Cruise Report, RV ATAIR, Cruise 0671021-1, Bremerhaven - Thyboron —- Aberdeen, 18 Jul - 10 Aug 2023 | 19 
5.7 Post-Cruise Processing of CTD Data 
Post-cruise comparisons were made between CTD sensor packs TC1 and TC2 and between 
CTD temperatures and a supplementary temperature sensor of type SBE35 installed at the 
height of the Niskin bottles. Furthermore, a comparison between CTD salinities and AU- 
TOSAL-derived salinities was performed (compare Tables 5.3 and 5.4 and Figures 5.5 and 
5.6). 
The following correction procedure is applied to temperatue and conductivity/salinity measure- 
ments as well as oxygen data: 
Assumed measurement accuracy: T-CTD: + 0.002, C-CTD: + 0.003, S: estimated: + 
7.004 
A temperature correction will be applied in case the temperature difference (CTD vs. 
SBE35) exceeds 2x 0.002 and/or the difference between temperature sensors T1 and 
T2 exceeds 2x 0.002. 
A conductivity correction will be applied in case the salinity difference (AUTOSAL vs. 
OTD) exceeds 0.003 and/or the difference between conductivity sensors C1 and C2 
8xceeds 2x 0.003. 
The random error of oxygen samples obtained from titration is approximately 1%. 
Corrections are made if the oxygen difference (bottle vs. CTD oxygen) exceeds 1%. 
Post-cruise quality control showed that differences between TC sensor pairs were small. All 
somparisons revealed small differences, which were within the error bounds and were slightly 
smaller for temperature, conductivity and salinity from the secondary sensors. In summary, 
any post-cruise correction was not applied, and the temperature, conductivity and salinity data 
were made available with flag 1 (good data). The usage of the data from the secondary sensor 
pair (TC2) is recommended. 
Sampling and titration of oxygen samples was performed on cruise 067/021-1 to the North Sea 
and on the subsequent cruise 067/021-2 to the German Bight. Throughout both cruises, the 
STD sensor setup remained the same, as did the use of the CTD-based SBE43 oxygen sen- 
sor. Oxygen analysis, including a comparison between CTD-derived oxygen and bottle oxy- 
gen, was thus generated for the entire dataset of both cruises. This extended the time period 
af the data and enabled the temporal drift of the sensor to be more accurately recorded and 
sorrected using a multi-parameter-fit (MPF) method. The analysed dataset comprised stations 
#002 — #100 of cruise 067/021-1 and stations #101 — #112, #114 — 138 and #140 and #142 of 
zruise 0671021-2. Obvious outliers were removed, and only data pairs with differences within 
+1 standard deviation were included in the comparison. This reduced the dataset to 89% of 
the matching data pairs. 
Regarding oxygen, corrections are made when the difference between the CTD-derived oxy- 
gen and titrated bottle samples exceeds 1% (titration measurement accuracy). For measured