feature request: calc_mld

[eleanorfrajka]

This is similar to compute_cline, but would work on ungridded profile data.

Common definitions of mixed layer depth would be finding when the 20m bin averaged density exceeds the top 20m density by some threshold (0.01 kg/m3 in the Labrador Sea, often larger elsewhere).

This would be a useful tool for some onward calculations like figure like Fig 3 in Giddy et al 2001. https://doi.org/10.1029/2020JC016814 (she uses density difference of 0.03 kg/m3 for a ref depth of 10m)

[MOchiara]

GliderTools has one and it works on ungridded data. Do we want to add it so some specific function?
https://glidertools.readthedocs.io/en/latest/physics.html#mixed-layer-depth

[eleanorfrajka]

Ok, found it. it does this: https://github.com/GliderToolsCommunity/GliderTools/blob/master/glidertools/physics.py

It looks like it does what Giddy says, but I don't think it handles well when data are slightly noisy. I'll try to find a reference. A method that handles noisy data better:

• bin the data in 20m averages and look for the density difference when the first 20m bin is greater than the bin-averaged density (0-20m) by at least the threshold. The 20m binning helps deal with small duration spikes.

Or there is Pickart et al. 2002. They use a standard deviation criterion.

Or there is manual inspection (Vage et al.)

Ah, actually those are both the same method. There is visual inspection to "pick" the depth, then the 2-standard deviation around the values above the visually chosen MLD is used to determine when the profile differs outside this 2-std window.

Ok - conclusion, implement/copy across the method from glidertools (probably defaulting to use sigma0 for density and the 0.03 threshold).

• But I'd like to find profiles that "break" this (i.e., where you'd pick a MLD by eye different from what this method chooses) and then implement something else (likely based on an accumulated buoyancy anomaly threshold) to see if it can produce a more stable/robust result.

[eleanorfrajka]

• Or this (but for a later issue): https://journals.ametsoc.org/view/journals/atot/26/9/2009jtecho543_1.xml?tab_body=supplementary-materials