For forecasters, it's no GOES for a while.

For forecasters, it's no GOES for a while

Mirrors, mirrors, for the sky: Are there any not awry?

That's the question haunting NASA as the agency finds itself beset with yet another major mirror flaw--this time in a $1 billion series of five weather satellites, the first scheduled for launch in February 1992.

Unlike the imperfection recently detected in the Hubble Space Telescope's primary mirror, this one was spotted on the ground. But mirror replacement could delay the launch of the first of the new weather satellites for a year, potentially hampering efforts to monitor hurricanes and other severe storms, says James Greaves, NASA's program manager for meteorological satellites.

These satellites constitute the third generation of Geostationary Operational Environmental Satellites (GOES), which provide visual and infrared snapshots of North America's weather every half-hour -- enabling the instrument to monitor huge weather patterns that cannot be tracked by radar, polar-orbiting satellites or airplanes. Only a single GOES now orbits Earth, and controllers expect it to run out of fuel sometime in 1993, says Elbert W. Friday, director of the National Weather Service in Silver Spring, Md. If that satellite fails before the launch of its successor, known as GOES-NEXT, it could force U.S. meteorologists to rely on less sophisticated tools that cannot reliably track hurricane development and progression, he adds.

The impaired optics of the weather satellites have no relation to Hubble's mirror flaw. Scientists found problems while testing a GOES prototype late last year, but attributed them to instrument misalignment rather than to defects in the pair of flat, oval mirrors that relay light to those instruments. However, a new computer simulation reveals that exposure to partial sunlight would cause the flat mirrors to arch "like a potato chip," says Joseph Dezio, GOES deputy program manager at NASA's Goddard Space Flight Center in Greenbelt, MD.

The temperature gradient created when sunlight strikes one edge of the mirror while the other edge remains in darkness -- a scenario occurring daily and lasting several hours -- temporarily causes the mirrors to deviate from flatness by about 300 millionths of an inch, an error large enough to significantly dim the images of clouds, oceans and storms, Dezio says. If new optical coatings can reduce the distortions, GOES-NEXT might be delayed by only two to four months, he says. If the remedy requires replacing mirrors, it could take a full year and several million dollars.

On the Hubble front, a NASA investigative team announced last week it had identified the likely cause of the "spherical aberration" that has crippled the telescope's ability to image faint objects. The panel reported finding an error in the "null corrector," a testing device used by mirror manufacturer Perkin-Elmer Corp. of Danbury, Conn. (now called Hughes-Danbury Optical Systems, Inc.) to check the shape of the primary mirror. An extra millimeter of space between a focusing lens and a small mirror in the null corrector apparently led the company's scientists to grind the primary mirror incorrectly, says Sarah Keegan, spokeswoman for the investigative panel.

Panel member Roger Angel, of the University of Arizona at Tucson, says Perkin-Elmer scientists discounted measurements taken with a second type of null corrector, which indicated the primary mirror might be misshapen, because that device had less precision.