United States Environmental Protection Agency
CBP/TRS 5/87
August 1987
A Comparison of Preservation
Techniques for Estuarine Water
Samples for Analysis of
Organic Carbon Fractions
Chesapeake
^ Ba^
Program
-------
A COMPARASION OF PRESERVATION TECHNIQUES FOR ORGANIC
CARBON ANAYSIS IN ESTUARINE WATER SAMPLES
ROBERT C. SIEGFRIED
CHESAPEAKE BAY PROGRAM
VIRGINIA WATER CONTROL BOARD
2111 NORTH HAMILTON ST.
RICHMOND, VA. 23230
JANUARY, 1987
-------
INTRODUCTION:
In order to determine whether freezing of water quality
samples was a acceptable method of preservation, the Chesapeake
Bay Program Monitoring Subcommittee funded a comprehensive
comparison study. Virginia Institute of Marine Science
preformed a comparison of five preservation treatments for eight
water quality parameters (TP, TDP, OP, N02, N023, TKN, Si,
TSS) . These results were reported in Salley et al 1984. The
State of Maryland conducted a comparison of three preservation
treatments for particulate nitrogen, phosphorus and carbon
samples. These results were reported in Vaas (1986) . Old
Dominion University provided the laboratory analyses for total
organic carbon (TOC) and dissolved organic carbon (DOC). The
statistical analysis and graphical presentation of these data is
presented in this report prepared by Chesapeake Bay Program
staff of the Virginia Water Control Board.
METHODS:
Four stations, two in the James and two in the York
rivers, were sampled to give a range of salinities and nutrient
levels. Five liter carboys of water were filled at each station
and returned to VIMS. The samples for carbon analysis were
transported to ODU. Five different combinations of preservation
and holding times where employed. Analyses for the 'Day 0'
samples (DO) were conducted 24 hours after collection. Samples
where frozen for 7 days (D7F) and 28 days (D28F) to examine
freezing as a preservation method. Samples where also preserved
with acid and analyzed within 48 hours of collection (Dl) and
after a 28 day (HT) holding time. See Salley et al (1986) for
further discussion of sampling.
The laboratory analyses were conducted following EPA
Method Reference Number 415.1. Inorganic carbon was purged from
the sample, which was then digested to convert organic carbon to
C02 , which was measured with an infrared detector. Five
replicates where prepared for analysis for each
treatment-station combination, but due to the loss of some
samples during the purging process, the number of replicates
ranged from three to five. The quality assurance and quality
control measurements included running standards, duplicates, and
spikes. In general, for each treatment three standards were
analyzed during the run, four samples where spiked with known
concentrations, and 30-40 duplicates were run. The following
are the QA\QC results.
-------
Standard= 1.4 mg/1 Duplicates Spiked Samples
Average % Diff. Average % Diff. Average % Recovery
from mean
HT 112 9.8 129.9
Dl 122 6.8 110.2
DO 102 7.0 81.9
D7F 114 5.6 89.0
D28F 108 14.4 126.7
Statistical methods similar to those used by Salley et al
(1986) were employed to test whether mean carbon concentrations
were different between the five treatments for each of the four
stations. As with the other preservation comparison reports,
comparisons between the same treatments from different stations
were deemed not appropriate. Comparisons between treatments from
the same station where the major interest in this study. \the
use of four different stations was to obtain a range of sample
concentrations and matrixes.
Due to the limited number of replicates (3-5) in each
treatment, rigorous tests of the data for normality and
homogenity of variance could not be performed. Since these
tests of the assumptions of parametric statistics were not
performed, both parametric and nonparametric statistics were
applied to the data.
For each of the four stations, a one way ANOVA was
performed to test the null hypothesis that all treatment means
were equal. A parametric Tukey's multiple comparison test was
preformed to identify which treatment means, if any, were
significantly different. Tukey's comparisons assume equal
sample size per treatment which was not always true within this
experiment. A Scheffe's comparison was conducted at both alpha
levels, since this test does not assume equal sample size, but
it is not as a powerful test as Tukey's if sample sizes are
equal.
A nonparametric ANOVA similar to the Kruskal-Wallis Test
was performed by assigning ranks to the data for each station
then performing a one-way ANOVA on the ranked data (Vaas,
1986). Tukey's multiple comparisons were also conducted on the
ranked data to identify significantly different treatments.
