Cloudy clues.

When atmospheric physicist Anthony DelGenio looks at clouds, he isn't planning today's weather report, or even the five-day forecast. He studies clouds to predict Earth's climate.

Clouds play a powerful role in warming and cooling our planet, DelGenio explains. "Even a one percent change in cloud cover can have dramatic effects on global temperature," he says.

In the next century, a slight increase or decrease in clouds could determine whether or not your hometown will be a "hot" place to live. That's because clouds have a dual role in balancing Earth's heat budget. On the one hand, clouds spread out like global umbrellas; blocking the sun's heat energy and reflecting much of it back to space. That cools the planet, says DelGenio.

But clouds can also act like giant, insulating blankets. The water droplets they're made of absorb and trap heat radiating from Earth, which helps to keep the planet warm and livable.

All life processes on Earth depend on a delicate balance between this warming and cooling. If the balance shifts, the effects could be drastic. Colder winters or hotter summers could threaten entire species of plants and animals, and force whole populations--humans included--to move to different latitudes, says DelGenio.

He and other scientists are now studying clouds in an attempt to forecast such climatic changes. Many of these scientists predict global warming--a rise in Earth's temperature--during the next century. They presume this warming will occur as levels of carbon dioxide and other greenhouse gases--spewed by the burning of fossil fuels--continue to build up in the atmosphere. When these gases accumulate, they trap heat, and cause temperatures to rise above normal.

But a looming question remains: If global warming does occur, how severe will it be? No one yet knows, says DelGenio, and clouds are a puzzling variable in that equation.

What is known is how clouds form: Water vapor in warm air rises. As it gains altitude, it cools and condenses on particles of dust. This condensation causes either water droplets or ice crystals to form. When they crowd together--presto! A cloud.

Applying this recipe for cloud formation, scientists hope to find answers to these questions: If Earth's oceans warm, as predicted, will more water evaporate, rise, and form more clouds? If so, will an increase in clouds cover block more sunlight and offset global warming? Or, will the added clouds blanket the Earth, trapping more heat and causing temperatures to rise even more?

These "unknowns" are the heart of DelGenio's studies. But finding answers raises another problem: How on earth do scientists even study clouds?

CLOUD VARIETY

Just by looking up in the sky (or at the pictures on these pages), you can begin to grasp why clouds make such difficult "lab specimens." For one thing, masses of clouds cover half the globe during any given month. In addition, individual clouds differ widely in shape and size. Some stretch out over the horizon and pile up 10km thick. Others are fleeting wisps.

What's more, each cloud type reflects and absorbs the sun's energy in varying amounts. For example, smooth sheets of stratus clouds, which bring light rain, tend to reflect more sunlight than they absorb. Result: a cooling effect on the land below. Yet feathery, high-altitude cirrus clouds, composed of tiny ice crystals, allow more sunlight through, thereby warming the Earth.

In addition, clouds can disappear as quickly--or slowly--as they form. "Some last 10 minutes, some 10 days," says atmospheric scientist Stephen Cox. "Some clouds last an entire season."

Finally, consider this: To make effective cloud/climate predictions, scientists must look at how the whole variety of clouds will perform in the future.

What tool can possibly take all these factors into account? Look no further than a personal computer.

CLEARING THE FOG

Computers can store reams of climate data--past, present, and future--and solve equations faster than any scientist, Cox says. Just as you might build a cotton-and-clay diorama to illustrate the science of cloud formation, scientists construct their climate models with computer programs.

Instead of using hobby shop materials, though, computer modelers simulate atmospheric conditions using numbers. Each number represents a variable in the cloud/climate equation-things like heat, air pressure, winds, and water vapor.

Next, the scientists write formulas to predict, for example, how increasing carbon dioxide levels will affect these variables, and therefore cloud formation, in the future. They then plug the variables into the formulas, and wait while the computer performs all the necessary calculations.

Soon the computer prints out its results--pages and pages with rows of the predicted numbers for each variable.

These numbers bear little resemblance to actual clouds. To make sense of this dizzying data, climate modelers display it visually: in maps, on graphs, or as pictures of clouds, says Cox. By identifying trends or patterns in those charts or pictures, scientists then try to make their climate predictions.

For instance, one computer model might show that if carbon dioxide levels rise, cloud cover will decrease by the middle of the next century. A scientist might then predict that with fewer heat-reflecting clouds, more heat will reach Earth, thereby worsening global warming. Based on such predictions, nations might increase efforts to reduce their use of fossil fuels.

Of course, computer modeling is not a perfect science. "There are hundreds of different processes that the models simulate at once," says Cox. Scientists must often base their formulas on estimates, which combined can oversimplify atmospheric conditions and throw off climate predictions.

Under these circumstances, says DelGenio, there are no guarantees. "I think it's possible that ... before the decade is over we will have the cloud problem sorted out." But, he adds, "you can never know for sure if you've gotten your prediction right until it happens."