Guest Post By Willis Eschenbach
Often I start off by looking at one thing, and I wind up getting side-tractored merrily down some indistinct overgrown jungle path. I was thinking about the difference in the strength of the sunshine between the apogee aphelion, which is when the Earth is furthest from the sun in July, and the perigee perihelion in January, when the Earth and the sun are nearest. On a global 24/7 average value, there is a peak-to-peak aphelion to perihelion swing of about twenty-two watts per square metre (22 W/m2). I note in passing that this is the same change in downwelling radiation that we’d theoretically get if starting in July the CO2 concentration went from its current level of 400 ppmv up to the dizzying heights of 24,700 ppmv by January, and then went back down again to 400 ppmv by the following July … but I digress.
Now, because the Earth and sun are nearest in January when the southern hemisphere is tilted towards the sun, there is a larger swing in the solar strength in the southern hemisphere than in the northern.
While I was investigating this, I got to looking at the corresponding swings of the ocean surface temperature. I split them up by hemisphere, and I noticed a most curious thing. Here’s the graph of the annual cycle of solar input and sea surface temperature for the two hemispheres:
Figure 1. Scatterplot, top-of-atmosphere (TOA) solar input anomaly versus ocean surface temperature. Northern hemisphere shown in violet, southern hemisphere shown in blue. Monthly data has been splined with a cubic spline. Data from the CERES satellite dataset.
So … what is the oddity? The oddity is that although the swings in incoming solar energy are significantly larger in the southern hemisphere, the swings in ocean temperature are larger in the northern hemisphere. Why should that be?
The difference is impressive. As a raw measure, the northern hemisphere ocean surface temperature changes about seven degrees C from peak to peak, and the TOA solar varies by 216 W/m2 peak to peak. This gives a change of 0.032°C per W/m2 change in solar input.
In the southern hemisphere, on the other hand, the ocean surface temperature only swings 4.7°C, while the solar input varies by 287 W/m2 peak to peak. This gives a change of .0162°C per W/m2, about half the change of the northern hemisphere.
So that’s today’s puzzle—why should the ocean in the northern hemisphere warm twice as much as the southern hemisphere ocean for a given change in solar forcing?
Part of the answer may lie in the depth of the ocean’s mixed layer. This is the layer at the top of the ocean that is mixed regularly by a combination of wind, waves, currents, tides, and nocturnal overturning. As a result, in any given location the mixed layer all has about the same annual average temperature. (However, monthly changes are still largest and the surface and decrease with increasing depth.) This mixed layer worldwide averages about 60 metres in depth. But the mixed layer is deeper in the southern hemisphere, averaging about 68 metres in the southern hemisphere versus about 47 metres in the northern.
However, two things argue against that conclusion. One is that the mixed layer in the southern hemisphere is about 45% deeper than in the northern … but the northern hemisphere sensitivity of temperature to incoming solar is double, not 40% larger.
The other thing that argues against the mixed layer difference is that the thermal lags in the two hemispheres are very similar, with peak temperatures in each hemisphere occurring almost exactly two months after peak solar. In a previous post entitled “Lags and Leads” I discussed how we can use the difference in time between the peaks of solar power and temperature shown in the scatterplot in Figure 1 to calculate the time constant “tau” of the system. This two month lag is equivalent to a time constant tau in both hemispheres of 3.3 months.
Then, using that time constant tau we can calculate the equivalent depth of seawater needed to create a thermal lag of that duration. A time constant tau of 3.3 months works out to be equivalent to 25 metres of seawater with all parts equally and fully involved in the monthly temperature changes (or a deeper mixed layer with monthly temperature swings decreasing with depth).
But since the time constant “tau” is the same for both hemispheres, this means that the equivalent depth of water that is actually involved in the annual cycle is the same in both hemispheres.
Or in other words, the more of the ocean that is involved in monthly temperature swings, the greater the lag there will be between solar and temperature peaks. But in this case, the thermal lags are the same in both hemispheres, meaning the equivalent depth of ocean involved is the same.
