Glimpsed Through The Clouds

Glimpsed Through The Clouds

Guest Post by Willis Eschenbach

In a recent post, I discussed the new CERES Edition 4.0 dataset. See that post for a discussion of the CERES satellite-based radiation data, along with links to the data itself. In that post I’d said:

I bring all of this up because there are some new datasets in CERES Edition 4. In the CERES TOA group, there are now measurements of cloud area, cloud pressure, cloud temperature, and cloud optical depth. These are quite interesting in themselves, but that’s another story for another day.

It turns out that “another day” is today, so here is an overview of the new cloud datasets.

Let me start with cloud area. This is expressed as the “cloud area fraction”, the percentage of the time that from the satellite’s point of view the surface is obscured by clouds. Figure 1 shows cloud area fraction around the globe.

CERES cloud area.png

Figure 1. Cloud area fraction. Red shows the areas with the most clouds. Blue shows the cloudless deserts, including Antarctica, the frozen desert.

Some items of interest. I can see why photos of the Southern Ocean are usually overcast. Also, the land on average only has 2/3 of the oceanic cloud coverage.

Next up is the cloud average visible optical depth. From the AMS Meteorology Glossary:

cloud optical depth

The vertical optical thickness between the top and bottom of a cloud.

Cloud optical depths are relatively independent of wavelength throughout the visible spectrum, but rise rapidly in the infrared due to absorption by water, and many clouds approximate blackbodies in the thermal infrared. In the visible portion of the spectrum, the cloud optical depth is almost entirely due to scattering by droplets or crystals, and ranges through orders of magnitude from low values less than 0.1 for thin cirrus to over 1000 for a large cumulonimbus. Cloud optical depths depend directly on the cloud thickness, the liquid or ice water content, and the size distribution of the water droplets or ice crystals.

CERES cloud optical depth.png

Figure 2. Average vertical optical thickness of the clouds

The most surprising feature of that map is the infamous “brown cloud” over China …

Now, the next two datasets are about the cloud tops. They give the pressure and temperature of the cloud tops. But those numbers don’t mean a whole lot to me. I can’t envision what a cloud top at -34°C and 300 hPa pressure means. What I really wanted to look at was the altitude of the cloud tops.

So I used those two to calculate the altitude of the cloud tops (details in the endnotes). Figure 3 shows how high the tops of the clouds are.

CERES cloud top altitude.png

Figure 3. Cloud top altitude (km).

We see that the tallest clouds are the towering cumulonimbus thunderstorm clouds over the Pacific Warm Pool to the north of Australia.

I was glad to see these new datasets because I hoped that they would provide further evidence in support of my hypothesis about the thermal regulation of the global temperature. I’ve espoused for many years now the hypothesis that one of the largest thermoregulating mechanisms involves the timing of the daily emergence and the extent of first the daily tropical cumulus field, and then the ensuing development of tropical thunderstorms.

I have said that these processes are threshold based. For example, in the daily cycle of tropical weather, when morning temperatures exceed some local threshold, the cumulus field starts developing. Within about half an hour to an hour, the cumulus field is fully developed.

Then, if the day continues to warm, when some temperature threshold is passed we start to see thunderstorms. And as with the cumulus field, once the threshold is passed, further thunderstorm development is quite rapid.

So my hope was that the data would support my hypothesis. To investigate that, I looked at a scatterplot showing how clouds respond to different ocean temperatures. First, here is how cloud area fraction responds to different temperatures. I use the response over the ocean because it is free of the dozens of other factors involved over the land (altitude, slope, soil moisture, mountains and valleys, plants, etc.).

CERES scatter cloud coverage vs sst.png

Figure 4. Scatterplot of sea surface temperatures versus cloud area fraction.

Can you say “non-linear”? I knew you could. This is why averages are often meaningless, or worse, misleading. But I digress.

There appear to be three separate regimes going on here. The first is on the left, below freezing, where we’re looking at clouds over sea ice. In that regime, cloud coverage increases with temperature.

In the middle section, from freezing to somewhere around 26°C (79°F), cloud coverage generally goes down with increasing temperature.

Finally, in the tropics, somewhere above 26°C, cloud area fraction starts increasing rapidly, just as my hypothesis predicts. As tropical temperatures rise, cloud area fraction increases at a very steep angle.

Next, the cloud optical depth is shown in Figure 5.

CERES scatter cloud optical depth vs sst.png

Figure 5. Scatterplot of sea surface temperatures versus cloud optical depth

The overall shape of the changes in optical depth is similar to the shape of the changes in the cloud area shown in Figure 4 above. However, optical depth doesn’t start rising at 26°C. It bottoms out a bit warmer, around 27°C, and increases from there. I suggest that this increase in optical depth is the sign of the increasing development of thunderstorms.

Finally, we have the cloud top altitude. As shown in Figure 3, in tropical thunderstorms the altitude of the tops can average 8 km (5 miles). Here is how cloud height varies with temperature.

CERES scatter cloud top altitude vs sst.png

Figure 6. Scatterplot of sea surface temperatures versus cloud top altitude

If you ever wanted a temperature hockeystick, there it is … you can see how rapidly the thunderstorms boil upwards towards the tropopause once the temperatures get warm enough.

My conclusion? These graphs absolutely support my hypothesis that tropical cumulus and thunderstorms act together to keep tropical temperatures, and hence global temperatures, within a fairly narrow range (e.g. ± 0.3°C over the entire 20th century).


ONCE AGAIN WITH FEELING: I’m tired of people accusing me of things I never said. I’m fed up with vaguely couched attacks on something I am claimed to have written sometime somewhere. I’ve had it with my ideas being taken out of context, twisted, and then fed back as something I’m supposed to have claimed. So please, QUOTE WHAT YOU ARE TALKING ABOUT! I am likely to get stroppy if you don’t quote whatever it is you are on about. Quote it, or don’t bother posting.

MATH: Here’s the R function showing the relationship between air temperature (Tair), surface atmospheric pressure (surfpress), pressure at altitude (pressure), and altitude. Hashmark (#) shows that what follows is a comment.

elevation.from.pressure =   # function name

function(Tair = -30, pressure = 300, surfpress = 1013.171)  {     # default values

-29.3*(Tair+273.15)*log(pressure/surfpress)/1000    # altitude calculation in km


Because I didn’t know the average surface atmospheric pressure for each gridcell, I used the global average surface pressure of 1013 hPa. This leads to a possible error of about ± 1% in the calculated altitude, far too small to be meaningful for this analysis.

Superforest,Climate Change

via Watts Up With That?

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