Simplistic Early Views
The earliest attempts to predict how changes in cloud cover would affect greenhouse warming concluded that they would have no net effect: clouds would neither speed nor slow an ongoing change in climate. That conclusion was based on the belief that any change that made clouds better at cooling would also make them more efficient at retaining heat near the surface. For example, if cloud cover were to increase (as many thought it would, assuming that warmer temperatures would speed evaporation), the amount of sunlight reaching Earth's surface would decrease, but then the thermal radiation trapped by the cloud would increase by the same amount.
Even such a simple scenario has difficulties, though. Because the decrease in solar heating would affect surface temperatures, and the change in the emission of thermal radiation would affect air temperatures at higher altitudes, additional cloud cover would change the temperature contrasts between the surface and the higher altitudes that drive the winds. Any reduction of winds might in turn inhibit the formation of clouds.
Another idea is that higher atmospheric temperatures could create denser clouds, since greater evaporation rates at higher temperatures would make more water vapor available in the atmosphere for cloud condensation. Because denser clouds reflect more sunlight, there could be an enhanced cooling effect. On the other hand, denser clouds might also lead to an increase in rainfall, possibly from storm clouds, whose tops are especially high and cold. Such clouds, which are particularly good absorbers of thermal radiation, could more than make up for their tendency to block sunshine. In that case warming would be intensified.
This kind of what-if discussion could go on indefinitely. All such changes are physically reasonable; the question is, Which ones actually will take place?
Another kind of complication is that clouds come in many forms, depending on the weather conditions that create them. Low, dense sheets of stratocumulus clouds hanging just above the ocean cool more than they heat. They make efficient shields against incoming sunlight, and because they are low — and therefore warm — they radiate upward almost as much thermal radiation as the surface does. In contrast, the thin, wispy cirrus clouds that soar 6,000 meters (20,000 feet) and higher reflect little sunlight, but they are so cold that they absorb most of the thermal radiation that comes their way. Hence they warm more than they cool. The net cooling effect of clouds is the sum of a large number of such specific effects, many of which cancel one another.
Atmospheric scientists have been aware for nearly two decades that the complex effects of clouds pose a major challenge to the understanding of climatic change. In 1974 an international conference of investigators in Stockholm highlighted the need for greater understanding of clouds as one of the two biggest obstacles to further progress in climate research. (The second was inadequate knowledge of ocean currents.) Recent comparisons of the predictions made by various computer climate models show that the problem has not gone away. In some models, for instance, clouds decrease the net greenhouse effect, whereas in others they intensify it.