This section outlines the considerations for selecting landing sites on Mars with the primary goal of returning to Earth samples collected from roving vehicles.
Candidate sites should provide samples of unambiguous origin and should include major geologic units.
Collectively, the several sites explored should address as many different scientific questions as possible; individual sites should address a reasonable subset of those questions. Ideal sites would provide ready access to volcanic rocks of all ages ("young", intermediate and ancient), waterlain sedimentary rocks, old impact breccias, ground ice, soils, and windblown sediments.
The ability to acquire a variety of sample types of known provenence depends on having mobility once the spacecraft has landed. Mobility permits acquisition of rocks even when the landing is on unconsolidated debris. It permits travel from safe landing sites to sites of greater diversity and geologic interest. It permits movement from one geologic unit to another, thereby ensuring a variety of samples of known provenance.
Table 9.1 indicates how range and likely success are related. This table is based mainly on Viking experience. The two Viking sites are judged from orbital measurements of daytime and nightime surface temperatures to be more rocky than most of Mars. Even so, only a few small rocks were within reach of the landers, and all attempts to acquire and analyze small rocks failed. The "rocks" analyzed turned out to be clods of unconsolidated debris, not true rocks. Thus, at a typical Mars site there may well be a high probability that no rocks are within reach of a lander's mechanical sampling arm. In addition, drifts of poorly consolidated material several meters across occur throughout the Viking 1 site. But because, as noted above, the Viking 1 site is atypically blocky, such drifts are probably more extensive at other, more typical sites. Thus, in most cases mobility of a few to tens of meters may be required to confidently acquire any rock samples at all.
Mobile landers -- rovers -- also enable more than one rock type to be sampled at a given site (which may contain many different rocks from local geologic variety and, in the form of impact ejecta blocks, from distant locations also). How much mobility is enough? Uncertainties in the location of geologic boundaries are typically a few kilometers. Thus, movement of several kilometers from the landing point will be required to ensure that two geologic units can be sampled. Given the expected need to land at relatively smooth and less interesting sites, travel of several tens of kilometers appears required for both safety and for diverse sampling.
Choices of where to sample and which samples to retain (and which to throw away) will be made based on all available information: remote sensing and other instruments on the rover, remote sensing from orbit, and previously acquired data.
Digging trenches or drilling cores of unconsolidated material (soil) will be carried at only a few locations because such surface operations will be highly time consuming. Those sites will be chosen on the basis of information about composition below the immediate, visible surface (as can be obtained using a neutron spectrometer) and on the basis of measurements on soil samples heated in a small oven to release gases such as water vapor.
A photographic record will be made of the geologic setting of all acquired samples.