United States
Environmental Protection
Agency
Office of Research
and Development
Gulf Breeze. FL
EPA/600/9 91/046b,
October 1991
Research and Development
&EPA
51
o o>
S£ •
Iw^
Alaska Oil Spill
Bioremediation Project
Science Advisory Board
Draft Report
Sections 7 through 13
and Appendices
.- •
U^N^rvT-n^vVx,
-------
EPA/600/9-91/046b
October 1991
ALASKA OIL SPILL BIOREMEDIATION PROJECT
SCIENCE ADVISORY BOARD DRAFT REPORT
SECTIONS 7 THROUGH 13 AND APPENDICES
*•*.
V"
P.H. Pritchard, EPA/ERL-Gulf Breeze, Scientific Coordinator
C.F. Costa, EPA/EMSL-Las Vegas, Program Manager
L. Suit, TRI-Rockville, Technical Editor
Contributors:
R. Araujo, EPA/ERL-Athens; D. Chaloud, Lockheed; L. Cifuentes, Texas A&M University;
J. Clark, EPA/ERL-Gulf Breeze; L. Claxton, EPA/RTP; R. Coffin, EPA/ERL-Gulf Breeze;
R. Gripe, EPA/ERL-Gulf Breeze; D. Dalton, TRI; R. Gerlach, Lockheed;
J. Glaser, EPA/RREL-Cincinnati; J. Haines, EPA/RREL-Cincinnati;
D. Heggem, EPA/EMSL-Las Vegas; F. Kremer, EPA/CERI-Cincinnati; J. Mueller, SBP;
A. Neale, Lockheed; J. Rogers, EPA/ERL-Athens;
S. Safferman, EPA/RREL-Cincinnati; M. Shelton, TRI;
A. Venosa, EPA/RREL-Cincinnati
Exxon Contributors:
J. Bragg, R. Chianelli, and S. Hinton, Annandale, N.J.;
S. McMillan, Houston, Texas; R. Prince, Annandale, N.J.
U.S. Environmental Protection Agency
Region 5, Library (PL-12J)
Prepared by: 77 West Jackson Boulevard, 12th Floor
Chicago, IL 60604-3590
ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
SABINE ISLAND
GULF BREEZE, FLORIDA 32561
-------
SECTION 7
PASSAGE COVE FIELD RESULTS
VISUAL OBSERVATIONS
Original oil contamination in Passage Cove was considered moderate. Following complete
physical washing, oil was well distributed over most of the surface of all cobble and all gravel under
the cobble. The oil was black, dry, and dull in appearance with considerable stickiness. It was spread
as a thin layer over the beach material. Relatively few patches of pooled oil or mousse were present,
but where they were, the oil was thick and viscus. Oil was also found at depth in the beach, generally
30 to 40 cm below the surface. It was well distributed within the beach material.
Approximately two weeks following the combined application of oleophilic and granular
fertilizer (Tern Beach), it became apparent that the treated beach was considerably cleaner relative
to the untreated control plots. In contrast to the observations at Snug Harbor, the rock surfaces looked
cleaner and the oil under the rocks in the mixed sand and gravel plots within the beaches had
disappeared to a greater extent. In another two weeks, oil was found only in isolated patches, and at
10 cm and below in the subsurface. At no time were oil slicks or oily material seen leaving the beach
area. During this time, no visual disappearance of oil from the surface of cobble on the untreated
control beach (Raven) occurred.
Unexpectedly, the beach treated with fertilizer solution from the sprinkler system (Kittiwake
beach) behaved in a very similar manner to the oleophilic/granular-treated plot: extensive
disappearance of the oil occurred compared to the untreated control plots (Figures 7.1 and 7.2). The
only difference was that it lagged behind the oleophilic/granular-treated beach by approximately 10
to 14 days. By the end of August, both beaches (Kittiwake and Tern) looked equally clean. In
contrast, the untreated control beach appeared very much as it did in the beginning of the field study.
Oil in the subsurface still remained in all plots within the beaches. However, in the fertilizer-treated
plots within the beaches, oil was visually apparent only below a depth of 20 to 30 cm.
Disappearance of oil from the rock surfaces on the beach treated with the fertilizer solution
provided the definitive proof that biodegradation (and not chemical washing) was responsible for the
oil removal, as there was no other reasonable mechanism to explain this fertilizer-mediated loss of
oil.
251
-------
Figure 7.1.
Kittlwake Beach at Passage Cove, Treated with Fertilizer
Solution from the Sprinkler System, Showed Extensive
Disappearance of OH Compared to Untreated Control Plots
(Figure 7.2).
Figure 7.2. Oiled Untreated Control Plots.
252
-------
PASSAGE COVE
NUTRIENT CONCENTRATIONS
Data on the nutrient concentrations in tidal waters following fertilizer application at Passage
Cove were lost by the contractor (SAIC) responsible for performing the analysis.
OIL CHEMISTRY
Oil Residue Weight
Cobble Surface Samples
Figures 7.3 to 7.5 show changes (decay curves), for the three test beaches (Tern, Raven and
Kittiwake), in residue weights of oil on the cobble surface. The residue weights are normalized to
a kilogram weight of cobble. Given the relatively smooth surface of the cobblestones, it was assumed
that a consistent relationship between rock surface area and rock weight existed. Variability in the
residue weights was a function of both the sampling of heterogenous beach material and the uneven
distribution of oil on the beach.
Highly variable oil residue weights at all sampling times (two orders of magnitude variation in
some cases) are conspicuous in the figures. Extensive decreases in oil residue weights through time
occurred on all beaches. The percent change over time in the median residue weight is shown in
Figure 7.6 and Table 7.1.
^
A statistical comparison of the residue weights for the cobble samples was conducted using the
nonparametric Mann-Whitney test (Zar, 1984). The results from this test show that at the 95%
confidence level, significantly less oil residue was present on the fertilizer solution-treated beach
(Kittiwake) than on the untreated control beach (Raven) at all sampling times except t-0 days (Table
7.1). Thus, initially these two beaches had statistically similar amounts of oil residue on the beach
cobble, but with time more oil was lost from the fertilizer solution-treated beach. By the last
sampling period in September, approximately four times more oil residue remained on the surface of
the untreated control beach than on the fertilizer solution-treated beach.
253
-------
100
10
0.1
0.01
0.001
22 Jut 69
06Aug89 21 Aug 88
Sampling
06Sep88
0 Anolviii Value
» Median
Figure 7.3. Change In Oil Residue Weight Through Time for Raven
Beach (Untreated Control) at Passage Cove (Cobble
Surface).
100
10
0.1
0.01
0.001
22 Jul 89 06Aug89 21 Aug 89
Sampling Data
* Median
05Sep89
Figure 7.4. Change In Oil Residue Weight Through Time for Tern Beach
(INIPOL + CUSTOMBLEN) at Passage Cove (Cobble Surface).
254
-------
I
100
10
0.1
0.01
0.001
22JUI89
06 Aug89 21 Aug 89
Sampling Data
06Sep89
0 Anolvaii Volu«
Median
Figure 7.5. Change In Oil Residue Weight Through Time for Klttlwake
Beach (Fertilizer Solution) at Passage Cove (Cobble
Surface).
Legend
% of 7/22
Median
-f>- Kittiwake
-0- Tern
HB- Raven
20
Figure 7.6.
0 5 10 15 20 25 30 35 40 45
Day
Change In the Median Residue Weight, Expressed as
Percent of the 7/22 Median Over Time for Klttlwake, Raven,
and Tern Beaches at Passage Cove.
255
-------
PASSAGE COVE
TABLE 7.1. MEDIAN VALUES AND STATISTICAL COMPARISONS OF OIL
RESIDUE WEIGHTS FROM DIFFERENT BEACH TREATMENTS'
Median Values (Percent of 7/22/89 Median)
Sampling Date
7/22/89
8/06/89
8/20/89
9/05/89
Days
0
15
29
44
Kittiwake
0.684
0.364 (53%)
0.228 (33%)
0.120(18%)
Tern
0.476
0.274 (58%)
0.194(41%)
0.249 (52%)
Raven
0.596
.61 1 (103%)
0.530(89%)
0.481(81%)
Mann-Whitney Test Results6
Sampling Date
7/22/89
8/06/89
8/20/89
9/05/89
Kittiwake vs. Tern
SAME
SAME
SAME
SAME
Kittiwake vs. Raven
SAME
KW
-------
PASSAGE COVE
Further data analysis demonstrated a statistically significant difference in the residue weights
at the 95% confidence interval between the INIPOL/CUSTOMBLEN-treated beach (Tern) and the
untreated control beach (Raven) included the t-0 sampling date (Table 7.1). Thus, the amount of oil
residue on each beach was initially different, and through time this difference was preserved.
Therefore, enhanced oil disappearance (relative to the untreated control) as a result of fertilizer
application on Tern beach cannot be established from this analysis.
Statistical analysis of the oil residue weight decay rates was also conducted. Regression analysis
of the natural log-transformed residue weights versus sampling time suggested that a linear first-order
approximation was sufficient to explain the decay curves for Kittiwake and Raven beaches, but not
for Tern beach. Examination of the data from Tern beach shows that the median oil residue weight
on the last sampling date actually increased. Plotting the percent changes in the medians shows this
rather dramatically (Figure 7.6). This increase could not be attributed to any anomalies in the data
(the residue weights for the last two sampling dates on Tern are not statistically different), although
some reoiling of the beach could have occurred.
It is also possible that the decay curve for Tern beach reflects an oil biodegradation rate
limitation by both substrate and nutrients because the nutrient release capabilities of the
INIPOL/CUSTOMBLEN fertilizers became spent by the last sampling date (at Kittiwake beach
nutrient supply was sustained at initial levels throughout). Accordingly, since the last two sampling
dates on Tern are not statistically different, it is possible that the biodegradation rate of the oil
leveled-off very suddenly.
Comparisons of the first-order decay constants (slopes of the lines) using a least squares fit are
shown in Table 7.2 (Tern results are based on only three sampling dates, as the September 5 results
are not consistent with the first order behavior seen in the first three sampling periods). Only the
slope for Kittiwake beach (fertilizer solution-treated) was significantly different from zero; p-0.0009
is much less than the 0.05 level for the traditionally used 95% confidence level. The slope for Tern
beach (INIPOL/CUSTOMBLEN-treated), however, is very close to being significantly different from
zero (93.8% confidence level), suggesting a potential significant effect of the fertilizer application.
The slope for Raven was significantly different from zero at a lower confidence level (<90%).
257
-------
PASSAGE COVE
TABLE 7.2. RATE ANALYSIS OF NATURAL LOG-TRANSFORMED OIL RESIDUE
WEIGHTS ON COBBLE SURFACES FOR TEST BEACHES IN PASSAGE COVE
Significance of Slope
Fertilizer Slope Greater than Zero Half Life, Time to Remove
Beach* Treatment (Std. Dev.) N T-value p days 90%, days
KW Fertilizer -0.030 80 -3.45 0.0009 23.1 76
Solution (0.0087)
T INIPOL/ -0.016 60 -1.90 0.062 43.3 143
CUSTOMBLEN (0.0087)
R None -0.006 82 -1.54 0.127 115.5 383
(0.0041)
Beaches Compared
KW/R
T/R
KW/T
Ratio of
Slopes
4.7
2.6
1.8
t Value
2.51
1.16
0.79
Degrees of
Freedom
158
138
136
Probability Level
<0.01
>0.10
>0.25
•KW - Kittiwake Beach; T - Tern Beach; R - Raven Beach.
Ratios of the slopes suggest that the addition of fertilizer solution to Kittiwake increased the
decay rates as much as 4.7 times over the untreated control beach. Addition of INIPOL/
CUSTOMBLEN produced a decay rate 2.6 times greater than the untreated control beach. However,
only the Kittiwake/Raven comparison was statistically significant at the 95% confidence interval.
The Tern/Raven comparison was significant at the 90% confidence level.
Table 7.2 also gives the calculated half-lives for the decay of the oil residue weights on each
beach. It was assumed that the decay rate remained constant over an extended period (into winter
months), although there is no evidence to support this assumption. Assuming that the first-order
258
-------
PASSAGE COVE
decay rate constant remains the same, the time to remove 90% of the starting material was calculated;
the values are 76, 143, and 383 days for Kittiwake, Tern, and Raven, respectively. From this we
would predict that by the following spring (1990), all of the oil on the fertilizer solution-treated beach
should be gone, but oil might still be found on the other beaches. The other beaches have large
uncertainties in their slopes, making predictions very uncertain.
The oil residue weight decay curves were also analyzed as zero-order rates by determining the
least squares fit of nontransformed medians to a straight line. From inspection it was noted that the
data from the last sampling date on Kittiwake and Tern (September 5) was not following a linear
trend. The Tern median was above the third sampling date (an anomalous increase) and the Kittiwake
median did not decrease enough for a linear rate of change (significantly different behavior from the
assumed zero-order decay). Thus, only the first three sampling dates were included in the analysis.
The results are shown in Table 7.3.
TABLE 7.3. RATE ANALYSIS OF NONTRANSFORMED OIL RESIDUE WEIGHTS
(ZERO-ORDER) ON COBBLE SURFACES IN PASSAGE COVE
Beach
Raven
Tern
Fertilizer
Treatment
None
INIPOL/
CUSTOMBLEN
Slope*
(mg/kg/day)
-7.2
-13.2
Std. Dev.«
1.3
2.4
T-Value
-5.5
-5.5
Significance
0.03
0.11
Estimated
Time (days)
for complete
removal
32
36
Kittiwake Fertilizer
Solution
-21.3
0.7 -29.0
0.02
82
'Linear regression fit.
The slopes and standard deviations are in units of mg oil residue lost per kg of beach material
per day. As expected, with only one degree of freedom per median point, the significance is not very
good, particularly for Tern beach. However, because each median represents the location of a
distribution of approximately 20 samples and the distribution is skewed, the derived slopes are
259
-------
PASSAGE COVE
probably more representative of the change in the central values of the oil residue than the
significance values would indicate. Direct extrapolation of this rate information to cleanup times
must, therefore, be considered with caution. Accordingly, the rate analysis would suggest that much
of the oil on the cobble surfaces from the fertilizer-treated beaches should have disappeared in four
to five weeks. This agrees with the visual observations. Oil on the cobblestone surface from the
untreated control beach should have disappeared much more slowly, and as observed, oil was still
present 11 to 12 weeks following initiation of the field study.
It can be confidently concluded that the application of fertilizer solution promoted the loss of
oil residue four to five times faster than no treatment, and with slightly less confidence,
approximately two to three times faster with the INIPOL/CUSTOMBLEN treatment.
Mixed Sand and Gravel Samples
Analysis of oil disappearance rates in the mixed sand and gravel samples showed that only the
fertilizer solution-treated beach (Kittiwake) had a slope significantly different from zero (Figures 7.7
to 7.9). The first-order decay rate constant based on regression analysis of log transformed medians
was -0.035 per day (std. dev. - 0.009, t - -3.73, significance - 0.0004), or a half-life of 20 days. Oil
disappearance in the mixed sand and gravel on Kittiwake beach, therefore, was as fast as the cobble
surface. The absence of any significant effect of fertilizer on Tern beach is difficult to explain. It
does not appear to be an effect of different oil concentrations; residue weights per unit of beach
material are generally the same. Availability of nutrients to the bacterial populations and to the oil
are unlikely to be much different from cobble surfaces, as the mixed sand and gravel was very porous.
However, because of the fertilizer release characteristics, it is possible that lower concentrations of
nutrients reached the mixed sand and gravel on Tern beach. It is also possible that greater retention
of oil residues within the matrix of the mixed sand and gravel may have occurred. Tidal action
should have had less of a physical scouring effect in this matrix relative to the surface of the
cobblestone.
260
-------
100
10
0.1
0.01
0.001
22JUI89
o
o
o
Analysis Volu«
i
06Aug80 21 Aug 80 06 Sep 89
Sampling Dito
Figure 7.7. Change In Oil Residue Weight Through Time for Raven
Beach (Untreated Control) at Passage Cove (Mixed Sand and
Gravel).
100
10
0.1
0.01
0.001
22 Jul 89 06 Aug 89 21 Aug 89
Sampling Dafc
0 Analysis Volm
Median
05 Sep 89
Figure 7.8. Change in Oil Residue Weight Through Time for Tern Beach
(INIPOL + CUSTOMBLEN) at Passage Cove (Mixed Sand and
Gravel). The Number of Samples Showing Concentrations
Below Detection Limit Is Shown Above the Sampling Date.
261
-------
100 r
10
0,1
0.01
0.001
0
8
22Jul89 06Aug89 21Aug89
Sampling Data
Anotv»i« Volue
• Medlon
—i
o
9
05 Sep 89
Figure 7.9. Change In Oil Residue Weight Through Time for Klttlwake
Beach (Fertilizer Solution) at Passage Cove (Mixed Sand and
Gravel).
262
-------
PASSAGE COVE
Oil Composition
Cobble Samples
Changes through time in the concentration of nC18, the sum of normal alkanes nC18 to nC27,
and phytane in cobble samples lend strong support the role of biodegradation. Decay curves for each
of these hydrocarbons are shown in Figures 7.10 to 7.18. All values of hydrocarbon concentration are
normalized to the weight of oil in the extracted sample. The normal alkanes nC18 to nC27 were
chosen because correlation analysis showed that these hydrocarbons tracked each other consistently.
In all cases, values below detection limits were treated as zeros.
For comparative purposes Tables 7.4 and 7.5 summarize the percent change in the medians of
individual hydrocarbons, and the number of samples showing a hydrocarbon concentration of zero
(below detection limits) with each sampling period, respectively. Figure 7.19 also provides a graphical
representation of the percent change in the medians.
