Water sampling details

Water samples were gravity filtered from Niskin bottles through 0.2 μm Whatman Polycap 36 AS nylon membrane cartridge filters, directly into combusted borosilicate glass bottles. Filtered sample waters were stored in the dark at 4°C until measurements. Sample waters were left in the dark at room temperature (~20°C) so they were back to room temperature before absorbance measurements.

These seawater samples were collected for photochemical degradation experiments, and water samples at the various BATS stations had similar optical properties, for example, similar absorbance at 325 nm. Therefore, which stations we collected samples from were largely determined by the cruise schedule, scientific tasks performed by the core BATS crew, and weather, i.e., where CTDs can be cast that allowed us ample time to directly filter water from the Niskin bottles.

We also wanted to compare the photoproduction of dissolved inorganic carbon (DIC) between seawater at the surface and at depth, as the deep water may have different photodegradation history and different dissolved organic carbon composition, and thus have different DIC photoproduction rates and efficiencies. Therefore, we collected water samples from the surface (1 m) and at depth (3000 m).

Measurement details

Absorbance (A(λ)) of the water samples were measured at 250–800 nm at 1.0 nm intervals, in triplicate. May 13th, 2017 samples were measured in 10-cm-pathlength quartz spectrophotometric cuvettes with an Agilent 8453 UV-visible spectrophotometer with ChemStation software; July 17th, 2018 samples were measured in a World Precision Instruments UltraPath capillary cell set at 50 cm pathlength using a Maya Pro 2000 spectrophotometer with SpectraSuite software, and the rest of the samples were measured in the same UltraPath capillary cell set at 200 cm pathlength using the same Maya spectrophotometer.

Milli-Q water was similarly measured to use as blanks. The absorbance spectra of the sample water were first corrected by subtracting the absorbance spectra of blanks (Milli-Q water) from the absorbance spectra of the sample water; the absorbance spectra of the sample water were further corrected for potential offsets and instrument drift by subtracting the average absorbance at 690–710 nm. The corrected absorbance spectra (A(λ)corr) of the sample water generated from these corrections were converted to Napierian absorption coefficients (ag(λ); m−1; represented in this dataset as CDOM_absorption_a_g), using the following equation: ag(λ)= A(λ)corr ln10/L, where L (m) is the pathlength.

The May 13th, 2017 samples were measured in 10-cm-pathlength cuvettes with the Agilent spectrophotometer because these samples were T0 samples of a photoirradiation experiments, the water was photoirradiated in these 10-cm-pathlength cuvettes, and the absorbance were also measured in these cuvettes, which only works on the Agilent spectrophotometer, which has a 10-cm-pathlength cuvette holder.

927300_v1_north_atlantic_cdom_absorption_coefficients.csv