QC script improvements

[bpbond]

1. Why generate those CV graphs for each core and print them (1-9) instead of just saving (10-25)? Ugh.
2. Script crashes if there's a NA core. Make more robust.

[bpbond]

1. Better diagnostics / ID for NA core observations, e.g. when analyzer is left running overnight.

[bpbond]

• Core mass plot - normalize all relative to 100% (initial starting mass)?
• CO2 and CH4 fluxes - plot by incubation day, not date?
• CV of CO2 and CH4 observations - by date AND incday
• Would like 9 most core means AND 9 most variable treatment means
• Make CH4 variability plots as for CO2

[bpbond]

I've added a new script (5-highvar.R) specifically aimed to investigating problematic data. See commit e101683.