Mars projection with shortlisted landing sites, two of which are in the southern hemisphere. Credit: MOLA SCIENCE TEAM

Six potential landing sites have been chosen for NASA's Mars Science Laboratory, a large rover set to assess the past habitability of sites on the planet's surface, when it lands in October 2010 (See image). The shortlist, chosen from dozens of possibilities, includes craters partly filled with sediment, an ancient flood channel and regions rich in clay minerals thought to date from an era when the martian surface was wetter than it is today. However, changes to the mission's scope mean that options that might offer excellent science could end up being dismissed as impractical.

If you ask an engineer, they'd like to land in a Walmart parking lot.

The mission has a 'landing ellipse' roughly 20 kilometres across to account for the uncertainties involved in guiding a spacecraft over millions of kilometres to a soft landing on a windy planet. The terrain in the ellipse needs to be smooth and flat. ?If you ask an engineer, they'd like to land in a Walmart parking lot,? says Jack Mustard, a planetary geologist at Brown University in Providence, Rhode Island.

As originally conceived, once landed, the rover would have been able to travel well outside this ellipse to places neither smooth nor flat ? the sorts of outcrop that geologists favour. But participants at last month's workshop to choose the candidate sites, hosted by NASA's Jet Propulsion Laboratory in Pasadena, California, found that sites where the rover would need to travel 10 kilometres or more to obtain samples were now being flagged as possibly problematic.

?I would have been screaming at that,? says Ken Edgett, of Malin Space Science Systems of San Diego, who is a principal investigator on the mission but was unable to attend the workshop because of the wildfires in California. ?It limits your expectations,? says Mustard, who favours a site in the Nili Fossae region from which the rover would be able, if all went well, to sally forth to a region of dramatic erosion that he has dubbed Monument Valley.

Changes in the way that the rover's moving parts will be lubricated raise issues for sites in the planet's southern highlands, as they reduce the rover's capabilities in winter conditions. Nevertheless, the scientists shortlisted two southern sites as worthy of further study. One of them, Holden Crater, contains what seem to be lake sediments and a delta. ?It's awesome,? says Mustard.

The shortlisted sites will now be scrutinized further by instruments on board NASA's Mars Reconnaissance Orbiter, which is currently circling the planet. At the same time, computer models will assess the risk of winds at the sites being strong enough to mess up the landing. It is very unlikely that all six will be considered too risky, says John Grant of the Smithsonian Institution's Center for Earth and Planetary Studies in Washington DC and co-chair of the site-selection committee. The final decision does not need to be made until nearer the launch in October 2009.