-------
RESULTS:
A listing of the raw data with treatment means and
standard deviations for TOC and DOC are presented in Tables I
and II, respectively. Table III and IV contains the results of
the statistical analysis of the TOC and DOC data, respectively.
Plots of the treatment means for each station for TOC and DOC
are depicted in Figures 1-4 and Figures 5-8, respectively.
In the discussion of the results an alpha level of 0.05 is
considered significant. Results for both an alpha level of 0.05
and 0.01 are presented in the .tables.
TOTAL ORGANIC CARBON;
For all stations the one-way parametric ANOVA was
significant, indicating that not all treatments were equal. The
parametric Tukey's multiple comparisons identified the HT
treatment to be significant different from the DO treatment for
all stations. For the James River samples, the HT treatment was
significantly different from all other treatments, with no other
significant differences observed in the other treatment
comparisons. The York 1 comparisons identified the HT treatment
as significantly different from the DO and D7F treatments. The
D7F treatment was also significantly different from the D28F
treatment. The York 2 comparisons identified the HT treatment
as significantly different from the DO and D28F treatments. The
D7F treatment was also significantly different from the D28F
treatment.
The Scheffe's comparisons identified eleven of the 14
significantly different treatment comparisons identified with
the Tukey' s comparisons. The James 1 and York 1 comparisons
results were the same as with Tukey's test. In the James 2
comparisons, the HT treatment was significantly different from
the other treatments except the D28F treatment. In the York 2
comparisons only the HT treatment was significantly different
from the DO treatment, in contrast to the Tukey's test which
identified three significant comparisons.
The nonparametric ANOVA (Kruskal-Wallis) procedures also
indicated that, at each station, not all treatments were equal.
The nonparametric Tukey's multiple comparisons of the ranked
data identified a few more significantly different comparisons
than the parametric Tukey comparisons. The James 1 comparisons
were the same as in the parametric Tukey's, i.e. HT treatment
was significantly different from all other treatments. The
-------
James 2 comparisons showed all treatments were significantly
different from the DO treatment except the Dl treatment. The Dl
treatment was significanlty different from the HT and D28F
treatments. The HT, D7F and D28F treatments were not
significantly different from each other. For the York 1
comparisons both the HT and D28F treatments were significantly
different from the DO treatment. The D7F treatment was
significantly different from all treatments but the DO
treatment. The York 2 comparisons indicated that both the HT
and D28F treatments where significantly different from the DO
treatment and the HT treatment was also significantly different
from the D7F treatment.
DISSOLVED ORGANIC CARBON;
The one-way parametric ANOVA was significant for only the
York 1 and York 2 stations. The treatment means for the James 1
and James 2 stations were not significantly different. The
Tukey's multiple comparisons for the York 1 station identified
the D28F treatment as significantly different from the DO, HT
and D7F treatments. The York 2 comparisons identified the D28F
treatment as significant different from the DO and HT
treatments. The Scheffe's comparisons identified the same
significant differences as the Tukey's comparisons.
The nonparametric ANOVA (Kruskal-Wallis) procedure
indicated the same results as the parametric ANOVA. The
nonparametric Tukey's multiple comparisons of the ranked data
identified a few more significantly different comparisons than
the parametric comparisons. For the York 1 comparisons the Dl
treatment was significantly different from the DO and HT
treatments. As in the parametric comparisons, the D28F
treatment was significantly different from all treatments but
the Dl treatment. York 2 comparisons indicated that the D28F
treatment was significantly different from the DO and HT
treatments, which is the same as in the parametric comparisons.
DISCUSSION:
The basic question of this research is if the preservation
methods employed preserve the original organic carbon content of
the samples. The DO treatment was the control against which the
other treatments where compared. The HT treatment is the
currently EPA approved and accepted method. The freezing
treatments were compared to both the control treatment (DO) and
the EPA approved treatment (HT) to determine if freezing could
be adopted as comparable to the EPA approved method.
