Sample Processing Protocols for Chemical Analysis of Surface-Water Samples: Choice of the Churn versus Cone Splitter 

To: "E  - All WRD Employees"<distribution@usgs.gov>
from: "David A Rickert, Chief, OWQ, Reston, VA" <drickert>,
        "Timothy L Miller, Chief, NAWQA, Reston, VA" <tlmiller>
Subject: Sample Processing Protocols for Chemical Analysis of 
 Surface-Water Samples: Choice of the Churn versus Cone Splitter 
Date: Fri, 28 Nov 1997 13:26:15 -0500


In Reply Refer To:
Mail Stop 412

                          MEMORANDUM

                                           November 26, 1997

To:       All WRD Employees

From:     Timothy L. Miller, Chief, NAWQA
          David A. Rickert, Office of Water Quality

Subject:  Sample Processing Protocols for Chemical Analysis 
          of Surface-Water Samples: Choice of the Churn versus
          Cone Splitter 

                     PURPOSE OF MEMORANDUM

An important goal of the Water Resources Division is that all data 
collected by national water quality programs are fully comparable. 
Therefore, any differences that exist in the collection and processing of 
samples should reflect differences in local site conditions rather than 
program design. At present, the significant difference between the 
surface-water protocols for the National Water-Quality Assessment (NAWQA) 
Program and the National Stream Quality Accounting Network (NASQAN) is 
choice of the sample splitting device and ensuing ramifications. 
Generally, NAWQA uses the cone splitter; NASQAN the churn. These choices 
were made because the protocols were developed at different times, with 
emphasis on different constituents. This memo: (1) reviews the 
differences, (2) documents that no fundamental impediment now exists for 
using either splitting device at most sites, and (3) provides guidance on 
how to optimize the choice of splitting device to suit hydrologic 
conditions.

            NAWQA AND NASQAN SURFACE-WATER PROTOCOLS

The NAWQA protocols were developed before the Part-Per-Billion (PPB) 
Protocol was established. In addition, for surface waters, NAWQA focused 
trace elements work on bed sediments and tissues, rather than on the 
water column. The NAWQA surface water protocols were developed to enable 
whole-water and filtered analysis on the same aliquot of water. Only one 
vertical transit is taken at each station across the river. Because 
pesticides are analyzed, a Teflon splitting device, the cone splitter, is 
used to prevent sorption by or contamination from the device. Except 
where "significant quantities" of sand are present, a split from the cone 
is also used for suspended-sediment (SS) analysis. Because of concerns 
about contamination by methanol (which is used to clean sampling 
equipment and the cone splitter prior to use), dissolved organic carbon 
(DOC) is measured in a separate centroid vertical sample, instead of in 
the width-integrated sample. Finally to measure suspended organic carbon 
(SOC) and DOC on the same sample, the protocol specifies measurement of 
SOC in the centroid vertical sample.

The NASQAN Program was redesigned after development of the PPB Protocol. 
NASQAN incorporated the PPB protocol, which uses the polyethylene churn 
splitter, partly because of the program's emphasis on large rivers, and 
partly because the protocol cleaning procedures had not been verified for 
the cone splitter. However, the redesigned NASQAN measures both 
pesticides and dissolved trace elements. This presented a design dilemma 
to NASQAN because pesticide samples cannot be exposed to polyethylene or 
other plastics. To solve the dilemma, the NASQAN redesign chose the churn 
splitter but incorporated multiple sets of transits at each vertical in 
the cross section to obtain the samples needed for the analysis of 
pesticides and trace elements. One set of width-integrated transits is 
composited in and split from the churn to provide subsamples for analysis 
of trace elements, nutrients, and major ions. A second set is composited 
in a glass carboy for analyses of pesticides. For a third set, each 
vertical transit is placed in an individual bottle and the bottles are 
sent to a laboratory for compositing and analysis of suspended sediment 
(SS). DOC is measured in a separate centroid vertical sample, similarly 
to NAWQA. However, because NASQAN emphasizes flux in large rivers, a 
width-integrated sample is considered mandatory to obtain a 
representative SOC concentration. Therefore, the SOC sample is split from 
the sample composited in the churn. 

                 RECENT TECHNICAL DOCUMENTATION

OWQ Technical Memorandum 97.03 documented that the cone splitter can be 
cleaned adequately in the office and in the field to prevent 
contamination of trace-element samples. Moreover, OWQ Technical 
Memorandum 97.06 documented that the cone and churn split the SS in 
samples with similar accuracy and bias at SS concentrations up to 1,000 
mg/L and mean particle size <250 micrometers. Thus, the choice of sample 
splitting device should have no effect on the quality of inorganic data, 
SS concentrations, or particle size distributions for a very high 
percentage of the samples that the Division collects. Therefore, there is 
now no reason for either NAWQA or NASQAN to mandate the use of a single 
splitting device. Choice of the splitting device should be based on the 
factors discussed below.

      GUIDANCE FOR SELECTING THE CHURN OR CONE SPLITTER

For each surface-water sampling station, the optimum choice of device 
depends upon the following factors:

1. SEDIMENT CONCENTRATION. At concentrations below 1000 mg/L, either the 
cone or the churn may be used. At concentrations between 1000 and 10,000 
mg/L, the cone must be used. Above 10,000 mg/L, neither device performs 
satisfactorily.

2. SAMPLE VOLUME. The largest churn splitter has a capacity of 14L, a 
limitation which may impact sample collection on large rivers by reducing 
the number of verticals that can be sampled or imposing an unacceptably 
high transit rate for the samples to limit the volume of water collected. 
The cone splitter has no such volume limitation. On the other hand, 
processing large volumes of water through the cone may be logistically 
difficult (for example, involving storing water temporarily on board a 
boat in multiple containers prior to processing on shore), and the 
additional handling may increase opportunities of contamination. Large 
sample volumes may also require multiple passes through the cone to 
reduce subsample volume to laboratory requirements. Because port biases 
are inherent in the manufacturing of the Teflon cone, port selection 
should be randomized to avoid compounding the port bias when multiple 
passes are required.

3. SAMPLE TYPE-The cone splitter can be used for all sample types except 
DOC (because of the methanol cleaning). However, use of the cone splitter 
in situations where leveling is difficult (such as a boat or cableway) 
requires additional Teflon containers to store water collected at each 
transit until processing can be accomplished. The churn may be less 
susceptible to airborne or other contamination as less sample handling 
and simpler cleaning is involved. However, the churn splitter is 
inappropriate for organic samples because it is not available in Teflon 
(The cost of developing a Teflon churn is prohibitive at this time). 
Another option for compositing samples for organic analyses is to collect 
a separate set of verticals and composite them directly into a glass 
carboy. The proper sample aliquots can then be removed directly (without 
splitting) from the carboy because the Division measures all water column 
organic analytes on filtered samples.

4. RIVER SIZE AND SAMPLING MODE. The churn splitter is distinctly more 
convenient for sites where sampling is performed from a boat or cableway. 
However, because the churn is constructed from plastic, multiple transits 
are required when organics constituents are measured. The longer time 
required to make multiple transects may be less important for 
boat-operated sites because the rivers tend to be larger and, presumably, 
exhibit slower changes in concentration than smaller, flashier rivers. 
However, during rapidly changing conditions, the shorter sampling time 
afforded by a single transit may be preferable on any size river despite 
the more complex processing.

                          CONCLUSIONS

The cone splitter provides superior performance over a broader range of 
SS concentrations and has the advantage of being constructed from Teflon. 
In contrast, the churn offers greater convenience as it is both a 
splitting and compositing device, but it is only available in plastic. 
There is no "right" answer in all cases for which device to use. The 
choice depends upon the described factors, and frequently compromises 
must be made between less-than perfect alternatives. This places a 
premium on obtaining an adequate amount of the appropriate types of QC 
data to substantiate the quality of the environmental data. For further 
details, please consult the National Field Manual for the Collection of 
Water-Quality Data (TWRI, Book 9), Chapters A2 and A5.

Based on the information in this memo, we ask that each of you please: 
(1) review your procedures for collecting and processing samples at each 
station, and (2) choose the procedure that best fits the conditions at 
each station. The optimum procedure may change at an individual station 
depending on the season and hydrologic conditions. If you have any 
questions concerning NASQAN stations, please contact Valerie Kelly 
(vjkelly, 503-251-3244). For NAWQA, contact Larry Shelton (lshelton, 
916-278-3099). For all others, contact Kathy Fitzgerald (kkfitz, 
703-648-6902.)