December 28, 1976


QUALITY OF WATER BRANCH TECHNICAL MEMORANDUM NO. 77.02

Subject: WATER QUALITY--Analytical methods for organic carbon

As many of you already know, close examination during the past 
several months of the results of determinations of total organic 
carbon (TOC) by WRD laboratories has cast doubt on the validity of 
some of those analytical values. A recent study of the situation 
for the Central Laboratories System by Jerry Leenheer has 
indicated that the primary problem lies in nonrepresentative 
subsampling of the original sample in the laboratory, and not in 
the analytical equipment itself or in the analytical procedures 
used. It is believed now that loss of organic carbon to the 
container walls through adsorption, precipitation, flocculation, 
or biological growth, and possibly some conversion of organic 
carbon to gaseous products through bacterial activity, can occur 
and has resulted in errors in an unknown fraction of this TOC data 
Inadequate cleaning of new sample bottles constituted an 
additional but irregular source of error until the problem was 
discovered

The errors described above are difficult if not impossible to 
detect for samples in which only TOC was determined. For such 
samples we have no basis for suspecting any particular 
determination to be in error, and consequently there is no basis 
for recommending that any particular value be deleted from the 
record.  The same can be said for values of TOC that are greater 
than either dissolved organic carbon (DOC) or suspended organic 
carbon (SOC) when only one or the other determination was made in 
addition to TOC. However, in the case of TOC values that are 
significantly less than values of DOC or SOC for the same sample,
there is good reason to believe that the TOC value is in error 
rather than the values of DOC and SOC

In order to insure that the organic carbon data in the water 
quality file are as nearly correct as possible, the following 
procedure has been agreed upon by the Quality of Water Branch and 
the Assistant Chief Hydrologist for Scientific Publications and 
Data Management.

A retrieval will be made in the next few weeks of all organic 
carbon data collected since October 1, 1973.  A printout of these 
data will be sent to the District Offices for review. In cases 
where DOC or SOC values exceed TOC by 20% or more, the district 
should delete the TOC values from the file through the district 
terminal. For samples in which only TOC was determined, there is 
no adequate basis for deleting any data except for cases in which 
the district has reason to believe that proper sample collection 
and preservation procedures 

Since all data believed to be suspect are to be deleted from the 
water quality file, we do not believe it is necessary to insert a 
statement in the remarks section of the next annual basic data 
report that qualifies TOC data published in previous reports.

Revised methods for determination of organic carbon will be 
released as provisional methods by the Quality of Water Branch in 
the near future. These hopefully will alleviate many of the 
problems referred to above. At the same time, separation of the 
dissolved and suspended phases by filtering in the field will be 
encouraged by the pricing policy of the Central Laboratories 
System. Like most other water-quality characteristics that can 
change while the sample is on its way from the field to the 
laboratory, it is safest to separate the component parts in the 
field and derive the total from the sum of the parts--in other 
words, determine TOC from DOC plus SOC. In addition, the process 
of filtering the sample through a silver filter helps to prevent 
changes resulting from bacterial action.

It is clear from the above that determination of organic carbon 
still must be regarded as a somewhat primitive measurement. This 
certainly is not without precedent in the history of water-quality 
measurement, but we will continue to strive to improve our 
methodology with time.



                                     R. J. Pickering

WRD Distribution: A,B,S,FO,PO