-------
Examination of the plots of the TOC results, (Figures 1 -
4) , shows a general trend for increasing TOC with increased
holding time, whether frozen or acidified. While a decrease in
organic carbon could be explained by poorly preserved samples,
an increase in not easily explained. Day to day variation in
the accuracy of the analytical method may represent as much
variation as that introduced with the use of different
preservation methods. Based on the Scheffe's comparisons, the
HT treatment was significantly different from the control
treatment in three out of four stations. The D7F treatment
appears to not be significantly different from the DO
treatment. Examination of the plots and raw data shows an
increase in variation and mean concentration when comparing the
D28F treatment to the control. The D28F treatment is not
significanlly different from the DO treatment in 3 out of 4
stations based on the Scheffe's comparisons.The EPA approved
holding time of 28 days with acidification (HT) is significantly
different from the control for three out of four stations, and
significantly different from the other treatments in six out of
the other twelve possible comparisons. Of the five treatments,
the EPA approved holding time appears to be the least favorable
for the preservation of TOC samples based on comparison with the
control treatment. When freezing is compared to HT, especially
the short term freezing treatment, this preservation method
appears comparable or better. When compared to the control
treatment, freezing is somewhat guestionable, especially when
held for 28 days.
The plots of the DOC results (Figure 5-8) do not depict
any general pattern to the data. The statistical analyses
indicate that there are no significant differences between the
treatments from the James 1 and 2 stations. The raw data
indicates a larger amount of variance in the D28F treatment than
the other treatments. The HT treatment for the James 2 station
was also highly variable (range 1.8 to 4.5). The EPA approved
28 day with acidification treatment (D28F) does not appear to be
significantly different from the control treatment (DO) , in
contrast with what was found in the TOC analyses. For the York
1 and 2 stations, the D28F treatments were significantly
different from the 'control1 treatment (DO), as well as the HT
treatment, yet not from the Dl treatment. All seven day
freezing treatments (D7F) where not significanly different from
the control. Out of 8 possible comparisons between the freezing
treatments and the control treatment, only 2 where significantly
different. When compared to the control or the HT treatment,
freezing appears to be an acceptable method of preserving DOC
samples despite a few significant differences. Since these
diffenences where detected in the long term freezing treatment,
the length of storage time for frozen samples should be as short
as possible.
-------
REFERENCES
Salley, B. A., J. G. Bradshaw and B. J. Neilson. 1986. Results of
Comparative Studies of Preservation Techniques for Nutrient Analysis on
Water Samples. Gloucester Piont, Virginia Institute of Marine Science.
Vaas, P. A. 1986. Freezing of Estuarine Nutrient Samples as a
Preservation Technique: The Analysis of Particulate Nitrogen, Carbon,
and Phosphorus Fractions. Maryland Office of Environmental Programs,
Ecological Modeling and Analysis Division. Technical Report No. xxx.
-------
TABLE I
TOTAL
JAMES 1
ORGANIC CARBON ANALYSES
HT Dl DO
MEAN
STD.DEV.
5.8
4.9
5.6
3.7
5.0
0.8
2.7
2.5
2.6
3.0
2.7
0.2
2.2
2.4
2.4
2.9
3.3
2.6
0.4
D7F
2.9
2.7
2.5
2.6
2.5
2.6
0.1
D28F
2.8
3.1
2.8
3.3
3.0
3.0
0.2
JAMES 2
YORK 1
HT
HT
Dl
Dl
DO
DO
D7F
D7F
D28F
MEAN
STD.DEV.
4.9
4.6
5.1
4.8
5.0
4.9
0.2
3.2
2.9
2.9
3.1
3.3
3.1
0.2
3.1
3.0
2.8
2.8
3.2
3.0
0.2
3.3
3.3
3.2
3.4
3.3
0.1
3.1
3.3
5.8
3.5
3.6
3.9
1.0
D28F
MEAN
STD.DEV.
5.4
3.9
3.8
4.3
4.4
0.6
3.4
3.6
3.6
3.1
4.2
3.6
0.4
3.4
3.5
3.2
3.1
3.1
3.3
0.2
2.8
2.8
2.9
3.2
3.3
3.0
0.2
3.6
3.3
3.9
4.5
4.7
4.0
0.5
YORK 2
MEAN
STD.DEV.
HT
3.6
3.6
3.7
3.6
0.0
Dl
3.6
3.0
3.3
2.8
3.2
0.3
DO
2.5
2.9
2.5
3.0
2.5
2.7
0.2
D7F
2.8
2.6
3.1
2.7
2.9
2.8
0.2
D28F
3.6
3.3
2.7
4.3
4.2
3.6
0.6
-------
TABLE II
DISSOLVED ORGANIC CARBON ANALYSES
JAMES 1
HT
Dl
DO
D7F
D28F
MEAN
STD.DEV.
