Water and ice clouds

Now we'll take a closer look at the way Earth's water and water-ice clouds effect the radiation budget, taking account of the balance between the shortwave albedo effect of clouds which act to cool the planet and the longwave cloud greenhouse effects which act to warm the planet. Much of the general behavior in evidence on Earth applies equally well to water clouds on other planets, or for that matter to any cloud-forming substance which is strongly absorbing in the infrared but fairly transparent in the solar spectrum.

To set the stage, we'll first discuss some calculations with the ccm radiation code which show how high clouds affect the albedo and OLR under typical tropical conditions. The results are shown in Fig. 5.9. These calculations include most of the radiative effects operating in the real Tropics, including solar and infrared absorption by water vapor and CO2, though we have left ozone out of the picture. The surface albedo has been set to zero, so as to focus on the reflective effect of the cloudy atmosphere itself. In this calculation, the tropospheric temperature profile is on the moist adiabat, and we place a geometrically thin cloud with specified water content in the upper troposphere, where the pressure is 283 mb and the temperature Tc of the cloud is 243 K. At these temperatures, the cloud is primarily composed of ice, and the ccm radiation model makes use of the complex index of refraction appropriate to water-ice particles. Results are given as a function of the condensed water path of the cloud. As long as the cloud is geometrically thin enough to be essentially isothermal, the actual geometrical thickness of the cloud is irrelevant to this calculation.

For 30 ym particles, which are typical of actual tropical ice clouds, it only takes a path of 50 g/m2 to make the cloud act like a blackbody. The OLR is about 20 W/m2 below the blackbody emission aT4 of the cloud itself, because there is some water vapor and CO2 greenhouse effect in the colder air above the cloud. This effect would disappear if the cloud were placed at the minimum temperature part of the atmosphere, and would increase if the cloud were lower. In essence, a cloud that is optically thick in the infrared acts like a new "ground", radiating upward into the upper part of the atmosphere with a blackbody temperature Tc. If the particles are made smaller, it takes less cloud water in order to make the cloud optically thick, because the same mass of water yields more aggregate cross-section area of cloud particles. As a practical matter, most high clouds occuring in the vicinity of deep convection in the Tropics can be considered optically thick in the infrared. The associated cloud greenhouse effect is enormous, and would lead to an uninhabitably hot planet if not compensated by shortwave albedo effects that are of similar magnitude.

Ice Clouds Greenhouse

Figure 5.9: Albedo and OLR as a function of cloud condensed water path, for a high ice cloud with temperature Tc = 242K at a pressure pc = 283mb. The temperature profile is on the moist adiabat corresponding to a surface temperature of 300K, patched to an isothermal 180K stratosphere. The relative humidity is 50% and the CO2 concentration is 300ppmv, but there is no ozone in the atmosphere. Calculations were done with the ccm radiation code, and results are shown for both 30 pm and 10 pm particles.

Figure 5.9: Albedo and OLR as a function of cloud condensed water path, for a high ice cloud with temperature Tc = 242K at a pressure pc = 283mb. The temperature profile is on the moist adiabat corresponding to a surface temperature of 300K, patched to an isothermal 180K stratosphere. The relative humidity is 50% and the CO2 concentration is 300ppmv, but there is no ozone in the atmosphere. Calculations were done with the ccm radiation code, and results are shown for both 30 pm and 10 pm particles.

The albedo effect of the cloud also increases monotonically with cloud water content, but at a much slower rate than the greenhouse effect, for the reasons discussed in Section 5.6. For 30 pm particles, the planetary albedo has only reached .2 when the cloud water path is 50g/m2. The albedo doesn't reach .5 until the cloud water content approaches 200 g/m2. On the other hand, the failure of cloud albedo to saturate until large cloud water paths means that the particle size can have a very important influence on albedo; reducing the particle size to 10 pm increases the albedo to .7 for a cloud containing 200 g/m2 of ice. As for compensation between shortwave and longwave cloud effects, taking a cloud with 30 pm particles and 100 g/m2, the albedo is about .4, which yields 170 W/m2 reduction in solar absorption based on typical annual average tropical insolation. This compares with a cloud greenhouse effect of 120W/m2, so such clouds have a moderate net cooling effect. If the cloud only had a water content of 50 g/m2, though, the cloud greenhouse effect would be nearly the same but the cloud albedo effect is reduced by nearly half, and the cloud would have a net warming effect. Similarly, if the ground were partially reflecting (owing to vegetation cover or low-lying clouds), the change in albedo due to high clouds would be reduced, shifting the balance again in favor of net cloud warming. On the other hand, reducing the particle size of the clouds makes them much brighter, making it easier for the clouds to have a net cooling effect.

For fixed particle size, the cloud altitude has relatively little effect on albedo for a given cloud condensed water path. Low altitude liquid water clouds tend to have smaller particles than ice clouds as well as larger water content (because there is more water around to condense), and are correspondingly more reflective. Because of this effect, the balance of power for mid-level and low-level clouds shifts decidedly toward a net cooling effect on the planet.

