The target compound for the comparison of the two methods was the dimethyl sulfide (DMS) sampled out of nine independent mesocosm enclosures. Both techniques used sub-samples taken from the same original aspirators. However, each method was performed by a different person, using a different

sample preparation process, type of calibration, calibration standard and analytical instrumentation. The NTD GC–MS sampling and analysis processes are described in detail in the Experimental section while detailed information about the P&T GC–FPD method can be found in earlier studies (Kiene, 1993, Zindler et al., 2012a and Zindler et al., 2012b). In short, there are three main differences between the NTD GC–MS (method GSK2118436 nmr A) and P&T GC–FPD (method

B) techniques: 1) method B used liquid nitrogen (LN2) for pre-concentration while selleck chemicals in method A sample tracers were trapped directly using three-bed NTDs, 2) method B used a potassium carbonate (K2CO3) column to trap the moisture while for method A the condensed water was used as an extracting medium in the desorption process and 3) immersion in hot water was used in method B for the injection of DMS into the GC where in method A desorption of the NTDs occurred directly into the injection port of the GC. The two techniques were calibrated independently. The NTD GC–MS method used a multi-component gas standard (5 % stated accuracy) while the P&T GC–FPD method used a liquid DMS standard for calibration (Kiene, 1993 and Zindler et al., 2012b). The liquid standard from the GEOMAR team was analyzed also using the NTD method. The difference between the two standards was found to be 7 % which was not considered significant as it is within the range of the NTD method

precision (RSD % 7–12.4) at the examined concentration levels (see Table 2). The NTD GC–MS method gave LODs as low as 0.04 nM and the P&T GC–FPD method 0.3 nM. Linearities (r2) for both techniques were > 0.99 for a concentration range of 0.5 to 10 nM. In Fig. 7, we present a visual comparison of the DMS measurements in each pCO2 group for the two analytical methods and a whole data method correlation. In Fig 7A, B, C, measurements provided by the NTD method are marked Abiraterone manufacturer with filled cycles while the ones provided by the P&T method with star symbols. On the whole, both methods are in good agreement, with similar DMS concentration ranges (0.3 to 6 nM by the NTD method and 0.34 to 6.18 nM by the P&T method), temporal variations and CO2 effect. Best agreement between the two methods was found for the higher DMS production group (low pCO2 treatment) with correlation coefficient r2 = 0.81. A linear regression ( Fig. 7D) for the whole data set gave a total r2 = 0.805 correlation between the two methods. The derived slope shows a 13 % overestimation of the NTD over the P&T method. This is mainly caused by discrepancies in the first period of the experimental study when the NTD method measured consistently slightly higher (i.e. days 0 to 10).