An approach for merging colored dissolved organic matter absorption measurements from complementary benchtop spectrometers

Abstract

Absorption of light by colored dissolved organic matter (CDOM) is often measured for in situ samples using benchtop spectrophotometers, whose valid wavelength ranges are set by the CDOM concentration of the sample and by instrument constraints such as linear response limit and instrument uncertainty. As tools for algorithm development, calibration, and validation, these measurements are most useful when they report CDOM absorption coefficients across the greatest wavelength range, which often means extending further into ultraviolet wavelengths. To maximize the reported spectral range while delivering the highest-quality measurements across ultraviolet–visible wavelengths we developed and evaluated two approaches, the value threshold and percent difference approaches, to combine CDOM absorption spectra of the same sample measured on two different spectrophotometers that have different but overlapping spectral ranges. Colored dissolved organic matter absorption measurements from 1001 samples were examined, which were sourced from nine different field campaigns with broad optical water type properties. The value threshold approach successfully produces a merged product with larger spectral range for coastal ocean and open ocean samples, transitioning between the input measurements over a span of 40–80 nm where both measurements are valid based on the instrument characteristics of linear response limit and instrument uncertainty. The percent difference approach relies on the percent difference between spectra from the two instruments and works for coastal ocean waters but fails for most open ocean samples where there is more uncertainty proportional to CDOM signal. The value threshold approach can be modified to reduce noise propagation in the merged product and can also be adapted for use with other benchtop spectrometer models, making it highly generalizable.