High clouds have both a warming and a cooling effect, and which one wins depends on the detailed of the cloud properties, including cloud temperature, particle size, and condensed water content. Colder cloud temperatures tend to favor net warming. Making particles smaller or increasing cloud water enhances the cooling effect except for very thin clouds, since it takes little cloud water to make the cloud act like a blackbody whereas the albedo continues to increase with cloud water increase or particle size decrease, for the reasons discussed in Section 5.6. The balance between albedo effect and greenhouse effect also depends on the intensity of solar radiation, since albedo matters not at all if there is no sunlight. In the polar night, clouds have an unambiguous warming effect as long as they are not right at the surface. Similarly, the cloud albedo effect depends on the albedo of the underlying surface; clouds over a reflective surface like ice (or surface clouds!) will tend to have a warming effect, as also discussed in Section 5.6. As the cloud is made lower, the cloud greenhouse effect is attenuated, because the cloud temperature is closer to the ground temperature and also because (especially in moist regions) the greenhouse effect of the clear air above the clouds masks the longwave radiative effect of the cloud itself.

Although the large and competing effects of clouds on OLR and albedo pose similar challenges on any planet whose atmosphere contains a condensible substance, Earth is the only case at present for which we have good observations of the net radiative effect of clouds. The first satellite mission to do this accurately was the Earth Radiation Budget Experiment (ERBE), and subsequent missions have taken a similar approach. We discussed some ERBE clear-sky results in Chapter 3, and now we will see what ERBE has to tell us about cloud effects. The ERBE mission measured the Earth's radiation budget using two sets of highly accurate broadband radiometers borne on satellites - one in the infrared spectrum and one in the shortwave (i.e. solar) spectrum. Moreover, the processing algorithm made use of the patchiness of Earth's cloud cover in order to estimate the effect of clouds on the longwave and shortwave radiation. Within each scene examined (think of a scene as a 50km square patch of the Earth's surface) the algorithm identified those pixels which represented cloud-free clear-sky conditions, and defined "clear sky" longwave and shortwave flux as the value the flux over the scene would have if the flux of all pixels in the scene were replaced by the average of the clear-sky pixels. In the longwave, for example, the ERBE retrieval reports the all-sky OLR, called OLRan and the clear-sky OLR, called OLRciear. The cloud longwave forcing is then defined as OLRclear — OLRall. Since clouds reduce the OLR by making the upper troposphere more optically thick, the cloud longwave forcing is positive, and represents a warming effect. Similarly, the cloud shortwave forcing is defined as Sabs,all — Sabs,clear, where Sabs is the top-of-atmosphere absorbed solar radiation - the difference between incoming and reflected solar radiation. Clouds reduce the solar absorption by increasing the albedo, and so the cloud shortwave forcing thus defined is generally negative, representing a cooling effect. The sum of the cloud longwave and cloud shortwave forcings is the net cloud forcing, with positive values representing a warming tendency and negative values representing a cooling tendency. Clear sky and all-sky albedo can be defined similarly.

Results for clear and cloudy albedo, and for the cloud radiative forcing are shown for the year 1988 in Figure 5.10. Other years show a similar pattern. The ERBE dataset contains information of this sort for each month, reported on a latitude-longitude grid. Here we show only annual-mean results averaged along latitude circles. The full monthly-mean dataset for all available years is provided as part of the dataset collection in the supplementary materials for this book. Turning attention first to the clear-sky albedo, we see that without clouds the albedo varies in a narrow range of .11 to .16 from 60S latitude to 42N. Poleward of 60S, the albedo increases sharply owing to the high albedo Antarctic ice. Still, the values indicate that the albedo of the partially snow-covered Antarctic ice must exceed .7, since the atmospheric absorption makes the planetary albedo lower than the surface albedo. The clear-sky estimates near the pole are somewhat unreliable, since it is hard to distinguish between low clouds and ice. Going towards the North pole from 42N the albedo increases somewhat more gently, owing to the patchy distribution of sea ice and its seasonal fluctuations; the rest of the Northern high-latitude albedo increase is due to winter snow-cover over land. Clouds have a strong reflective effect, approximately doubling the tropical albedo and increasing the midlatitude albedo to .4 or more. The area-weighted mean albedo is .19 for clear

Shortwave Albedo Cloud Radiative Forcing (W/m2)

Figure 5.10: Zonally averaged annual mean clear and cloudy sky albedo(left panel) and cloud radiative forcing (right panel) measured by ERBE for the year 1988.

sky and .33 including cloud effects. Area-weighting doesn't take into account the seasonal and latitudinal distribution of sunlight though; a more appropriate mean albedo is based on taking the ratio of global net reflected solar radiation to the incident radiation. This estimate yields somewhat smaller values: a mean clear-sky albedo of .16 and a mean all-sky albedo of .30.