2.6
2.0
2.5
2.6
2.4
0.2
2.3
2.2
2.2
2.5
2.3
0.1
2.1
2.2
2.2
2.5
2.6
2.3
0.2
2.8
2.5
2.3
2.3
. 2.1
2.4
0.2
2.4
2.6
2.2
3.9
2.5
2.7
0.6
JAMES 2
MEAN
STD.DEV.
HT
2.5
1.8
4.3
2.4
2.7
0.8
Dl
3.5
2.9
2.8
3.0
3.6
3.2
0.3
DO
2.6
2.4
2.5
2.5
2.6
2.5
0.1
D7F
2.7
2.6
2.5
2.8
2.7
0.1
D28F
2.2
2.5
3.4
2.4
3.2
2.7
0.5
YORK 1
HT
Dl
DO
D7F
D28F
MEAN
STD.DEV.
2
2
2
1
2
0
.1
.3
.2
.7
.1
.2
2
2
2
2
3
2
0
.5
.5
.7
.6
.2
.7
.3
2.
2.
2.
2.
2.
2.
0.
3
4
2
2
2
3
1
2.
2.
2.
2.
2.
2.
0.
2
1
3
4
6
3
2
2
2
3
3
4
3
0
.8
.3
.3
.8
.6
.4
.8
YORK 2
MEAN
STD.DEV.
HT
2.9
3.0
3.1
3.0
0.1
Dl
3. 1
2.7
2.9
2.5
2.8
0.2
DO
3.2
2.6
3.1
3.0
2.5
2.9
0.3
D7F
2.5
2.2
3.0
2.3
2.5
2.5
0.3
D28F
2.6
2.4
2.0
2.4
1.9
2.3
0.3
-------
TABLE III.
TOTAL ORGANIC CARBON
James 1 James 2 York 1 York 2
One-way ANOVA 0.0001 0.0001 0.0013 0.0038
PARAMETRIC
TUKEY'S Mult. DO Dl HT D7F DO Dl HT D7F DO Dl HT D7F DO Dl HT D7F
Comparisons Dl . Dl . Dl . Dl .
HT # # HT # # HT*. HT*.
D7F . . # D7F^, . # D7F'^ . # D7F . . .
D28F . . # . D28F ',.' . * . D28F ... * D28F $ . . *
Scheffe's DO Dl HT D7F DO Dl HT D7F DO Dl HT D7F DO Dl HT D7F
Comparisons Dl . Dl . Dl . Dl .
HT # # HT # # HT*. HT..
D7F . . # D7F . . # D7F . . # D7F . . .
D28F . . # . D28F .... D28F ... * D28F * ...
James 1 James 2 York 1 York 2
KRUSKAL-WALLIS 0.0018 0.0001 0.0001 0.0041
NONPARAMETRIC
ANOVA
NONPARAMETRIC ''
TUKEY'S Mult. DO Dl HT D7F DO Dl HT D7F DO Dl HT D7F DO Dl HT D7F
Comparisons Dl . Dl . Dl . Dl .
HT # * HT # # HT # . HT*.
D7F . . # D7F * . . D7F . * # D7F . . *
D28F . . * . D28F # * . . D28F * . . # D28F * ...
# = Significant difference between means at alpha=0.01.
* = Significant difference between means at alpha=0.05.
. = No significant difference between means at alpha=0.05.
D0= DAY 0 (CONTROL)
Dl = DAY 1
HT = EPA APPROVED HOLDING TIME, 28 DAYS WITH ACIDIFTCATION
D7F = SAMPLE FROZEN FOR SEVEN DAYS
D28F = SAMPLE FROZEN FOR 28 DAYS
-------
TABLE IV;
DISSOLVED ORGANIC CARBON
James 1
0.4568
James 2
0.4136
York 1
0.0026
York 2
0.0072
One-way ANOVA
PARAMETRIC
TUKEY'S Mult. DO Dl HT D7F DO Dl HT D7F DO Dl HT D7F DO Dl HT D7F
Comparisons Dl . Dl . Dl . Dl .
riJL riP rxP * Hi
D7F ... D7F . . . D7F ... D7F . .. .
