By Willis Eschenbach – Re-Blogged From http://www.WattsUpWithThat.com
After the turn of the century, I became interested in climate science. But unlike almost everyone else, I wasn’t surprised by how much the global temperature was changing. As someone with experience with heat engines and engine governors, I know how hard it is to keep a heat engine stable under a changing load. As a result, I was surprised at how little the temperature was changing.
Over the 20th Century, for example, the temperature changed by a trivially small ±0.3°C. Since the average temperature of the planet is on the order of 287K, this means that the global temperature varied only about a tenth of one percent in a hundred years … that that is amazingly stable.
So I started my climate science investigations by looking for some kind of long-term mechanism that would keep the temperature stable. I read about the slow weathering of the mountains that constrains the CO2 levels. I thought about long slow changes in the ocean overturning. I looked at all kinds of long-term mechanisms … and found nothing that could constrain the temperature for 100 years. I thought about this for over a year. No joy.
Then one day I had an insight. I’d been looking at the wrong end of the time spectrum, the long-term, century-long end. I should have been looking at hours and days instead. I realized that if there were some mechanism that kept each day from getting too hot or too cold, it would keep the week from getting too hot or cold, and the month, and the year, and the century.
I was living in Fiji at the time. Every day, I watched the daily parade of tropical clouds and thunderstorms, and one day I realized that I was looking right at the very mechanism that I sought. But I still didn’t understand it. Where in all of the comings and goings of the clouds and the thunderstorms was the understanding that I sought? I thought that it might have something to do with the timing of the clouds and storms, but what?
I finally had another insight, that there was a point of view from which it all made sense. This was the point of view of the sun, which is a most curious point of view.
From the point of view of the sun, it’s always daytime, and there is no night. Not only that, but there is no earth time. From the sun, the left edge of the earth is always at dawn. Right under the sun, it’s always noon. And the right edge of the earth is always at dusk.
Intrigued by this, I went and got a series of pictures from the GOES-West satellite, taken at the same time of day. I averaged all of these photos, and this was the result.
Figure 1. Average of one year of GOES-West weather satellite images taken at satellite local noon. The Intertropical Convergence Zone is the bright band in the yellow rectangle. Local time on earth is shown by the black dashed lines on the image. Time values are shown at the bottom of the attached graph. The red line on the graph is solar forcing anomaly (in watts per square meter) in the area outlined in yellow. The black line is the albedo value in the area outlined in yellow.
It’s an oddity. By looking from the point of view of the sun, I’ve traded the time dimension for a space dimension. This lets me look at the evolution of the tropical day.
As you can see in the black line at the bottom of Figure 1, at around 10:30 in the morning the clouds start to build up. Within an hour the cumulus field is fully formed, and it maintains that level throughout the rest of the day.
My hypothesis was that this combination of cumulus clouds and thunderstorms formed a moveable sunshade. On warmer days, the sunshade moves to the left in Figure 1, and the clouds and thunderstorms start earlier in the day. This cools the day down. And on cooler days, the sunshade moves to the right, and the sun warms the surface.
As I said, all of this took some years. Finally, in 2009 I published my hypothesis as The Thermostat Hypothesis here at WUWT. Then I re-wrote it and submitted it to Energy and Environment, where it was peer-reviewed and published in 2010.
Since then I’ve been gathering supporting evidence for my hypothesis and developing it further. I realized that there are other emergent phenomena that contribute to the planetary thermoregulation. These include inter alia the El Nino/La Nina pump that moves warm water to the poles where it is freer to radiate to space; dust devils that move heat on land from the surface to the troposphere; the PDO and other ocean current shifts that alternately impede and assist the polar movement of warm water; cyclones moving heat out of the ocean and into the atmosphere; and squall lines that increase the efficiency of thunderstorms in refrigerating the surface.
I wrote a number of posts on various aspects of these emergent phenomena. However, what I didn’t have was data on the response of thunderstorms to surface temperatures. According to my hypothesis, both clouds and thunderstorms should increase with increasing surface temperature. In 2015, I was able to approach demonstrating this indirectly using the CERES data, by showing the correlation of tropical albedo and temperature. Figure 2 shows that relationship.
Figure 2. Correlation, total albedo and surface temperature.
As you can see, towards the poles the two are negatively correlated. This is because ice and snow melt with increasing temperatures and the albedo goes down. But in the tropics, as my hypothesis predicted, they are positively correlated—clouds increased with increasing surface temperature.
