scatter it all around them: half of that is back-scattered to the ground.
          There is a certain frequency (around eleven micron) where GHGs are transparent. That band is called the infrared window in which Earth surface is directly linked to the outer space.
          Meteorological satellites use that band for day and night vision of the surface. You have to understand that the source of that radiation is from below, no source above the satellite.
          Clouds close that window. In fact, meteorological satellites, contrary to your claim, cannot see Earth surface in the IR but just the cloud tops and they see cloud tops as a colder or very cold matter (less IR brightness, i.e. darker in a direct image).
          Depending on the thickness of a cirrus layer, it can be difficult to discriminate a cirrus cloud through a vis image whereas it is much more easy with an IR image.

        • Erl Happ
          Posted Oct 26, 2008 at 3:21 PM | Permalink

          Re: Paolo M. (#150),
          Agreed. (except for the bit about transparency of GHG, its the transparency of the atmosphere due to all constituents) Transparency is a question of specific wave length. Now can we get to the nitty gritty re the cirrus.

        • Erl Happ
          Posted Oct 26, 2008 at 4:01 PM | Permalink

          Re: Paolo M. (#150),

          If there were significant greenhouse type response to OLR in the troposphere we would see what is plain and obvious in the stratosphere i.e. as soon as OLR hits a reactive greenhouse gas (in this case ozone) the temperature would rise. Look what happens in the stratosphere starting below the tropopause a 150hPa. There is an August peak corresponding to the August peak in OLR from the land masses of the northern hemisphere. By August global cloud cover has fallen 3% and this also affects the tropics of the southern hemisphere. So, it manifests strongly at 0-10°north lat and just as strongly at 10-20° south.

          Does it look as if the August warming at 100hPa is backscattered 50% towards the surface. I can’t see any warming response at 150hPa. Can you? Does it just skip that level?

          The atmospheric response to ozone is telling us that GHG theory as it is supposed to affect the troposphere is invalid.

          Notice the double peak at 20hPa and even better expressed at 10hPa. What is causing that? Its not OLR. Its not TSI?
          Any ideas. Its important because it also appears at the surface.

  137. Posted Oct 27, 2008 at 2:48 AM | Permalink

    131 (Erl):

    If one looks at geomagnetic data there is very strong variability in March and September. If one looks at temperature data for all levels of the atmosphere down to 500hPa (I have not gone further) that same strong variability occurs in March and September.

    As we have discussed so many times, the same two peaks do not occur at all levels of the stratosphere. The peaks change phase [drifts from month to month] as one descends, such that at the lowest altitude the March peak has become the September peak. We have also discussed the role of planetary waves in creating and maintaining those peaks.

    • Erl Happ
      Posted Oct 28, 2008 at 3:25 AM | Permalink

      Re: Lei Svalgaard (#153),
      Good to see you are taking an interest and stamping on the heretical stuff with your usual vigour. I am not talking about traveling planetary waves. These peaks do not travel. I am talking about the twin peaks that appear in the graph at post #150 at 20hPa, 10hPa (and 1hPa). The data is an average of all years 1948-2008.

      Leif, this is a good question for you to address. If there were a significant greenhouse gas content in the troposphere (CO2+ water vapor+ anything) like we have with ozone in the stratosphere why do we not see a peak temperature in August below 100hPa like that at 100hPa, 50hPa,70hPa and 30hPa? The Earth itself is our laboratory. It produces a pulse in OLR that peaks in August each year with the warming of the northern hemisphere land masses and according to Palle this is associated with a 3% reduction in global cloud cover. This accounts for the strength of the pulse in OLR and the strong reaction in the stratosphere. As we know the sun is most distant in June and irradiance is almost 7% less.

      And here is an associated question: Why is there no evidence of downward propagation of energy between 100hPa and 150hPa in August. The graph to the right in post #150 shows the data?

  138. Posted Oct 27, 2008 at 7:46 PM | Permalink

    Here is an interesting chart from the current NOAA weekly ENSO update:

    The threshold for La Nina is -0.5K for a period of months.

    NOAA forecasts ENSO-neutral conditions into 2009 but their CFS ensemble plot is beginning to suggest otherwise.

  139. Posted Oct 27, 2008 at 8:10 PM | Permalink

    This paper of solar influence in the mesosphere may be of interest [Erl?]
    REVIEWS OF GEOPHYSICS, VOL. 46, RG3002, doi:10.1029/2007RG000236, 2008
    Overview of the temperature response in the mesosphere and lower thermosphere to solar activity
    G. Beig, J. Scheer, M. G. Mlynczak, P. Keckhut
    The natural variability in the terrestrial mesosphere needs to be known to correctly quantify global change. The response of the thermal structure to solar activity variations is an important factor. Some of the earlier studies highly overestimated the mesospheric solar response. Modeling of the mesospheric temperature response to solar activity has evolved in recent years, and measurement techniques as well as the amount of data have improved. Recent investigations revealed much smaller solar signatures and in some cases revealed no significant solar signal at all. However, not much effort has been made to synthesize the results available so far. This article presents an overview of the energy budget of the mesosphere and lower thermosphere and an up-to-date status of solar response in temperature structure based on recently available observational data. An objective evaluation of the data sets is attempted, and important factors of uncertainty are discussed.