By qualitatively assessing the extent of compositional change, several important trends were
identified. Overall, the results mirror trends observed with the oil residue weights: a substantial
decrease in hydrocarbon concentration through time occurred for both fertilizer-treated and untreated
control beaches. However, Kittiwake beach was consistently and distinctly different from the other
two beaches. Percent change in the medians for Kittiwake was always the largest and the quickest
to occur, and many more samples had hydrocarbon concentrations below detection limits. As
observed with the residue weights, there was little question that the application of the fertilizer
solution on Kittiwake beach affected changes in oil composition that were substantially greater than
those observed in samples from the untreated control beach (Raven).
Effects resulting from the application of INIPOL/CUSTOMBLEN to Tern beach are less clear
and distinct. In general, it appeared that changes in hydrocarbon composition were in-between those
observed for Kittiwake and Raven. Tern beach appeared to behave differently on many sampling
dates compared to Raven beach: percent change in the medians and the number of below detection
limit values were considerably greater than Raven beach in many cases. These differences were most
prominent on the August 6 and August 20 sampling dates.
263
-------
10
5 0.1
"8
00
5
0.01
0.001
22 Ju! 89 06 Aug 89 21 Aug 89
Sampling Date
I ° Anplyjis Voiut •
13
05Sep89
Figure 7.10. Change In nC18 Alkane Concentration Through Time for
Raven Beach (Untreated Control) at Passage Cove (Cobble
Surface). The Number of Samples Showing Concentrations
Below Detection Limit Is Shown Above the Sampling Date.
264
-------
100 J
10
= 1
5 0.1
0.01
22 Jul 89 06AUQ68 21 Aug 89
SmptigDrti
06 Sap 89
0 Anolvtli Volua
• Mtdlan
Figure 7.11. Change In the Sum of Alkane Concentration nC18 to nC27
Through Time for Raven Beach (Untreated Control) at
Paeeage Cove (Cobble Surface). The Number of Samples
Showing Concentrations Below Detection Limit Is Shown
Above the Sampling Date.
10 \
-j.
0.1-
0.01-
0.001-
22 Jul 89 06 Aug 89 21Aug89
Sampling DM
9 Anolviii Valut
05Sep89
• M.d,an
Figure 7.12. Change In Phytane Concentration Through Time for Raven
Beach (Untreated Control) at Passage Cove (Cobble
Surface).
265
-------
10-
i
O Q^\
"B
5
i
c
0.001 L^
22 Jul 89 06 Aug 89 21Aug89
Sampling Date
° Anolysis Volue
Median
16
05Sep89
Figure 7.13. Change in nC18 Alkane Concentration Through Time for
Tern Beach (INIPOL + CUSTOMBLEN) at Passage Cove
(Cobble Surface). The Number of Samples Showing
Concentrations Below Detection Limit Is Shown Above the
Sampling Date.
100 f
10-
= 1-
8
0.1-
0.01
o
o
22 Jul 89 06Aug89 21 Aug 89
Sampling Date
° Anolyais Volue
* Medion
05Sep89
Figure 7.14. Change In the Sum of Alkane Concentration nC18 to nC27
Through Time for Tern Beach (INIPOL + CUSTOMBLEN) at
Passage Cove (Cobble Surface). The Number of Samples
Showing Concentrations Below Detection Limit is Shown
Above the Sampling Date.
266
-------
101
^•
3 0.1
•8
0.01
0.001
22Jul89
o
o
06 Aug 89 21 Aug 89 06 Sep 89
Sampling Data
° Anolviii Vqlue
• M«dion
Figure 7.15. Change In Phytane Concentration Through Time for Tern
Beach (INIPOL + CUSTOMBLEN) at Passage Cove (Cobble
Surface). The Number of Samplea Showing Concentratlone
Below Detection Limit It Shown Above the Sampling Date.
10 {
1
<3 0.1
T5
eo
5
i
c
0.01
0.001
19
22 Jul 89 06 Aug 89 21 Aug 89
Sampling Data
Anolviii Volut
* M«dion
21
06 Sep 89
Figure 7.16. Change In nC18 Alkane Concentration Through Time for
Klttlwake Beach (Fertilizer Solution) at Paeeage Cove
(Cobble Surface). The Number of Samplee Showing
Concentratlone Below Detection Limit ie Shown Above the
Sampling Date.
267
-------
100
10
5 0.1
0.01
22 Jul 89
06 Aug 88 21 Aug 89
Sampling
0 Anolvili Valut
Mtdton
OSSepBQ
Figure 7.17. Change In the Sum of Alkane Concentration nC18 to nC27
Through Time for Klttlwake Beach (Fertilizer Solution) at
Paeeage Cove (Cobble Surface). The Number of Samples
Showing Concentrations Below Detection Limit le Shown
Above the Sampling Date.
10
0.1
0.01
0.001
22 Jul 89
06 Aug 89 21 Aug 89
Sampling Date
• Mlon
12
05Sep8i
Figure 7.18. Change In Phytane Concentration Through Time for
Klttlwake Beach (Fertilizer Solution) at Passage Cove
(Cobble Surface). The Number of Samples Showing
Concentrations Below Detection Limit Is Shown Above the
Sampling Date.
268
-------
PASSAGE COVE
TABLE 7.4. CHANGE IN HYDROCARBON COMPOSITION THROUGH TIME
EXPRESSED IN PERCENT OF THE MEDIAN CONCENTRATION OF
INDIVIDUAL HYDROCARBONS ON THE 7/22 SAMPLING*
Alkane
nC18
nC19
nC20
nC21
nC22
nC23
nC24
nC25
nC26
nC27
nC18 to
nC27
Phytane
Beach Code
K
T
R
K
T
R
K
T
R
K
T
R
K
T
R
K
T
R
K
T
R
K
T
R
K
T
R
K
T
R
K
T
R
K
T
R
8/6
11
43
60
18
54
52
27
41
68
24
39
59
32
47
59
23
35
83
31
44
92
59
41
92
43
53
75
43
57
82
30
44
77
76
67
85
8/20
0
34
27
0
33
34
0
23
24
0
15
40
11
27
41
10
32
34
0
28
46
0
33
56
6
40
59
37
51
55
10
29
44
19
41
86
9/5
0
0
0
0
13
11
0
0
6
0
0
1
0
0
7
0
12
14
0
0
9
0
0
12
0
0
6
0
0
0
1.4
7
8
0
41
33
•R-Raven (untreated control beach); T-Tern (INIPOL/CUSTOMBLEN-treated beach); K-Kittiwake
(fertilizer solution-treated beach).
269
-------
PASSAGE COVE
TABLE 7.5. OF APPROXIMATELY 21 SAMPLES, NUMBER OF SAMPLES TAKEN AT
EACH SAMPLING TIME WITH ALKANE CONCENTRATION BELOW DETECTION LIMIT'
Alkane
nC18
nCI9
nC20
nC21
nC22
nC23
nC24
nC25
nC26
nC27
nC18 to
nC27
Phytane
Beach Code
K
T
R
K
T
R
K
T
R
K
T
R
K
T
R
K
T
R
K
T
R
K
T
R
K
T
R
K
T
R
K
T
R
K
T
R
7/22
I
5
0
0
2
0
1
1
0
0
0
0
0
0
0
0
0
0
1
0
0
1
1
0
1
1
0
1
1
0
0
0
0
0
1
0
8/6
6
4
1
0
3
1
0
0
1
0
2
1
0
2
1
0
2
1
1
2
1
1
3
1
2
2
2
0
2
1
0
0
1
0
0
0
8/20
19
7
6
13
6
1
12
4
0
13
5
1
5
3
1
6
2
2
14
6
1
13
5
1
10
5
1
2
3
3
1
2
0
7
2
0
9/5
21
16
13
17
9
6
17
15
7
14
13
10
13
11
9
19
7
1
20
17
8
20
15
5
20
14
9
15
14
13
9
4
0
12
2
0
•R-Raven (untreated control beach); T-Tern(INIPOL/CUSTOMBLEN-treated beach); K-Kittiwake
(fertilizer solution-treated beach).
270
-------
LEGEND
H R • Ravtn Beach
I I T.Tem Beach
PI K.KittiwakaB«ach
I?
Sum nCl8tonC27
Figure 7.19. Change in Composition Expressed as
Percent of the Median Concentration of
Individual Hydrocarbons on the 7/22/89
Sampling for Raven, Tern, and Klttiwake
Beaches at Passage Cove
271
-------
PASSAGE COVE
These qualitative differences are further enhanced in Figures 7.20 and 7.21. On the August 6
sampling day, it is apparent that for each of the individual hydrocarbons, Kittiwake samples
consistently showed the greatest percent changes in the medians, Raven samples the least, and Tern
samples in-between. Samples from both Kittiwake and Raven also show a general upward trend in
the percent change in the medians, suggesting, as one might expect, that the lower molecular weight
alkanes degrade more rapidly than the higher molecular weight alkanes. It is interesting that the
upward trend is not apparent in the Tern samples. It is tempting to argue that initially the
IN1POL/CUSTOMBLEN fertilizer application caused the higher molecular weight alkanes to be
degraded as rapidly as the lower molecular weight alkanes. This is also reflected in the percent
change of the phytane medians for this sampling date (data not plotted) in which the greatest change
occurred in the samples from Tern beach. Svem, 1987, speculated that the mechanism of action for
INIPOL might involve increased bioavailability of oil hydrocarbons, and this could affect the
biodegradation rate of the higher molecular weight alkanes.
Results from the August 20 sampling day again show that the extent of the oil compositional
change on Tern beach was generally in-between Raven and Kittiwake. However, in the samples from
Tern it appeared that further decreases of the high molecular weight alkanes were small in contrast
to the lower molecular alkanes, thus generating a pattern of compositional change similar to the other
test beaches. This could be related to a waning of the INIPOL effect, and consequently a "catching
up" by degradation in the absence of fertilizer.
Historically, the definitive indication that changes in hydrocarbon composition are due to
biodegradation has been based on examining the weight ratio between a hydrocarbon that is known
to readily biodegrade (generally the nC17 or nC18 alkanes) and those that are slower to biodegrade
(generally the branched alkanes, pristane and phytane, which chromatograph very close to the nCI7
and nC18 alkanes, respectively). This is based on the concept that most nonbiological fate processes
(physical weathering, volatilization, photolysis, etc.) do not produce differential losses of aliphatic
hydrocarbons that have similar gas chromatographic and chemical behavior. Consequently, any
decrease in the ratio can be confidently assigned to the process of biodegradation. The changes in
the nC18/phytane ratio are shown in Figures 7.22 to 7.24. The nC17/pristane ratios are not shown,
but they provide essentially the same information.
272
-------
Raven
Ttrn
Klttlwaki
UJ
u.
O
N-ALKANE
Figure 7.20 Hydrocarbon Composition on August 6,1989, Expressed as
Percent of the Median Concentration of Individual
Hydrocarbons on the 7/22/89 Sampling for Raven, Tern, and
Klttlwake Beaches at Passage Cove.
Rav«n
T«rn
Klttlwikt
N-ALKANE
Figure 7.21. Hydrocarbon Composition on August 20,1989, Expressed as
Percent of the Median Concentration of Individual
Hydrocarbons on the 7/22/89 Sampling for Raven, Tern, and
Klttlwake Beaches at Passage Cove.
273
-------
.2
5
i
c
1.5
1
0.51
22 Jul89
06 Aug 89 21 Aug
Sampling Date
0 Anolysis Value
• Median
13
05Sep89
Figure 7.22. Change in the nC18/phytane Ratio Through Time for Raven
Beach (Untreated Control) at Passage Cove (Cobble
Surface). The Number of Samples Showing Concentrations
Below Detection Limit is Shown Above the Sampling Date.
2
1.5
DC
« 0.5
5
i
c
0
2
o
o
o
o
o o
j i • i
8 i — r~"~~-~— ~-— _
5 * 7 16
2JUI89 06 Aug 89 21 Aug 89 06 Sep 89
Sampling Date
1 ° Analysis Value * Median |
Figure 7.23. Change in the nC18/phytane Ratio Through Time for Tern
Beach (INIPOL + CUSTOMBLEN) at Passage Cove (Cobble
Surface). The Number of Samples Showing Concentrations
Below Detection Limit is Shown Above the Sampling Date.
274
-------
1.5
1
0.5
0
19
21
22 Ju! 89 06 Aug 89 8 Aug 89
Sampling Dite
06 Sep 89
0 Anolv»a Volvt
Figure 7.24. Change In the nC18/phytane Ratio Through Time for
Kittiwake Beach (Fertilizer Solution) at Passage Cove
(Cobble Surface). The Number of Samplee Showing
Concentratlone Below Detection Limit le Shown Above the
Sampling Date.
275
-------
PASSAGE COVE
Several important pieces of information can be derived from these graphs. First, the initial
ratios indicate that the oil samples taken from all beaches were significantly biodegraded. The ratio
of unweathered Prudhoe Bay crude oil should be around 2.0. Median ratios were beow 1.0 at the
initiation of the test. Thus, oil on the beaches was more than 50% biodegraded in terms of the
aliphatic hydrocarbons. The extent of biodegradation was not statistically different between the three
test beaches.
Under the assumption that oil on the beaches at the initiation of the fertilizer application test
was generally uniformly colonized by oil-degrading bacteria, it is important to know if the variability
in oil residue weights among beach samples was a function of the initial heterogenous distribution of
the oil, or a function of the amount of biodegradation that had already taken place. In other words,
was biodegradation prior to fertilizer application sufficient to remove significant amounts of oil
residue from the beach material. Figures 7.25 and 7.26 show there was a very weak correlation (r
square - 0.36 and 0.26, respectively, at the 95% confidence level) between standing oil residue weight
and the nC18/phytane ratio at t-0 and t-0 plus one sampling. Thus, the variability in both the extent
of biodegradation of the oil and the amount of oil on the beach at t-0 was essentially random.
Second, it is clear that the ratios decreased significantly on all beaches during the first two
sampling periods, with the fertilizer solution-treatment showing the most effect. Considering just
the medians on the second sampling (August 6), the percentage decrease in the ratios relative to the
July 22 sampling was 85% for Kittiwake, and 33% for Tern and Raven. In many of the later samples,
as the nC18 fell below detection limits, the ratio could not be calculated. In addition, significant
decreases in the concentration of phytane occurred on all beaches (Figures 7.12, 7.IS, and 7.18). It
is assumed that this decrease is the result of biodegradation, although there is little direct proof other
than it is easy to isolate bacteria from oiled beach samples that grow on phytane as a sole source of
carbon and energy. The results further support that fertilizer-enhanced biodegradation does not
result in a preference by the natural microbial community to degrade only the easily degradable
hydrocarbons. In fact, just the opposite appears to be true; enhanced biodegradation, as evidenced
on Kittiwake beach, perhaps leads to a more extensive degradation of oil.
276
-------
1.5
1-
00
5
I
0.5
0 0
10 100 1,000 10,000
OH Residue Weight (mg/kg)
o o o Analytic Vtt»
... D**ton Urn* Reported
Figure 7.25. Relationship Between the nC18/phytane Ratio and Oil
Residue Weight for All Beaches at Passage Cove (Cobble
Surface) at Time Zero. R Value Day U0.36.
277
-------
1.5
oo
5
0.5
08
oo
10 100 1,000 10,000
Oil Residue Weight (mg/kg)
o o o AralyHs
• • •
LJfnlt
Figure 7.26. Relationship Between the nC18/phytane Ratio and OH
Residue Weight for All Beaches at Passage Cove (Cobble
Surface) at Time Zero + Time One Sampling. R Value for
Days 1 and 2=0.26.
278
-------
PASSAGE COVE
Third, if changes in phytane concentration were due to biodegradation, then the persistence of
phytane is quite different between the three beaches as shown in Figure 7.27. Up to and including
the August 20 sampling, phytane in samples from the untreated control beach had changed very little,
showing only a 15% decrease in the median concentration. The large change subsequent to this date
is unexplained. Phytane concentrations were substantially reduced on the fertilizer-treated beaches,
completely absent in samples from Kittiwake on the last sampling date, and reduced approximately
60% on Tern beach by the August 20 sampling. On Tern, reductions in median phytane
concentrations leveled-off by the last sampling, suggesting perhaps a possible depletion of the active
ingredient in the IN1POL/CUSTOMBLEN fertilizer materials. The difference in phytane reduction
between Tern and Raven beaches (Figure 7.27) suggests that the nC18 ratios are misleading, and, in
fact, there is probably considerably more oil biodegradation occurring on Tern beach than is
indicated. However, we can not rule out the possibility that the INIPOL/CUSTOMBLEN fertilizer
application was physically or chemically removing both phytane and nC18 together, resulting in no
greater change in the ratios than that seen on Raven beach.
However, for the fertilizer solution-treated beach and the untreated control, where the question
of chemical or solvent effects can be ignored, based on the ratios it appears that extensive
biodegradation of the oil occurred, and that the degradation was significantly increased by the
addition of nutrients. Therefore, it is argued that all observed changes in hydrocarbon composition
discussed above were also due to biodegradation. The relationship of these measures of
biodegradation to the losses of oil residue weight are portrayed in Figures 7.28 to 7.30. The results
are very interesting. Changes in hydrocarbon composition for Raven beach, as modeled by changes
in the summed alkanes, were generally weakly dependent on the loss of oil residue weight (Figure
7.28). Over time there appears to be, on average, much more change in hydrocarbon composition than
in oil residue weight. However, it can be argued that the enhancement of biodegradation by the
application of the fertilizer solution (greater rate of decay in the nC18/phytane ratio and in the
summed alkanes relative to the untreated control), caused concomitant or corresponding decreases in
oil residue weights (Figures 7.29 and 7.30). The results from Tern beach, as seen so many times, fell
in-between the results of Raven and Kittiwake. Thus, the visual removal of oil from the fertilizer
treated-beaches is a direct result of enhanced biodegradation, and it can be confidently assumed that
measurements of changes in oil residue weight are good indicators of the extent of oil biodegradation.
