What about the operational attenuation 2K in the simplified RTE?
This all important question is discussed in detail by Maritorena et al, 1994, in their paper "Diffuse reflectance of oceanic shallow waters: influence of water depth and bottom albedo".
I have come to suspect that shallow areas in a deglinted remote sensing image allow the practioner to derive a very good estimation of the value of spectral operational 2K in units of m-1 :
Radiances in units of native image DNs may used: conversion into reflectances (0-1) is not a requirement when operating a "ratio method".
Retrieved depths are computed in units of m, without the need/use of any field data for calibration of the simplified RTE.
Water column corrected spectral bottom signatures are computed in units of native image DNs.