Guest Post by Willis Eschenbach
Like anyone else, I’m not fond of being wrong, particularly very publicly wrong. However, that’s the price of science, and sometimes you have to go through being wrong to get to being right. Case in point? My last post. In that post I looked at what is known as “net cloud radiative forcing”, and how it changed with surface temperature. Net cloud forcing is defined as the amount of downwelling upwelling longwave radiation (ULR, or “greenhouse radiation”) produced by the cloud, minus the amount of solar energy reflected by the cloud (upwelling shortwave radiation, or USR). If net cloud forcing is negative, it cools the earth below.
I found out that indeed, as temperature goes up, the net cloud radiation goes down, meaning the clouds have a greater cooling effect. I posted it, and asked for people to poke holes in it.
What could be wrong with that? Well, I forgot a very simple thing, and none of the commenters noticed either. The error was this. Net cloud forcing is cloud DLR ULR minus shortwave reflected by that same cloud. But what I forgot is that reflected shortwave is the cloud albedo times the total insolation (downwelling solar shortwave radiation).
The catch, as you probably have noticed, is this. If the cloud doesn’t change at all and the total insolation rises, the net cloud forcing will become more and more negative. The upwelling reflected solar is the cloud albedo times the insolation. As insolation rises, more and more sunshine is reflected, so the net cloud forcing goes down. That’s just math.
The problem is that as insolation rises, temperatures also rise. So by showing net cloud forcing goes down with increasing temperature, all I have done is to show that net cloud forcing goes down with increasing insolation … and duh, the math proves that.
However, recognizing that as the problem also gave me the solution. This is to express the net cloud forcing, not as a number of watts per square metre, but as a percentage of the insolation. That way, I could cancel out the effect of the insolation, and extract the information about the clouds themselves. Figure 1 shows the results of that analysis.
Figure 1. Net Cloud Forcing (W/m2) as a percentage of gridcell insolation (W/m2), monthly averages from 1985-1989. Average percentage results shown above each map are area-averaged. Missing data shown in gray. Cloud forcing data from ERBE, insolation data from NASA.
This is an interesting result, for a variety of reasons.
First, it is quite detailed, which gives me confidence in the geographical accuracy of my calculations. For example, the cooling effect of the thunderstorms in the Inter-Tropical Convergence Zone (ITCZ) is clearly visible in the Pacific as a horizontal blue line slightly above the equator, and can be seen in the Atlantic Ocean as well. The ITCZ is the great band of equatorial thunderstorms around the planet that drive the Hadley circulation. Remember that the majority of the energy entering the climate system is doing so in the Tropics. Because of that, a few percent change in the equatorial net cloud forcing represents lots and lots of watts per square meter.
Second, the differing responses of the clouds over the land versus clouds over the ocean are also clearly displayed. In general, land clouds warm more/cool less than ocean clouds. In addition, you can see that while the clouds rarely warm the NH ocean, they have a large warming effect on the SH ocean.
Third, and most significant, look at the timing of the seasonal changes. Take December as an example. In the Northern Hemisphere this is winter, the coldest time of year, and the clouds are having a net warming effect. In the Southern Hemisphere summer, on the other hand, clouds are cooling the surface. But by June, the situation is reversed, with the clouds having a strong cooling effect in the warm North, while warming up the winter in the South. (Note that the NH warming effect is somewhat masked by the fact that there are large areas of missing data over the land in the NH winter, shown as gray areas. The effect of this on the global average is unknown. However, by using a combination of gridcells which are adjacent temporally and gridcells which are adjacent spatially, it should be possible to do an intelligent infill of the missing areas and at least come to a more accurate estimate of the net effect. So many paths to investigate … so little time.)
I have hypothesized elsewhere that the earth has a governor which works to maintain a constant temperature. One of the features of a governor is that it cannot be simple fixed linear feedback. By that, I mean it must act in two directions—it must act to warm the earth when it is cold, and to cool the earth when it is warm. This is different from linear negative feedback, which only works to cool things down, or linear positive feedback, which only works to warm things up. A governor has to swing both ways.
Figure 1 clearly shows that, as I have been saying for some time, including both the longwave and shortwave effects clouds act strongly to warm the earth when it is cold (red areas in Figure 1) and to cool the earth when it is warm (blue areas in Figure 1). In addition, as I have also said (without much evidence until now to substantiate my claim), the ITCZ has a large net cooling effect.
