Switch to a tide API?

[dbuscombe-usgs]

Tidal correction in coastseg using the 'coastsat method' is problematic:

1. accessing fes2014 data is a pain - it requires making an account, waiting, large downloads. It is not clear if all the needed data is even included in the download. I spent most of 1 day trying to understand it, and I failed.
2. Even with the download, each user has to make an ini file (which is not well documented), and run the pyfes package, which has no docs, no maintainers, and needs to be installed in its own conda env
3. then, cvs files need to be generated by each user and applied to the raw timeseries using coastsat codes

Ideally, the above would be replaced with an automatic process, querying tide height using an API based on location and time.

If we were to limit CoastSeg to just the USA, we could potentially use NOAA's CO-OPS API for data retrieval https://api.tidesandcurrents.noaa.gov/api/prod/

• there are well-maintained python codes hosted on github
• data is open, without need for user accounts, large downloads etc
• tidal height estimation is automatic and not prone to user error
• stations are less spatially well resolved than fes14, and obviously not available globally
• other non-tidal quantities would be available, such as wind and waves, which could provide context to observations and might even be used explicitly in ML routines for segmentation or future advances such as runup correction

We'd be keen to get any thoughts on this @kvos - I'm sure you've thought a lot about this issue. I accept that we might take an accuracy hit this way, but there are tradeoffs to consider. A lot of the USA has relatively small tides. Our design brief is to make CoastSeg easy to use by USGS researchers without expertise in python programming

[dbuscombe-usgs]

Of course, now I find the FES14 API: does it work? https://github.com/CNES/aviso-fes/blob/main/PYTHON_API.md

It still has the limitation of the data access

[dbuscombe-usgs]

This works, is very easy, and is very fast https://pytmd.readthedocs.io/en/latest/api_reference/aviso_fes_tides.html

[dbuscombe-usgs]

We are adopting pyTMD and this issue is being tracked here: #124