The DIMES system involves adding a lightweight camera to the lander that would take three successive photos of the surface in the 18 seconds before the braking rocket is ignited. The photos would then be immediately analyzed by the spacecraft's computer to locate surface features and calculate an estimate of the lander's sideways drift that would be added to the gyro package's tilt data to decide which TIRS rockets to fire up.

The DIMES camera system seems to work well in tests - but the decision as to whether it's worthwhile to add it to the MER spacecraft to further reduce landing risks won't be made until November or December.

And even with the whole TIRS system added, winds still remain a serious problem for the engineers building the MER landers to fully factor in.

Already, new data on possible wind speeds in different places on Mars has forced two of the top six candidates for MER landing sites to be rejected at last March's MER landing site selection workshop - both of the very scientifically interesting sites within the great Marineris Valley, which unsurprisingly channels winds to higher speeds.

Since then, another site - in the Athabasca Valley, which may be the site of recent volcanic eruptions and geothermal liquid-water outbursts less than 20 million years ago - has been rejected because Earth-based radar studies indicate that its ground may be unacceptably rough and rocky. And of the remaining three sites, studies are still going on to determine whether two of them have horizontal winds too strong for the landing risk to be acceptable.

Gusev Crater

Those two sites are the floor of Gusev Crater - which has a giant, apparently water-carved channel draining into it, suggesting that the crater may have been a water-filled lake during Mars' early days - and another landing ellipse located in the southern part of the Isidis Basin, whose southern rim has many smaller possible water-carved channels draining down it on to the flat basin floor, so that a lander would again have a good chance of finding ancient water-formed sediments.

The only original candidate site which scientists are certain has acceptably low winds and smooth terrain is the giant field of hematite - again, probably deposited by ancient liquid water - which the Mars Global Surveyor's IR spectrometer has mapped in the Terra Meridiani region.

It's now a near-certainty that one of the two MER rovers will land there - but, while Gusev is the other favorite judged strictly by scientific interest, it may be that both it and Isidis will ultimately also have to be rejected due to excessive wind risks.

This has led to the landing site selection committee hastily locating one more backup landing site -- a flat, scientifically rather dull site on the Elysium Plain that is a candidate only because it has low winds.

The final selection of the MERs' two landing sites won't be made until next January. And - even with all this - it is still not entirely certain that it will be considered safe to launch either of the MERs next year.

The final tests of the ultimately selected airbag design are still underway, and three candidates for the MER's big landing parachute also remain to be tested to determine which of them best combines good braking ability with the minimum tendency to be blown sideways by winds.

Officially JPL remains highly confident that the final designs for both the chute and the airbags will have been picked by the end of October - but if more unexpected problems turn up and satisfactory airbag and parachute designs still haven't been picked by next February, it will become impossible to launch the MERs to Mars in 2003.

Moreover, if additional problems rear their heads in the next few months, NASA will have to seriously consider delaying one of the MERs to a later launch window, allowing the remaining project funding to be fully focused on finishing adequate design and testing for the one remaining MER so that it can meet the 2003 deadline.

The cost-effectiveness of this is questionable, however. And in any case, NASA would like to avoid delaying either rover to a later launch if at all possible, because 2003 is an unusually favorable launch opportunity to Mars with a higher payload capacity available.

By contrast, the next opportunity in 2005 is one of the worst launch windows in the cycle. If one or both rovers do have to be delayed, the current feeling is that their landing on Mars will be set not to occur until 2008.

This doesn't necessarily require that their launch be delayed all the way into 2007, however - especially since the U.S. already plans one other Mars launch in both 2005 and 2007.

Instead, the delayed MERs may be launched in 2006, onto a path that allows them to complete one and a half orbits of the Sun before finally intercepting Mars. (The same strategy was followed with the Magellan Venus probe.) They might even be launched in 2004 and carry out three and a half solar orbits before reaching Mars, if the Mars Program's overall funding profile favors such a deployment.

Moreover, any delayed launches -- since 2003 was an unusually favorable Mars launch opportunity -- will probably have to use the bigger and considerably more expensive Atlas 3 booster (which is already planned for the 2005 U.S. Mars Reconnaissance Orbiter)." It should be emphasized, however, that this is strictly worst-case planning. MER's managers feel that they finally do have a good handle on the project's problems, that launch of both rovers in 2003 is now highly probable, and that simulations show that each one has a 90 percent or so chance of surviving landing even without the DIMES sensor added.

But this project has given JPL - and NASA - a vastly harder time than anyone dreamed possible when it was picked, and has played a major role in convincing engineers that Pathfinder's "simple" hard landing system is much less preferable to soft landing systems than had been thought.