  140. Posted Oct 29, 2008 at 9:15 PM | Permalink

    156 (Erl):

    I am talking about the twin peaks that appear in the graph at post #150 at 20hPa, 10hPa (and 1hPa).

    Those peaks do not appear at any levels of the stratosphere below 20 hPa, so do not extend downwards.

    • Erl Happ
      Posted Oct 30, 2008 at 1:57 AM | Permalink

      Re: Leif Svalgaard (#158),

      Those peaks do not appear at any levels of the stratosphere below 20 hPa, so do not extend downwards.

      But the strong variability in temperature in March and September from year to year is certainly in evidence all the way into the troposphere. One can infer that although the force that produces the twin peaks at 1hPa and 20hPa is insufficient to produce twin peaks below 20hPa it certainly influences the temperature of the upper troposphere. There is a lot of space for cloud between 250hPa and 100hPa. It amounts to the upper 40% of the tropical troposphere and its all colder than minus 40°C.

      The data below is for equator to 20°S which is a critical zone for heat accumulation. The ocean is cool enough to gain energy without promoting a lot of evaporation. I believe the same zone north of the equator is more reactive but sea surface temperatures are considerably warmer.

    • Erl Happ
      Posted Oct 30, 2008 at 10:29 PM | Permalink

      Re: Leif Svalgaard (#158),

      Looks to me, (eyeball only) as if the degree of variability at any level depends upon its ozone content. I suggest that the temperature depends upon ozone interception of UVB.

      To derive the cirrus response one looks at the change in specific humidity over time. Data for SH and even RH is available for the 300hPa level at http://www.cdc.noaa.gov/cgi-bin/Timeseries/timeseries1.pl

  141. Posted Nov 4, 2008 at 7:43 PM | Permalink

    The RSS lower troposphere anomaly estimate is +0.18C, which is 0.01C lower than their September estimate.

  142. Erl Happ
    Posted Nov 5, 2008 at 8:02 AM | Permalink

    Here is my contribution to the speculation in relation to where ENSO (and its temperature effects in the PDO, NAO and so on, involving oscillations to the nth degree) is going.

    My reasoning goes like this:
    1. 200hPa temperature is driven by ultraviolet light intensity that responds to solar cycle influences including the solar wind. See the top two graphs. Anyone who maintains that 200hPa temperature is driven by surface conditions has not looked at the data. Perhaps they don’t want to look because they might have to give up some cherished beliefs.
    2. The driest air with the greatest ozone content is located in the low latitudes of the southern hemisphere in the rain shadow zones of the major continents and in the high pressure cells that are the southern leg of the Hadley Cell. Here the reaction to UV light is greatest as ozone reacts to UVB. As the air is heated its relative humidity falls and the cirrus cloud disappears enabling more sunlight to stream through to the ocean. This initiates a tropical warming event. In other words the ocean absorbs energy. Lots of it. This will eventually show up as very warm water at the ITCZ about 10°N of the Equator. This part of the tropics is particularly susceptible to lose humidity and cloud from April onwards as the northern hemisphere heats up and cloud cover is lost from about 20°S latitude all the way to the Arctic.
    3. High cloud reflects solar radiation, low cloud is composed of liquid droplets rather than multi branching ice crystal filaments of high altitude cirrus. Low cloud tends to absorb solar radiation. Hence cloud top temperature is warmer in low cloud. Low cloud is below 825hPa. High cloud is extremely dispersed and is composed of ice crystals. It is above 500hPa all the way to the tropopause and beyond. At high altitude any light deflection at all will increase interception rates and reduce surface flux. That is simple geometry.

    All the graphs relate to 10°S to 20°S latitude.

    Notice in particular the transition from 1997 to 1998 and 2007 to 2008 in the lower diagrams. Looking at the last 12 solar cycles the average delay between an upturn in sunspot numbers and a swing towards El Nino warming (not a peak) as reflected in the SOI index (annual data not monthly) is just over 6 months.

    Solar cycle 24 may be a special case because of the weakness of the actual sunspots as described by Livingstone and Penn. Perhaps the X Ray flux will be less. Leif may be able to enlighten us in that respect.