279
-------
% of 7/22
Median
100
90
80
70
60
50
40
30
20
10
\\
5 10 15 20 25 30 35 40 45
Day
Legend
Kittiwake
-9- Tern
-*-~ Raven
Figure 7.27. Change In Phytane, Expressed as Percent of the 7/22/89
Median Over Time for Kittiwake, Raven, and Tern Beaches at
Passage Cove.
280
-------
100
io
8
= 1
8
I
c
E 0.1
0.01
10 100 1,000
Oil Residue Weight (mg/kg)
10,000
oooDtyl
ooo Diy 4
DDaD«y3
Mtdtan
Figure 7.28. Relationship Between Oil Reeldue Weight and the Sum of
Alkane Concentration nC18 to nC27 for Raven Beach
(Untreated Control) at Passage Cove (Cobble Surface).
Values Less than Detection Limit Were Set to Zero. R Value
for Day 1»0.16, Day 2.-0.18, Day 3»0.30, Day 4s0.41.
281
-------
100 \
10
0.1
0.01
A
A
aaaD«y3
10 100 1,000 10,000
Oil Residue Weight (mg/kg)
Figure 7.29. Relationship Between OH Residue Weight and the Sum of
Alkane Concentration nC18 to nC27 for Tern Beach (INIPOL
+ CUSTOMBLEN) at Passage Cove (Cobble Surface). Values
Less than Detection Limit Were Sat to Zero. R Value for Day
1«0.66, Day 2.-0.23, Day 3*0.03, Day 4*0.19.
282
-------
100
10
S
i
00
5
I
c
0.1
0.01
ooo D«y4
aaaD«y3
10 100 1,000 10,000
Oil Residue Weight (mg/Kg)
Figure 7.30. Relationship Between Oil Reeldue Weight and the Sum of
Alkane Concentration nC18 to nC27 for Klttlwake Beach
(Fertilizer Solution) at Paeeage Cove (Cobble Surface).
Values Less than Detection Limit Were Set to Zero. R Value
for Day U0.60, Day 2&-0.31, Day 38-0.05, Day 4&-0.72.
283
-------
/'ASSACK COVE
Internal Markers
Finally, many of the beach samples are currently being analyzed by selective ion mass
spectroscopy to further establish the extent of compositional change. Not enough analyses have been
performed to establish statistically significant trends at this time. However, it is clear that in many
samples changes in the composition of the alkanes were accompanied by substantial changes in other
hydrocarbons. This is illustrated in Figure 7.31.
i a-cie
2 Total Alkanes
3 Phytane
^ ..'4 Phenanthrtn»/Anthracen«-Ci
5 Phananthrtnt/Anthracene-C2
6 Phtnanthrant/Anthrac»nt-C3
7 Dibtnzothiophtne-Cl
8 Dib«nzothioph«ne-C2
9 Dib«nzothiophtne-C3
10 Chrysan*
11 Chrys«ne-Cl
12 Chryi»na-C2
2
2 3 4 J « 7 I • 10 It 12
CHEMICAL NUMBER
HOPANE STRUCTURE
23 24
30
Figure 7.31. Percent Remaining (Hopane Normalized) of different
Hydrocarbons at Klttlwake Beach at Passage Cove on
8/20/89.
284
-------
PASSAGE COVE
The analysis is based on hopane, an internal marker hydrocarbon. As mentioned above, phytane,
the persistent branched alkane, has been used in the past as a stable internal marker. However,
biorcmediation appears to accelerate the biodegradation of this hydrocarbon. Hopanes are pentacyclic
molecules derived geologically from bacteria. Through diagensis of the bacterial kerogen, hopanes
are modified, eventually producing specific isomers that are uniquely characteristic of different oils.
Prudhoe Bay crude oil has significant quantities of 17a,21b-hopane (structure shown in Figure 7.31).
Cyclic alkanes in general are very difficult to biodegrade, and we believe that the multi-ring cyclic
alkanes are almost totally recalcitrant to biodegradation. Thus, 17a,21b-hopane is a good internal
marker.
However, unlike the nC18/phytane ratio, which is based on two hydrocarbons that chemically
behave very similarly, use of hopane ratios has to be carefully considered because of very different
chemical behavior. Most crucial in this regard is the assumption that as oil is degraded and portions
of residue are sloughed off, the hopane will not leave, but instead becomes enriched in the remaining
"biodegraded" oil. All of the samples analyzed to date from Kittiwake beach, where the most oil
degradation was observed, were substantially enriched for hopane. The standard in this case is
Prudhoe Bay crude oil that has been artificially weathered (evaporated at 52 PC to remove 30% of the
original weight) to represent the oil that washed up on the beaches in Prince William Sound after the
spill. Our assumption, therefore, appears to be correct in • very conservative way (i.e., some hopane
is undoubtedly lost as oil residue is removed from the beaches).
Any change in the mass ratio of a specified hydrocarbon to hopane represents differential decay
of hydrocarbons and we assume, based on the recalcitrance of hopane and other aspects of
nonbiological decay, that this can only occur through biodegradation. Accordingly, Figure 7.31 shows
that very extensive biodegradation of the straight chain alkanes is accompanied consistently by
substantial biodegradation of many other hydrocarbon groups. Thus, we can be assured that
fertilizer-enhanced biodegradation is doing more than simply removing the readily degradable
alkanes.
This can be further illustrated by examination of different groups of hydrocarbons in the
residual oil. For example, if the phenanthracene/anthracene group is compared with the
dibenzothiophene group in undegraded oil, it can be shown in the mass spectral analysis that the
different substituted isomers in each grouping relative to hopane, C-l's, C-2's, and C-3's,
respectively, were present in decreasing concentrations. Cl was the highest. As oil is degraded, the
285
-------
PASSAGE COVE
respective concentrations become reversed; the isomer groupings are found with increasing
concentrations. C3 was the highest. This change in degraded oil is shown in Figures 7.32a and 7.32b.
Differential removal of the lower molecular weight materials first (presumably more soluble,
teachable, and volatile), could theoretically be due to either biological or nonbiological processes.
However, because of the substantially different chemical behavior of the benzothiophenes relative
to the phenanthracenes/anthrancenes, as evidenced by their gas chromatographic behavior, it is highly
unlikely that this reverse concentration gradient will occur in the oil sample simultaneously if the
process is due strictly to chemical or physical processes. Biodegradation, on the other hand, which
could involve the simultaneous degradation of these chemical groups by bacteria with different
substrate specificities, is a much more reasonable explanation of simultaneous reversal of these
concentration gradients. Thus, in the samples analyzed to date, changes in alkane composition by
biodegradation were almost always accompanied by simultaneous reversals in the concentration
gradients for the phenanthracene/anthracene and dibenzothiophene groups. These results leave little
doubt that biodegradation of oil is quite comprehensive and simply does not reflect a selective use of
the most easily degradable hydrocarbons.
Assessing the rates of hydrocarbon composition change for cobble samples was also complicated.
Because the concentration of certain alkanes in many samples was zero (below detection limits),
indications of significance were disproportionately affected it the latter sampling dates where fewer
data points were available for analysis (with log transformations, any value set to zero automatically
drops out of the analysis). However, plots of the summed normal alkanes, nC18 to nC27, could be
analyzed statistically to some extent because during the course of the test the total mass of these
hydrocarbons did not fall below detection limits even though individual hydrocarbons within the
summed hydrocarbons did. in essence, this approach creates the second bias of producing artificially
low values for the summed alkanes.
Regression analyses of the median values of the summed alkanes (nC!8 to nC27) plotted on
linear and logarithmic axes are shown in Figures 7.33 and 7.34. Depending on the treatment, it is
clear that a zero-order model provides a better fit to the data in some cases, while a first-order model
is better in others. For example, Kittiwake decay rates appear to be first-order, Raven appears to
be zero-order, and Tern somewhere in-between. In fact, Tern appears to be biphasic; on the
logarithmic display the median of the last sampling date was significantly different from a regression
line drawn through the median for the first three sampling dates. The same is also true for Raven
beach. Thus, some unexplained change in decay rate appears to have occurred between the third and
286
-------
C1-PHENAN-
THERENES
C2-PHENAN- C3
•PHENAN-
THERENES THERENES
••SHU [ V/////////A
10
8
LD
§6
— 4
a
2
o
m
m
' n rta
% %
y/ w
• i i
_. !_• 1
K1 K2
K3
i
—
J
1
i
\
J
R1 R2
J
i
&
%
w
K
%
%
-------
Median
Legend
-•- Kittiwake
-9- Tern
Raven
0 5 10 15 20 25 30 35 40 45
Day
Figure 7.33. Change In the Median Concentration of Summed Alkanee
nC18 to nC27 (Arithmetic Scale) Over Time, for Kittiwake,
Raven, and Tern Beaches at Passage Cove.
Median
Legend
Kittiwake
Tern
Raven
0 5 10 15 20 25 30 35 40 45
Dey
Figure 7.34. Change in the Median Concentration of Summed Alkanes
nC18 to nC27 (Log Scale) Over Time, for Kittiwake, Raven,
and Tern Beaches at Passage Cove.
288
-------
PASSAGE COVE
final samplings on these two beaches. Again, it is quite possible that the effect of INIPOL application
was short-lived (nutrient depletion or degradation of the added carbon).
A comparison of first-order decay constants (slopes of the linear regressions of log transformed
medians) using all sampling dates in one case and excluding the last sampling time in another case,
is shown in Table 7.6. In the former case decay rate constants were not significantly different
between Tern and Raven; both were significantly less than Kittiwake. From these calculations it can
be seen that the enhancement effect of fertilizer solution application on the rate of hydrocarbon
composition change, as measured by the summed alkanes nC18 to nC27, was approximately 1.2 times.
This contrasts dramatically with the enhancement of rates observed with the residue weights (4.7
times). However, if the last sampling date is excluded, the rate enhancement effect of fertilizer
application was 2.1 times for the fertilizer solution. There was no enhancement for the INIPOL/
CUSTOMBLEN application. The differences in rates are a little more consistent with the oil residue
weights.
Mixed Sand and Gravel Samples
Data on hydrocarbon composition for the mixed sand and gravel samples are shown in Figures
7.35 to 7.46. Significant changes occurred only in samples from Kittiwake beach. Less reduction
occurred in the mixed sand and gravel samples than in samples from the cobble surface. Decreases
in the nCI8/phytane ratios were also smaller.
Analysis of the relationship between the summed alkanes nC18 to nC27, and the oil residue
weight for mixed sand and gravel samples shows several interesting trends (Figures 7.47 to 7.49).
Pearson's correlation coefficients were generally higher than those observed in samples from the
cobble surface. Again, there appears to be substantial differences between beaches. Raven beach,
which showed no significant decreases in either oil residue weight or the concentration of the summed
alkanes, had the highest correlation coefficients with each sampling time. If the two outliers on the
first sampling are excluded, the correlation coefficient improves to 0.77, and is then consistent with
the other sampling dates for Raven beach (Figure 7.47). Fertilizer treatment causes the correlations
to be considerably reduced with the greatest effect seen with the fertilizer solution treatment
(Kittiwake). INIPOL/CUSTOMBLEN treatment was in-between. The correlation breaks down
because oil residue weight appears to have decreased, while the concentration of the summed alkanes
remained relatively unchanged. In fact, a relatively strong inverse relationship was observed with the
289
-------
PASSAGE COVE
TABLE 7.6. COMPARISON OF ZERO-ORDER AND FIRST-ORDER RATES
(LINEAR REGRESSION FIT TO MEDIANS) OF CHANGE IN
HYDROCARBON COMPOSITION, BASED ON DECREASES THROUGH TIME IN
HYDROCARBON CONCENTRATIONS OF THE SUMMED ALKANES, nCI8 TO nC27
Fertilizer
Beach Treatment
Kittiwake Fertilizer Solution
Tern INIPOL/CUSTOMBLEN
Raven None
Beaches Compared
KW/R
T/R
KW/T
Slope (standard deviation)
First-Order-I'
-0.032 (.0027)
-0.022 (0.0033)
-0.025 (0.0025)
Ratio
First-Order-I
1.27(A)d
0.85
1.50(A)
First-Order-IIb
-0.033 (0.004)
-0.016(0.005)
-0.016(0.003)
of Slopes (standard
Zero-Order0
-0.19(0.0-6)
-0.11(0.02)
-0.16(0.01)
deviation)
First-Order-II Zero-Order
2.1(A)
1.0
2.1(A)
1.2
0.7
1.8
* 1 - All sampling dates used; units « days"1
6 II - Last sampling date (9/5) excluded; units - day"1
0 All sampling dates used; units - mg/kg/day
d (A) - Statistically different (95% confidence level).
290
-------
5
CC
6
"o
o
-•.
0.1
!>
•* 0.01
co
5
i
c
0.001
22 Jul 89
06 Aug 89 21 Aug 89 05 Sep
Sampling Date
° Anolysis Volue
Medion
Figure 7.35. Change in nC18 Alkane Concentration Through Time for
Raven Beach (Untreated Control) at Passage Cove (Mixed
Sand and Gravel). The Number of Samples Showing
Concentrations Below Detection Limit is Shown Above the
Sampling Date.
100
10
I
c
oo
5
i
c
1
0.1
0.01
o
o
22 Jul 89 06 Aug 89 21 Aug 89
Sampling Date
05 Sep 89
° Anolysia Volue
• Median
Figure 7.36. Change in the Sum of Alkane Concentration nC18 to nC27
Through Time for Raven Beach (Untreated Control) at
Passage Cove (Mixed Sand and Gravel). The Number of
Samples Showing Concentrations Below Detection Limit is
Shown Above the Sampling Date.
291
-------
10 f
0)
6 0.1-
o
0.01-
0.001-
22 Jul 89 06 Aug 89 21 Aug 89
Sampling Date
0 Anolysis Value
05Sep89
Medion
Figure 7.37. Change in Phytane Concentration Through Time for Raven
Beach (Untreated Control) at Passage Cove (Mixed Sand
and Gravel).
10
2?
tr
Q 0.1
"5
E
** 0.01-
00
5
i
c
0.001
§ o
§
° O
§
8
g °
™ O
§
0 °
5 3
22 Jul 89 06 Aug 89
Sampling
I o Anolysis Volue
i
§
8
o
o
9
21 Aug 89
Date
• Medion 1
O
O
0
e
S
7
1 —
OSSepf
Figure 7.38. Change in nC18 Alkane Concentration Through Time for
Tern Beach (INIPOL + CUSTOMBLEN) at Passage Cove
(Mixed Sand and Gravel). The Number of Samples Showing
Concentrations Below Detection Limit is Shown Above the
Sampling Date.
292
-------
ioo r
10
i
c
CO
5
1
1-
0.1
0.01
22 Jul89
° Analysis Value
Median
06Aug89 21Aug89 05 Sep
Sampling Date
Figure 7.39. Change in the Sum of Alkane Concentration nC18 to nC27
Through Time for Tern Beach (INIPOL + CUSTOMBLEN) at
Passage Cove (Mixed Sand and Gravel). The Number of
Samples Showing Concentrations Below Detection Limit
is Shown Above the Sampling Date.
10 4
1
0.1
0.01
0.001
9
o
o
o
T
8
o
o
o
o
22 Jul 89 06 Aug 89 21Aug89
Sampling Date
05 Sep 89
° Anolysis Volue
Median
Figure 7.40. Change in Phytane Concentration Through Time for Tern
Beach (INIPOL + CUSTOMBLEN) at Passage Cove (Mixed
Sand and Gravel). The Number of Samples Showing
Concentrations Below Detection Limit is Shown Above
the Sampling Date.
293
-------
10
o 1-
5 o.i
0
"* 0.01
CO
5
c
0.001-
2
§ 8
0 0
$~~ _J S 8
8 8 1
o
217 12
2 Jul 89 06 Aug 89 21 Aug 89 05 Sep 89
Sampling Date
° Anolysis Value
Median
Figure 7.41. Change in nC18 Alkane Concentration Through Time for
Kittiwake Beach (Fertilizer Solution) at Passage Cove
(Mixed Sand and Gravel). The Number of Samples Showing
Concentrations Below Detection Limit is Shown Above the
Sampling Date.
100 ]
10
co
5
I
0.1
0.01
22 Jul 89 06 Aug 89 21 Aug 89
Sampling Date
05 Sep 89
0 Anolysis Volue
Median
Figure 7.42. Change in the Sum of Alkane Concentration nC18 to nC27
Through Time for Kittiwake Beach (Fertilizer Solution)
at Passage Cove (Mixed Sand and Gravel). The Number of
Samples Showing Concentrations Below Detection Limit is
Shown Above the Sampling Date.
294
-------
10
QC
3
o
0.1
a01
0.001 {
0
8
22 Jul 89 06 Aug 89 21 Aug 89
Sampling Date
05Sep89
° Anolvsis Volue
• Medion
Figure 7.43. Change in Phytane Concentration Through Time for
Kittiwake Beach (Fertilizer Solution) at Passage Cove
(Mixed Sand and Gravel). The Number of Samples Showing
Concentrations Below Detection Limit is Shown Above the
Sampling Date.
o
5
I
c
2\
1.5
1
0.5
0
Z
o
o
o
o
8
B •
a *
0 8
1 2
> Jul 89 06 Aug 89
Sampling
° Analysis Volue
O
i 0
8 0
8 o
a 8
1 o
4 3
21 Aug 89 05 Sep 89
Date
• Median ]
Figure 7.44. Change in the nC18/phytane Ratio Through Time for Raven
Beach (Untreated Control) at Passage Cove (Mixed Sand and
Gravel). The Number of Samples Showing Concentrations
Below Detection Limit is Shown Above the Sampling Date.
295
-------
22 Jul89
06 Aug 89 21 Aug 89
Sampling Date
05Sep89
° Anolysis Volue
Medion
Figure 7.45. Change in the nC18/phytane Ratio Over Time for Tern Beach
(INIPOL + CUSTOMBLEN) at Passage Cove (Mixed Sand and
Gravel). The Number of Samples Showing Concentrations
Below Detection Limit is Shown Above the Sampling Date.
