Posted by: erl happ | October 23, 2011

High level cloud and surface temperature

The orbit of the Earth’s around the sun is slightly eccentric. The closest point is called the perihelion. On January 4th the Earth is just 147,098,291 km away from the sun. Aphelion occurs July 4th when the Earth is 152,098,233 km away from the sun, a difference of +3.3%. Naturally the power of the sun falls away with distance. It’s radiation is 7% weaker in July than in January.  The near surface air temperature for the Earth as a whole is 3.3°C warmer in July than in January. The northern hemisphere is rich in land and the southern hemisphere rich in sea. The land returns energy to the atmosphere. The sea stores energy.

As the air warms in mid year it is likely to be drier and less cloudy because evaporation lags the temperature increase. Cloud cover increases in the northern hemisphere as the atmosphere  warms but the loss of cloud in the southern hemisphere as the south cools is much greater than the gain of cloud in the northern hemisphere.  So, on a global basis cloud cover falls  in mid year by 3%.  Total cloud cover tells us very little about global albedo because different types of clouds vary in their albedo. Some cloud is said to trap radiation and warm the surface and this type changes a lot. So the cloud cover statistic tells us little about the state of cloud albedo or its radiative impact.

This post explores where,  why and what sort of  cloud is lost as the global atmosphere warms in mid year. There is a heavy loss of high level cloud in the southern hemisphere. The manner in which this occurs  is interesting. It informs us as to the artificiality of our notions of what constitutes the troposphere and the stratosphere and the dynamics of the system that determines surface temperature and its variability over time.

All data for the graphs from

Figure 1 Surface air temperature (°C) and precipitable water (kg/m^2). Percentage change between minimum and maximum is indicated.

The increase in precipitable water lags the temperature increase by a couple of months. There is plainly more variability in the land rich northern hemisphere.

The maps below come from the invaluable JRA-25 Atlas at:

Figure 2 Total cloud cover January

Figure 3 Total cloud cover July

It is evident that all the driest parts of the land  in both hemsipheres have less cloud in July than they do in January. These locations are sources of potent surface radiation.  There is less cloud as a whole in July (more dark blue) than in January. There is more dark blue between the equator and 30° south in southern winter (July) than in summer. But what type of cloud is lost?

Cloud is classified according to elevation:

High cloud 7.6km to 12 km (300hPa to 150hPa)

Medium level cloud 2.4 to 5.5km  (700 to 400hPa)

Low cloud below 2.4km. Below 700hPa.

Figure 4 Annual cycle of relative humidity at 10-30°south and 10-30°north at 925hPa (near surface) and at 300hPa (8km)

In the topmost figure we see a marked reduction in relative humidity at 300hPa at 10-30° south in mid year.  The same latitudes in the northern hemisphere experience an increase in relative humidity in mid year.

In the lowest figure we see only a slight reduction in relative humidity at 925hPa affecting the last half of the year. However there is a loss of humidity at 700hPa, 500 hPa and 300 hPa in the middle of the year increasing with altitude. Notice that humidity at 300hPa generally exceeds that at 700hPa and 500hPa.

Figure 5 Relative humidity between 50°north latitude to 50° south latitude

Figure 5 aggregates data for all latitudes between 50° north and 50° south.  We see a marked trough in relative humidity at 300hPa between April and November. This establishes that the main inter-seasonal dynamic occurs in the upper troposphere. But at what latitude?

Figure 6 Relative Humidity by latitude and altitude

It is plain that the mid year loss of humidity at 300hPa that characterizes the latitude 50°north to 50° south as a whole is is driven by marked change in the southern hemisphere.

Why is it so?

Figures 7 and 8 illustrate the point that the great high pressure cells of the Hadley circulation produce copious amounts of thermal (infrared) radiation colored red. This is particularly the case in the winter hemisphere. Here is a conundrum. Why do the subtropical latitudes of the winter hemisphere give off more radiation when the surface is cooler than when it is warmer?

First, what’s a Hadley cell?  At the equator the air ascends. As it ascends latent heat is released, the air becomes less dense is driven upwards and in the process it cools via decompression. Hence the paucity of outgoing radiation in near equatorial latitudes. The warm waters between India and New Guinea give off little radiation but they deliver much evaporation.    Air that ascends at the equator ultimately returns to the surface at 10-40° of latitude. It descends over cool surfaces that support the process of descent by cooling the surface air. Extensive high pressure cells circulate anticlockwise in the southern hemisphere and clockwise in the northern hemisphere giving rise to the trades and the westerlies. These cells are largely free of low and middle troposphere cloud. The air in these cells warms via compression, the bike pump effect. So, as these high pressure cells occupy more space over land and sea in the winter hemisphere the surface must receive more direct sunlight and the winter hemisphere at these latitudes must be warmer than it otherwise would be.

