Science Priorities for Mars Sample Return

VI.B.Number of Samples.

The number of samples needed to capture appropriate heterogeneity depends on the local Martian environment and geological history. Field experience on Earth has taught us the importance of acquiring sufficiently diverse samples to evaluate whether or not a specific result is representative as well as to extrapolate interpretations of processes from variations among and within samples. In many cases, carefully selected suites related rocks allow one to reasonably evaluate: 1) how representative each sample may or may not be of the geologic unit; 2) the consistency of processes creating and altering the samples; and 3) abundances of specific attributes such as minerals and geochemical signatures.

Without pre-characterization of a specific Martian site, it is not possible to define the number of samples required to capture local to regional diversity in geological materials. However, an estimate of sample number is necessary for mission planning. For many studies, a suite consisting of about five to eight samples would be sufficient for a first-order evaluation of the heterogeneity of units, the consistency of processes, and the abundances of representative features. Two examples demonstrate this. In Endurance Crater, Mars, the analysis of seven stratigraphically distributed sites in the Burns Formation allowed the Opportunity Rover team to identify several significant diagenetic events, some of which were associated with variations in groundwater (McLennan et al., 2005). In a second example, APXS analyses of a eight separate samples of alkaline volcanics revealed that they were formed under different conditions or from very different starting composition compared to the bulk of Martian rocks; thus, shedding new insight on the complexity of the Martian interior (McSween et al., 2006). In both of these examples, a smaller number of the “right” samples could have provided sufficient information for the resulting interpretations, but pre-selection of the smaller set of samples would have required significant characterization. Thus, for Mars sample return, either extensive in situ characterization capabilities would be needed or a suite of at least five to eight samples should be collected from each geological unit. More samples would provide better information, but a suite of five to eight samples should provide sufficient diversity to provide substantial scientific return.

Quantity of samples to be cached by MSL

The MSL cache is designed to accept rock samples 0.5–1.5 cm in size provided by MSL's soil scoop, collected over 5–10 separate caching events (Karcz et al., 2007). The cache will have mesh sides to allow fines to filter out, leaving behind rocks. The strategy for employing the scoop to acquire either individual targeted rocks, or rock-bearing regolith would depend on further experience with prototype scoops (note that the volume of the scoop is roughly half the volume of the cache). The empty cache is specified to have a mass less than 52 g. However, the mass of the latest revision of the design (as estimated by the CAD software) is 29 g (Karcz, writ. comm., 01-07-08). A draft specification (as of this writing) is for the mass of the cache container when full to be 200 g or less. Because the mass of the contents will be uncertain, it is likely that the science team will fill it to somewhat less than capacity--to, say, 180 g instead of 200 g.

• 
  Would a total mass of 700 g inside the returned container (instead of 500 g) break the mission?
  
• 
  Because it might not be possible to determine whether the MSL cache is recoverable until MSR gets there (for example, MSL may end its life in an inaccessible location), it would be advantageous for MSR to carry enough sample vials for the full set of samples described in Case B of Table 6, not just those in Case A.  That way either scenario could be accommodated.
  
• 
  The value of the MSL cache would need to be explicitly compared to an incremental eight rock and one regolith sample having a known context and far better curation--these are mass-equivalent.  This comparison could be meaningfully made only after the cache has been loaded and the MSL science team knows its contents.