Then, I thought “Well, maybe it’s because one pole is underwater and the other pole is on land”. So I repeated the calculation of the temperature and solar swings using just the range from 60° North to 60° South, in order to eliminate the effect of the poles and the ice … no difference. The northern hemisphere non-polar ocean warms twice as much for a given change in solar energy as does the southern non-polar ocean. The difference is not at the poles.
So my question remains … why is the northern hemisphere ocean surface temperature twice as sensitive to a change in solar input as is the southern hemisphere ocean temperature?
My best to all. Here, we have had a rare June rain, most welcome in this dry land, so for all of you today, I wish you the weather of your choosing in the fields of your dreams …
w.
My Usual Request: Misunderstandings can be minimized by specificity. If you disagree with me or anyone, please quote the exact words you disagree with, so we can all understand the exact nature of your objections. I can defend my own words. I cannot defend someone else’s interpretation of some unidentified words of mine.
My Other Request: If you believe that e.g. I’m using the wrong method or the wrong dataset, please educate me and others by demonstrating the proper use of the right method or the right dataset. Simply claiming I’m wrong about methods or data doesn’t advance the discussion unless you can point us to the right way to do it.
UPDATE: Here are two maps of the same data, which is the change in ocean temperature per 0ne watt/metre squared (W/m2) change in top of atmosphere (TOA) solar radiation …
The gray contour lines show the global average value of 0.02 °C per W/m2.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Willis, Congratulations; if I am right you have found something very important. These plots are servo diagrams of the feedback systems of the North and South hemispheres. Notice the southern hemisphere plot is very controlled and is tight indicating high feedback and the northern plot is loose indicating a lower feedback ratio. This is probably due to the higher percentage of the south being in contact with the Ocean.
Willis, I am sorry that my comment was so short; I was in a hurry. I believe that you have found the master thermostat for the Earth’s climate. Many researchers have hunted for this over the years. The comments on your post do not recognize that both the high and low temperatures are being controlled.
I will look primarily at the Southern Hemisphere curve and compare it to the Northern Hemisphere. The curve is an extended ellipse with the temperature being controlled about the major axis. Both high and low temperatures are controlled with respect to the axis. An ellipse has two centers of rotation, where a thermostat or reference should be only one point. However, in this case, the center could be moving with the input causing the effect of an ellipse. The question would be what would serve as a reference. Bob Tisdale has been commenting and showing on his graphs that the Earth’s temperature does not change smoothly; but rather moves in jumps primarily in concert with El Nino warming and remnants. This would indicate that the Earth’s temperatures follow the Ocean surface temperatures. Since the Oceans’ surface temperature changes with seasons the reference would change and the ellipse form would be expected.
How would the Ocean surface temperature control the Earth’s temperature? The Hadley cells, The Ferrel, and the Polar cells move heat from the Equator to the polar regions. The horse latitudes form a mixing of the high side winds from both the Hadley cells and the Ferrel cells; somewhat like the summing junction of an operational amplifier. The mixed winds, which could also be considered an error signal, flows back to the Equator on the low side of the Hadley cells and is in contact with the Ocean’s surface. The surface temperatures add or subtract energy and prepare them to enter the uplifting at the Equator. The energy and moisture they carry affects the thunderstorms and uplifting at the ITZ. More energy increases the wind flow and serves as a negative feedback cooling the Earth.
I hope this is complete enough for you to use.
Thanks, John
I am intrigued by your comment
Samuel C Cogar
June 19, 2016 at 9:00 am
Thanks, Samuel. Several other people have made the same point. You are correct when you say:
However, what you and others seem to be missing is that because the SH has a larger ocean area (43% NH, 57% SH), it receives proportionately more solar than does the northern hemisphere ocean. Remember, the incoming sunlight is measured in watts per square metre … so the ocean with more square metres gets more sunlight.
This leaves only the depth to vary between the two hemispheres … but as I pointed out, there’s not much difference in the mixed layer depth. In addition, the two hemispheres have the same lag (peak insolation to peak temperature) and thus appear to have no difference in the equivalent mass of water involved.
Best regards,
w.
Willis: “the two hemispheres have the same lag (peak insolation to peak temperature) and thus appear to have no difference in the equivalent mass of water involved.”