Figure 7 July outgoing long wave radiation

Figure 8 January outgoing long wave radiation

It is apparent that atmospheric processes determine where thermal radiation is released by the Earth system. It is released from the atmosphere rather than directly from the surface. The area of cloud free sky tends to be enhanced in winter. This must be considered a positive. We like to be warmer in winter. If this is what the greenhouse effect is all about I am all for it. But hang on, the greenhouse effect must be quite weak because there is little chance of water vapor amplification in dry air. What a bummer.

Figure 9 Air temperature at various elevations at 20-30°south

Figure 9 shows that the temperature of the upper troposphere at 20-30° south responds to enhanced radiation in winter just like the stratosphere at 50hPa.  The response depends upon the ability of ozone to trap long wave radiation at a quite specific wave length, 9.6 micrometers. Infrared spans 4-28 micrometers. We see a strong response to just a small part of the total spectrum by a mass of tiny little radiators that populate this part of the atmosphere in the parts per billion range but sufficient to invert the seasonal temperature profile. Hang on, this is not in the rule book, the troposphere is supposed to be warmed from the surface and should move with surface temperature! But here we see the upper troposphere acting like the stratosphere in that it responds to long wave radiation. Do we need to alter our ideas of where the stratosphere starts? Where is the ozone tropopause?

A winter warming response at 250hPa,  involving  a marked loss of relative humidity in the ice cloud zone involves a disconnection of moisture supply from the temperature dynamic in the upper southern troposphere. Quite possibly, the supply of moisture to the upper troposphere in the southern hemisphere is cut off by the northward migration of the inter-tropical convergence and the cooling of the tropical seas at this time.

Climatologists have long wondered why a 1°C increase in temperature at the sea surface relates to as much as a 3° increase in temperature of the upper troposphere. They call this phenomenon ‘amplification’ as if the temperature of the upper troposphere in some way depended on the temperature at the surface.  The presence of a long wave absorber namely ozone, is responsible for this phenomenon. The warming of the upper troposphere results in cloud loss and because of this, the surface temperature increases.

In the mid and high  altitude parts of the troposphere cloud is present as highly reflective interlacing micro-crystals of ice that we describe as cirrus and stratus. When the air warms these crystals sublimate. Ice cloud is the dominant cloud of the subtropical region. In IPCC climate science high altitude ice cloud is supposed to warm the surface by enhancing back radiation.  But when radiation from the atmosphere increases in winter relative humidity falls. This radiation it is not bounced back by the cloud, the cloud disappears and lets the sun shine through. The surface temperature response is due to the disappearance of the cloud, not back radiation. Oops.

Figure 11 High cloud cover in January and July Source JRA-25 Atlas at

Figure 11 shows the latitudes of the southern hemisphere where high cloud is evaporated in July. There is a marked expansion in the area that has less high cloud  by comparison with January.

Figure 11 The advance of global temperature in January and July

Figure 11 indicates that there is more year to year variability in the minimum global temperature  (January) than the maximum (July).

The minimum is experienced when the Earth is closest to the sun. The Earth is coolest at this time because the southern oceans absorb solar radiation without warming the air and the atmosphere is cloudier in January. January is characterized by a relative abundance of high ice cloud in the southern tropics. Relative humidity peaks in April (figure 6) when tropical waters are warmest. I suggest the variation in the minimum global temperature is likely due to change in high altitude cloud.  The southern hemisphere experiences the largest flux in ice cloud.That flux in high cloud is likely to be due to change in the ozone content of the upper troposphere.

You would have to be very naive to think that the inter-annual  change in temperature that is most obvious between November and March could be due to something other than a change in cloud cover in response to a change in upper tropospheric ozone. Lets look therefore at the temperature of the upper troposphere at 20-30° south in the context of temperature change over time/ Is this the canary in the coal mine?

Figure 12 Temperature of the sea and the upper troposphere at 250hPa at 20-30° south in January.

Figure 13 Temperature of the sea and the upper troposphere at 250hPa at 20-30° south in July.