So that’s where I am up to right now in my investigation of the ERBE data. Always more to learn, I’ll continue to report my results as they happen, the story of the ERBE data is far from over. I’ll be in and out of contact for a bit, I’m around today but I’m hitchhiking up to Oregon tomorrow for a friend’s bachelor party, so don’t think I’m ignoring you if I don’t answer for a bit.
w.
PS – there are some interesting results that I’ll post when I have time. These involve looking at the phase diagrams for cloud forcing, temperature, and insolation. Having the insolation available allows the phase of both the temperature and the forcing to be compared to what is actually the underlying driving mechanism, the insolation.
Regarding temperature and insolation, the ERBE data shows what is well known, that the temperature changes lag the insolation changes by about two months in the Southern Hemisphere, and by one month in the Northern Hemisphere. This is because of the thermal inertia of the planet (it takes time to warm or cool), along with the greater thermal inertia of the greater percentage of ocean in the south.
The interesting part is this: the phase diagram shows that there is no lag at all for the changes in the clouds. They change right in step with the insolation, in both the Northern and Southern Hemispheres.
This means, of course, that the clouds move first, and the temperature follows.
I’ll post those phase diagrams when I have some time.
[UPDATE: The phase diagrams, as mentioned. First, Figure 2 shows the temperature versus the insolation:
Figure 2. Insolation vs absolute temperature, from the equator to 65 N/S. The poles are not included because the ERBE cloud data only covers 65 N/S. This does not affect the phase diagrams. Black line shows no lag, gold line shows one month lag, red line shows two months lag between maximum insolation and maximum temperature. Numbers after month names show months of lag.
Since the driving signal (insolation) peaks in June and December, those months will be in the corners when the two cycles are aligned. In the Northern Hemisphere (upper panel), December is in the lower left corner with a lag of 1 month (gold line).
The Southern Hemisphere is half a cycle out of phase, so December is maximum insolation in the upper right corner. This occurs with a lag of two months (red line).
This verifies that temperatures lag insolation by a month in the Northern Hemisphere (the warmest time is not end June, when the insolation peaks) and two month in the southern hemisphere.
However, the situation is different with the clouds, as Figure 3 shows.
Figure 2. Insolation vs cloud forcing %, from the equator to 65 N/S. The poles are not included because the ERBE cloud data only covers 65 N/S. I suspect that the odd shape is a consequence of the missing gridcell data in the ERBE dataset, but that is a guess.
For the cloud forcing in both Hemispheres, there is no lag with regards to the insolation.
w.
Legatus, interesting side note. As a rule in the tropics you would find a cooling process from clouds regardless, but this would tend to go from a large effect to a minor effect depending on dry or wet season. (2 in each year) But regardless, the effect in the tropics is always going to be strongly cooling..with variations going from strong cooling effect to weak cooling effect simply depending on season.
But the effect at these locations although they do change have lesser changes then lots of others around the tropics. Combined with lower sea temps. could very well tell us something interesting.
A data artifact is possible as always, but I am probably thinking we will see an increased effect from clouds from yea more clouds of course. The question to ask yourself is this: What mechanisms could cause those two factors? (and no (or very little) variance in cloud cover between dry/wet seasons.
Just trying to get you to think, I thought of a couple that could cause this mechanism, and yes one of em does involve currents, but their involvement is rather unknown whether it causes the increased cloud cover or is possibly caused by said increased cloud cover. Just an idea really and letting you do the thinking on your catch there.
“It is not the critic who counts; not the man who points out how the strong man stumbles, or where the doer of deeds could have done them better. The credit belongs to the man who is actually in the arena, whose face is marred by dust and sweat and blood; who strives valiantly; who errs, who comes short again and again, because there is no effort without error and shortcoming; but who does actually strive to do the deeds; who knows great enthusiasms, the great devotions; who spends himself in a worthy cause; who at the best knows in the end the triumph of high achievement, and who at the worst, if he fails, at least fails while daring greatly, so that his place shall never be with those cold and timid souls who neither know victory nor defeat.”
– Theodore Roosevelt.
