Endeavour's payload bay with the radar mast extended earlier in the mission. Photo: NASA

Space shuttle Endeavour's astronauts are setting their sights on returning to Earth on Tuesday after mapping a vast majority of the planet, but where the shuttle lands in anyone's guess.

Gusty crosswinds and low clouds could conspire to keep Endeavour from a planned Florida homecoming at the Kennedy Space Center. The Cape is NASA's desired landing site because that is the home base of the shuttle fleet.

Mission controllers will have two shots at sending Endeavour to KSC on Tuesday. The first would begin with a firing of the shuttle's twin orbital maneuvering system engines at 3:53 p.m. EST (2053 GMT) to slow Endeavour slightly, slipping the craft from orbit to start an hour-long guide back to Earth. Touchdown on KSC's three-mile-long runway would occur at 4:50 p.m. EST (2150 GMT).

If weather conditions are unacceptable on the first try, NASA will keep Endeavour in space for one additional orbit. The second and final landing option at KSC on Tuesday would start with a deorbit burn at 5:23 p.m. (2223 GMT) and landing at 6:22 p.m. EST (2322 GMT).

"The problem is there is a high pressure center up in the northeast part of the United States," said entry flight director John Shannon. "That flow is coming right down in the Florida peninsula, picking up the some low-level clouds just off shore and bringing them up onto the landing site."

Should the weather in Florida not cooperate, the space agency could divert the shuttle to the alternate landing site at Edwards Air Force Base in California, if conditions there are acceptable. The landing opportunity would begin with a deorbit burn at 6:50 p.m. (2350 GMT) and landing at 7:48 p.m. EST (0048 GMT).

But there are concerns at Edwards too, with the forecast calling for clouds and a strong headwind at the landing site.

"It is not by any means in the bag," Shannon said of the chances to bring Endeavour to Mojave desert military base.

Overall, it is suspected that Edwards is the best chance for landing Tuesday.

If all three landing opportunities are waved off on Tuesday, Endeavour's six international astronauts will get another 24 hours in space, extending the mission to 12 days. On Wednesday there will be two chances to land at both KSC and Edwards.

However, forecasters are predicting weather at KSC and Edwards to be "no go" on Wednesday and Thursday.

With such a gloomy forecast for the rest of the week, NASA officials are considering activating the backup landing site in New Mexico for use on Wednesday.

The White Sands Space Harbor, or Northrup Strip, was used once for a space shuttle landing -- STS-3 on March 30, 1982 by Columbia.

Conditions are expected to be favorable at White Sands on Wednesday. The site will not be "called up" on Tuesday to support landing.

Space shuttle Columbia approaches the dry lakebed at White Sands for landing in 1982. Photo: NASA

"That option has not been entirely ruled out for Wednesday or Thursday," Shannon said. "Right now, it's a little premature. We're not doing anything with White Sands (on Tuesday). But if the weather degrades and the forecast is for it to continue to degrade, if that happens we'll consider Northrup. But that is still a decision we would not make before going through all three opportunities (on Tuesday)."

"Since we haven't sure it in a long time, we want to make sure that everybody is think of everything if we had to call it up," Shannon said.

Furthermore, the space agency says White Sands no longer has the capability to provide live television via NASA TV, meaning Endeavour's homecoming on the gypsum runway would not be seen until later during a video tape replay.

The last 20 shuttle landings have occurred at KSC, a streak started in May 1996. The most recent landing at Edwards was in March 1996 by shuttle Atlantis after the third docking with Russia's space station Mir. In addition, 27th of the last 28 shuttle missions have concluded in Florida.

"We've gotten real lucky with the weather. But sometimes we run into those days where you have a weather system off the East Coast and a weather system off the West Coast, and it clobbers both sites," Shannon said. "And that is kind of what we're looking at (this week)."

NASA likes to avoid landing shuttle anywhere but Florida to save the $1 million cost of ferrying the spaceships atop a modified 747 jumbojet back to KSC.

NASA says Endeavour only has enough fuel and power to stay in space until Thursday.

The STS-99 mission used a complex radar system to map the Earth's surface where 95 percent of the world's population lives. The data will be used over the next two years to create the highest resolution 3-D surface elevation map of the planet.

SRTM image of Drakensberg, South Africa. Photo: DLR

On Monday, the astronauts shut down the radar mapper after 222 hours and 23 minutes of data collection. The radar images filled 332 high density tapes and covered 99.958 percent of the planned mapping area - land between 60 degrees north latitude and 56 degrees south latitude - at least once and 94.594 percent of it twice.

"We'd like to congratulate you on a flawless operation of this most sophisticated mapping instrument in the universe," Mission Control radioed the astronauts. "This topographic database will be a real treasure of the human race for many years."

The crew on Monday also retracted the 197-foot long accordion-like radar mast protruding from Endeavour's payload. However, the boom failed to collapse all the way into its 9-foot canister, sticking out by about an-inch-and-a-half.

After three more tries and more torque, the canister was closed and the lid latched into place. Three locks were engaged, securing the $35 million mast -- the longest rigid structure ever extended in space -- for the return to Earth aboard Endeavour.

If the efforts to stow the mast were unsuccessful, NASA would have tossed the mast and canister into space.

"I feel like I have to almost apologize for the almost boring nature of this flight. We didn't rendezvous with anything, there were no spacewalks," said project manager Michael Kobrick. "But it shows you what a real science measurement experiment is like. It's tedious, time consuming, you have to work very carefully."