What is the exact lag for each hemisphere ? Are you saying it’s 2mo+/-0.5 mo ?
Could you say how are you evaluating the lag and to what accuracy?
What would the implied difference in mixed layer be if SH was 2.5mo ?
It seems that you are saying the lag is about equal ( with no indication of just how different the could be on available information ) . It is now “the same lag …. and thus appear to have no difference in the equivalent mass of water involved.”
Totally tossing out this option seems a bit radical on the evidence. It is not a binary choice. It seems unlikely that this is not a factor. Quite likely not the only one.
could you comment on the accuracy of lag in each hemisphere, and how much 2.5mo would mean for the depth calculation?
Best regards, Greg.
Greg, as mentioned in the caption I splined the monthly values to give what is shown in Figure 1. It’s splined to the nearest degree, and I used it to calculate the lag. The two hemispheres have the same lag to the nearest degree.
Best regards,
w.
The smaller pot is not only smaller, but forces polewards currents moving some heat North, where West Wind Drift does not do that on the SH. Also asymmetry may arise from places of upwelling and sinking being not symmetrically positioned.
I would think that a difference of 45% would not be negligible.
Werner Brozek June 19, 2016 at 3:32 pm
Thanks, Werner. I’d said.
The issue is not that the difference is “negligible”, that’s your word, not mine. There are two issues. First, the difference in the mixed layers is not large enough to explain the difference.
Second, and more important, if there were a difference due to different masses of water involved in the cycle in the two hemispheres, then the lag in one hemisphere would be larger than in the other … but the lags are the same.
w.
Thank you!
Perhaps different numbers need to be compared. Instead of comparing 0.032 to 0.016 since these numbers talk about areas and not volumes, the number 216 should be compared to 287. 287 is 33% more than 216. Perhaps this explains 33% of the 45% and we just need to explain the other 12%.
Willis Eschenbach — you state that “but as I pointed out, there’s not much difference in the mixed layer depth”. A comment above presented a significant difference between NH and SH mixed layer depth with NH is smaller than SH.
Dr. S. Jeevananda Reddy
The S Hem would not lose heat to the land as much as the N Hem does.
Willis Eschenbach June 19, 2016 at 11:02 am
However, what you and others seem to be missing is that because the SH has a larger ocean area (43% NH, 57% SH), it receives proportionately more solar than does the northern hemisphere ocean. Remember, the incoming sunlight is measured in watts per square metre … so the ocean with more square metres gets more sunlight
———————————————————–
You know, Willis, that does make sense.
But in reality there are other processes affecting how much and where.
The SWARM satellites are doing more than just measuring magnetic fields on Earth.
First results from the Swarm Dedicated Ionospheric Field Inversion chain
http://link.springer.com/article/10.1186/s40623-016-0481-6#Sec7
Fig. 12
Evolution with UT and season of the maximum absolute value of the primary current function Ψ1 in the Northern ( ΨN ) and Southern ( ΨS ) hemispheres. ΨN and ΨS represent the total currents (in kA) flowing in the dayside vortex for each hemisphere. F10.7=100 SFU
Thanks Carla, I was not aware of this work.
I have always wondered why the line of cloud indicating the ITCZ drifted upwards between S.Am and Africa. The following plot seems to show the reason.
Willis,
Very cool. I have two suggestions to make re the hemispheric difference.
First, the southern hemisphere has more clouds than the northern hemisphere. The difference is sufficient to give both hemispheres virtually identical albedo, even though surface albedo is higher in the NH (more land, more desserts, more snow and ice). So a smaller fraction of sunlight makes it to the surface in the SH, reducing the magnitude of the seasonal swing in insolation and therefore the temperature change.
Second, the NH has more land than SH and land heats up more readily than water. The effect is large enough that the global average temperature is highest during NH summer, when the earth is farthest from the sun. That effect is mainly over land, but heat does not stay in one place. So some of the effect should be transferred to the ocean, producing a larger temperature swing in the NH.
I could dig up references to support the above, if needed.
Thanks Mike. That makes a lot of sense. It seemed that the albedo being the same went counter to my initial intuition that the difference was due to cloud. What you say shows they are not contradictory.