“This means, of course, that the clouds move first, and the temperature follows.”
Now, how about some data linking total global cloud quantities to the average latitudinal position of the surface air pressure systems and especially the jet streams ?
I think the whole lot shifts poleward and/or becomes more zonal from solar and oceanic changes to allow more energy into the oceans so that the system warms which accelerates energy to space through a more active water cycle so that the faster or larger water cycle offsets the warming.
The whole lot shifts equatorward and/or becomes more meridional from solar and oceanic changes to allow less energy into the oceans so that the system cools which slows energy loss to space through a less active water cycle so that the slower water cycle tries to offset the cooling as best it can until the system starts to warm again from further solar and oceanic variations.
Ithink that accords with Willis’s observations as above.
Willis: Re: Fig. 2 Phase diagramm: “Black line shows no lag, gold line shows one month lag, red line shows two months lag between maximum insolation and maximum temperature. Numbers after month names show months of lag.”
No, it does not. NH: Black line shows a 1-2 months lag, gold line shows a 0-1 month lag and red line shows max temp _preceding_ max insolation by 1 month. SH: black shows 2 month lag, gold shows 1 month lag, and red shows 0 month lag but also more preceding than lagging.
Thanks, Willis, for another clear and informative post. Understanding how our climate can self-regulate over long periods to avoid temperature extremes is important to our overall understanding of climate. It could also provide some important supporting evidence for the Svensmark hypothesis.
I’d be interesting in hearing your thoughts as to why ocean clouds cool more – warm less than land clouds? Also why the SH lag is around one month longer that that seen in the NH?
Joel Heinrich says:
October 12, 2011 at 2:30 am
I don’t understand. As far as I can tell, and I just checked the program, the black line is zero months lag, that is to say, comparing one months insolation to that same month’s temperature. Gold compares one month’s insolation to the following month’s temperature. Finally, red shows temperature lagging the insolation by two months. Just like the label says.
w.
Very apt quote Eternal, very apt!
Willis, the missing data could well bring the numbers closer to zero. This is an extremely clear exposition and shows what thought, as opposed to blindly “chucking card decks at the problem” (yeah, I know that no one uses card decks anymore) can do. With all the available data, anyone really interested in the climate should be peering at the sky, the ocean, and back at the data to make sense of it.
Your time in the tropics was certainly well spent in observations, now we get to see a thinker apply that knowledge. It’s a privilege and a joy to ride along.
Note bene, I still have to think that looking at June, July, and August there is something funny about the SH near to Antarctica. June and August have larger missing data regions than July. WUWT.
“black line is zero months lag, that is to say, comparing one months insolation to that same month’s temperature. Gold compares one month’s insolation to the following month’s temperature.”
Thank you, now I understand what you did there. I just don’t know why you did it. The black line shows for the SH that the peak insolation is in Dec. while the peak temp. is in Feb. So there is a 2 month lag, which is what you said. Likewise for the 1 month lag in the NH. Just what did you need the gold and red line for?
Willis : I’m totally uninterested in annual averages, Rob, they are a snare and a delusion. The clouds are warming the surface during the cold part of the year, and cooling the surface in the warm part of the year.
Willis, are yyou are still confused.
For starters, clouds are warming the surface during the night, and cooling the surface during the day. Neither of which tells anything about cloud forcing (which talks about cooling to space rather than warming the surface), but either way, why would you think that taking the monthly average (as you did) is all fine but taking the annual average is “snare and a delusion” ? Does radiation have some magical property that shows up only at monthly averages, but not at daily or annual average ?
Second, you keep on using the wrong definition of “cloud forcing”. You define it in terms of “downwelling longwave radiation” (DLR), but in reality, cloud forcing refers to “upwelling” longwave radiation, or the IR radiation that escapes to space (cooling the planet) rather then to the surface (warming the surface). Maybe this is why you got so confused with your math and get conclusions that contradict observations.
Willis : You keep thinking of it as feedback. It is not. It is a shift in cloud types and altitudes and colors as the earth warms and cools. The important thing is, when the earth cools, the clouds act to warm it, and when the earth warms the clouds act to cool it.