1.5
oo 0.5 \
5
I
c
o
o
22Jul89 06Aug89 21 Aug 89
Sampling Date
0 Anolysis Volue
Median
12
05Sep89
Figure 7.46. Change in the nC18/phytane Ratio Over Time for Kittiwake
Beach (Fertilizer Solution) at Passage Cove (Mixed Sand and
Gravel). The Number of Samples Showing Concentrations
Below Detection Limit is Shown Above the Sampling Date.
296
-------
100
10
0.1
0.01
10 100 1,000 10,000
Oil RMldu* Weight (mg/kg)
oooD«y1
ooo Oiy4
oaaDiy3
Figure 7.47. Relationship Between the Sum of Alkane Concentration
nC18 to nC27 and Oil Residue Weight for Raven Beach
(Untreated Control) at Passage Cove (Mixed Sand and
Gravel). Values less than Detection Limit were set to Zero.
R Value Day 1s-0.11, Day 2=0.86, Day 3=0.61, Day 4=0.63.
297
-------
100 f
10
0.01
10 100 1,000 10,000
Oil Residue Weight (mg/kg)
oooD«y1
ooo Dcy 4
Figure 7.48. Relationship Between the Sum of Alkane Concentration
nC18 to nC27 and OH Residue Weight for Tern Beach
(INIPOL 4- CUSTOMBLEN) at Passage Cove (Mixed Sand and
Gravel). Values less than Detection Limit were set to Zero.
R Value Day 1*0.28, Day 2=0.62, Day 3=0.57, Day 4=0.22.
298
-------
100
1 10
8
c 1.
00
5
I
c
E 0.1
0.01
o A
a
\J|^ o
a^^i A
G& a
° *
-------
PASS ACE COVE
last two samplings on K ittiwake. Similar responses were also seen in samples from the cobble surface,
and the most pronounced effect was always seen on K ittiwake beach. It is possible that this is an
artifact of scaling the results to reflect the sample size (samples were very small), and the effect of
samples falling below detection limits.
Thus, fertilizer treatment of oil in the mixed sand and gravel under the cobble caused oil residue
to be lost without large decreases in the readily biodegradable alkanes. The only reasonable
explanation to account for the results is an effect due directly to biodegradation. Physical scouring
in the mixed sand and Rravel should be greatly reduced, and chemical washing can be eliminated as
a factor because the effect was greatest with the fertilizer solution application, which was essentially
sea water.
Significant decay of the summed alkanes occurred only on Kittiwake beach. The decay rate
from the slope of a first-order plot was -0.0072 per day, or approximately 4 times less than what was
observed on the cobble surface. This contrasts with the oil residue weight comparisons in which
decay rates in mixed sand and gravel and cobble surface samples were the same. Thus, as noted
above, significant changes in oil residue weights in mixed sand and gravel samples from the fertilizer
solution-treated beach were accompanied by considerably less change in oil composition than oil
samples from the cobble surface. We currently have no explanation for this difference.
WINTER 1989/1990 SAMPLING
On November 19, 1989 the beaches at Passage Cove were sampled a fifth time. Despite winter
conditions, it was apparent that there was very little oil remaining on the cobblestone surface of the
two treated beaches (Tern and Kittiwake). The untreated control beach. Raven, was still visually
oiled, although the oil was very patchy. The amount of oil on the surface, however, was much less
than the previous sampling in early September. In the mixed sand and gravel below the cobble, it was
very clear that many samples taken on the treated beaches were visually void of oil. This was not the
case on the untreated control beach. Oil on the treated beaches appeared to be reduced to
approximately the same extent; there was no visual evidence that one fertilizer worked better than
another. Samples from the beach subsurface (approximately 15 to 18 cm) generally showed the same
pattern; patchy oil distribution, but visually much less oil on the two treated beaches as compared to
the untreated control.
300
-------
PASSAGE COVE
Concentrations of oil residue, nCI8, nC27, the summed alkanes nC18 to nC27, phytane. and
differences in the nC18/phytane ratio are shown in Figures 7.50 to 7.61. Sampling was conducted in
the same plots used during the summer sampling. Due to short winter days and windows of good
weather, sampling was performed only in every other block starting at the edge of each plot. Samples
were taken from the upper (blocks 1 to 7) and lower (blocks IS to 21) tide zone. The data were not
combined because of significant differences between the two areas (see below); this spatial effect was
not apparent in the summer sampling.
A general overview of all the data clearly portrays the long-term effects of enhancing oil
biodegradation by the addition of fertilizers. In most cases, the sampling showed significantly less
oil residue and more change in hydrocarbon composition on the treated beaches than on the untreated
control beach. Where the statistical significance of INIPOL/CUSTOMBLEN enhancement of oil
biodegradation was difficult to establish during the summer sampling, there is little question, based
on the winter sampling results, that the enhancement effect was real.
Cobble Surface Samples
The amount of oil remaining on the cobble surface was generally very low, showing
concentrations approximately ten fold less than concentrations at the time of initial fertilizer
application. This remaining material probably represents the fraction of oil that is slowest to
biodegrade. The distribution of the remaining oil varied between the two sampling zones(Figures 7.50
and 7.56); in the low tide zone oil concentrations on all beaches were approximately the same (Figure
7.50), regardless of fertilizer application. This may reflect less resolution of differences because of
the relatively small number of samples taken, or it could be a true spacial effect, possibly related to
greater wave energy or more consistent coverage by tidal waters on the lower part of the beach.
Based on the half-lives calculated for the decay rate of oil on each beach (Table 7.2), the degree
of reduction in oil concentration seen in the winter sampling was less than would have been predicted,
suggesting that the winter conditions slowed the decay rates.
Changes in hydrocarbon composition were best interpreted by examination of the summed
alkanes, nC18 to nC27, and phytane; concentrations of nC18 and nC27 were not as useful because
many samples had concentrations below detection limits. In general, the summed alkanes and phytane
were reduced to the greatest extent in the low tide zone of the beaches (Figures 7.53 and 7.54). It was
301
-------
10
s
0001
o
a °
a o
o
8 S •
§ °
o
0 °
g
8 °
o
• 0
080
o 8
g
1
2 o
n •
8
O
^ @
y
o o
o
ftowt-C *n-C Moto-C RMn-MSQ *m-M8Q KMM-M8Q RMn-88 Tmi-tS KMto-SS
Beech and Sample Depth Identification
1 ° Anolysis Vol • Median
C=Cobble MSG=Mixed Sand and Gravel SS-Subsurface
Figure 7.50. Oil Residue Weights (Log Scale) for the Low Tide Zone at
Passage Cove During the Winter of 1989.
I
•5
_
"9 1"
1
UH
5
•s
"^ 0.1-
•2-
03
5
i
c
0.01
O °
O
O °
O
564
O
0
O
8
•
O
o
° 8
8 0
o
1 1
0
O 0
o •
•
0
o
o
0
o
0
1 3
RMV-C *m-C Kta*«-C RMn-M3G *m-MBO KMM-MBQ RMn-8S 'km-SS «•«•-•
Beach and Sarnpte Depth Wentifteaflon
[ ° Analysis Vol • Median
C=Cobble MSG=Mixed Sand and Gravel SS=Subsurface
Figure 7.51. nC18 Concentration (Log Scale) for the Low Tide Zone at
Passage Cove During the Winter of 1989.
302
-------
10
f
5
•8
j«.
8
c
0.01
777
0
0
o
•
o
3 4 7
0
° 0
8
o
0
2 2 4
HMn-C •ftm-C NMM-C NMn-MH *m-IW KMto-MH NMn-H *m-M «•«•-•
Batch *id Svnpb Depth IdentHlotfon
0 AnolviiiVol • Mtdian
C-Cobble MSG-Mixed Sand and Gravel SS-Subsurtace
Figure 7.52. nC27 Concentration (Log Scale) for the Low Tide Zone at
Passage Cove During the Winter of 1989.
8
i
c
100
10 -
1
0.1
OJH
0 8
o
i .
0
2 6 1
0
8
0
0 0
8 • °
0
A
o •
0
o
2
8
8
o
o
8
™
0 °
0
1 2
1 1 1 — '
BMCh md Simpto Dipih
M «m-
0 AfiQlyiliVol
• Mtdian
C-Cobble MSQ-Mixed Sand and Gravel SS-Subsurface
Figure 7.53. Sum of Alkane Concentration nC18 to nC27 (Log Scale) for
the Low Tide Zone at Passage Cove During the Winter of
1989.
303
-------
5
75
0.1
0
0 g
0 Q
8 o 1
o
8
4
0
0
O
0
1
0 8
0 f
I 8
1
1 1
0 0
o
o
s
0
2
i-C
Vn-C MaM-C
Bnoh md Svrpto Depth Wentftefton
0 AnolvtliVQl
• Mtdiqn
C-Cobble MSG-Mlxed Sand and Qravel SS-Subsurface
Figure 7.54. Phytane Concentration (Log Scale) for the Low Tide Zone at
Paeeage Cove During the Winter of 1989.
8
c
10
1-
0.1
0.01
0 Q
o °.
Q
0
564
0 0
o
8
• o 9
0 •
o
e
1 3
8 Q
1 O
o 8
0 •
1 3
1 1 1 1 I 1 III
^KM» f* ^MH r* " — *— f* ite^M_ufln '^•t ftjvs H^^te^uen I^^M^AA ^•—•B K^^^.M
^•Mnv^ IVT)^^ HHBV v ^Wl •••!• IVfl^HVII NH^^P MVI0 mMn^OT IVrl^OT RHHV •
Beech end Sempto Depth Wentfteaflon
1 ° AnalviiiVol ' Mtdion
C-Cobble MSQ-Mixed Sand and Qravel SS-Subsurtace
Figure 7.55. nC18/phytane Ratio (Log Scale) for the Low Tide Zone at
Passage Cove During the Winter of 1989.
304
-------
w-T
8
I
0.1
aoi
o
8
o
0
8
8
8
I
0,001
-c
«n-C MM-C
Bert and Swpto D«plh
0 Anolvtii Val
C-Cobble MSG-Mixed Sand and Qravel SS-Subsurface
Figure 7.56. Oil Residue Weights (Log Scale) for the High Tide Zone at
Passage Cove During the Winter of 1989.
1-
0.1
I
c
o
o
o
-c
o
o
0
0
o
0
0
*n-c NMM-C
Bwoh md Svnpto Dtplh Idmtffloatfon
0 Anglviii Vql
• Midlqn
C-Cobble MSQ-Mlxed Sand and Qravel SB-Subsurface
Figure 7.57. nC18 Concentration (Log Scale) for the High Tide Zone at
Passage Cove During the Winter of 1989.
305
-------
3
•8
ai
a
c
am
8
0
-c
n-C
Beach md Sample Depth Identification
0 AnolyiiiVol
• Midlgn
C-Cobble MSG-Mixed Sand and Gravel SS-Subsurface
Figure 7.58. nC27 Concentration (Log Scale) for the High Tide Zone at
Passage Cove During the Winter of 1989.
I
c
T5
ai
0
•
-c
8
§
o
o
Btach rd Sampta D«pth
0 AnalYili VQ|
Mtdlon
C-Cobble MSG-Mlxed Sand and Gravel SS-Subsurface
Figure 7.59. Sum of Alkane Concentration nC18 to nC27 (Log Scale) for
the High Tide Zone at Passage Cove During the Winter of
1989.
306
-------
8
•8
0.1
0.01
0
o
i-C
0
0
8
0
0
8
8
o
8
•
o
0
1
tm-C MoM-C towi-MM *n-fc
Beach and Sample Depth Identification
1° Anoly»i» Vql
Madion
C-Cobble MSG-Mixed Sand and Gravel SS-Subsurface
Figure 7.60. Phytane Concentration (Log Scale) for the High Tide Zone at
Passage Cove During the Winter of 1989.
0.1
001
0
9
i-C
o
0
o
*n-C MM-C
Bwch md Sample Depth Identification
0 Anolvtii Vol
• Mtdign
8
C-Cobble MSG-Mlxed Sand and Gravel SS-Subsurface
Figure 7.61. nC18/phytane Ratio (Log Scale) for the High Tide Zone at
Passage Cove During the Winter of 1989.
307
-------
PASSAGE COVE
unexpected that in this area of the beach many samples from Tern had values below detection limits;
this suggests that 1NIPOL/CUSTOMBLEN may have had a greater long-term effect than the fertilizer
solution application. For samples taken from the high tide zone, the effect of fertilizer was very
pronounced. Significantly greater reductions at the 95% confidence level in the summed alkanes
(Figure 7.59) and phytane (Figure 7.60) were apparent as a result of fertilizer application. In
addition, the effect of the fertilizer solutions was greater than the effects of the
INIPOI./CUSTOMBLEN. However, as noted above, there were many cases where samples from the
INIPOL/CUSTOMBLEN-treated beach had concentrations below detection limits and, thus, over
time, the 1N1POL/CUSTOMBLEN fertilizer may have been as effective as the fertilizer solution.
Mixed Sand and Gravel Samples
Examination of the oil concentration in the mixed sand and gravel below the cobble clearly
depicts the long-term effects of fertilizer application and enhanced oil biodegradation. At the end
of the summer sampling, only the fertilizer solution application had significantly affected
concentrations of oil residue in the mixed sand and gravel. It is clear from the winter data that the
1N1POL/CUSTOMBLEN application was also exerting an effect, but to a lesser extent (Figure 7.53).
Therefore, the effect was only visible after an extended incubation period. This contrasts
dramatically with the observed residual oil concentrations on Raven beach, which were only slightly
different from those observed at the initial sampling on July 22, 1989. The importance of enhancing
oil biodegradation by fertilizer addition before the winter conditions slowed biological activity is
clearly illustrated in these results.
These trends occurred despite residue weights in the low tide zone of the beach (Figure 7.50)
that were approximately 10 times higher than those recorded in the high tide zone (Figure 7.56). This
difference in oil residue weights may have been due to consistently colder temperatures resulting from
more coverage by tidal waters over time in the low tide zone. Otherwise, there is no reasonable
explanation for the difference.
Changes in hydrocarbon composition are again best expressed in terms of the summed alkanes
(Figure 7.53) and phytane (Figure 7.54). The most pronounced effects of the fertilizers are shown
in samples taken from the low tide zone. Again, the greatest compositional change occurred on the
treated beaches with Kittiwake consistently demonstrating greater change than Tern beach. This
trend is also reflected in the nCI8 alkane as well in Figure 7.51 (fewer samples below detection
308
-------
PASSAGE COVE
limits). The trend is not as evident in samples from the high tide zone; alkane concentrations were
not significantly different between Kittiwake and Raven beaches. However, all of the samples with
hydrocarbon concentrations below the detection limits were from the fertilizer-treated beaches
(particularly the 1NIPOL/CUSTOMBLEN treatment), and, thus, the absence of these data points (set
to zero) in the statistical analysis creates a bias that is conservative in terms of fertilizer effect.
Subsurface Samples
Beach samples taken at a depth of IS to 18 cm below the beach surface provide the first
evidence that fertilizer-enhanced biodegradation was also occurring in the beach subsurface.
Unfortunately, only the low tide zone was sampled. The results clearly show that penetration of
nutrients to the subsurface and the concomitant enhancement of oil biodegradation was most effective
with fertilizer solution application. All measurements shown in Figures 7.50 to 7.55 (oil residue
weight, individual alkanes, summed alkanes, phytane, and the nCl 8/phytane ratio), were significantly
different at the 95% confidence level from the untreated control. The application of
IN1POL/CUSTOMBLEN fertilizer did not appear to have an effect at this time. It is therefore
concluded that oil in the subsurface was adequately colonized by oil-degrading microorganisms, but
the oil biodegradation was nutrient limited. The ineffectiveness of the INIPOL/CUSTOMBLEN
application was probably the result of insufficient total nutrient release; in other words, there were
only enough nutrients provided in a single application to significantly affect oil biodegradation below
the beach surface.
SPRING 1990 SAMPLING
Following the field demonstration during the summer and late fall, 1989, subsequent sampling
was conducted in the early summer, 1990. Oil residue weight results from these samplings are shown
in Figure 7.62.
On June 12, 1990, visual examination of the Passage Cove beaches showed there was no oil on
the cobble surface or in the mixed sand and gravel below the cobble on any beach, including the
untreated control beach. However, after excavating many holes to a depth of approximately 15 to 18
cm below the beach surface throughout the high, mid, and low tide zones, it was visually obvious that
only the untreated control beach, Raven, still contained oil. There were no visual quantities of oil
on the two treated beaches. Although many areas of Raven beach had become free of oil in the
309
-------
o
1(00
1600
1400
1200
1000
MO
WO
400
200
0
Q
-0-
D
a
-B-
Klttlwake
Ttm
Riven
TREATMENT
Figure 7.62. OH Residue Weight In the Subsurface for the Three
Treatments In the Spring of 1990.
310
-------
PASSAGE COVE
subsurface over the winter, 5 to 6 large pockets of oil-contaminated beach material remained. It is
possible that these areas were those heavily contaminated with oil the previous summer.
Determination of oil residue weight in samples from the June 12, 1990 sampling verified that
subsurface samples on Kittiwake and Tern beaches contained essentially no oil (Figure 7.62).
Subsamples on Raven, where oil was observed, had significant concentrations of oil, some as high as
the average starting concentration the previous summer.
Revisiting Passage Cove in August, 1990, showed that very little oil could be found on Raven
beach in the subsurface. Thus, it required essentially another summer of natural biodegradation
activity to adequately remove all oil contamination from Raven beach. Fertilizer application probably
accomplished this end point in the very early spring, if not sooner. It is therefore believed that by
general extrapolation bioremediation, as a secondary cleanup process, potentially reduced beach
cleanup time by as much as 50%.
