Switch data source for the buoys that are no longer being updated

[andrew-edwards]

Compare with Andrea's plots - looks like some of the DFO-based ones are not getting updated now, so switch data-raw code to use the ECCC ones for those.

[andrew-edwards]

Leaving open as still want to compare the SST for when the DFO and ECCC data overlap. For example, for buoy C46146 it looks like (when comparing plots on the vignette before/after redoing data) some of the previous years may have changed? Can test by making plot, reverting above commit, making plot again.

[andrew-edwards]

• [ ] In data-raw code, change opp to eccc since now we're not just using the two opp buoys from eccc data source.
• [ ] Incorporate the flags for ECCC data. @schckngs to look into a little first, Andy can do some checking also. Hadn't used them before since just the opp data. See https://github.com/IOS-OSD-DPG/Pacific_SST_Monitoring/issues/2
• [ ] Check if that automatically eliminates the dodgy looking data for which we concluded the buoy was probably switched on then lay on the deck of a boat for a few days.
• [ ] Also allow two weeks for the buoys to be in the water (though maybe not needed once we include the flags); so maybe change the switch_dfo_to_opp value.
• [ ] Andrea had checked overlapping times and found DFO and ECCC data to agree. Andy can do a quick check once we've used the flags.
• [ ] Add to help that buoy SST is actually 1 m depth. And the satellite data are calibrated to also be 1 m, and OI SST actually incorporates buoy SST (i.e. not totally independent data sources - important, and others hadn't realised that for a while). Also ROMS SST is the top box of the grid, < 1m, maybe down to 0.2 m I think.