You call this conclusion a “a governor which works to maintain a constant temperature”, I call it “negative feedback”. As I pointed out, your conclusion is incorrect as a result of your now abundant amount of mistakes. To show “a governor which works to maintain a constant temperature” or “negative feedback”, you cannot simply divide the cloud forcing by insolation, but you need to take the temperature derivative : You need to show that if the surface temperature goes up that the forcing goes down (becomes more negative). And ERBE results show the opposite (positive feedback) from what you present, as was already noted by Ramanathan, 22 years ago.
I really like your observation of how clouds warm only in the respective hemispheric winters. Of course I already knew this, after all I have been camping outdoors in Norwegian mountains in January and know very well that a starry night is much, much colder than a cloudy night… But it reminded me of something I’ve wondered about before: If AGW is right, shouldn’t CO2 warming be most obvious in those clear winter nights? I had a look at the GISS anomaly maps, with all the flaws of the GISS temperature sets, and indeed Arctic winters have warmed more than Arctic summers. But then the Arctic is the difficult case because of those oceanic cycles that makes it virtually impossible (Tamino would of course disagree here ;-)) to isolate the contributions of the “antropogenic forcings”. So I thought, why not look at the Antarctic continent during the austral summer and winter? And looking at http://data.giss.nasa.gov/cgi-bin/gistemp/do_nmap.py?year_last=2011&month_last=08&sat=4&sst=1&type=anoms&mean_gen=0603&year1=2001&year2=2011&base1=1951&base2=1980&radius=1200&pol=pol (and the corresponding chart for the Antarctic summers), it looks to me like the last 10 winters show a cooling tendency in the heart of Antarctica, while the summers show a slight warming. So where’s the CO2 warming then?
Legatus: “The question is, what would cause this to be always true, all the year round? Is there a certain type of cloud always at these two spots which always creates this net negative effect? Is there some other effect, say a current or some such, which creates this year long pattern? How do these spots effect the weather/climate/currents etc?”
Yes, there are cold water currents (Benguela in Africa and Humboldt in South America) that result in low level clouds reflecting sunlight.
http://commons.wikimedia.org/wiki/File:Ocean_currents_1943.jpg
http://www.sat24.com/image2.ashx?region=af&time=201110120945
(c) by sat24.com
wsbriggs:
Re. your comment October 12, 2011 at 2:55 am;
SECONDED.
Richard
Rob:
I have a question concerning your post addressed to Willis at October 12, 2011 at 3:38 am.
Among other sureal and unsubstantiated assertions, your post says to Willis;
“Willis, are yyou are still confused. ”
and
“you keep on using the wrong definition of “cloud forcing”. ”
and
“As I pointed out, your conclusion is incorrect as a result of your now abundant amount of mistakes.”
My question is;
Are you joking or dissembling?
Richard
Does this help anyone! I have observed this many times on clear nights with stable air, the temperature drops, when cloud passes overhead the temperature rises, it does not stay the same i.e. cloud cover not only stops the temperature from falling further, but increases it, just a simple observation.
Keith Gordon
@willis
No, the major mistake you make and that you continue to make is that you treat all radiation as generic power sources measured in Watts/m2 and all surfaces as generic absorbers. The fact of the matter is that when a cloud covers a snowfield the shortwave that is blocked has no effect because it would be been reflected in any case. Albedo matters. The ocean on the other hand has an albedo close to 0 so when a cloud covers that it deprives the surface of a tremendous amount of shortwave energy. But on yet another hand there’s a huge difference between rock (albedo 15%) and ocean (albedo 0%) that goes beyond albedo. The ocean can d)store (buffer) huge amounts of energy for undetermined (not well understood) periods of time from months to perhaps thousands of years while rock retain daytime heating for just a matter of hours. DLR also has game changer considerations with different surfaces. Rocks are great absorbers and emitters of DLR. Water is not. Water is heated by shortwave and near infrared from the sun just like rocks but it doesn’t give up the heat predominantly by DLR as is the case for rocks. The ocean gives it up by evaporation which has the effect of reducing the role of DLR to insignificance when compared to DLR on rocks.
Nothing is going to make sense in this controversy and no one’s predictions are going to pan out under the vast over-simplification represented by expressing energy purely in Watts and surfaces as generic gray bodies. The form of the energy and the nature of the surface makes all the difference in the world.
