Detection Limits: Limits of detection are not reported because they are not applicable to the Th-234 beta counting method. A 'non-detect' for Th-234 or a case where there is no Th-234 present (initially or after 6 months of decay) will still result in a measurable amount of background radioactivity due to the beta decay of long-lived natural radionuclides that are also carried by the Mn precipitate used in the shipboard sample processing. These background values are utilized and therefore, they are not reported as non-detections of Th-234.

Uncertainty: At the start and conclusion of each cruise, high activity U-238 standards and background counts (empty detectors) were measured to confirm correct operation of the RISØ detectors and to determine detector to detector variability. The reported uncertainty on each total Th-234 measurement represents the propagated counting uncertainty, volume (pipettes) and weighing (scales) uncertainties, and detector to detector normalization. Counting uncertainty is generally the largest source of uncertainty so whenever possible samples were counted until errors were below 5%.

Inter-calibration Efforts: Results from the GEOTRACES Th-234 inter-calibration efforts are published in Maiti et al. (2012). Fifteen labs participated in two cruises that centered on particulate, total and dissolved Th-234 inter-calibration. Total Th-234 was assessed on the 2009 GEOTRACES inter-calibration cruise. A short excerpt from Maiti et al. (2012) is included below that summarized the results and recommendations for total Th-234 shipboard analyses:

'Total 234Th inter-calibration results for deep water samples showed good agreement amongst laboratories. The mean 234Th activity of 2.433 ± 0.035 dpm L-1 for all laboratories was also found to be in very good agreement with the salinity derived expected 234Th activity of 2.442 ± 0.0003 dpm L-1, assuming secular equilibrium. However in comparison, the 234Th activities from SBB [Santa Barbara Basin] surface water had much more variability because of lower initial activity and large ingrowth corrections from 238U. It is thus recommended to keep the time between collection and filtering to a minimum in order to keep in-growth corrections to a minimum. This also reduces the error associated with the uncertainties in both salinity measurements and the 238U-salinity relationship as the error is propagated when ingrowth corrections are made.'

Because of the short half-life of Th-234 (~24 days), the nature of radionuclide analyses, and the recommendations stated above, inter-calibration between multiple labs was not possible during the 59-day GP16 cruise. Therefore, internal quality control was a priority and addressed in the following ways. First, five replicate deep water samples were analyzed to determine the degree of reproducibility. These samples were collected from the same niskin bottle and processed simultaneously. The results of this analysis are incorporated into the sample measurement uncertainty. Second, the shallow, intermediate, and deep pump casts were made to overlap to provide some means of assessing cast to cast differences and reproducibility. While conditions can change during the duration of a pump cast (~4 hours) as well as from one cast to the next, noticeable differences (larger than expected for natural variability) in overlapping casts were only observed at one station during GP16. This suggested a high degree of reproducibility was achieved for mid-water column total Th-234. Third, all filters measured for total Th-234 were counted twice aboard ship (second count within a few days of the first) and immediately compared to salinity-derived U-238 values. Any abnormal deviations were flagged (>10% between counting results) and the filter was run a third time if necessary. These two counts were averaged for each sample to get a final 'total Th-234' for that depth and location. Lastly, the same internal uranium standards were used by the Buesseler lab during the NAZT and EPZT GEOTRACES cruises and will continue to be used on future cruises. These standards provide a consistent means for calibration and comparison.

Problem reporting:
Data were flagged using the SeaDataNet quality flag scheme. For more information on SeaDataNet flags, see: https://www.geotraces.org/geotraces-quality-flag-policy/ and https://www.seadatanet.org/Standards/Data-Quality-Control

SeaDataNet quality flag definitions:
0 = No quality control;
1 = Good value;
2 = Probably good value;
3 = Probably bad value;
4 = Bad value;
5 = Changed value;
6 = Value below detection;
7 = Value in excess;
8 = Interpolated value;
9 = Missing value;
A = Value phenomenon uncertain.

Two samples were marked as 'bad values' due to issues encountered during sample filtration. Because of the obvious poor quality of these data, the total Thorium-234 values were not included. A few samples were denoted as 'probably good' because of a pipetting issue affecting individual sample recovery calculations. Average recoveries for the surrounding water column depths were used for these samples. Accordingly, uncertainty values were increased for these samples.

BCO-DMO Processing:
- renamed fields:
- added date/time fields in ISO8601 format;
- replaced blanks (missing data) with 'nd' ('no data');
- 11 June 2020: replaced with GEOTRACES DOoR-formatted/IDP version.