However, this still didn’t show that the thunderstorms were also correlated with temperature. However, in the most recent edition of the CERES data, Edition 4.0, there are four new datasets. These are cloud area, cloud top temperature, cloud top pressure, cloud area (percent), and optical depth.
Now, I got to thinking the other day about the El Nino 3.4 area of the ocean. This is one of the most variable areas of the Pacific as far as temperature goes because it is at the heart of where the El Nino/La Nina phenomenon occurs.
I also found that I could convert the CERES data to actual cloud top height. To do the conversion you need the sea level pressure, the cloud top temperature, and the cloud top pressure. I had two of these, so I got the HadSLP2 gridded dataset of sea level pressure. Using those three I calculated the cloud top altitude.
Now, where is the Nino 3.4 area? It’s in the mid-Pacific. It’s the blue rectangle in Figure 3 below.
Figure 3. Average cloud top altitude, CERES data, Mar 2000 – Feb 2017
You can see the preponderance of the tallest thunderstorms over the “Pacific Warm Pool” above Australia. A couple of months ago I posted about my first look at the CERES cloud dataset in a post called “Glimpsed Through The Clouds“. At that time I made a movie of the cloud height overlaid with contours of the sea surface temperature. I repost that movie below to show the close correspondence of temperature and thunderstorms.
That showed the general agreement between thunderstorms and temperature, but nothing in detail. So to return to the El Nino 3.4 area, the area shown as the blue rectangle above, I graphed the sea surface temperature of that area alone. Figure 4 shows those temperatures.
Figure 4. Sea surface temperature in the Nino3.4 Region.
As I said, the Nino 3.4 area has some of the most variable sea surface temperatures in the tropical Pacific. In Figure 4 you can see the large El Nino of 2015/16, along with the three smaller El Ninos in 2002/3, 2006/7, and 2009/10.
Next, I took a look at the cloud heights in that area over the same period. Figure 5 shows both the sea surface temperature and the heights of the cloud tops.
Figure 5. Sea surface temperature (black) and cloud top heights (red) in the Nino3.4 Region.
Wow! I expected a correlation, but I never expected something that close. I figured that there might be other factors involved such as CAPE or wind shear, but they seem to be very minor players.
These CERES cloud datasets have provided the first clear evidence supporting my hypothesis that tropical thunderstorms are critical parts of the global thermoregulatory system, not in a general sense, but in a clear, step-by-step, month after month sense.
Finally, I wanted to take a look at the tropical cumulus cloud area. As with the thunderstorms, my hypothesis requires that the cumulus field begin earlier in the day and cover more area. Here is the temperature, along with the cloud area as a percentage of the sky.
Figure 6. Sea surface temperature (black) and cloud area (blue) in the Nino3.4 Region.
Once again, we have an extremely close correlation between the two variables, temperature and cloud area. Since thunderstorms do not generally cover a large amount of the sky, these would be mostly cumulus clouds.
• As I hypothesized a decade ago, tropical cumulus clouds and thunderstorms do form an active governing system that acts to oppose any temperature variations by changing the timing and the strength of the daily emergence of the cumulus field and the associated thunderstorms.
• The thunderstorm connection is demonstrated by the very close correspondence between the temperatures and the strength of the thunderstorms as measured by average cloud height.
• The thunderstorm connection is demonstrated by the very close correspondence between the temperatures and the strength and timing of the thunderstorms as measured by average cloud height.
• The cumulus field connection is demonstrated by the very close correspondence between the temperatures and the strength and timing of the cumulus as measured by average cloud coverage.
• More clouds and thunderstorms when the ocean is warm cool the surface in a variety of ways, including cloud albedo changes, increased evaporation, cold rainfall, and as I’m writing this I remember one more thing … surface albedo changes. Hmmm … hadn’t thought of that in a while.
In my original hypothesis, I said that one of the ways that the thunderstorms increased the albedo was by forming breaking waves and spume, both of which are white and reflect more sunlight. In addition, the albedo of rough water is greater than that of calm water. So I hypothesized that thunderstorms would increase the surface albedo in a couple of ways … I should look at that as well. Hang on while I pull up that data … OK, here’s Figure 7, hot off the presses …
Figure 7. Sea surface temperature (black) and surface albedo (purple) in the Nino3.4 Region.
More good news. This is the first evidence I’ve found for this minor part of my hypothesis, the claim that thunderstorms cool the surface in part by increasing surface albedo. And while the change is small, about half a percent, it represents a change of ~ 4 W/m2 in absorbed solar energy. This is more change in absorbed energy than would result from a doubling of CO2.
So that’s what I found out today about the situation in the Nino 3.4 region …