washington

Cape Canaveral from above: Landsat 7 will capture around 450 high-quality images every day.

The US government is about to keep a seven-year-old promise by offering new Landsat images of the Earth at low prices, following a failed attempt to commercialize remote-sensing data in the 1980s and 1990s.

The Landsat 7 satellite, launched in April, will also provide much more thorough coverage of the Earth than its predecessors, with increased spatial resolution.

Project managers at the US space agency NASA and the US Geological Survey (USGS), which initially will share responsibility for operating the satellite, have validated its onboard systems and had hoped to begin routine image collection by 15 August.

But technical problems with outside users connecting to the Earth Observing System (EOS) data gateway have pushed back the date, probably to late August or early September, according to Ralph Thompson, programme manager at USGS for Landsat 7.

Like previous versions, Landsat 7 will return multispectral images with a resolution of 30 metres. But it adds a panchromatic channel with a resolution of 15 metres and a thermal infrared channel with a resolution of 60 metres. Users will be able to combine the high-resolution black-and-white images with the lower-resolution data to produce synthetic colour images with greater detail.

The satellite marks the return of Landsat to government ownership. In 1992, Congress admitted that attempts to privatize Landsat had failed and passed a law calling for the government to operate the next satellite in the series and to sell its data at cost price.

But the advent of low-cost imagery was delayed by the loss of Landsat 6 shortly after its launch in 1993, and by problems with the power-supply hardware of the Enhanced Thematic Mapper on Landsat 7 before the launch of the latter.

Now that Landsat 7 is operating in orbit as planned, images will be much cheaper. The Earth Observation Satellite Company (EOSAT), which marketed Landsat images in the mid-1980s, charged up to $4,400 for a single 185 by 170 kilometre ‘scene’.

Even a modest land-use study could therefore rack up a hefty bill for pictures. “You really couldn't afford it,” says Samuel Goward, chairman of the University of Maryland geography department, who heads NASA's 14-member Landsat science team.

Raw Landsat 7 images will be sold through the USGS EROS Data Center in South Dakota for just $475 a scene. This should put Landsat imagery in the hands of many more researchers, says Goward.

Equally important, he says, is the programme's commitment to collecting cloud-free images of the entire land surface of the Earth, with regular seasonal updating. Targets of interest can be photographed every 16 days, the period between Landsat passes over a particular location.

Around 150,000 Landsat 7 scenes will be collected each year — 250 every day at the EROS Data Center, and another 200 at the dozen or so international Landsat stations. These will store images of regional interest and set their own pricing policies. International users will also be able to buy images from the EROS centre.

Commercial software packages and more powerful desktop computers have combined to make the data easier to interpret and manipulate, he says. But he estimates that “we're only halfway” to the point where an untrained scientist can download a Landsat image and start using it for research.

According to Goward, a decade of high prices and patchy data availability (EOSAT had reduced the number of pictures taken by Landsat to save costs) have slowed the development of easy-to-use image analysis packages.

He adds that the Landsat science team and others can develop software to bridge the gap between the relatively small group accustomed to using remote-sensing images and the wider scientific community.

Thomas Lillesand, director of the University of Wisconsin at Madison's Environmental Remote Sensing Center and current president of the American Society for Photogrammetry and Remote Sensing, says the drastic reduction of Landsat data costs is partly a recognition that the commercial market in remote sensing is about to shift to higher-resolution data.