MICROBIOLOGY
Numbers of Oil-Degrading Bacteria
The number of oil-degrading bacteria present on beach materials was also determined for
Passage Cove. Samples of beach material were taken from grids 1, 3, 5, 7, 8, 10, 12, 14, 15, 17, 19,
and 21. Numbers of degraders were assessed by using modification of the dilution to extinction
method used for Snug Harbor. Five replicate dilution series were prepared from the initial 1:10
dilution. The relative numbers of bacteria in each sample were an average of the five replicate
dilution series.
Results from these studies are shown in Table 7.7. The values reported are the logj,, normal
mean and standard deviation of 11-12 dilution series for each mixed sand and gravel sample. Results
suggested no consistent increase in oil-degrading microorganisms occurred as a result of fertilizer
application. This means that even in the plot treated with nutrient solutions from a sprinkler system,
where nutrient exposure to the bacteria should be optimized, an increase in oil-degrading
microorganisms did not occur. This could be due to a relatively constant sloughing of microbial
biomass from the surfaces of the beach material, perhaps caused by tidal flushing action. Grazing
by protozoans could also keep the microbial numbers at a specific density. The presence of high
311
-------
PASSAGE COVE
numbers of oil-degrading microorganisms was much greater than that observed for unimpacted
beaches (Table 7.8). It is clear that the number of oil degraders in uncontaminated areas was 1,000
to 100,000 times lower than in contaminated areas. Thus, the presence of oil causes a significant
enrichment of oil-degrading microorganisms.
TABLE 7.7. RELATIVE CONCENTRATION (LOG,« OF THE CELL NUMBER/G OF BEACH
MATERIAL) OF OIL-DEGRADING MICROORGANISMS IN PASSAGE COVE
Sampling Date Fertilizer-Treated Plots
Untreated Water- Oleophilic +
Before Application Control Soluble Water-Soluble
07/22/89 6.44 6.31 6.44
±1.44 ±1.36 ±1.33
After Application
08/06/89
08/19/89
5.32
±1.12
6.60
±1.83
5.78
±1.45
5.47
±1.34
5.71
±0.67
5.66
±0.35
TABLE 7.8. RELATIVE CONCENTRATION (LOG10 OF THE CELL NUMBERS/G
OF BEACH MATERIAL AND STANDARD DEVIATION) OF OIL-DEGRADING
MICROORGANISMS IN SAMPLES FROM BEACHES THAT WERE NOT IMPACTED BY OIL.
Site
Tatitlek
Fish Bag
Snug Corner Cove
Hell's Hole
Commander Cove
High Tide
2.41
±.58
(1.51
2.31
±.54
(2.11
4.51
±1.14
Mid Tide
4.31
±1.14
(1.31
2.51
±.55
2.51
±.89
(1.31
Low Tide
6.11
±2.05
(2.71
(1.11
(.91
3.11
±.45
312
-------
PASSAGE COVE
Microbial Activity
Oiled Venus Unolled Beaches
To compare oiled with unoiled beaches, samples were collected from Raven beach, an oiled
beach that served as a control in the Passage Cove field test, and from a beach in Hell's Hole, an area
not impacted by the oil spill. The activity of bacterial populations was assessed by analyzing the
evolution of 14CO2 from radiolabeled hexadecane, naphthalene, and phenanthrene. Each compound
was incubated in 50 mL of artificial sea water medium with 5.0 g of sample for periods up to 120
hours.
Hexadecane was mineralized at the same rate by both samples (Figure 7.63). Naphthalene was
also mineralized in both samples but at a lower rate in samples from Hell's Hole. Phenanthrene was
mineralized only in the samples from Raven beach. Degradative activity, indicative of the presence
of bacteria capable of using the compound as substrate, was clearly present at Raven. Activity in
samples from the unimpacted site may indicate either that the site had a prior history of oil exposure
(perhaps from oil navigational traffic) that resulted in an acclimated bacterial population, or that
unexposed microbial communities may have broad enough metabolic capabilities to utilize oil
hydrocarbons without a lengthy acclimation period. If the former were the case, all of the compounds
should have been degraded at both sites.
Given that hexadecane appears to be more readily degradable than naphthalene or phenanthrene,
it is likely that indigenous bacteria can utilize straight chain hydrocarbons more readily than
aromatics. Both phenanthrene and hexadecane were used as test compounds in subsequent
experiments, but phenanthrene may be a better indicator of adapted populations.
Passage Cove
For the Passage Cove tests only mineralization of radiolabeled hexadecane and phenanthrene was
measured in material from the test beaches (Kittiwake, Tern, Raven) three days prior to the
application of the fertilizer treatments, and four and six weeks after application. Comparison of tidal
regions for pre-application samples had demonstrated that the mid-tide zone was representative of
mean mineralization activity for a beach. Therefore, in post-application samples it was assumed that
this would also be true. Due to the lengthy incubation times needed to determine mineralization rates
313
-------
25
20
15
10
5
0
40 60
HOURS
80
Figure 7.63. Mineralization of Radlolabeled Phenanthrene (*,o),
Naphthalene (D,B), and Hexadecane (A,*), In Samples From
Oiled (Open Symbols) and Unolled (Closed Symbols)
Beaches.
314
-------
PASSAGE COVE
and the limited number of samples that could be processed daily, it was felt that mineralization
measured for several time points using a few samples was more accurate than using single time points
for numerous samples. The concentration of radiolabeled substrate incubated with beach material was
therefore reduced, reflecting the decrease in the amount of these compounds present in the oil, as
determined by gas chromatography.
In the pre-fertilizer application samples, mineralization commenced before the first sampling
time (20 hours), indicating little or no lag period. The mineralization curves were S-shaped, which
is typical of a growing population. A representative curve is shown in Figure 7.64. No significant
differences in the rates of mineralization were detected among the three beaches (Table 7.9). In
addition, the extent of mineralization was somewhat lower in the high tide zone.
TABLE 7.9. MINERALIZATION OF 10 ftG "C-PHENANTHRENE PER G PASSAGE COVE
BEACH MATERIAL PRIOR TO APPLICATION OF FERTILIZER (7/22/89)
Sample
Treatment* Zone
Water-
Soluble
(KW)
Untreated
Control
(R)
Oleophilic
(T)
Low
Mid
High
Low
Mid
High
Low
Mid
High
Lag Timeb Max Rate
(h) %/g/h ng/g/h
42
18
37
±5° 1.0 ±0.2 103
±9 2.1 ±0.4 205
±7 2.3 ± 0.5 230
NSd
11 ±1
18±7
16
19
19
± 8
±17
±10
±16
±41
±50
NS NS
.6 ±0.1 160 ± 7
.7 ±0.2 169 ±21
.8 ±0.4 183
.5 ±0.8 147
.9 ±0.9 186
±40
±75
±91
Extent
%
42
50
40
±8
±2
±5
NS
50 ±4
44 ±6
45
31
39
±9
±3
±5
*KW - Kittiwake Beach; R - Raven Beach; T - Tern Beach
'Time required for 10% mineralization of phenanthrene
cMean and standard deviation of four samples
dNo sample analyzed
315
-------
60"
50H
S 4°
* 30-
20-
10
0
20
High Tide Zone 7/22/89
40
HOURS
60
80
Legend
z
2
60-
50H
40
30
20-
10
0
Low Tide Zone 7/22^9
20
40
HOURS
60
80
Legend
z
3
3?
60
50-
40-
30
20-
10
0
20
High Tide Zone 7/22/89
40
HOURS
60
80
Legend
BL-Block
Flgura 7.64. Mlnarallzatlon of Phananthrana In Samplaa From
(A) Watar-Solubia Fartlllzar-Traatad; (B) Untraatad Control;
and (C) Olaophlllc Baachaa at Paaaaga Cova.
316
-------
PASSAGE COVE
At four and six weeks following application of the fertilizer, the rate of mineralization was
lower than the pre-application rate when expressed on a mass basis (Table 7.10), but similar when
expressed as percentage rate. This suggested that the mass transfer of the compound may be rate-
limiting. At four weeks post-application, the mineralization of phenanthrene on the untreated control
beach (Raven) was greater than on the treated beaches (Kittiwake and Tern). No difference was
detected between the water-soluble and oleophilic beach. By six weeks, the previously noted
difference between the treated beaches and the untreated control beach had disappeared.
TABLE 7.10. MINERALIZATION OF 0.32 pG OF 14C-PHENANTHRENE AND 0.44 pG
14C-HEXADECANE PER G PASSAGE COVE BEACH MATERIAL
No. of Weeks Cmpd Lag Max Rate
Post Application* (h) %/g/l ng/g/h
Water-Soluble (K.W1 Treatment**
4(8/22/89) Hex > 50
4(8/22/89) Phe > 50
6(9/5/89) Hex >100
6(9/5/89) Phe >100
Untreated Control (R)
1.9 ±0.8 8.4 ± 3.4
0.8 ±0.2 2.8 ±0.6
1.4 ±0.4 4.4 ±1.3
2.4 ±0.8 7.6 ± 2.4
4 (8/22/89)
4 (8/22/89)
6 (9/5/89)
6 (9/5/89)
Oleoohilic (J)
4 (8/22/89)
4 (8/22/89)
6 (9/5/89)
6 (9/5/89)
Hex
Phe
Hex
Phe
Hex
Phe
Hex
Phe
39 ± 1
42 ±10
>100
>100
>50
>50
>100
>100
5.6 ±0.1
6.4 ± 0.5
1.4 ±0.4
2.3 ± 0.9
1.8 ±0.4
2.6 ± 0.9
1.8 ±0.3
2.3 ± 0.6
24.6 ± 4.4
21.2 ±1.6
4.4 ±1.3
7.5 ±2.8
8.0 ± 1.8
8.6 ± 3.0
5.7 ±0.9
7.8 ± 2.0
*KW • Kittiwake Beach; R - Raven Beach; T • Tern Beach
bFour blocks from mid-tide zone
317
-------
PASSAGE COVE
The observed differences between the extents of mineralization for the pre- and post-fertilizer
application samples can not be readily explained. The abundance of bacteria in the water column at
Passage Cove declined during this interval, although no differences in population size were reflected
in the numbers of hydrocarbon degraders in the beach material. However, not all of the hydrocarbon
degraders are necessarily phenanthrene degraders.
A radiolabeled MPN technique indicated that the number of phenanthrene degraders was
4.95±0.30 x IOB, which is less than the total number of hydrocarbon degraders (approximately 106 per
gram of beach material). It is possible that as the more labile components of the oil mixture on the
beaches were degraded, the fraction of the bacterial population adapted to the utilization of those
compounds declined, and thus there was less mineralization activity with time. An alternative
explanation is that the bacteria may utilize the compound at different rates or with different
efficiencies at different concentrations. Either cometabolism or a high percentage of incorporation
into cell material might explain the relatively small fraction of labeled compound converted to 14CO2
at the lower concentration.
Regardless of the reason for the differences in the rate and extent of mineralization before and
after fertilizer application at Passage Cove, it can be concluded that organisms capable of degrading
hexadecane and phenanthrene existed on the beach prior to treatment, and that the application of
either fertilizer treatment did not significantly alter the rate and extent of mineralization compared
to the untreated control beach.
Mineralization in Microcosms
Microcosms were designed to test the efficacy of bioremediation for the treatment of subsurface
oil. Clean beach material from an unaffected beach at Hell's Hole and oiled beach material from
Raven beach at Passage Cove was used to construct the microcosms. Containers were filled by either
layering oily material over clean material or vice versa. The material from Hell's Hole was tested
prior to the construction of the microcosms and no mineralization activity was detected.
Each set of microcosms was treated with water-soluble fertilizer, oleophilic fertilizer (INIPOL),
or no fertilizer (untreated controls). All of the microcosms were flushed with sea water in accordance
with the tidal cycles. Twenty-four microcosms were constructed, allowing duplicate microcosms to
be sacrificed on two sampling dates. During sampling, the contents of each microcosm was
318
-------
PASSAGE COVE
fractionated into thirds, giving clean, oily, and interface layers. Subsamples were collected from the
well-mixed material derived from each layer and were analyzed for numbers of oil-degrading
organisms, chemical composition of the oil, and bacterial activity. Bacterial activity was measured
as mineralization of phenanthrene and hexadecane.
The results for both compounds were similar, so only the phenanthrene data is presented.
Mineralization of phenanthrene was measured for incubation periods of approximately 120 hours.
The compound was 5 to 20 percent mineralized in all samples except for the top layer in the
oleophilic-treated microcosms where clean material overlay oily material (Figure 7.65). No
mineralization was detected in these samples, and this was true for the replicates and both sampling
times. Bacterial activity is vertically translocated either up or down from oily to clean material for
all treatments except for the case noted above. The oleophilic fertilizer would be a possible treatment
for a beach only containing subsurface oil because fertilizer will enhance activity in the subsurface
if nutrients filter down into the subsurface material. When oleophilic fertilizer is applied to an oily
surface, however, bacterial activity is evident throughout the beach material.
Biodegradation rates, as determined by mineralization of radiolabeled substrates in natural
samples, are subject to interpretation, especially in light of the amount of residual hydrocarbon in the
sample and the processes that affect the rate of availability of the compound to bacteria. The rate of
mineralization of the labeled substrate is proportional to the total rate of mineralization of all the
substrates utilized by a given population or metabolic pathway. Therefore, caution must be exercised
in extrapolating such rates to unknown concentrations of compounds and to periods of time over
which both concentrations and bacterial populations may change. Similarly, radiolabeled compound
added to a sample is almost certainly more available to bacteria than an in situ compound which may
be limited by insolubility or the presence of a complex matrix. On the other hand, the absence of
mineralization of labeled substrate is a strong indication of the absence of acclimated bacteria.
Mineralization of radiolabeled phenanthrene and hexadecane in samples of beach material
demonstrate the activity of acclimated populations of hydrocarbon-degrading bacteria prior to the
application of fertilizers. Rates of mineralization were lower after fertilizer application, although the
decrease of rates at the control and treated beaches suggests this was a general phenomenon instead
of a response to fertilizer. The depletion of labile substrate and the attendant shift in bacterial activity
over time would explain both the overall decrease in rates and the lower rates observed for treated
beaches one month after fertilizer application. Thus, a lower instantaneous rate may reflect a higher
activity in the preceding interval.
319
-------
z
2
10
O Tub 1 Top
A Tub 2 Bottom
x Tub 2 Middle
Figure 7.65. Mineralization of Phenanthrene In Oleophilic-Treated
Microcosm with Oiled Layer Over Clean Layer, Showing
Lack of Bacterial Activity In Upper Layer.
320
-------
PASSAGE COVE
ECOLOGICAL MONITORING
The same environmental parameters monitored at Snug Harbor were monitored at Passage Cove,
but a somewhat modified strategy was used for sample site location. Sample stations were located
along the central axis of the embayment and along three nearshore areas where fertilizers were applied
(see Sections 3 and 8). Reference sites for the Passage Cove study were established outside the
embayment along the eastern shore of northern Knight Island. Water from the central sites of Passage
Cove was sampled at 0.5 m and 5 m depths, while the nearshore stations (1 m offshore of low tide)
were sampled at 0.5 m depths. Fertilizers were applied on July 25 and 26, 1989, to selected plots
along the shoreline. Samples were collected prior to application of fertilizer along the shoreline, 3
days after application, and then at weekly intervals for 6 weeks after application.
Nutrients
Data from analyses of water samples from Passage Cove for ammonia, nitrite, nitrate, and
phosphorus in water off the test beaches were unfortunately lost.
Chlorophyll Analysis
Phytoplankton chlorophyll data showed a slight increasing trend over the course of the study
period for all treated and control stations (Figure 7.66). No trends consistent with nutrient effects
were observed. The differences through time were not significant except for an increase observed
on August 27 in the 0.5 m sample from all mid-channel stations and the reference site. These results
are consistent with expected range of plankton chlorophyll for Prince William Sound and seasonal
increases leading to a fall plankton bloom.
Phytoplankton Primary Productivity
Results from the pre-treatment sample (July 21), Day 3 (July 28), and Weeks 1 (July 31) and 2
(August 7) are presented in Figure 7.67. These data, and results from other dates not presented, show
no trends for Passage Cove stations with greater primary productivity as a result of nutrient additions,
except for the results from July 31. Primary productivity estimates on this date show greater values
for stations 5, 6, and 7, the nearshore stations along the treated shoreline. This trend was not
observed one week later, nor was it borne out in the chlorophyll data. If primary productivity was
321
-------
w
_J
_l
I
g
7-81
7-H 7-31 *-7 *-14 Ml
M7
• 1.0 meters
D 10.0 mttera
741
7-t«
STATION 4
741 7-M 741 »-7 «-U Ml M7 »-4
SAMPLE DATES
Figure 7.66. Mean Chlorophyll a/L Measurements (+ SD) from 4 Replicate
Plankton Samples Taken at Passage Cove Study Site
Following July 25,1989, Fertilizer Applications to Shorelines.
Values are Means (+ SD) of 4 Replicates; Dark Bars for
Samples Collected at 0.5 m, Open Bars for Samples
Collected at 5 m. Refer to Figure 3.3 for Sample Locations.
322
-------
741
741 7-J1
741 1-tt 7-11 t-7 M4 141 M7 *-4
§
741 74* 741 (-7 t-14 Ml *47
SAMPLE DATES
Figure 7.66. (Continued)
323
-------
Legend
STA1-
STA2-
STA3-
STA4-
STA5-
STA6-
STA7-
Station 1
Station 2
Station 3
Station 4
Station 5
Station 6
Station 7
s.t
10
4.1
4.0
1.S
10
II
to
1.1
1.0
Figure 7.67. Mean Primary Productivity Measurements (+ SD) From
"C-Uptake (mg C/m'/hour) From 4 Replicate Plankton
Samples Taken at Passage Cove Study Site Following July
25,1989, Fertilizer Applications to Shorelines. Refer
to Figure 3.3 For Sample Locations.
32A
-------
PASSAGE COVE
enhanced along the shoreline because of nutrient input, the effect on plankton growth was not
sufficient to overcome dilution and transport that resulted from tidal exchange (i.e., there was no
concomitant increase in plankton chlorophyll).