D28F .... D28F .... D28F # . # * D28F * . * .
Scheffe's DO Dl HT D7F DO Dl HT D7F DO Dl HT D7F DO Dl HT D7F
Comparisons Dl . Dl . Dl . Dl .
HT . . HT . . HT . . HT . .
D7F ... D7F . . . D7F ... D7F . . .
D28F .... D28F .... D28F * . * * D28F * . * .
James 1
0.6715
James 2
0.1034
York 1
0.0002
York 2
0.0075
KRUSKAL-WALLIS
NONPARAMETRIC
ANOVA
NONPARAMETRIC
TUKEY'S Mult. DO Dl HT D7F DO Dl HT D7F DO Dl HT D7F DO Dl HT D7F
Comparisons Dl . Dl . Dl *' ^ Dl .
HT . . HT . . HT . | ' HT . .
D7F . . . D7F . . . D7F ... D7F . . .
D28F .... D28F .... D28F # . # * D28F * . * .
* = Significant difference between means at alpha=0.05
. = No significant difference between means at alpha=0.05
D0= CAY 0 (CONTROL)
Dl = DAY 1
HT = EPA APPROVED HOLDING TIME, 28 DAYS WITH ACIDIFICATION
D7F = SAMPLE FROZEN FOR SEVEN DAYS
D28F = SAMPLE FROZEN FOR 28 DAYS
-------
FIGURE 1
COMPARISON OF TOTAL ORGANIC CARBON
CONCENTRATION BY TREATMENT
e.o
3.3
a.o
2: 4.0
O
^ " ^
o
0 3.3
3.O-
2.3-
2.O-
AT JAMES 1
\
\ '.
\\
\ \\
\\v
-tl^^^^
'--.. ..-''
HT D1 DO D7F D28F
LEGEND
1- SD
- SD
TREATMENT
FIGURE 2
COMPARISON OF TOTAL ORGANIC CARBON
CONCENTRATION BY TREATMENT
AT JAMES 2
UJ
o
LEGEND
mean
-e SD
- SD
HT
D1
DO
D7F D28F
-------
FIGURE 3
cor
a.o-
o.o-
a.o-
CONCENTRATION
(MG/L)
V f .ť
a o a
3.O-
a.o-
a.o-
\/IPARISON OF TOTAL ORGANIC CARBON
CONCENTRATION BY TREATMENT
AT YORK 1
\
/
X \ 7
X Vx \ //
'"'"^^^-"r^'''
HT D1 DO D7F D28F
TREATMENT
LEGEND
mean
-t- SD
- SD
FIGURE 4
COMPARISON OF TOTAL ORGANIC CARBON
CONCENTRATION BY TREATMENT
AT YORK 2
CJ
z
o
HT
Dl DO D7F
TREATMENT
D2SF
LEGEND
mean
-t- SD
- SD
-------
FIGURE 5
cc
<.oo-
J.BŤ
J.38
CONCENTRATION
(MG/L)
-MM
ť H a
00-
1.13-
.00-
0-
DMPARISON OF DISSOLVED ORGANIC CARBON
CONCENTRATION BY TREATMENT
AT YORK 1
/
/ /
; /^^^^/>
.-'
HT D1 DO D7F D28F
TREATMENT
LEGEND
+ SO
. SD
FIGURE 6
COMPARISON OF DISSOLVED. ORGANIC CARBON
CONCENTRATION BY TREATMENT
AT YORK 2
4.3O
O
(=
UJ
O
I o
LEGEND
moan
+ SD
- SD
HT
D1 DO D7F
TREATMENT
D28F
-------
FIGURE 7
COMPARISON OF DISSOLVED ORGANIC CARBON
CONCENTRATION BY TREATMENT
4.SO
3.84-
3.38-
CONCENTRATION
(MG/L)
M M
'Ť u a
ID 0 -
1.13-
.sŤ-
o-
AT JAMES 1
x
X
r____^r^:^
HT D1 DO D7F D2BF
LEGEND
maon
-t- SD
- SD
TREATMENT
FIGURE 8
COMPARISON OF DISSOLVED ORGANIC CARBON
CONCENTRATION BY TREATMENT
AT JAMES 2
3.94-
3.38-
O
(=
o
O
LEGEND
moon
-t- SD
- SD
HT D1 DO D7F
TREATMENT
D28F
------- |