Marking "noisy" acquisitions/output products

[scottstanie]

Q. How could we mark "noisy" acquisitions that aren't worth pulling?

We will be pulling $k$ old SLCs along with each new SLC to estimate/produce a new product. We almost certainly to not want to just pull the $k$ most recent SLCs (c.f. the Norwegian InSAR service, who basically skips the entire winter).

This is a two part issue:

1. What quality metrics can we produce that best tell is what are the old SLCs that are worth re-pulling?
2. Operationally, how can we mark our output products (or label Sentinel products) so we know that certain historical ones aren't worth using to estimate the latest phase?