The cycle here is 4 time units, let’s call them seasons. So the lag you are referring to here is the range between 4 and 5 time units on the graph. 2/3 of a season ( 2mo) seems to match a tau of about 3 or 4 on that scale ( estimating 4.67 by eye ).
2.5/3 ( 2.5 mo ) looks like it could be tau of 8 or 9. That is what I mean about it moving quickly in that region, so even a week or two of difference is the lag between N and S could correspond to a significant change in tau , and hence depth calculation.
That is why I think you need to be a bit clearer how much “the same” they are before dismissing the impact of differences in mixed layer depth.
I would start by asking ‘why are the mixing layers of different depth?’ Ocean currents can circumnavigate the globe in the Southern Ocean and we know that cold water at depth migrates north. Just some questions
So my question remains … why is the northern hemisphere ocean surface temperature twice as sensitive to a change in solar input as is the southern hemisphere ocean temperature?
Willis, I think this is a quantitative question. You have to use a climate model to answer this question. I have used a simple gridded EBM model (solar insolation and reflection and other radiative parameters from CERES, horizontal and vertical convective as adjustable parameters). I found for the monthly hemispheric ocean temperatures in °C
Month ;tomNH ;tosNH ;Diff ;tomSH ;tosSH ;Diff
Jan ; 15,1; 15,9; -0,8; 17,7; 18,8; -1,1
Feb ; 14,9; 14,6; 0,3; 17,8; 20,3; -2,5
Mar ; 15,5; 14,3; 1,2; 17,6; 20,9; -3,3
Apr ; 16,9; 14,9; 2,1; 16,8; 20,3; -3,5
May ; 19,0; 16,2; 2,8; 15,7; 18,7; -3,1
Jun ; 20,7; 17,8; 2,9; 14,6; 16,6; -1,9
Jul ; 21,9; 19,4; 2,5; 13,7; 14,4; -0,7
Aug ; 22,4; 20,7; 1,7; 13,4; 12,8; 0,5
Sep ; 21,7; 21,4; 0,3; 13,6; 12,3; 1,3
Oct ; 19,8; 21,1; -1,2; 14,4; 13,0; 1,4
Nov ; 17,8; 19,8; -2,0; 15,6; 14,6; 1,0
Dec ; 16,0; 17,8; -1,8; 16,8; 16,7; 0,2
Jan-Dec; 18,5; 17,8; 0,7; 15,6; 16,6; -1,0
tomNH are the measured Ocean temperatures (HADCRUT 4) and tosNH the simulated temperatures, and so on. The largest swing is between the temperatures of Feb and Aug. It is 6.1°C for NH (measured 7.5°C) and 7.5°C for SH (measured 4.5 °C). So the difference still exists. My proposal: It is caused by the different circulation patterns of the atmosphere and the oceans in both hemispheres. This is not described very well by my simple model.
One cannot begin to answer that question until one looks not only at oceanic volume, but also currents. Since currents operate in 3 dimensions, one needs to know at what rate near surface volume is sequestered to depth, and at what rate oceanic currents distribute a given volume of ‘warm’ water in a poleward direction. Of course, it is much more complicated than that since ‘warm’ water pools in various areas such as the Gulf of Mexico, off the West coast of Africa etc.
As I mentioned to years ago when discussing Radiating the Oceans, one cannot look at the radiative budget of the oceans to determine whether the oceans freeze. The oceans will not freeze (even absent DWLWIR) because there is sufficient solar irradiance being inputted in to the equatorial/tropical oceans to keep that area ice free, and thereafter, ‘warm’ waters from that area are distributed globally via ocean currents.
As mentioned to you, if you look at the radiative budget of oceans at the same latitude, eg. in and around the coast of Iceland and say equivalent latitude in the Baltic, one ocean freezes in Winter, the other does not. the reason is ocean currents which operate differently in these two different areas.
Just a thought: Given the much larger land area of the NH, it follows that the ocean areas above the continental shelves will also be much larger, proportionately, than in the SH. So, the total volume of seawater in the Southern Hemisphere is much larger than the ocean area itself suggests. The rest of this discussion is well above my pay grade- but interesting, nevertheless…
Rivers. There are twice as many rivers emptying warm water into the NH Oceans as there are into the SH oceans.