Bacterial Abundance
The mean number of bacterial cells per liter of water at Passage Cove sample sites ranged from
approximately 0.4 to 1.2 x 109 over the seven-week sample period (Figure 7.68). All stations followed
the same general pattern of greater numbers on the first three sample dates and lesser abundance
thereafter. No trends were observed for nearshore and offshore comparisons, treated versus untreated
control comparisons, or 0.5 m to 5.0 m sample comparisons. Fertilizer additions had no stimulatory
effect on bacterial numbers.
Bacterial Productivity
The functional measure of bacterial incorporation of tritiated thymidine demonstrated
considerable variability between sample dates with no consistent trends through time or with fertilizer
treatments (Figure 7.69). The prominent samples with increased productivity usually occurred on the
same dates for all samples, with similar trends at upper and lower depths. There were no trends
consistent with effects of nutrient addition.
Caged Mussels
Only three of the samples had total PAH concentrations above the 0.20 /ig/g detection limit.
One of the replicate samples from the Passage Cove untreated control site (station 10) on August 8
and August 14, and from station 3 on August 14 had total PAH residues of 0.27, 0.23, and 0.25 /ig/g
wet weight, respectively. The other replicates from these stations on these sample dates did not have
detectable PAHs. These results are given in Table 7.11. Residues of this order would not be
unexpected, given the magnitude of boat traffic and residual oil in the study area. The overall results
show no trend of enhanced hydrocarbon release and mussel bioaccumulation attributable to fertilizer
additions. Problems in analysis were encountered using a small sample size (3 mussels/sample
collected) and should have been larger (10 mussels/sample collected). In addition, the number of
"time zero" mussels collected should have been equal to the number of mussels set out in all the test
sites at the beginning of the test. This would have ensured enough tissue for analytical method
325
-------
c
UJ
m
c
UJ
o
<
CD
14
STATION 2
7-21 7-28 7-31 *-7 8-14 *-21 (-27
1.4 .
1.2 .
1.0
0.»
0.6
04
0.2
0.0
STATION 3
7-21 7-36 7-11 *-7 *-14 *-21 *-27
1.2
1.0
O.t .
a*
0.4
0.2
0.0
• 0.5 meters
D 5.0 meters
7-21 7-26 7-31 t-7 «-14 6-21 »-27
STATION 4
7-21 7-2« 7-31 »-7 »-14 »-21 »-27
SAMPLE DATES
Figure 7.68. Abundance of Bacteria (cells x10fl/L) From Water Samples
Collected at the Passage Cove Study Site. Fertilizer was
Applied on July 25,1989. Values are Means (+ SD) of 4
Replicates; Dark Bars For Samples Collected at 0.5 m, Open
Bars For Samples Collected at 5 m. Refer to Figure 3.3 For
Sample Locations.
326
-------
1.4,
•b
X
^
1
DC
UJ
CD
c
I
ffi
7-21
7-26 7-31 *-7 6-14 »-21 6-27
SAMPLE DATES
7-21 7-26 7-31 »-7 »-14 «-21 M7
Figure 7.68. (Continued)
327
-------
CO
">»
O
^
Q
O
flC
D.
ffi
5
<
CD
7-21 7-M 7-31 »-7 »-14 »-21 »-27
7-21 7-2» 7-31 »-7 »-14 »-21 W7
7-21 7-2( 7-31 «-7 0-14 *-21 »-27 «-»
SAMPLE DATE
Rgure 7.69. Bacterial Productivity Measurements From Tritiated
Thymidine Uptake (mM Thymidine/L/day) in Water Samples
Collected at Various Sites in Passage Cove Following
Nutrient Application to Shorelines on July 25,1989. Values
are Means (+ SD) of 4 Replicates; Dark Bars for Samples
Collected at 0.5 m, Open Bars for Samples Collected at 5 m.
Refer to Figure 3.3 for Sample Locations.
328
-------
BACTERIAL PRODUCTIVITY (mM Thy/L/day)
OOO;-*-*-»-»-»M OOOOO- — -• — —M
CO
ro
5
o
o
m
-------
PASSAGE COVE
development and validation at the beginning of the test, and enough for quality assurance analysis for
the duration of the test.
TABLE 7.11. TOTAL PAH'S (/iG/G) IN CAGED MUSSELS AT PASSAGE COVE
AT 5 STATIONS OVER TIME
Date
8/5
8/8
8/14
8/21
8/27
9/3
10
ND
0.27, 3 ND
0.23, 3 ND
ND
ND
ND
Station
30 50
ND
ND
0.25, 3 ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
60
ND
ND
ND
ND
ND
ND
#
70
ND
ND
No sample
No sample
No sample
No sample
Replicates
Analyzed
4
4
4
4
4
4
Detection limit - 0.20 /tg/g (wet wt.)
ND - None detected
All values are reported as wet weight
Field Toxicity Tests
A risk assessment indicates that application of INIPOL poses a potential toxic risk to marine
biota if water concentrations approach 15 mg/L, the LC50 for the most sensitive species tested in
laboratory toxicity tests. To characterize the extent to which toxic concentrations might develop
during or immediately after application to oiled shorelines, a series of toxicity tests were conducted
with field water samples using a testing scheme similar to that used to test acute toxicity of industrial
effluents. These data provide some insight into the rate at which INIPOL enters the marine
environment and the dilution required to mitigate toxic effects.
Water samples collected at specified intervals before and after INIPOL application were sent to
a consulting laboratory for 48-hour toxicity tests with oyster larvae, Craxsostrea gigas. Endpoints
330
-------
PASSAGE COVE
monitored for these tests were larval survival to test termination, and percentage of larvae that
exhibited abnormal development. The results for these tests are tabulated in Table 7.12.
TABLE 7.12. 48-HOUR SURVIVAL AND DEVELOPMENT OF LARVAE TESTED WITH
100% SITE WATER (UNDILUTED EXCEPT FOR SALINITY ADJUSTMENT)
Sample Designation
Survival
Abnormal Tidal Stage
Lab Seawater Control
Hypersaline Control (28 ppt)
Field Control (10 a.m.)
Pre-Application (Time 0, 10 a.m.)
Application 10 a.m. - 2 p.m.
1-hour Post Application (3 p.m.)
3-hour Post Application (5 p.m.)
6- hour Post Application (8 p.m.)
12-hour Post Application (2 a.m.)
18- hour Post Application (8 a.m.)
92%
75%
70%
74%
62%
87%
77%
58%
39%
9.5%
7.8%
8.4%
10.4%
14.2%
16.1%
10.5%
10.1%
31.4%
2-hour pre-low
2-hour pre-low
3-hour post-low
near high tide
mid-tide, outgoing
mid-tide, incoming
mid-tide, outgoing
Test acceptability criteria dictate that for each series tested, control survival must be greater than
70% with abnormality equal to or less than 10%. All laboratory control, field control, and pre-
application samples met these criteria. The greater survival of larvae in laboratory seawater controls
relative to hypersaline controls, field controls, and pre-application samples may be related to minor
toxic components in field samples or the brine solution. These differences were not statistically
significant when compared by Dunnett's procedure. The percentage of abnormal larvae varied little
among the four control samples.
Tests with water samples collected at the field site after INIPOL application resulted in survival
of less than 70% and rates of abnormal development greater than 10%, suggesting the presence of
toxic components. Because the survival of larvae was greater than 50% for all water samples except
the 18-hour sample, an LC50 could only be computed for the 18-hour sample. Toxicity associated
with the other samples was assessed through the use of Dunnett's procedure to determine if observed
effects were significantly greater than mortality or abnormal development rates for the field control
and pre-application samples, which are the proper samples for comparisons with test site treatments.
331
-------
PASSAGE COVE
None of the values for post-application samples, except the 18-hour test samples, were significantly
different from the field control survival (70%) and pre-application survival (74%). In addition, none
of the percentages of abnormally developed larvae for these samples were significantly different from
those of the field control sample (8.4%) and pre-application sample (10.4%). Comparison with
laboratory seawater controls showed significant effects for several samples, but these comparisons
combine INIPOL toxicity with residual toxicity in site water at Passage Cove before fertilizer
additions.
The water sample taken at 18-hour post treatment killed 61% of the oyster larvae during the
toxicity test. Using the dilution series of 100% site water, 56%, 32%, 18% and 10%, a 48-hour LC50
of 58% of full-strength water was calculated. In other words, when full-strength site water was
diluted to 58% of its original concentration, it would kill 50% of the oyster larvae during a 48-hour
test. The 95% confidence interval for the LC50 is 46% to 75%. The full-strength site water collected
at 18 hours had significantly greater numbers of abnormal larvae compared to site and laboratory
controls. Dilution to 56% of the full strength concentration produced abnormal larvae at a rate not
significantly different from the field control and pre-application samples.
Toxic effects from potential over-application of INIPOL or immediate release of INIPOL from
the shoreline during initial tidal flooding were not seen. Test results indicate that application of
INIPOL to oiled shorelines at the Passage Cove test site resulted in water concentrations that caused
abnormal development and mortality of oyster larvae only during the sampling that occurred 18 hr
after application. The 48-hour LC50 for this sample was 58% of full-strength site water. The
increase in abnormal development associated with this sample was mitigated by dilution to 56% of
full-strength.
Apparently, more toxicity was associated with the second flooding of the INIPOL-treated
shoreline than with the initial flooding. This is unexpected, and may be related to a different degree
of wave action or to additional floating oil that moved into the test area. However, no unusual
weather or oil movements were observed following the INIPOL application. In the absence of
INIPOL additions, test site water produced survival and abnormality rates that were marginally above
acceptance criteria. This may demonstrate residual toxicity problems that exist along oiled shorelines
unless definitive clean-up actions are taken.
332
-------
PASSAGE COVE
If we attribute all the observed toxicity to release of INIPOL from the treated shoreline upon
re-flooding by incoming tides, then the release rate can be estimated.
a) Using the application rate of 293 g INIPOL/m1, a concentration of 4,500 mg/L would be
expected if we assume 100% of the applied INIPOL is released into water immediately over
the treated beach, with minimal dilution.
b) The LC50 for the most toxic sample, the 18-hour post application sample, was 58% of full-
strength, i.e., an exposure resulting in 50% mortality from the field sample.
c) Using 50 mg/L as the LC50 for oyster larvae and INIPOL, any field sample that caused 50%
mortality should have 50 mg INIPOL/L. Thus, we can assume that a 58% dilution of 18-
hour water would lower concentrations to 50 mg/L.
d) Thus, the initial concentration in the 18-hour sample may have been 90 mg/L INIPOL
(dilution to 58% gave 50 mg/L). This concentration is 2% (90 mg/L divided by 4,500 mg/L)
of the "no-dilution and 100% release" assumption.
e) This crude estimate of the release rate (2%) is within the expected range for initial releases
of INIPOL following application.
SUMMARY AND CONCLUSIONS
The following conclusions can be drawn from the Passage Cove study (beaches previously
washed) conducted during the latter part of the summer of 1989:
a) The visual reduction in oil due to application of the oleophilic/granular fertilizer
combination was similar to Snug Harbor results. This visual reduction became apparent
approximately two to three weeks following application of the fertilizers; the untreated
control beach, on the other hand, essentially did not change visually. The effect was perhaps
more dramatic in Passage Cove since oil from both the cobble surface and the subsurface
mixed sand and gravel visually disappeared in a shorter timeframe. It is possible that when
the beaches in Passage Cove were physically washed, oil was distributed over a large surface
area, subsequently creating improved conditions for the biological degradation of oil.
333
-------
PASSAGE COVE
b) Application of fertilizer solutions from a sprinkler system also caused oil to visually
disappear in approximately the same general timeframe as Snug Harbor results (3 to 4
weeks). This observation provided definitive proof that biodegradation (and not chemical
washing was likely responsible for the oil removal, since there was no other reasonable
mechanism to explain this effect of nutrient addition to the oil. The application of fertilizer
solutions, therefore, proved to be the most efficient system for exposing oil-degrading
microorganisms to nutrients in a controlled and reproducible manner.
c) Application of the fertilizer solution produced a statistically significant enhancement of oil
biodegradation relative to the untreated control beach. Rates of total oil residue loss were
greater than four-fold faster than rates of removal on the untreated control beach. The loss
of oil residues was accompanied by extensive changes in oil composition. This included
large decreases in the nC18/phytane ratio. Thus, enhanced biodegradation was probably
responsible for changes in oil residue and composition.
d) Results from the fertilizer solution treatment further support that oil biodegradation in
Prince William Sound was limited by the availability of nutrients and not by the availability
of the oil itself. In addition, reapplication of nutrients (the extreme in the case of the
fertilizer solution) is probably important for sustaining enhanced biodegradation.
e) Application of the INIPOL/CUSTOMBLEN combination also substantially enhanced oil
biodegradation. At a slightly lower degree of statistical confidence (90% instead of 95%),
the INIPOL/CUSTOMBLEN fertilizer produced asignificant two to three-fold enhancement
in the removal of total oil residues relative to the untreated control beach. This was
accompanied by an extensive change in the composition of the oil as well.
f) Mechanistically, there is no evidence to suggest that the application of 1NIPOL/
CUSTOMBLEN worked differently than the application of the fertilizer solution; each
process provided enough nutrients to the oil-degrading microbial populations to enhance
biodegradation. Results from changes in oil composition during the initial two weeks
following fertilizer application suggest that the oleophilic fertilizer uniquely caused
simultaneous degradation of the higher and lower molecular weight hydrocarbons. Results
from the untreated control beach and the fertilizer solution-treated beach showed that
during the same time period a more typical response was observed; that is, the lower
334
-------
PASSAGE COVE
molecular weight hydrocarbons degraded faster than the higher molecular weight
hydrocarbons. It is also believed that the eventual greater response from the fertilizer
solution application was due to higher nutrient concentrations sustained over a longer period.
Reapplication of the 1NIPOL/ CUSTOMBLEN fertilizer combination every three to four
weeks might produce the same effect observed with the fertilizer solution application.
g) Although it is difficult to prove at this time, we believe that the physical washing process
provided conditions that greatly enhanced the effectiveness of oil bioremediation. In
addition, we believe that after a certain amount of biodegradation, the physical consistency
of the oil changes, making it less likely to adhere to surfaces. As a result, the turbulence
created by tidal action is sufficient to remove the degraded oil residues from the beach
material, thus producing visually cleaner beaches.
h) Further monitoring of the fertilizer-treated beaches through early summer 1990, revealed
that even subsurface oil (to a depth of approximately 0.3m) was virtually completely
removed within approximately 10 months. However, significant, but patchy amounts of oil
remained on the untreated control beach after this time period. This suggests that
bioremediation greatly reduced beach cleanup time.
i) Due to high variability in the numbers of oil-degrading bacteria in each sample, it was not
possible to show statistically significant increases in the oil-degrading microbial populations
as a result of the fertilizer addition.
j) No widespread or persistent adverse ecological effects were observed from the monitoring
program that was designed to measure toxic responses, eutrophication, and bioaccumulation
of oil residues. Ammonia, the only component in the oleophilic fertilizer that was
potentially toxic to indigenous species, never reached toxic concentrations outside the
immediate zone of application (as inferred from the toxicity test results). Measurements of
chlorophyll, primary productivity, and bacterial production indicated eutrophication did not
occur. The absence of oil residues in caged mussels, held just offshore of the fertilizer-
treated areas, supported the tenet that oil was not released from the beaches into the water
column as a result of the fertilizer treatment.
335
-------
SECTION 8
DISK ISLAND FIELD RESULTS
NUTRIENT RELEASE TEST
To assure that the application of fertilizer granules would release nutrients into a defined plot
size, a short test was performed on an uncontaminated cobble beach. Three plots of different sizes
were established and subsurface sampling wells were placed in the middle of each of four quadrants
within the plots. Thus, distance between wells was scaled to the plot size. In addition, on the first
incoming tide interstitial water was also sampled with root feeders, the same method used to monitor
nutrients in beach material during the summer of 1989. This was done to compare the sampling
efficiency of this method with the well method. Fertilizer granules were applied to each plot at a rate
of 100 g/m2. The results showed significant nutrient release occurred on each sampling day, but
higher concentrations of ammonia and phosphate were present in the interstitial water during the
outgoing tide than the incoming tide (Figures 8.1 to 8.14). Thus, it would appear that some "soaking"
of the granules was required to obtain the greatest nutrient release.
Other than a few outliers, it appeared that nutrient release was reasonably uniform and not
affected by scaling; in other words, nutrients in the wells were generally similar (not statistically
different) on all plot sizes. This means that all sampling baskets placed within a plot will likely
experience the same nutrient exposure. The sporatically high concentrations of ammonia and
phosphate in certain wells is not readily explained, but could represent channeling effects in the beach
material. However, since the higher levels were not consistent from well to well or day to day, any
channeling would have been temporary, such as that created and removed by tidal action.
Compared to the well method, the use of root feeders to sample interstitial water appeared to
be equally effective. In fact, in some cases, higher nutrient concentrations were seen in root feeder
samples, suggesting that nutrient-containing water was drawn from points distal to the immediate
sampling area.
This test showed something of the nutrient release pattern expected from the fertilizer granules:
a large, short-lived, pulse within the first 24 to 48 hours after application, followed by further slow
release to give nutrient concentrations barely above background levels. Based on studies at Elrington
Island (Section 9), this pulse effect may be the primary way in which oil biodegradation rates are
enhanced by this method of fertilizer application.
336
-------
•
a
i
$
2
I
A
iiJi
I
d
1 2 3 4 Ited 1 2 3 4 Ml
1—
— W 1 \ 20 \
I
d 1
|
|
I
I
I
,
2 3 4 ttod W»i
1 30
\ PM
Figure 8.1. Ammonia Concentrations in Wells in Each Plot on the
Incoming Tide for the Scaling Experiment on 6/19/90.
i,
234
10
i a
3 4
20 —
M*d 1 2
3 4 ttodWW
30 I Plot
Figure 8.2. Ammonia Concentrations in Wells in Each Plot Using Root
Feeders on the Incoming Tide for the Scaling Experiment
on 6/19/90.
