CERES Edition 4 and the Cloud Radiative Effect

Guest Post by Willis Eschenbach

One of the enduring questions in climate involves what is usually called “cloud feedback”. When the earth warms up a bit, the clouds change in response. The question is the direction of that response. Does the change in clouds amplify a warming, or does it reduce a warming? Here’s Dr. Roy Spencer on the subject:

The total amount of CO2 humans have added to the atmosphere in the last 100 years has upset the radiative energy budget of the Earth by only 1%. How the climate system responds to that small “poke” is very uncertain.

The IPCC says there will be strong warming, with cloud changes making the warming worse. I claim there will be weak warming, with cloud changes acting to reduce the influence of that 1% change.

The difference between these two outcomes is whether cloud feedbacks are positive (the IPCC view), or negative (the view I and a minority of others have).

So far, neither side has been able to prove their case. That uncertainty even exists on this core issue is not appreciated by many scientists!

After some shenanigans, I’ve beaten my computer back into shape, and I’ve found the time to download the latest version of the CERES satellite radiation datasets. This version is called “Edition 4”. “CERES” is an acronym for Clouds and the Earth’s Radiant Energy System. The CERES project is collecting radiation data from three different satellites. The CERES project is online here.

There are two groups of CERES files—the top-of-atmosphere (TOA) datasets and the derived surface datasets. See the “DATA” note at the end of the post for information on where to get the two datasets.

The CERES TOA datasets are the actual measurements of the upwelling and downwelling shortwave (solar) radiation and longwave (thermal infrared) radiation.

The surface datasets, on the other hand, are derived from the TOA datasets along with a variety of other both satellite and surface data.

In previous Editions of the CERES data there have been TOA datasets for what is called the “cloud radiative effect”, or “CRE”. Figure 1, for example, shows the net effect of clouds on TOA radiation. This is at the inherent resolution of the CERES data, which is given on a one-degree latitude by one-degree longitude basis.

Figure 1. TOA net cloud radiative effect. Negative values show net cloud cooling; positive values show net cloud warming.

Note that overall, the effect of the clouds is to cool the system by about -18 W/m2 (negative values show cooling).

Now, clouds have two opposing effects on the radiation in the Earth’s climate system. The first of these effects is that clouds cool the surface by reflecting sunlight back into space. We experience this cooling effect when a cloud covers the sun on a hot afternoon.

The second effect is more subtle. Clouds increase the thermal radiation coming down from above us. We can experience this as well, but usually only on clear winter nights. On those nights, when a low-lying cloud comes over, it instantly replaces our view of outer space with a relatively warm cloud.

Warm relative to what, you might ask? Relative to the infinite heat sink of outer space. From outer space, we get on the order of 3 W/m2 of what is called “cosmic background radiation”. From a low-lying cloud, on the other hand, which is at say a temperature of about freezing, we get on the order of three hundred W/m2 of thermal radiation … and as a result, when clouds come over on a clear winter night you can feel the temperature difference. On such nights, you can feel how the advent of the clouds leaves the surface warmer than it was without the clouds.

So clouds have two opposite effects—less sunshine due to cloud reflections cools the surface, but more downwelling thermal radiation warms the surface. These opposing cloud radiative effects are referred to as the “shortwave” (solar) CRE and the “longwave” (thermal infrared) CRE. The difference between radiation gains and losses is referred to as the “net” CRE.

And the question, as Dr. Roy pointed out above, is this: what happens to the net CRE when the surface warms?

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.

More to the current point, in the derived CERES Surface datasets, there are now datasets for surface cloud radiative effect. In past CERES Editions, there have been TOA cloud radiative effect datasets (shortwave, longwave, and net). But this is the first Edition with surface cloud radiative effect datasets. Here’s the net effect of clouds at the surface.

Figure 2. Surface net cloud radiative effect. Negative values show net cloud cooling, positive values show net cloud warming.

The difference between looking at the net CRE from the top (Fig. 1) and from the bottom (Fig. 2) of the atmosphere is instructive. From the surface, the changes due to the clouds along the intertropical conversion zone (ITCZ, blue line above the Equator) are much larger than from the top of the atmosphere.

And to return to the main question, I thought I’d look at the changes in surface CRE with respect to changes in surface temperature. To do that, on a 1° latitude x 1° longitude gridcell basis I first removed the month-by-month seasonal variations in both the local cloud radiative effect and the local temperatures. Then, again on a 1°x1° gridcell basis, I calculated the trend of the local net CRE with respect to the local changes in temperature. Figure 3 shows that result:

Figure 3. Change in surface net cloud radiative effect for a 1°C increase in surface temperatures. Negative values show increased cooling in response to surface warming.

Now, this is a most fascinating result for a variety of reasons. First, note that the global average change of CRE with temperature is 0.0 W/m2 per degree C. This might explain why there is so much debate even as to the sign of the change in cloud radiative effect with increasing temperature.