@Pamela
the connection of this problem to bio growth looks remote to me.
I am sitting with a similar conundrum
my results indicate that the SH never warmed in the past 40 years, even though maxima were rising sharply, even more sharply then in the NH.
It looks like all global warming happened in the NH (during the past 40 years)
yet there never was any warming in the SH. For example, here in southern Africa.
minima have been falling rather than rising (as AGW would suggest if GW were a global problem)
It looks to me that as other writers have suggested, the (extra) heat produced in the SH during the past 40 years was shipped to the NH by the weather, .e.g. by thunderstorms…..
I am sure Willis knows all about that?
I don’t follow you. Bio growth was not a listed factor in my posts, and the central issue of Willis’ post is the hemispheric asymmetrical nature of seasonal changes in SST as a function of the orbital influence on Solar heating.
@John the retired engineer
I know what controls maxima
I am still puzzled as to what controls minima?
Henry, as far as I can tell, it is the same mechanism. The lower water temperature will be the reference that the system corrects to, and the actual resultant temperature will be determined by the feedback ratio. The system corrects based on the ratio of the open and closed loop gains and there will be some uncorrected error in the cold direction.
The widely divergent, often self-contradictory, speculation here about the annual hemispheric SST inequality is symptomatic of the inadequacy of climate-science paradigms in explaining geophysical realities. Concentrated upon over-simplified equilbrium responses to radiation, the complex effects of modulation of surface insolation and of thermal energy storage–particularly by the ever-evaporating, ever-flowing oceans–are often greatly overlooked.
Sailors have known for centuries that there is nothing in the NH to compare with the roaring forties and whistling fifties for persistence and scale of high winds and strong currents. Physics tells us that evaporation increases with windspeed.. Oceanographers have long recognized not only the unmatched scale of the mass transport by the Circumpolar Current, but also the upwelling all along the edge of the Antarctic ice-shelves. Contrary to the mere horizontal redistribution of hemispheric heat by surface currents, or the limited downward mixing of surface heat, it is the perpetual upwelling of near-freezing water in those and higher southern latitudes that is largely responsible for the persistent hemispheric inequality of SST.
Thank you for this very accurate point of view. From these few lines I learned much more than from the whole rest!
Does that mean (yes: oversimplified) that this hemispheric inequality would be by far lower if the Antarctic was not a continent?
Since the presence of a coast is essentail for truly deep upwelling (in contrast to the shallow circulation of Langmuir vortices in equatorial upwelling) , the absence of the continent Antarctica would certainly reduce the inequality in the forties and fifties. The question remains, however, whether the additional mass of near-surface water thus made available at higher latitudes for mixing by the ferocious winds would fully compensate the hemispheric average for that zonal reduction. .
thinking about the weather…
is it not that the amount of weather is determined by the rate of change in the delta T between the poles and the equator?
perhaps I should refer to
https://wattsupwiththat.com/2016/06/18/hemispheric-ocean-temperature-sensitivity/#comment-2241766
to illustrate
….
to illustrate that I have the same problem as Willis about the transfer of heat from SH to NH
As promised above, a more detailed look at the issues …
I’ve added a grey line at the global average sensitivity, which is 0.02 °C per W/m2.
I’ve also added the graphics to the head post.
Comments welcome,
w.
Then it appears that greater NH mixing from chilled Arctic water versus less mixing in the SH from chilled Antarctic water could be plausible. It also looks like the asymmetrical split of the equatorial current favoring the NH is also plausible. These issues would show up using season as the independent variable.
Willis,
You have shown the incoming TSI in your opening diagrams. Unless the outgoing TSI is constant, this gives only part of the picture. Roughly the SS temperature is set, with a lag, by the difference between incoming and outgoing.
Just as incoming TSI peaks in January in the SH , on the dark side of the globe the outgoing TSI has a feature that sits in the dark for months, namely the South Pole plateau and its surrounding sea ice. Six months later the Arctic ice mass does likewise, but it is smaller and warmer. Because of their size differences, the ice masses leave less ocean area to give SSTs in the dark SH winter and the waters are further from the Pole and thus warmer than the NH summer case. The warmer they are, the more outgoing radiation they give to cause overall cooling.