337
-------
1
1
1-
i
2
!
^
I
P
I
I
3 4 M*d 123
10 1 1 20
I
I
!
V7\
Y.
'/
y.
^
\
v\
y.
'/
y.
f
f
1
4 MM 1 2 3 4 ItodMW
1 I » - I Ptol
Figure 8.3. Ammonia Concentrations In Wells in Each Plot on the
Outgoing Tide tor the Scaling Experiment on 6/19/90.
D'
I.
?
2
' n
^
^
%
\
6
f.
I
1234 IM 1234 M(d 1234 M*d WM
| 10 1 1 20 1 I — X I Plot
Figure 8.4. Ammonia Concentrations In Wells in Each Plot on the
Incoming Tide for the Scaling Experiment on 6/20/90.
338
-------
•
I.
!
I
0'
life
y.
J
^ ^
1
i W$
1 t • 4 MM 1 t 3 4
^ ^^
5; X- X/
11
2 ^ Z
MM 1 1
I — « — i 1 10 -H I —
3
30
1
/v
I
4 KM WM
I Hoi
Figure 8.5. Ammonia Concentrations In Wells In Each Plot on the
Outgoing Tide for the Scaling Experiment on 6/20/90.
i.
1 2 3 4 Mid 1 ( 3 4 Mtd t i 3 4MMWM
h—10—I I—ao—| |- so | PW
Figure 8.6. Ammonia Concentrations In Wells In Each Plot on the
Incoming Tide for the Scaling Experiment on 6/21/90.
339
-------
•
*
I.
1
1
fs
i a
1- —
I
I
Id
^
2
1
J
1
3 4 Mid » 1 3 4 Mid
«--H | — to — |
J
1
1
1
i a 3 4 n
1 - 30
Figure 8.7. Ammonia Concentrations In Wells In Each Plot on the
Outgoing Tide (or the Scaling Experiment on 6/21/90.
at
OJ
ar
OJ
L
0.4
OJ
u
ai
HA
I
1
rara
1
I
I
E3
/^
j
|
I
m
I « 4 MM 1 t « 4 MM 1 I I 4MtdWH
w H (•—v -H I 30 I n«
Figure 8.8. Phosphate Concentrations In Wells In Each Plot on the
Incoming Tide for the Scaling Experiment on 6/19/90.
340
-------
04
04
0,7
04
04
04
04
04
0,1
tu>
I
I
m
'/
y,
i
i
I
ll
i
!
0
8
1
^
1
^
^
!
I
\
1 » » 4\
11041
|_—H —
1104
•» —
Figure 8.9. Phosphate Concentrations In Wells In Each Plot Using Root
Feeders on the Incoming Tide for the Scaling Experiment
on 6/19/90.
1.0
04
04
0.7
04
I
04
0.4
04
OJ
0.1
0.0
II34liM1l34Mtd1t34
I - - 10 H | —»- H | w |
Figure 8.10. Phosphate Concentrations In Wells In Each Plot on the
Outgoing Tide for the Scaling Experiment on 6/19/90.
341
-------
••V
OJ
OJ
0.7
0.0
IM
5
04
OJ
OJ
Ol
M'
R2
t
I 1 V
V f/ V
/ ;
'i
Y,
/,
-------
1.0
04
0.1
0.7
OJ
06
0.4
OJ
OJ
at
0.0
1 2 3 4 IM 1 f 1 4 IM t » 3 4 MM WM
| — 10 H (•— SO 1 1-30 I PM
Figure 8.13. Phosphate Concentrations In Wells In Each Plot on the
incoming Tide for the Scaling Experiment on 6/21/90.
1JW
OJ
OJ
07
0.0
{"
M
OJ
OJ
ai
0.0
I
fj
ft
1 t
\ —
WL
9 1
y y
y y
II
3 4
li
I
p
\
H R
^
5 y
^
y
^1 ^
/
-------
DISK ISLAND
NUTRIENT CONCENTRATIONS
The lack of any fertilizer effect on oil degradation cannot be attributed to a lack of nutrient
exposure to the microbial communities. The pattern of nutrient concentrations (Figures 8.15 to 8.19)
was very similar to the scaling test; nutrients were readily detected in the beach subsurface, and the
application resulted in an initial pulse of nutrients followed by the maintenance of concentrations
slightly above background levels over a 21-day period, especially on the beaches receiving the higher
fertilizer loadings.
For ammonia release, the incoming tide generally produced higher interstitial water
concentrations than the outgoing tide (Figures 8.15 and 8.16). Concentrations above toxic thresholds
(approximately 10 ppm NH4) did not occur. Following the initial release, phosphate concentrations
were actually maintained at higher levels than ammonia over the remainder of the test (Figures 8.17
and 8.18). This verifies, rather convincingly, the long-term release characteristics of the fertilizer
granules. The much lower concentration of ammonia over the 20-day test period suggests that either
it was all released during the initial burst (rough calculations show this is not unreasonable) or that
it was rapidly taken up by the oil-degrading microbial community in the beach subsurface (algal
uptake is also a possibility).
Although fewer data points were taken for nitrate, nitrate release also appeared to mimic
phosphate in terms of the longer-term release characteristics (Figure 8.19). Initial burst
concentrations were higher than ammonia. If it is assumed that nitrate is as toxic as ammonia, then
the combined nitrate/ammonia concentrations at the highest fertilizer application approached the
toxicity threshold of approximately 10 ppm. This does not consider that the exposure was very short.
At the next highest fertilizer application (500 g/ma), total concentrations were more acceptable and,
thus, normal application rates of fertilizer granules (100 g/m1) could be increased five-fold without
any ecological effect.
OIL CHEMISTRY
In an attempt to determine the specific activity of fertilizer-enhanced biodegradation, or the
extent of rate enhancement per quantity of nutrients released, sampling baskets containing
homogenized, oiled beach material were placed in six beach plots and exposed to different
concentrations of fertilizer. Baskets were sampled periodically and subsamples analyzed for changes
in oil chemistry as a function of four different fertilizer application rates. Two untreated control
344
-------
Q.
z
0
\ v\
oA * >
22 Jun 90 29 Jun 90 06 Jul 90
Sample Collection Date
13 Jul 90
Legend
Control 1 «- * * Control 2 ODD Rate 1
Rate 2 A-A-A Rate 3 e-e-e Rate 4
Figure 8.15. Change in Ammonia Concentration Over Time for the
Incoming Tide for all plots for the Disk Island Fertilizer
Application Rate Study.
/A
22 Jun 90
29 Jun 90 06 Jul 90
Sample Collection Date
Legend » « x Control 1 »• * -» Control 2 ODD Rate 1
«-e-» Rate 2 A-A-A Rate 3 e-e-a Rate 4
13 Jul 90
Figure 8.16. Change in Ammonia Concentration Over Time for the
Outgoing Tide for all plots for the Disk Island Fertilizer
Application Rate Study.
345
-------
0.5
0.4
'0.3-
)
0.2
0.1-
0.0
22 Jun 90 29 Jun 90 06 Jul 90
Sample Collection Date
Legend
Control 1 «-•»•-» Control 2 B-B-B Rate 1
Rate 2 &-a-ARate3 e-^-e Rate 4
13 Jul 90
Figure 8.17. Change in Phosphate Concentration Over Time for the
Incoming Tide for all plots for the Disk Island Fertilizer
Application Rate Study.
0.5-
0.4
•0.3
>
0.2
0.1
0.0
22 Jun 90
29 Jun 90 06 Jul 90
Sample Collection Date
Legend
Contrail «-*-» Control 2 a a a Rate 1
Rate 2 A-*-* Rate 3 e-e-eRate4
13 Jul 90
Figure 8.18. Change in Phosphate Concentration Over Time for the
Outgoing Tide for all plots for the Disk Island Fertilizer
Application Rate Study.
346
-------
Control 1
Control 2
50 g/m2
100 g/m2
500 g/m2
1,000 g/m2
6/24/90 7/5/90 7/9/90
Sample Collection Date
Rgure 8.19. Change In Nitrate Concentration Over Time for the Incoming
Tide for all Plote for the Dlek Island Fertilizer Application
Rate Study.
347
-------
DISK ISLAND
plots within the beach were used to assess background oil degradation rates and any possible beach
effects.
Changes in oil residue weights and oil composition (concentration decreases in nC18, nC27, the
summed alkanes nC18 to nC27, phytane, and the nC18/phytane ratio) are shown in Figures 8.20 to
8.55. A broad overview of the data shows that unfortunately, the experiment did not work. Losses
of oil residue were generally insignificant and compositional changes were not different from the
controls.
Trends in the oil residue weight decay over time (Figures 8.20 to 8.25) were confusing and
complex. On the two untreated control plots, triplicate baskets removed from the beach on the same
date showed relatively small variations in oil residue weight. Variation between sampling dates,
however, was quite large, showing a very fluctuating pattern. While a regression line through the
median values for untreated control plot number one indicated a slight, but not statistically
significant, decrease in residue weights through time, untreated control plot number two showed
virtually no decrease.
The most reasonable explanation for these results is that oil physically moved within the entire
beach, perhaps as a function of different tidal or weather conditions, thus producing this high
variability. Groundwater surges due to heavy rains could have caused oil to move up or down in the
beach material. This would have occurred uniformally across each of the untreated control beach
plots because the differences between sampling baskets was low. Visual observations of the oil
revealed it was considerably more "fluid" than on other beaches and consequently seemed to be
removed from the beach material quite easily. Any water accumulating in holes dug on the beach
invariably had very pronounced oil slicks. After disturbance of the beach material following insertion
of the sampling baskets, a large slick accompanied the incoming tide line as it passed over the
disturbed area. These observations contrasted sharply with Elrington Island where slicks were never
seen following physical reworking of the beach material. In addition, oil on this beach was drier and
thicker in consistency, with more glacial till incorporated. It is not clear what conditions preventing
oil on the Disk Island test site (DI-067a) from becoming the more typical consistency of oil on
Elrington Island. It may be due to the lower energy of the beach, it's relative shallow slope, or the
presence of humic material throughout the beach material.
348
-------
10000
9000
8000
7000
6000
5000
20Jun90 28Jun90 06Jul90 14 Jul 90 22Ju(90 30Jul90
Sample Coflecdon Date
ooo Oiled Homogenate
A Basket Sample
Median Value
Figure 8.20. Change in Oil Residue Weight Over Time for Untreated
Control Plot 41 for the Disk Island Application Rate Study.
349
-------
10000
9000
8000
7000
6000
5000
A '
•
A
20Jun90 28Jun90 06 Jul 90 14Jul90 22JU90 30Jul90
Sample CoUecton Date
L«o«na
Otod Homogtnut
A BnkM Sampto
Median VWut
Figure 8.21. Change In OH Residue Weight Over Time for Untreated
Control Plot *2 for the Dlek Island Application Rate Study.
350
-------
10000
. 9000
>
8000
5
6000
5000
T ' 1 1 1 1 r
20Jun90 28Jun90 06 Jul 90 14 Jul 90 22 Jul 90 X Jul 90
Sample Cofectfon Date
L*gtnd ooo OHtd HomogtnaH A a A BMtat Smpto •-«-« Median
Figure 8.22. Change In Oil Residue Weight Over Time for the 50 g/m2
Fertilizer Application for the Disk Island Application Rate
Study.
10000
9000
8000
7000
6000
5000
A
A
A
A
20Jun90 28Jun90 06 Jul 90 14Jul90
Sample CoBecton Date
22JLJ90 30 Jul 90
ooo Otad HomoganiM a A A Otttt* Samplt
Rgure 8.23. Change In Oil Residue Weight Over Time for the 100 g/m2
Fertilizer Application for the Disk Island Application Rate
Study.
351
-------
10000
9000
8000
7000
6000
5000
8
•
20Jun90 28Jun90 06Jul90 14 JuJ 90 22JU90 30Jul90
Sample Collection Date
utgtna ooo Oiled Homogenale & & & Baakat Sample
Median Value
Figure 8.24. Change In Oil Residue Weight Over Time for the 500 g/m2
Fertilizer Application for the Disk Island Application Rate
Study.
10000
9000
8000
7000
6000
5000
o
8
20Jun90 28Jun90 OSJuiQO 14Jul90 22Jul90 XJul90
Sampto CoHactton Date
o o n OHad Homoganate & a a Baalm Sample
Median VaJue
Figure 8.25. Change In Oil Residue Weight Over Time for the 1,000 g/m2
Fertilizer Application for the Disk Island. Application Rate
Study.
352
-------
1000
800
600
400
200
I
c
-i 1 i 1 1 '—i
20Jun90 28Jun90 06Jul90 14 Jul 90 22 Jul 90 30Jul90
Sample Cotecton Date
Ligtna o o o
Otod Homogtntf* A a a BMktt Swnpto
Figure 8.26. Change In nC18 Concentration Over Time for Untreated
Control Plot #1 for the Disk Island Application Rate Study.
1000
800
600
400
200
I
c
20Jun90 2BJun90 06 Jul 90 14 Jul 90 22 Jul 90 30 Jul 90
Sample CoHecton Date
ooo OU«d HomogtnMt A A A Dn)«> Sanpto •-•-« Mtctan Vriut
Figure 8.27. Change In nC18 Concentration Over Time for Untreated
Control Plot #2 for the Disk Island Application Rate Study.
353
-------
1000
800
600
400
§
I
c
200
9
o
20Jun90 28 Jin 90 06Jul90 14JulBO 22Jul90 30Jul90
Sample Cotoctton Date
ooo Odad HomogwwM & a a BMkrt Svnpto
Figure 8.28. Change In nC18 Concentration Over Time for the 50 g/m2
Fertilizer Application for the Disk Island Application Rate
Study.
1000
800
GOO
400
§ 200
I
c
20 Jin 90 28 Jin 90 06Ju)90 14Jul90 22Jul90 30Jul90
o o OM Homogmto A A t, BMM torpto »-»-« Midtan VWi»
Figure 8.29. Change In nC18 Concentration Over Time for the 100 g/m2
Fertilizer Application for the Disk Island Application Rate
Study.
354
-------
1000
800
600
S
T5
I
400
eo
5 200
I
c
20 Jun 90 28 Jun 90 06 Jul 90 14 Jul 90 22 Jul 90 30 Jul 90
Sample Collection Date
Legend u <>
Oiled Homogenate a A a Batket Sample »-•-• Median Value
Figure 8.30. Change In nC18 Concentration Over Time for the 500 g/m2
Fertilizer Application for the Disk Island Application Rate
Study.
1000
;
i 800
»
t
600
;
!> 40°
{
i 200
0
20
u
o
x
\
\
\
\
\ A
*— ••--»-- -• 4
A
Jun 90 28 Jun 90 06 Jul 90 14 Jul 90 22 Jul 90 30 Jul 90
Sample Collection Date
Legend o o o oiled Homogenate A A A Baakat Sample •-»-• Median VaK»
Figure 8.31. Change in nC18 Concentration Over Time for the 1,000 g/m2
Fertilizer Application for the Disk Island Application Rate
Study.
355
-------
1000
800
600
400
5
•5
f
§ 200
I
c
20Jun90 2BJun90 06Jul90 14JU90 22Jul90 30Jul90
Sample Cotodlon Date
ooo OUtd Homoganrtt
Sampte
Figure 8.32. Change In nC27 Concentration Over Time for Untreated
Control Plot #1 for the Disk Island Application Rate Study.
1000
800
5
•5
600
400
§ 200
I
c
20Jun90 28Jun90 06Jul90 14Jul90 22JU90 30Ju)90
Sample Collection Date
o o i, OU«d HomogwwM A & A BMtat Svnpto •-»-• MKJtan VWut 1
Figure 8.33. Change In nC27 Concentration Over Time for Untreated
Control Plot #2 for the Disk Island Application Rate Study.
356
-------
1000
800
3
"5
600
400
§ 200
I
c
A
A
1
20Jun90 28Jun90 06Jul90 14Jul90 22Jul90 30Jul90
Sample CottectJon Date
Ltgtno ooo OHtd Homogwwto
Svnpfc
Figure 8.34. Change In nC27 Concentration Over Time for the 50 g/m2
Fertilizer Application for the Disk Island Application Rate
Study.
1000
800
5
"5
600
400
8 200
I
c
0
A
-«•.
A
20Jun90 28Jun90 06Jul90 14Jul90 22Jul90 30Jul90
Sample CoitectJon Date
ooo OHtd Homogwwft A A A BMkM S«rpt*
MectanVWut
Figure 8.35. Change In nC27 Concentration Over Time for the 100 g/m2
Fertilizer Application for the Disk Island Application Rate
Study.
357
-------
1000
800
S
600
400
§ 200
I
c
20Jun90 28Jun90 06Jul90 14J-J90
Sample Cotoction Date
22JUI90 30JUI90
L*gind ooo Otod Homogantft a A A BMkM Sampto
Figure 8.36. Change in nC27 Concentration Over Time for the 500 g/m2
Fertilizer Application for the Disk Island Application Rate
Study.
1000 \
800
600
400
§ 200
I
c
20Jun90 28 Jin 90 06Jul90 14Jul90 22JU90 30Jul90
Sample Cotecton Date
ooo Otad HomoovMM A A & BMkM Swnpto
Figure 8.37. Change in nC27 Concentration Over Time for the 1,000 g/m2
Fertilizer Application for the Disk Island Application Rate
Study.
358
-------
9000
7000
1-^
8
i
c
CO
5
I
c
•5
5000
aooo
1000
20Jun90 28Jun90 06Jul90 14Jul90 22Jul90 30Jui90
Sample CoHectton Date
L.gtnd ooo Otod HornogwwM a a A BMkM Swpto
M«ttn VWui
Figure 8.38. Change In the Sum of the Alkane Concentration nC18 to
nC27 Over Time for Untreated Control Plot #1 for the Disk
Island Application Rate Study.