Next, I have argued for a long time that the effect of clouds and thunderstorms in the tropics is to regulate the temperature. I have provided a variety of evidence showing that on a daily basis, when the tropical ocean is cool, clouds form later in the day and in fewer numbers. This allows the sun in to heat up the ocean. On the other hand, when the tropical ocean is warm, clouds and thunderstorms form earlier in the day and in greater numbers. This both cuts down on incoming sunshine and cools the surface directly in a variety of ways. See “Air Conditioning Nairobi, Refrigerating The Planet” for a discussion of surface cooling.

Figure 3 clearly shows that my hypothesis is correct. In the tropics, as the surface warms, the clouds act to increase their cooling effect. Not only that but in some areas, the effect is very strong. Over the “Pacific Warm Pool”, the warmest part of the world’s oceans shown in blue in Figure 3, there is a very strong cap on further warming. For each degree of warming, the clouds cool the surface by up to -20 W/m2.

Next, Figure 3 shows the problem with global averages. As a global average, the change in the surface net CRE per degree of surface warming is zero … but in the all-important tropics, the change in the CRE is strongly cooling.

Next, Figure 3 shows an additional reason why it has been hard to determine even the sign of the change in the CRE with changes in temperature—it’s negative over the ocean and positive over the land.

In assessing all of this, note that these month-long averages obscure what is actually happening on the ground. The temperature regulation actually occurs on a minute-by-minute basis, and only where and when the surface temperature exceeds or drops below a critical local temperature threshold. So when the cumulus cloud formation temperature is exceeded at say eleven AM in the tropics, the emerging clouds immediately reflect hundreds of watts per square metre of solar energy back to space. The cooling and warming occur as needed, when needed, where needed in order to maintain the temperature within a narrow range. The figures above show the month-long averages of these very precise, threshold-based, and discrete interventions in the evolution of the temperatures.

Anyhow, I think that’s all the scientific fun I’m allowed to have in a twenty-four-hour period … here, it’s been raining off and on all week. I’ve been marveling at the plants growing up between the bricks on the patio.

The surprising part about those plants is that after the rain, they’ve captured an amazing amount of water on their leaves.

Here’s a closeup:

Almost every leaf has managed to intercept and hold some water. I rate that as an amazing trick, a solar collector that doubles as a water catcher … note how the leaves are arranged so that the upper surfaces are flat and level, which keeps the drops from rolling off.

One last fun fact to make it all even stranger. The flat and level arrangement of the leaves shown above is not permanent, far from it. At night, this same plant folds its leaves up. The three two-lobed leaves of each head fold inwards, and the lobes fold back outwards so that what is the upper surface in the picture above becomes the inner surface. This cuts the surface exposed to the outer air in half, thus conserving both heat and internal water throughout the night.

Then the three-leaved head folds over so that the leaves hang down vertically. This reduces the profile of the plant and buries the heads down among the lower stems and foliage, which cuts down on wind-chill and thus also reduces overnight heat loss.

Astounding.

My best to you all on this amazing planet, where even the commonplace is extraordinary,

w.

My Usual Request: When you comment, please quote the exact words you are discussing. I can and am happy to defend my own words. I cannot defend you attacking something you think I said.

DATA: The CERES TOA dataset is available here and the surface dataset is available here  I simply downloaded everything, which is about three-quarters of a gigabyte for the TOA dataset and about 0.9 gigabytes for the surface dataset. However, each individual dataset is only about 50 Mbytes, and you can download just what you wish.

DATA INFO: The CERES TOA datasets are as follows:

toa_sw_all

toa_lw_all

toa_net_all

toa_sw_clr

toa_lw_clr

toa_net_clr

toa_cre_sw

toa_cre_lw

toa_cre_net

solar

cldarea_total_daynight

cldpress_total_daynight

cldtemp_total_daynight

cldtau_total_day

The CERES surface datasets are as follows

srf_sw_down_all

srf_sw_down_clr

srf_sw_up_all

srf_sw_up_clr

srf_lw_down_all

srf_lw_down_clr

srf_lw_up_all

srf_lw_up_clr

srf_net_sw_all

srf_net_sw_clr

srf_net_lw_all

srf_net_lw_clr

srf_net_tot_all

srf_net_tot_clr

srf_cre_net_sw

srf_cre_net_lw

srf_cre_net_tot

There is no CERES surface temperature dataset. In lieu of that, I use a direct conversion of the CERES surface longwave upwelling all-sky radiation dataset to temperature by means of the Stefan-Boltzmann equation. I have checked this converted temperature dataset against various other temperature datasets (HadCRUT, Reynolds SST, UAH MSU, TAO buoys) and found it to be in very good agreement.

Abbreviations: sw – shortwave; lw – longwave; clr – clear sky conditions; all – all sky conditions; cldarea – cloud area; cldpress – cloud pressure; cldtemp – cloud temperature; cldtau – cloud optical depth; cre – cloud radiative effect; up – upwelling; down – downwelling; tot – total lw + sw