Estimates what fraction of each cell is the first of two named water masses, by projecting its temperature and salinity onto the mixing line between the two endmembers in T-S space.
Usage
water_mass_fraction(
env_dat,
endmembers,
temperature = "SST",
salinity = "SSS",
residual = FALSE,
name = NULL
)Arguments
- env_dat
an
sfPOINT object with one row per location and time step, as datamatch's access functions return- endmembers
a list of exactly two named
c(temperature, salinity)vectors. The fraction returned is of the first.- temperature
name of the temperature column, in degrees C
- salinity
name of the salinity column, in PSU
- residual
also add the distance from the mixing line, as
<name>_residual- name
name for the fraction column.
NULLuses the first endmember's name, lowercased, with_frac.
Details
The Gulf of Maine is the standard setting for this. Its deep water is a mixture of Warm Slope Water and Labrador Slope Water in proportions that vary episodically, and the proportion sets the nutrient supply: Labrador Slope Water is colder, fresher, and nutrient-poor, Warm Slope Water warmer, saltier, and nutrient-rich (Townsend et al. 2015). Surface water is instead a mixture of Scotian Shelf Water with slope water, which is the same calculation with different endmembers.
How the fraction is computed
A cell's (T, S) is projected onto the line joining the two endmembers, and
the fraction is where it lands: 1 at the first endmember, 0 at the second.
Temperature and salinity are on different scales, so each axis is normalised by the endmember separation in that variable before projecting. Without that the axis with the larger numerical spread would dominate the answer for reasons of units rather than of oceanography.
Values are clamped to [0, 1]. A cell beyond an endmember is either a third
water mass or an endmember choice that does not bracket the data, and reporting
a fraction of 1.4 would dress up that problem as a measurement.
Check the residual
Projection always returns a fraction, even for water that is not a mixture of
these two masses at all. residual = TRUE adds the distance from the mixing
line in the same normalised units, which is the number that says whether the
fraction means anything: near zero, the cell sits on the line and the mixture
is a good description; large, and it does not.
This is why the endmembers are an argument with no default. They vary by region, season, and year, and a wrong pair produces confident nonsense.
References
Townsend DW, Pettigrew NR, Thomas MA, Neary MG, McGillicuddy DJ, O'Donnell J (2015). Water masses and nutrient sources to the Gulf of Maine. Journal of Marine Research 73, 93-122.
Examples
if (FALSE) { # \dontrun{
# Deep water: which slope water is filling the basins?
env <- water_mass_fraction(env, endmembers = list(
LSW = c(temperature = 6, salinity = 34.4),
WSW = c(temperature = 12, salinity = 35.4)
), residual = TRUE)
# Surface water: how much of this is Scotian Shelf Water?
env <- water_mass_fraction(env, endmembers = list(
SSW = c(temperature = 2, salinity = 32.0),
slope = c(temperature = 10, salinity = 35.0)
))
} # }