If uncompensated by the cloud greenhouse effect, the high albedo of clouds would probably be sufficient to throw the Earth into a Snowball state. In reality, the reduction of OLR by clouds cancels most of the cloud cooling effect, as shown in the right hand panel of Figure 5.10. The cloud longwave forcing - i.e. the reduction in OLR due to clouds - is anti-correlated with the cloud shortwave forcing, and has sufficient magnitude to cancel most of the cloud shortwave forcing. The distribution of cloud forcing takes us into some consideration of aspects of the general circulation we have not introduced previously. Somewhat North of the Equator there is a region of deep convection yielding deep, thick clouds, which is manifest in the Figure as a peak in both the cloud longwave and shortwave forcing (marked "ITCZ" in the figure for Inter-Tropical Convergence Zone, in honor of the winds which converge moisture into this region and feed the convection). The ITCZ is flanked by two subtropical regions where convection is suppressed by downward motions in the atmosphere, and is shallow or absent. Here one encounters local minima in both the cloud longwave and shortwave forcing. Throughout the tropics, the two terms sum to a net cooling effect of about -20W/m2, which is stronger in the subtropics than near the ITCZ. The subtropical cloud cooling is in part due to near-surface clouds which are associated with the boundary layer rather than deep convection. In the midlatitudes, there is another region of deep cloud activity. This one is associated with the large scale organized storm tracks, which loft water from the subtropical ocean and move it poleward and upward. The albedo effect of clouds more strongly dominates the greenhouse effect in this region, and even more so towards the Antarctic region, where there is strong cloud shortwave forcing associated with low-lying marine stratus clouds. As a result there is strong net cooling in the midlatitude and polar regions.

The area-weighted global mean cloud longwave forcing is 28W/m2, while the mean cloud shortwave forcing is —47W/m2, which nets out to a cooling influence of —19W/m2. Using a sensitivity factof of 2.2W/m2K from Section 4.5, we conclude that the Earth would be about 8.6K warmer if there were no clouds.


In circumstances under which the clear-sky regions do not absorb much solar radiation, high clouds have a potent net warming effect, though there must be enough convection around to loft water to sufficiently high altitudes to make a high optically thick cloud. As we have already mentioned, clouds can contribute significantly to deglaciation of the high-albedo Snowball Earth, though the main question there is whether it is possible to make sufficiently high clouds with sufficiently great water content in an atmosphere with low water content (because of low temperature) and sluggish convection (because of low solar absorption). Another situation in which the clear-sky solar absorption is low is the high-latitude winter. Here, there is little solar radiation to reflect from clouds simply because it is night or twilight most of the time, and so if there are high clouds they will have a pronounced winter warming effect, and perhaps even inhibit the formation of sea ice in open water conditions. This effect may play a role in the Arctic during the Cretaceous hothouse climate, since there is open water in the Arctic Ocean which can maintain a supply of relatively warm water throughout the winter to feed deep convection. It seems plausible that this mechanism would help explain the mysterious low-gradient climate of Cretaceous and similar hothouse climates, described in Section 1.9.1. General circulation models to date do not support a sufficiently strong cloud effect for clouds to be the answer to the Cretaceous puzzle, but there is much remaining to be learned about clouds, so the last word has not by any means been uttered on this topic. This potential mechanism is only viable when there is open water in the polar ocean. Over a polar continent, such as Antarctica, the ground would cool off rapidly in the Winter, foreclosing any serious possibility of deep convection and the associated deep clouds.

The effect of clouds on the water-vapor runaway greenhouse represents one of the most vexing and important unresolved issues in planetary climate. The observed behavior of Earth clouds can provide little guidance as to cloud effects in a much warmer atmosphere in which water vapor is the dominant component. Given the availability of water throughout the depth of the atmosphere, and how little water it takes to make a highly reflective cloud, it seems almost inevitable that the albedo will become very high. There is no simple physics, however, that can be employed to estimate the fraction of the atmosphere which will be cloudy; this is an intrinsically dynamical question. High clouds can also reduce the OLR, however, which has the potential to offset the albedo effect. A strong cloud greenhouse effect is likely, given that the top 100 mb of a near-runaway steam atmosphere contains far more water vapor than Earth's entire atmosphere. In order for clouds to make the radiating temperature cold enough to offset the strong cloud albedo increase, one would need optically thick clouds in regions of the atmosphere which are very cold. This is not impossible, though: a thick cloud at an upper atmospheric temperature of 200 K would reduce the OLR to 91 W/m2, which is just about the same as the solar radiation Early Venus would absorb if it had an albedo of 80 %. In Earth's atmosphere, optically thick tropical cirrus clouds occur at similar temperatures, so one can hardly rule them out for Venus. The question of whether the cloud greenhouse or cloud albedo effect wins out in a runaway situation simply cannot be answered by back-of-the-envelope calculations, even if we have a rather large envelope. A definitive answer must await attainment of a far better understanding of convection, cloud microphysics and cloud fraction under near-runaway conditions.

Was this article helpful?

0 0

Post a comment