So, the outgoing TSI is different, NH versus SH, in their respective winters.
There is a further asymmetry because of the elevation of the Antarctic plateau causing it to be cooler than the air above it. This means that its outgoing flux to space starts in the air as opposed to the ground everywhere else. This causes a different heat loss scheme in the SH dark side, compared to NH dark side six months later. One consequence is that different regimes exist to set the SSTs in those two places.
One can speculate about the directions of these effects but a better scenario becomes possible if diagrams of incoming TSI are complemented with similar ones for heat loss from outgoing radiation.
Willis has described simple event diagrams. It is not hard to imagine many other effects to explain them, as others have done above, but hard to know of their interdependence and relative weights.
Geoff, don’t put too much trust in TSI. The instruments measuring it might be less reliable than you think.\ not only that:
The incoming solar energy has a chi-square distribution but the top is not constant. Depending on the amount of the polar solar magnetic field strengths, it shifts. Currently, they are at its lowest field strengths, releasing more of the most energetic particles, causing more ozone, peroxides and N-O oxides being formed TOA. In turn, that affects all UV incoming. Peroxides are being formed preferentially to ozone, if OH radicals are available.
IMHO that affects the total difference T between the poles and that might explain the problem that both Willis and myself have picked up – trying to explain what we are seeing is happening.
As others (Frank, Steve Fitzpatrick) have correctly pointed out already, one cannot divorce ocean temperatures from its effective “heat capacity.” That is plainly affected in situ by the winds, which determine the depth of mixing and the strength of evaporation and upwelling This is not a direct TSI “forcing” problem and one should not expect any simple functional relationship between the those two variables. Furthermore, because winds drive ocean currents, which advect heated seawater, it should be apparent from the Stokes derivative that local insolation alone is far from adequate in determining temperature changes at any fixed location. In short, global maps of the “ocean sensitivity” to TSI fail to address the physical essence of the problem and show mostly the advective/convective effects of ocean circulation.
Geoff Sherrington June 22, 2016 at 11:24 am
Indeed, that was one of my points—forcing is only part of the story.
w.
Dear Willis
in conclusion
I should perhaps just point that your first graph is marked incorrectly, on the X-axis. Surely it should indicate delta K rather than absolute temperature (in degrees C)?
delta K= delta C= T anomaly (difference in T, rather than absolute T)
Alternatively,
Figure 1. Scatterplot, top-of-atmosphere (TOA) solar input anomaly versus ocean surface temperature. Northern hemisphere shown in violet, southern hemisphere shown in blue…..
Should read
Figure 1. Scatterplot, top-of-atmosphere (TOA) solar input anomaly versus ocean surface temperature anomaly. Northern hemisphere shown in violet, southern hemisphere shown in blue…
Willis I think the problem is related to the fact that you have taken the intensity of the solar insolation variation due to the elliptical orbit into account but have not allowed for the fact that the earth is exposed to the elevated insolation for a shorter time. Kepler’s law says that as the earth is closer to the sun it orbits faster and so spends less time close to the sun than it does far from the sun.
I was thinking about the difference in the strength of the sunshine between the apogee aphelion, which is when the Earth is furthest from the sun in July, and the perigee perihelion in January, when the Earth and the sun are nearest. On a global 24/7 average value, there is a peak-to-peak aphelion to perihelion swing of about twenty-two watts per square metre (22 W/m2).
See for example: http://hyperphysics.phy-astr.gsu.edu/hbase/kepler.html
Thanks, Phil. In fact I’ve allowed for both the intensity and the time of the insolation. As you point out, the earth spends less time near the sun when the intensity is higher.
As a result, the total insolation over the year is identical in the northern and southern hemisphere.
w.
Still, though, local effects matter. Where I live, peak high (daytime highs and nighttime lows) temperatures are seen 15 to 45 days after the solstice. July is my hottest month, not August. Winter, being much more variable where I am relative to summer, the peak low (daytime highs and nighttime lows) temperatures cover more days, yet January isn’t my coldest month. It wobbles between January and February.