BELLAGIO, Italy -- Will Planet Earth be able to handle the mega-surge of people pouring into the cities of Africa, Asia and Latin America?

Back in 1950, there were 2.2 billion of us, mostly spread across the world's rural areas. Today the United Nations estimates world population at 6.6 billion. Half live in cities where an accelerating human flood of rural people -- many desperately poor -- generates slums, endangers water and sewage systems, and breeds local misery and potential pandemics.

If today's birthrates continue unaltered, U.N. figures suggest there could be 11.7 billion people by 2050.
There is some good news here. Birth rates have declined as rural people migrate into cities and have fewer children than farm and rural families typically do. The mid-range population expectation for 2050 is 9.1 billion.

And humans have the power to effect huge change on our future numbers, Joel Cohen, head of the Laboratory of Populations at Rockefeller and Columbia universities, told a global Urban Summit, assembled by the Rockefeller Foundation, here last week.

If women, on statistical average, have half a child more than now predicted, Cohen noted, then the world population will soar to 10.6 billion. Conversely, if they choose to have a half child less, then the global population will rise to a comparatively more manageable 7.7 billion.

Already, decisions by families to have less children have brought us back halfway from the unprecedented fertility surge that increased world population by over 2 percent a year in the late 1960s.

The problem is that the global population base has increased so radically that even seemingly modest birthrates can have momentous consequences. Cohen calculates that if we do add 2.5 billion people by 2050, and virtually all this population increase, as expected, happens in poor countries, then the world will have to build one city of 1 million people every week for the next 43 years. "Is this," he asks, "feasible -- physically, environmentally, financially, socially?"

One sort of shudders at the answer. But there is a first step: Get a handle on growth of the world's cities. Without that, how can city leaders estimate the peripheral areas they'll have to urbanize or, alternatively, how much they'll have to "infill" their current territory with higher density development?

Breakthrough research on that very point, by Stephen Sheppard and his Williams College colleagues, was presented at the Bellagio conference. Using Landsat satellite images of a sampling of 120 world cities -- one set taken around 1990, another around 2000 -- they were able to show global cities' dynamic form of growth -- how much they move to the urban periphery ("outspill"), or find space inside ("infill").

On top of that, the Landsat readings permitted intensity light readings indicating types of land use, pixel by pixel, down to very small areas. Then the Williams team, operating with a remarkably small $775,000 budget from the World Bank and the National Science Foundation, matched its images with census-type information from each city to estimate actual population and per capita income.

What are they discovering? It turns out new residents don't cause a city to expand physically at quite the rate of population growth. But they still cause, on a worldwide average, seven times more outspill than infill. America's suburban sprawl may look different, but the pattern of geographic expansion is worldwide.

Now Sheppard can make a prediction mayors across the globe should prize: how fast the population growth will force them to infill or expand onto new territory. Just take the average of a region's population growth and income growth, he has found, and you can approximate the answer.

One wonders how cities will receive the news. At current trends, Sao Paulo, Brazil, for example, will need a stunning 7.5 square miles and Shanghai an even more extreme 14 square miles of new development every year.

The bottom line is clear: The developing world's cities -- and the developed world's cities still expanding significantly -- must plan early, much more carefully, or expect to be overwhelmed by a virtual growth tsunami.

Good planning, for example, can recycle underused urban land, or schedule better use of expansion areas, to achieve much greater people-carrying capacity. Good planning can avoid some of the worst modern traffic jams, put public transit first, make walking and biking convenient, and preserve pockets of green space critical to humans' physical and emotional health.

Sheppard sees a frightening tide of population growth enveloping cities. He urges they take their thousands of planning documents, too often focused on some ideal future, and update them to reflect realistic growth scenarios. If the 20th century was the century of freewheeling development worldwide, the 21st century needs to be one of much earlier, careful planning.

Is planning a boring word? Maybe? But how else can there be any hope for a sustainable century?