In January we observe a close relationship between the temperature of the upper troposphere at 20-30°south and the temperature of the sea. The so called ‘amplification factor’ is plainly there.

In July we see a decline in 200hPa temperature between 1948 and 1978 in line with the decline in the temperature of the northern hemisphere during that interval. There is a strong increase in 200hPa temperature after 1978 as the northern hemisphere warmed strongly. We know that the temperature of the stratosphere at 20-30°south is linked to the extent of warming in the northern hemisphere in mid year. The greater the convectional updraft that occurs north of the equator in mid year, the more voluminous is the stream of air that descends in the winter hemisphere. So, as the north warms the greater will be the outgoing radiation and the warmer will be the stratosphere and the upper troposphere in the southern hemisphere. The warmer it is, the less extensive must be the reflective ice cloud umbrella.

Figure 14 Anomalies in temperature at 200hPa, 300hPa and at the sea surface 20-30° south. Three month moving averages of monthly data.

Looking now at the monthly anomaly (departures from the 1948-2011 monthly average) the dependance of surface temperature upon  upper troposphere temperature is plain to see. In a cooling cycle we see 200hPa temperature falling below 300hPa and sea surface temperature and rising above in a warming cycle. It is plain that the upper troposphere leads the way. The shift of in 200hPa temperature between 1976-1980 had its origins in the increase in the temperature of the entire Antarctic stratosphere at that time led by change over Antarctica. It is this that is responsible for the  warming in the tropics and in part the northern hemisphere.


The $64,000 question is what causes the change in the ozone content of the high cloud zone between November and March when the greatest variability in global surface temperature is seen.

The $164,000 question is what is causing cloud cover to rise and fall on  decadal and centennial time scales.

The answer to both questions lies in the activity of the coupled circulation of the stratosphere and the troposphere at the poles that feeds ozone into the troposphere. The upper troposphere warms or cools depending upon the feed rate of ozone. The feed rate changes over time.

The ozone content and temperature of the upper stratosphere depends in the first instance upon the activity of the night jet at the poles that introduces NOx from the mesosphere. Less NOx means more ozone.  The activity of the night jet depends upon surface pressure and the concentration of NOx in the jet depends upon solar activity. In Antarctica, surface pressure has been falling for sixty years indicating a continuous increase in the ozone feed into the troposphere, the second major influence upon the ozone content of the polar stratosphere.

In that the coupled circulation of the stratosphere and the troposphere over Antarctica changes surface pressure at 60-70° south it changes the strength of the westerly winds in the southern hemisphere,  cloud cover and surface temperature on all time scales. Stratospheric ozone is wasted above and below the stratosphere, processes referred to as ‘unknown dynamical influences’ in the more respectable polar ozone studies.

These phenomena are the very essence of the Southern Annular Mode, arguably the fundamental mode of global climate variation on all time scales.

One thing is plain. High altitude ice cloud in the southern hemisphere is plainly a reflector of solar radiation. It does not promote warming (positive feedback). It promotes cooling (negative feedback). It’s presence depends upon the flux in ozone in the upper troposphere as governed by processes in the Antarctic and the Arctic stratosphere. These processes are tracked as the Arctic Oscillation and the Antarctic Oscillation indices that are acknowledged in more enlightened circles as prime modes of inter annual and longer term climate variation.

Suggested reading



  1. […] High level cloud and surface temperature ( […]

  2. Regarding the relative humidity and precipitable water graphs, for instance, what longitude was used for


    • Thanks for the question. Answer is 0-360° The entire globe.

  3. As a layman I wondered how much increased vapour clouds could offset the reduced albedo effect from arctic sea ice melting. When I look at a current satellite image I see what appears to be an increase of cloud formation/cover in the southern Antarctic region compared to the Northern Hemisphere. This appears to be contrary to this article’s premise that warmer air (as is now increasing in the Northern Hemisphere) creates more cloud formation. Is this an anomaly or am I misreading the data?

  4. Hi Mike, thanks for your comment. I suggest that the intensity of cloud in high latitudes depends upon the feed of moisture from the tropics. That in turn depends on pressure relativity between the equator and the poles. The hemispheres compete for tropical moisture depending on the relative pressures of the polar sinks the dominant sink is always Antarctica due to the planetary low in surface pressure on the margins of the continent where polar low pressure systems circulate.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s


%d bloggers like this: