Even flow: Brazil plans to reroute a river to bring water to the parched north, but some say it's wasting money.

Until recently, Joao Alves do Nascimento's family drank dirty water, filtered through a cotton rag, from a mud pit located a two-hour hike across the scrub from their cinderblock home. "There was nothing else to drink," the father of two says, "though there's water just outside our door."

The Nascimento home, in the hamlet of Riacho Grande in Brazil's sandy, arid northeast, sits by a lake formed from a dam on the Rio Silo Francisco, the country's second-biggest river after the Amazon. But the lake flooded a salt mine, rendering its water undrinkable.

So, the home also sits amid a great geographic irony: While Brazil boasts more freshwater than almost any other country on the planet, a shortage of the means to collect and distribute it means that legions of poor Brazilians, especially in the northeast, go thirsty.

The Nascimentos are lucky: A local charity recently installed a cistern, a circular, concrete tank that collects rain from a rooftop, storing fresh water outside their house. But the federal administration hopes to obviate such remedies through a controversial US$1.80 billion project that would divert water through two long canals from the Silo Francisco River to reservoirs elsewhere in the northeast.

The project, early construction on which was scheduled to begin late 2005, is at the heart of a tropical-style New Deal the president--a beleaguered but pragmatic leftist--promised but hasn't delivered since elected in 2002. The rerouted water, the government predicts, will stem migration from northern Brazil south and encourage investment, reviving the fortunes of a Venezuela-sized region long afflicted by at least nine months of drought a year. "It will ease the hardship of millions of Brazilians;' says Pedro Britto, the project's coordinator. "This is the beginning of a transformation for the entire northeast."

Some Brazilians are skeptical. Critics--including water-rights groups and many of the waterless themselves--see it as a boondoggle for the construction industry, a costly venture that will benefit only a select fraction of the region's inhabitants. Opposition is built around one key criticism: The more than 640 kilometers of proposed canals fail to address the broader problem of distribution.

In theory, the nine states afflicted by drought have enough water from existing dams and reservoirs to meet basic needs. Even after supplying cities, the reservoirs suffer more from evaporation than lack of supply. And the brief wet season provides rain that, if stored, could buoy reserves.

What's lacking, critics charge, is a network to collect and carry water to the region's poorest, most remote corners. To redirect part of the Sao Francisco to the same reservoirs, opponents argue, is to move water toward those who already have it--namely, urbanites and wealthy, industrialized farmers.

Hardship. "It's construction for the sake of construction," says Roberto Malvezzi, national coordinator for the Pastoral Land Commission, a Catholic organization that espouses property rights for the poor. "It will not alleviate the hardship."

Projects to hydrate the northeast--many variants of the Silo Francisco plan--have existed for centuries, including a proposal in the 1840s by Dom Pedro II, a Brazilian monarch, who vowed, unsuccessfully, to do so even if it meant selling the crown jewels.

For President Luiz Inficio Lula da Silva, the project hits close to home. As a boy, his family hopped on the back of a truck and made the days-long journey from the northeastern state of Pernambuco to Silo Paulo, then Brazil's burgeoning industrial heart, in search of work. Millions made similar migrations during the last century, depriving the northeast of human capital and cramming today's overcrowded cities of the south.

Though Lula opposed previous versions of the project, the president, with a year remaining before elections, now seeks a legacy of progress through massive public works. When an ongoing corruption scandal last June prompted a shuffle among advisors, he ordered Dilma Rousseff, his new chief of staff, to put it high on the administration's agenda.

"The idea of a big project overshadows the reality that this is not a smart plan;' says Fernando Bezerra Coelho, mayor of Petrolina, a Pernambuco city on the river's northern bank and a member of the Socialist Party, which backs Lula's governing coalition. Petrolina is a rare pocket of prosperity because of a successful irrigation program that turned the city into a hub for mango growers. Officials here offer sentiments common elsewhere along the river. Even before dams upstream diminished the Sao Francisco's downstream flow, the river, on its own, did little to quench the thirst of its baked shores.

"The very edges suffer here, so how will it cure ills elsewhere?" asks Vitorio Rodrigues, a Petrolina environmentalist. The project's planners dismiss such criticisms as selfishness among those who already enjoy the Silo Francisco's bounty. But critics abound on the receiving end of the canals, too. Joao Abner, a former director of water services for the government of Rio Grande do Norte, a would-be recipient state, in a report earlier this year attacked the math the administration uses to promote the project. Of the 12 million intended beneficiaries, for instance, some 75% already live in cities or other areas with access to the reservoirs, according to the report. "The numbers don't add up," says Abner, now an engineering professor at the Federal University of Rio Grande do Norte.

What does add up, opponents charge, is the expense. While the government says it would be more expensive not to build the canals--social costs from two years of drought would exceed the project's cost, it argues--independent studies question the investment.

After a 2001 analysis of a similar but larger proposal for the canals, the World Bank told the Brazilian government in a letter that secure supplies of household water for the entire northeast could be guaranteed by alternatives at a fraction of the cost.

Subsistence. One of those alternatives--the construction of cisterns to store rainwater--helped the Nascimentos end their daily trek to the mud pit. With drinking water now assured, the cistern enables them to use the time they once spent looking for water to tend to the few crops that take root in the sandy ground, such as manioc.

Part of a project by ASA Brasil, a non-governmental organization dedicated to the region's water problems, the Nascimentos' cistern is one of a handful in Riacho Grande, a two-hour journey by road and by stream from Petrolina.

The Nascimentos and most of their neighbors, also subsistence farmers, get by on far less than Brazil's monthly minimum wage of $122. So remote is the hamlet that some residents still speak of distances in terms of leagues--not kilometers--and ferry their leftover produce, mostly ground manioc flour, to markets in the nearest town by canoe and horseback.

Driving there in a pickup truck along the side of a bumpy road, ASA Project Inspector Antonio Paiva Mangabeira points to a big fruit farm. Its green orchards, irrigated with water pumped from the Sao Francisco, end abruptly, giving way to singed, ocher bushes. "There's no real shortage of water," he says. "Just a shortage of means to pay for it."

PARTING THE WATERS

Rerouting a large river like the Silo Francisco is no easy task,
but the government says it's necessary to bring drinking water
to the northeast. Some facts and figures on the waterway:

Length          2,700 kilometers
Size of basin   634,000 square kilometers
Tributaries     168
States served   Minas Gerais, Bahia, Pernambuco, Alagoas, Sergipe
Population      13 million
Project cost    US$1.80 billion

SOURCE: Brazilian Ministry of National Integration

PAULO PRADA * RIACHO GRANDE, BRAZIL