Edward Peltzer

Total Organic Carbon Methodology
 Peltzer, Edward T. (1993).  Shipboard determination of total organic carbon by
   a high temperature combustion/direct injection technique.  U.S. Joint Global 
   Ocean Flux Study - Equatorial Pacific Protocols, 1993, section 21A.

  Alternatively, see:  Chapter 16 of the JGOFS Data Protocols.  Determination
     of Dissolved Organic Carbon by a High Temperature Combustion/Direct Injection
     Technique.  JGOFS Report Nr. 19

  The EqPac methods were followed explicity on the Ross Sea Process cruise 3
   with three exceptions:

   1.  Low carbon water (LCW) prepared onboard ship using a commercial NanoPure (R)
       UV/TOC system was used to measure the instrument blank as opposed to the
       carbon free distilled water (CFDW) that was used in EqPac.

   2.  The LCW prepared onboard was checked against LCW prepared at WHOI using a
       Milli-Q UV/TOC (R) system and shipped to New Zealand in glass bottles.  This
       water was assigned a background TOC concentration of zero.

   3.  Arabian sea deep water was used to prepare the daily calibration standards.
       This water has a background TOC concentration of 42.3 uMC.

 PLEASE NOTE:
   1.  These samples were NOT filtered; ergo they are TOTAL ORGANIC CARBON (TOC) analyses,
       not dissolved organic carbon analyses (DOC)

   2.  During this cruise some very low level contamination was encountered.  This
       contamination was introduced during the sampling process.  It could be minimized by
       drawing the TOC samples immediately after the oxygen samples.  Compare the two casts
       from Minke (Station 5, cast 1 and cast 4).  Cast 4 is noticeably "cleaner", with fewer
       wiggles and a lower deep water mean TOC concentration than cast 1 which was collected
       after several other samples were drawn.

   3.  In the data table, obviously "bad samples" have been removed.  However, given the
       low level contamination we were experiencing, it is possible that some contaminated
       samples are still present.  Since it is impossible to know for certain which samples
       are truly correct, all data have been included.  Those which our best guess suggests
       to be clean are flagged '1'; those we are suspect are flagged '0'.  To obtain our
       best guess at a clean data set, multiply the data by the data flag and omit all the
       rows where the product is zero.