The abnormally low dissolved oxygen concentrations of samples collected during the HOT cruises conducted between November 23, 2002 and April 23, 2003 represent a potentially significant problem since these bottle values are the basis of the CTD oxygen sensor calibrations. Possible causes of the problem have been investigated and are discussed in the following section (Sect. 4.2.4.1).
There are three areas of the titration process used to measure the dissolved oxygen concentrations of HOT bottle samples that, if faulty, could result in significant errors in the final dissolved oxygen titrations: reagent preparation and possible subsequent contamination, volume measurements, and electrode stability.
· Proper preparation of all reagents and standards involved is imperative. Incorrectly calculated concentrations of any one of the solutions involved would result in systematic error that would carry through to the final calculations. Optimum reagent concentrations are given by Carpenter (1965) and Culberson (1991), and the concentrations utilized for the HOT program are given in the Field and Laboratory Protocols, Chapter 5 (Karl et al. 1990). Experiments conducted by Dale Hebel and Lance Fujieki (2003, final report pending) ruled out this particular source of error as being the cause of the observed anomalous dissolved oxygen concentrations.
· The reliability of the dissolved oxygen titrations is also dependent upon the proper calibration of all volumetric devices involved in the process, e.g. pipettes, flasks and burettes. One example is that if the burette of the computer controlled Dosimat were to deliver a volume of titrant greater than expected, whether as the result of a poorly calibrated burette or due to diffusion through the burette tip in the time between aliquot additions, the endpoint of the titration would be reached too quickly resulting in an erroneously low dissolved oxygen concentration. The experiments conducted by Hebel and Fujieki (2003, report pending) also ruled out improperly delivered reagent volumes as the cause of the problem.
· The redox probe and the associated electronic pH meter to which it is attached represent the third area of the titration procedure that could introduce a significant amount of error into the final dissolved oxygen calculations. Redox probes have a finite life span of reliability that will vary depending upon the individual probe and on the maintenance and care of the probe. The investigation by Valenciano (2003, Appendix A) revealed that historical care of the main redox probe used for the dissolved oxygen determinations deviated from the manufacturer’s recommendations, and that the questionable probe was disposed of thereby preventing any drift analysis to be conducted on the probe, which could potentially have provided data to correct the suspicious dissolved oxygen concentrations.
Based upon the results of the investigations conducted by Hebel and Fujieki and by Valenciano (Appendix A), the primary source of error would seem to be derived from the redox probe. Beginning with HOT-148, the redox probe was replaced along with a new dosimat , pH meter, and controlling computer, and the dissolved oxygen values measured from that time forward agree well with historical values.
Back to Table
of Contents 
