Explicit biogeochemical units

[nanophyto]

@jagoosw I would love to pick your brain on how to implement more explicit biogeochemical units in OceanBioME.jl

Oceananigans.jl has a nice Units module with syntax that could be extended for common biogeochemical units:

volume, concentration, fluxes, others? e.g. 1 / (W / m² / s), gC / dm² / s / einstein, 1 / dm² , etc.

Options could be to either suggest extending Oceananigans.Units to include relevant units like above, or to add something like OceanBioME.BGCUnits

An alternative is to rely on something like Unitful.jl and extensions like UnitfulMoles . The upside would be features like converting e.g. mol C to grams, but the downside might be unneeded complexity/ additional dependencies - so this seems like the less desirable option?

Happy to put a PR together as appropriate/ if of interest.

[jagoosw]

Hi Joost!

This is a really good idea, I think the easiest thing todo would be add a module that defines units like Oceananigans does, but I have thought about using Unitful too.

Maybe we should try and come up with a list of the units we'd like to put in then one of us can open a PR?

[nanophyto]

Great, I've started putting together a list already so I will share that shortly

[nanophyto]

Length (SI reference: meter)

nanometer(s) = 1e-9 (spectra, pigments, optics)
micrometer(s) = 1e-6 (ESD, cell size, particle size)
millimeter(s) = 1e-3 (larger particles, aggregates)
meters, kilometers are already in Oceananigans.Units

Areas (SI reference: square meter)

square_micrometer(s) = 1e-12 (cell surface)
square_meter(s) = 1
square_kilometer(s) = 1e6 

Volumes (SI reference: cubed meter)

cubic_micrometer(s) = 1e-18 (cell volumes)
cubic_meter(s) = 1
cubic_kilometer(s) = 1e9 
liter(s) = 1e-3 
milliliter(s) = 1e-6 
microliter(s) = 1e-9 

Molar concentrations (SI reference: mol)

picomole(s) = 1e-12 
nanomole(s) = 1e-9
micromole(s)  = 1e-6 
millimole(s)  = 1e-3 
mole(s) = 1

Parts-per concentrations

ppt = 1e-3
ppm = 1e-6

Mass (SI reference: kilogram)

microgram(s) = 1e-9
milligram(s) = 1e-6 (often used for Chl)
gram(s) = 1e-3
kilogram(s) = 1
teragram(s) = 1e9 (global budgets)
petagram(s) = 1e12 (global budgets)

Light (reference: einstein)

einstein = 1
microeinstein = 1e-6

Energy transfer (SI reference: watts)

watt(s) = 1
#unsure if others are useful?

Presure (reference: pascal)

pascal(s) = 1
bar(s) = 1e5
atm = 101325

[jagoosw]

The only thing is we might need to be careful defining what an einstein is because I gather that it doesn't actually have a numerical definition and technically depends on the spectrum of light its referring to. In my projects I've used:

const μE_s_to_W = 1 / 4.56

from the daylight definition here: https://en.wikipedia.org/wiki/Photosynthetically_active_radiation

[nanophyto]

Yes, this is a good point. I would prefer not including any conversion Einstein/watt in the Units module, and instead keeping it local to a given model/project

Conversion within e.g. Einsteins would be valuable however:

microeinstein = 1e-6 einstein

[nanophyto]

const μE_s_to_W = 1 / 4.56

If including it in the module is useful, maybe adding something like *_default or *_daylight to highlight that this is not a universal constant?

[nanophyto]

Do you think including aliases for things like concentrations would be useful? e.g.:

const mmol_per_L = millimole / liter

Adding semantic aliases for bookkeeping could also be an option

mmolC = mmol
mmolN = mmol
mmolP = mmol
etc.