Analyses of dissolved organic carbon in seawater: the JGOFS EqPac methods comparison

Abstract

Abstract Results of a dissolved organic carbon (DOC) methods comparison are presented here in which five high temperature combustion (HTC) instruments and a wet chemical oxidation (WCO) method were used on a series of oceanic samples. The samples were collected during US JGOFS Equatorial Pacific Ocean cruises (EqPac) and most of the authors were involved with DOC analyses for the EqPac Program. Samples were collected with a “clean” protocol and were immediately quick frozen in replicate sample bottles. They were distributed by the first author to the other authors for “blind” analyses later on land on the stored samples. Comparable results (±7.5%) were found by three HTC instruments and the WCO method. There were difficulties with the other two HTC methods for which explanations and improvements are offered. The single most critical element for comparable DOC values appears to be assessment and subtraction of the total instrument blank (or reagent and handling blank for WCO methods). A “zero” carbon (very low C) water sample assisted in having all analysts achieve a uniform assessment of individual instrument or methods blanks. “Conditioning” of the catalyst bed in the combustion tube is critical to achieve consistent low instrument blanks. Failure to thoroughly condition the catalyst bed may be a significant error that can give erroneously high DOC values for oceanic samples. Reference standards available to all analysts also allowed comparison of instrument and methods performance. Contamination problems were demonstrated and it was shown that careful preparation and handling can reduce the potential for errors from contaminated samples. Results indicate that Equatorial Pacific oceanic DOC values in near surface waters are on the order of 60–70 μM C and deep water values on the order of 35–40 μM C. Since the “zero” carbon water contained a small, but measurable, amount of DOC, the sample values reported here may be slightly low. Because the lowest instrument blanks were equivalent to about 10 μM C, it is suggested that even if there were no instrument blank at all and all this “blank” were in the “zero” carbon water, the oceanic sample concentrations could not be underestimated by more than 10 μM C.