Points to ponder:
We have had 12,000 years where the average surface temperature of the earth has been around 16C. Yet the global ocean has an average temperature of 3.9C. I can explain that and if you cannot then you can’t even begin to understand the earth’s climate. The earth’s climate is dominated by water in all its phases. Non-condensing greenhouse gases play only a very small role except for when the water cycle is shut down by global freezing. The liquid and vapor phases keep the earth’s temperature capped at a maximum that is comfortable for living things and so long as it remains liquid puts in a temperature floor as well. However, while the liquid and vapor phases exhibit negative feedback during rising temperatures the solid phase exhibits a positive feedback in falling temperatures. There IS a tipping point for the earth’s climate but the tipping point is from comfortably warm to unbearably cold. These are the facts and they are, being observed facts not theoretical predictions, beyond dispute.
I am really sorry, and the last thing I want to do is upset anyone, but since the ‘clouds warm’ is an integral part of the conclusions of this post, I have to humbly confess that I am with DP on this. To have stood outside on a cloudy night and to have felt the warming effect of clouds is anecdotal and not scientific. As a minimum, I would want to see the following: A cloudy, quiet day turning to a cloudy quiet night. Temperature reading at (say) 9pm. Somehow exclude wind/convection or any artificial mechanisms as a vehicle for imported heat. Temperature reading at (say) 5am the next morning turns out higher than the 9pm reading. Then one could postulate (but remember correlation is not causation) that the clouds caused warming. Unless proven otherwise, I am pretty much convinced that the 5am temp will be lower than the 9am temp. But the difference will be less than if the sky was clear. So, let’s be specific: clouds can slow down cooling but cannot ‘warm’. As such, they can have a nett warmer outcome, but they have not ‘warmed’ anything. Sorry if I am being too pedantic.
Willis, you’re the best! While you’ve learned a lot, you’ve taught far more. I think you’re on to something, really.
For the impatient –
A’la Joni Mitchell and her song “Both Sides Now”
We’ve looked at clouds from both sides now,
From up and down, and still somehow
It’s cloud illusions we recall.
We really don’t know clouds at all.
Sometimes ’tis better to ponder what our eyes and ears have taken in than to talk or raise small questions. Think about it. Please?
You could also extend the data beyond the annual cycle.
Plot each data type across the whole time horizon.
Net cloud forcing, temperature, insolation by month (and then the same by different latitudes). See what type of changes have occured over time. [Pinatubo is in the middle and likely to impact the data].
I might also pick two near-by regions that have similar insolations (the Sahara at 15N versus Equatorial Africa). Clouds cool off the more cloudy equatorial region while leaving a smaller diurnal/seasonal cycle. Sahara is warmer overall but has a greater diurnal change from day to night and a greater seasonal cycle.
Empiric observation: Willis is as cool as net cloud forcing on a hot afternoon in the Pacific Inter – Convergence Zone. 🙂
Willis – thanks for all your hard work to keep genuine science moving…
Fair play to you, Willis. When you first postulated the Thunderstorm thermostat, it sounded intuitively right. Now you’re doing the hard yards mathematically and it is strengthening your theory.
In a “governed” system, using percentages also seems more intuitive because the governing is being done in real time, rather than when a set point climatically is reached.
As for mistakes, we are the sum of our mistakes ultimately. We learn far more from mistakes than getting things right the first time. It’s how we react to them is what defines our character.
Thanks for keeping on top of this Willis. One thing I noticed in the graph in your first post was that the cooling when warm, and warming when cool effect was split by temperature. It appeared that sub-0°C (ice crystals), warm; while above 0°C (water vapor) cools. Yet these occur in the counter-intuitive seasons, ie. ice crystals form clouds when the surface is warm, and water vapor forms clouds when it is cold. Does this imply a difference in altitude of cloud formation seasonally?
Doubly interesting. The elegance of the exposition also the thought that some of the comments gave me with regard to a conversation I had with Ulric Lyons some years back.
What if we took every yearly average: Jan > Dec, Feb > Jan, Mar > Feb through Dec > Jan (i.e. disregarding the calendar year for 11 of them).
What would those “averages” disclose compared one with another?
Not just for this topic either. e.g. I wonder what then would be “the warmest year evah”?
Just throwing it out there as I have zero experience with stats 8)