  143. Posted Nov 7, 2008 at 10:44 AM | Permalink

    This week’s CFS ensemble forecast continues to show us entering another La Nina phase:

    Other ensembles forecast neutral conditions.

  144. Posted Nov 7, 2008 at 12:46 PM | Permalink

    165 (David):
    How can the models forecast ENSO conditions that according to Erl are caused by the Sun [and not incorporated in the models}?

    • Erl Happ
      Posted Nov 8, 2008 at 8:45 AM | Permalink

      Re: Leif Svalgaard (#166),
      Good to see you taking an interest in tropical warming events.

      Look at the variation between the models and the way they flop around from week to week. I guess the modelers just can’t agree on the parameters that should be included and how to weight them.

      Perhaps a meteorologist can answer the following question for us? What determines the flux of temperature at 200hPa in the Pacific near Peru and Chile and how does it affect cirrus cloud cover and the strength of the South East Trade winds between Tahiti (17°S Lat.) and Darwin (8° S Lat.).

      I have another question. Why is there a very close parallel between the SOI (sign inverted) and global 200hPa temperatures between 10°-20° south latitude and in no such close parallel with other 10° latitude bands?

      A bit of advertising: I have a post at http://climatechange1.wordpress.com that might be of interest in relation to greenhouse theory.

  145. Posted Nov 7, 2008 at 3:01 PM | Permalink

    Re #166 Hmmm… that would be a good question for Erl.

    Most ENSO models have some forecasting skill. So, any solar/ENSO connection would likely involve the atmospheric parameters used by those models. If I was a solar/ENSO explorer I’d venture there.

    • Posted Nov 8, 2008 at 5:56 AM | Permalink

      Re: David Smith (#167),
      ENSO models have some forecasting skill to do what?
      To forecast, anticipating, the early stage of a new ENSO phase or just to recognize that early stage and to persist with the new tendency?

  146. Posted Nov 8, 2008 at 7:55 AM | Permalink

    Re #168 Paolo, my understanding is that the dynamical ENSO models do have moderate skill at forecasting SST anomalies some months in advance, including changes in SST anomaly trend. Their abilities are poorest around the boreal Spring “persistence barrier”.

    • Posted Nov 8, 2008 at 11:09 AM | Permalink

      Re: David Smith (#169),
      what I meant was that the early signal of a new ENSO phase (if I understood well what I read long time ago) is an eastward travelling Kelvin wave along the thermocline. Are the dynamical models you talk about able to forecast the start and the strenght of a Kelvin wave some months in advance?
      Whenever a Kelvin wave is in the initial analysis, I’m not surprised that a model can forecast its effect on the SST some months later.
      But what about the forecast of the heat content of the deeper water?
      I have no expertise in ENSO forecasts, but, at a first sight, models don’t seem to predict deep water behaviour some months in advance. However, I can very well be wrong.
      The April barrier make me think of a sort of strong non linearities in the troposphere circulation at the end of the boreal winter that can overcome any initial state of water temperature anomalies in the Pacific.

  147. Posted Nov 8, 2008 at 1:45 PM | Permalink

    Re #171 Paolo, I think the short answer for the dynamical coupled models is “yes”, but I believe that their skill is modest, especially in boreal Spring.

    • Posted Nov 9, 2008 at 4:35 AM | Permalink

      Re: David Smith (#172),
      what I know is that almost all of AOGCMs don’t have something resembling the ENSO mode of natural variability.
      So I don’t think that the heat content of Pacific water is forecasted in advance.
      Anyway, David, if you know some works assessing this issue, please let me know.

  148. Erl Happ
    Posted Nov 9, 2008 at 9:07 AM | Permalink

    At http://bobtisdale.blogspot.com/ Bob Tisdale has some interesting observations re the way subsurface temperature change occurs before surface temperature change. Is this a light penetration factor? The surface exchanges energy with the atmosphere to some extent. The bodies of warm wet topical air that stream away from the equator keep the sea surface warm. To see this compare a SST anomaly map with Precipitable water like here: http://www.coaps.fsu.edu/~maue/extreme/gfs/current/plan_water_000.png.

    If it is a light penetration factor, change must be related to albedo.

  149. Posted Nov 11, 2008 at 6:47 PM | Permalink

    Erl: Here’s a link to a video from the NASA Scientific Visualization Studio titled “Visualizing El Nino”.
    Hope it helps with this discussion. Sorry to jump in so late.

  150. Erl Happ
    Posted Nov 22, 2008 at 1:53 AM | Permalink

    At http://climatechange1.wordpress.com/ my latest description of the ENSO driver.What is increasingly apparent is that most of the action is between the equator and 30°S east of Tahiti. ENSO 3.4 looks increasingly like an artifact of something that happens elsewhere.

%d bloggers like this: