lukegre / pyseaflux Goto Github PK

We see SeaFlux as both a data product and an accompanying code base.

We thus need to be able to update the data product at least annually.

The following configuration will be followed:

• a 1°x1° by monthly resolution
• from 1982 to present
• variables:
  • SST = NOAA AVHRR only data
  • SSS = EN4 monthly salinity
  • Press = ERA5 msl
  • ATM = NOAA Marine Boundary Layer
  • Gas transfer velocity
    • scaled to 16.5 cm/hr
    • scaled for period 1988 to 2017 (overlaps all wind products)
    • Wind products
      • JRA55
      • ERA5
      • CCMP2
      • NCEP R1
      • NCEP R2

The data pipeline will be created with the fetch-data package

Currently, each of the kw parameterisations has basic documentation (article + wind product).

More detail in the documentation would help the users understand when to use a particular parameterisation. Further, the method used to arrive at the estimate could be helpful.

Lastly, a dedicated page briefly outlining the theory of gas transfer velocity and when to use which product would be immensely useful. Perhaps a table showing the wind product used for a parameterisation.

The sphinx-autodoc generator is quite sensitive to doc string formatting.
This needs to be cleaned up accordingly.

This uses the numpydoc setting
https://developer.lsst.io/python/numpydoc.html#py-docstring-short-summary

Not sure if you're already aware of this, but I just came across another tool that could be used to validate/compare these conversions - https://github.com/dpierrot/fCO2_Calc

We currently use the square of the wind product parameterisation, as in Wanninkhof (1992). Several members of the community have commented that additional parameterisations should be scaled. This could include Nightingale et al. (2000), Ho et al. (2006), McGillis et al. (1999).

One of the problems with optimising these parameters is how to perform the optimisation when more than one parameter (alpha) needs to be scaled.

Documentation is currently sparse. So a bit of an overhaul required.

I've migrated the documentation to sphinx which can also take Jupyter-notebooks.

The following sections that should be covered.

• Flux calculations using raw inputs
• Scaling the Kw for the wind products
• A high level interface for 1°x1° monthly data calculations
• What's new
• How to contribute
• Authors