Use a smoothing function for MAC?

[bpbond]

It's not a visual artifact -- those lines are the actual Value_MAC values.

It seems to occur because of soil salinity is highly variable in time and between sensors:

and as a result the MAC (mean annual cycle) used is a little nuts:

Originally posted by @bpbond in #453

Thought for the future: maybe the MAC should be computed using a loess or other smoothing/windowing function. That would help with this and similar problems.