9000
97000
8
ri5000
i
c
|aooo
1000
20
o
*\
a
-
a
*\
^
8
i i i i i i
Jun90 28Jun90 06JJ90 14Jul90 22Jul90 X Jul 90
Sample Collection Date
ng.no ooo OiM HomogenM a A a BMtat Svnpto »-^» Medtan VWue |
Figure 8.38. Change In the Sum of the Alkane Concentration nC18 to
nC27 Over Time for Untreated Control Plot #2 for the Disk
Island Application Rate Study.
359
-------
9000
57000
S
^5000
5
I
c
ISOQO
1000
20
o
§
• A
•
A
Xf A
H §
1 ' ) l i r
Jun90 28Jun90 06Ju>90 14Jul90 22Jul90 30Jul90
Sample Collection Date
utg.no ooo Otod Homogmto a a a BMtat Swnpto •-»-« Mtdtan Vriu*
Figure 8.40. Change In the Sum of the Alkane Concentration nC18 to
nC27 Over Time for the 50 g/m2 Fertilizer Application for the
Disk Island Application Rate Study.
9000
§7000
I
5 5000
r:
I
C
g3000
3>
1000
20
i. J
\
\
*^^ A
" ~ ••.,_
" N
A " ..
V
A £
A
Jun90 28Jun90 06 JJ 90 14JUJ90 22Ju!90 30Jul90
Sample CoHecbon Date
Ltgmd o o o Otttd Homogwwto A A A DMlcX Sampte •-»-• Mcctan VWu*
Figure 8.41. Change In the Sum of the Alkane Concentration nC18 to
nC27 Over Time for the 100 g/m2 Fertilizer Application for
the Disk Island Application Rate Study;
360
-------
9000-
'7000
8
I
CD
5
i
c
•5
5000
3000
1000
20Jun90 28Jun90 06JUJ90 14JU90 22JUI90 30Jul90
Sample CoNectton Date
ooo 0**J Homogvwli A A A BMtat Samp*
Figure 8.42. Change in the Sum of the Alkane Concentration nC18 to
nC27 Over Time for the 500 g/m2 Fertilizer Application for
the Dlek Island Application Rate Study.
9000
'7000
5000
i
i™
1000
\
-A
A
20 Jin 90 28Jun90 06Jul90 14Jul90
Sample Cotectton Date
22JJ90 30JUI90
oo Otod HomoganM A a a taktt Svrpto
KtodtoVMu
Figure 8.43. Change In the Sum of the Alkane Concentration nC18 to
nC27 Over Time for the 1,000 g/m2 Fertilizer Application for
the Disk Island Application Rate Study.
361
-------
3000
§2500
1 2000
5 1500
•5
f 1000
I 500
0
20
A
o
8
"~*\
^ - 1
' 1 I t 1 1
Jun90 28Jun90 06JU90 14Jul90 22Jul90 30Jul90
Sample Cotecton Date
L.gt no ooo Otad Homogtrnn A a A Botat Svnpto »-•-• Mtdtan VWut
Figure 8.44. Change In Phytane Concentration Over Time for Untreated
Control Plot #1 for the Dlek Island Application Rate Study.
3000
«
>2500
2000
3 1500
•8
1000
500
20Jun90 28Jun90 OeJulOO 14Jul90 22Jul90 30Jul90
Scmpte CodacHon Data
Li0»no ooo Otod
Figure 8.45. Change In Phytane Concentration Over Time for Untreated
Control Plot *2 for Untreated Disk Island Application Rate
Study.
362
-------
3000
'2500
2000
<3 1500
T5
-5
91000
500
0
o
*-
B
20Jun90 28Jun90 06JUOO 14Jui90 22Jul90 30Jul90
Senate Coftactton Date
Lig«na ooo Otod HomoQtna» a a a Bwtot Svnpto
Figure 8.46. Change In Phytane Concentration Over Time for the 50 g/m2
Fertilizer Application for the Disk Island Application Rate
Study.
3000
2500
2000
5 1500
1000
500
20Jui90 28Jun90 06Jd90 14Jul90 22Jul90 30Jul90
Sample Collection Data
Lig«nd ooo Otod Homoowt* A A ji BtikO Swnpto
Figure 8.47. Change In Phytane Concentration Over Time for the 100
g/m2 Fertilizer Application for the Disk Island Application
Rate Study.
363
-------
3000
i,
»2500
2000
5 1500
15
^1000
500
0
20Jun90 28Jun90 06Jul90 14JUI90 22JulOO 30Jul90
Sample Collection Data
ooo Otod itomogtnai a * A BM<(* Sampto
Figure 8.48. Change In Phytane Concentration Over Time for the 500
g/m* Fertilizer Application for the Disk Island Application
Rate Study.
3000
5 2500
•§2000
1
3 1500
T5
§1000
500
**s
8-
20Jun90 28Jun90 06Jul90 14Ju)90
Sampto CoHadton Data
22JU90 30JUI90
Ltgtna ooo
HOfDOQinUt & A a BMtal SHT^)to
MtdtanVriu*
Figure 8.49. Change In Phytane Concentration Over Time for the 1,000
g/m2 Fertilizer Application for the Disk Island Application
Rate Study.
364
-------
0.1
0.0
20Jun90 28Jun90 OSJulOO 14Jul90
Sample CoHedlon Otto
22JU90
30JUI90
Ltgind o o o OHtd HomogtTWto
BMtal Sampto
Figure 8.50. Change In the nC18/phytane Ratio Over Time for Untreated
Control Plot #1 for the Disk Island Application Rate Study.
1.01
0.9
0.8
0.7
0.6
0.5
0.4
T0.3
c
0.2
0.1
0.0
20
JunQO 26Jun90 06Jul90 14Jul90
Sample CoflecUon Date
22JUI90 30JUJ90
no«no ooo OHad HomoginiH ^ A ^ Drtiat Sample
Medl«nVak«
Figure 8.51. Change In the nC18/phytane Ratio Over Time for Untreated
Control Plot #2 for the Dlek Island Application Rate Study.
365
-------
1.0
0.9
0.8
$0.7
JB0.6
|r 0.5
«0.4
I 0.3
c
0.2
0.1
0.0
20
* •
- , -^._
— i — — - i i i i i
Jun90 28Jun90 06Jul90 14Jul90 22Jul90 30Ju)90
Sample CoBecton Date
Oted Homogwwtt
A BnMt Scmpto
Median V*M
Figure 8.52. Change in the nC18/phytane Ratio Over Time for the 50 g/m2
Fertilizer Application for the Disk Island Application Rate
Study.
1.0
0.9 \
0.8
SB 0.6
£0.5
00.4
I 0.3
c
0.2
0.1
0.0
20Jun90 28Jun90 06Jul90 14Jul90
Sample Collection Date
22Jul90
30Jul90
OH«d HomogwwM a A a Botat Sampto
MwltanVMu*
Figure 8.53. Change In the nC18/phytane Ratio Over Time for the 100
g/m2 Fertilizer Application for the Disk Island Application
Rate Study.
366
-------
1.0 ]
0.9
0.8
0.7
0.6
'0.5
«0.4
I 0.3
c
0.2
0.1
0.0
20 Jung) 28Jun90 06Jul90 14 Ju! 90
Sample Collection Data
A
A
*
A
A
1 r
22 Jul 90 30 Jut 90
ooo Otod Homogwwtt a A A Bntet Swnpto
Median
Figure 8.54. Change In the nC18/phytane Ratio Over Time for the 500
g/m2 Fertilizer Application for the Disk Island Application
Rate Study.
1.0
0.9
0.8
0.7
0.6
0.5
0.4
T0.3
c
02
0.0
20Jun90 28Jun90 08Jul90 14JU90
Sample Cotocbon Data
22JulOO
30Jul90
L»o»na ooo OHKl HomogmM A A A QitlH* Svnpto
Figure 8.55. Change In the nC18/phytane Ratio Over Time for the 1,000
g/m2 Fertilizer Application for the Disk Island Application
Rate Study.
367
-------
DISK ISLAND
On the fertilizer-treated plots, trends in oil residue decay were obscured by the large variability
between the sampling dates, and in some cases between the sampling baskets. Variability between
sampling baskets was particularly apparent on the plots receiving the 50 g/m3 and 100 g/m2
application of granules. When comparing day zero with the last sampling date, the only fertilizer-
treated beach which showed a net loss of oil residue was the 100 g/m2 application. The lack of any
net change on the other beaches would suggest that this decrease was not real but due more to the
observed variability in oil concentrations.
Changes in the nCI8/phytane ratio strongly suggest that oil biodegradation was occurring on
each test plot, although slowly (Figures 8.26 to 8.49). The nC18/phytane ratios showed slow but
steady decreases through time despite relatively low values initially (i.e., significant hydrocarbon
degradation previous to this test) (Figures 8.50 to 8.SS). Decreases were generally seen with all other
n-alkanes and to a smaller extent with the branched alkane, phytane. However, most importantly,
it appeared that the fertilizer application had virtually no effect on the hydrocarbon decay rates and
therefore on biodegradation. The only possible effect was seen with the nC18 alkanes and the
summed alkanes (nC!8 to nC27), where a somewhat faster initial drop in concentration of these
hydrocarbons occurred at the 500 g/m3 application. We would conclude that biodegradation was not
nutrient limited on this Disk Island beach. It is possible that the low energy conditions and/or the
packed nature of the beach material on this beach (more sand than gravel) could have caused oxygen
to become the limiting factor. This condition could have been compounded by the presence of
additional oxygen demand from humic and/or other organic materials which were visually apparent
in the beach matrix.
It is also important to note the considerably reduced variation in the hydrocarbon concentrations
between sampling baskets (low variation in replicates) and between sampling dates (relatively
consistent downward trend) compared to that seen in oil residue weights. Since hydrocarbon
concentrations were normalized to the oil residue weights, the results suggest that percentage
composition of the oil was very consistent, regardless of the amount of oil present. If the methylene
chloride was extracting other organic matter besides the oil (a possible explanation for the variable
residue weights), this should also affect the hydrocarbon percentage composition. In fact, it does not
appear that this was the case. Thus, it is concluded that oil on this Disk Island test beach was quite
uniform in terms of the extent of biodegradation, but very heterogeneous in terms of distribution over
time and space.
368
-------
DISK ISLAND
MICROBIOLOGY
As samples were brought back to the laboratory for chemical analysis, subsamples were taken
to measure rates of total CO2 production, or oil mineralization by microbial communities. In addition,
following incubation for the CO2 analysis, the samples were spiked with 14C-phenanthracene and the
amount of radiolabeled CO2 was also measured. In all cases incubations were performed in biometer
flasks with a 12:12 low:high tidal cycle. Results from these studies are shown in Figure 8.56. It is
quite evident that considerable oil biodegradation was occurring in these samples. Production of large
amounts of radiolabeled COa from phenanthracenc strongly suggests that this was due to
biodegradation of oil and not other types of organic matter. The amount of total CO2 produced was
also in the same range as samples taken from Elrington beach (Section 9), where oil chemistry results
showed more rapid oil degradation than on the Disk Island test beach.
The effect of fertilizer application rate on oil biodegradation, as measured by the CO2
mineralization, was fairly consistent. Adding fertilizer generally caused significant increases in
mineralization rates relative to the untreated control plots, and the greatest stimulation occurred with
the 500 g/m2 application (approximately 2 to 5 times greater than the untreated controls). The highest
application of 1,000 g/m3 actually seemed to inhibit mineralization to some extent. Graphing a
double reciprocal plot of the mineralization rate versus the fertilizer application rates (Figure 8.56,
insert) showed a generally linear relationship, at least up to the 500 g/m3 concentration of granules.
Based on this relationship it would appear that background mineralization rates were very high, and
to approximately double the oil degradation rate a six-fold increase in fertilizer granules was required.
It appears that application of almost any amount of fertilizer below 500 g/m2 will stimulate oil
mineralization. However, similar studies performed during the winter of 1989/90 (Section 10) where
oiled beach material was exposed to different fertilizer concentrations in shake flasks showed that
biodegradation rates could be stimulated at nitrogen concentrations at least 10-fold higher (35 ppm)
than those actually measured in the interstitial water at the Disk Island test site (approximately 3
ppm). It is possible that oxygen availability in the field became a limiting factor, which would not
be the case in the idealized conditions of a shake flask test. Therefore, extrapolation of laboratory
data on fertilizer specific activity to the field must be considered carefully, as it will be very
dependent on beach conditions.
369
-------
0.005 0.01 0.015 0.02 0.025
Ftrtlllzar Application Rat*
100 200
300 400 500 600 TOO 800 900 1000
Rate of Fertilizer Application (grams/square meter)
Figure 8.56. Plot of Rate of CO, Production Versus Rate of Fertilizer
Concentration.
370
-------
DISK ISLAND
It is possible that the mineralization study results were due to the carryover or fertilizer granules
from the lab to the flasks. Despite attempts to remove subsamples from the sampling baskets that
were free of the granules, their exclusion could not be guaranteed (to avoid the granules, subsumples
were removed from below the surface of the beach material in the baskets). However, since nutrients
in the fertilizer granules are depleted over time, we would expect that mineralization differences
between beach plots would become less with time if they were due strictly to carryover of fertilizer.
This, in fact, was not the case.
Mineralization of radiolabeled hexadecane (performed by the ADEC Oil Spill Response
Laboratory and the University of Alaska, Fairbanks) also showed similar effects of the fertilizer
application (Figures 8.57 and 8.58). A summary of the combined results is shown in Figure 8.59. The
samples showed increases in mineralization relative to fertilizer application in the field. The highest
application rates, 500 and 1,000 g/ma, gave the greatest effect with the peak occurring between day
4 and 12. The 1,000 g/m2 application did not seem to have the inhibitory effect seen with
measurements of total CO2 production and phenanthracene mineralization. Unfortunately, samples
taken prior to fertilizer application were lost.
Finally, measurements of total heterotrophic and oil-degrading bacterial populations are shown
in Figure 8.60. Other than some anomalous results shortly after fertilizer application (first sampling),
there was a strong indication that the number of oil degraders increased substantially on the fertilizer-
treated plots, especially at the two highest application rates. On the 19th day after fertilizer
application, the number of oil degraders on the treated beaches (500 and 1,000 g/m3) were
significantly higher than the untreated control plots (at the 95% confidence level). The difference
was greater than 10-fold. There was also some indication that total heterotrophs may have also
increased on the 1,000 g/m2 plot. It is interesting that as the effectiveness of fertilizer release from
the granules presumedly diminished through time, the enrichment effect also diminished with
essentially no statistically difference in oil degrader numbers on the last sampling date. This suggests
that the oil-degrader populations were quite dynamic, responding quickly to the presence or absence
of fertilizer.
371
-------
40"
Untreated Control Plot 1
30'
w
8 20
10
-I-
I
6/23
6/27
7/2
DAY
I
7/9
I
7/27
I HIGH I LOW I- MEAN
MEDIAN
Untreated Control Plot 2
o
40"
tn»
ou ™
20"
10"
0
,
6/23 6/27 7/2 7/9
DAY
-I-
1
7/27
I HIGH I LOW I- MEAN -J MEDIAN
Figure 8.57. Radiolabled Hexadecane Mineralization Over Time
(Standard Deviation) for both Untreated Control Plots
at Disk Island.
372
-------
RA1:50g/m2
4U -
CM 3°
O
Oon
tU •
m .
T
~f } ,
+ "I"
6^3 6/27 7/2 7/9 7/
DAY
-
27
40 •
CNJ 30 -
O
10 -
0
^
6/23 6^7 7
D
_ -j
/2 7/9 7/
AY
27
JHIGH JLOW JMEAN £ MEDIAN
JHIGH JLOW JMEAN JMEDIAN
RA3:500g/m2
RA4:1,000g/m2
•Jrt .
OJ
O20 -
55
in -
i- i J
* T _.. t.J
6^3 6^7 7/2 7/9 7/2
DAY
>7
40 •
CM 3°-
o
O on
^* eM '
&
10 -
o •*
T ^
^
6/23 6^7 7/2 7/9 7/
DAY
-
27
JHIGH JLOW JMEAN J MEDIAN
JHIGH JLOW JMEAN JMEDIAN
Figure 8.58. Radiolabeled Hexadecane Mineralization Over Time
(Standard Deviation) for the Four Treated Plots at
Disk Island.
373
-------
DISK ISLAND DOSE RESPONSE STUDY
HEXADECANE MINERALIZATION
%CO2
RA4
RA3
CTR2
CTR1
9 16 34
DAYS AFTER FERTILIZER APPLICATION
CUSTOMBLEN
DOSE
RA1 SOg/m2
RA2 lOOg/m2
RA3 SOOg/m2
RA4 1000g/m2
Figure 8.59. Summary of Radlolabeled Hexadecane Mineralization Over
Time for All Plots at Disk Island.
374
-------
-------
DISK ISLAND
SUMMARY AND CONCLUSIONS
The following conclusions can be drawn from the Disk Island study conducted during the
summer of 1990:
a) Effects of different fertilizer concentrations could not be assessed because of an overall lack
of general oil biodegradation. Field samples did show increased mineralization (total CO2
production) in proportion to increased fertilizer application, but oil chemistry results showed
essentially no effect.
b) Analysis of ammonia, nitrate and phosphate in interstitial beach water showed nutrient
release was generally proportional to fertilizer application rate. However, nutrient release
represented more of a pulsed high concentration during the first 2 to 3 days following
application rather than a slow release over time.
c) Monitoring of wells in the test area revealed that nutrient release from fertilizer granules
was not totally slow-release. Highest nutrient concentrations occurred shortly after fertilizer
application and then decreased to background levels 2 to 3 days later.
d) Analytical chemistry results showed that the addition of fertilizer failed to enhance oil
biodegradation, regardless of the application rate. Changes in oil composition occurred,
although slowly, but were similar for both treated and untreated plots. It is not known what
conditions at the Disk Island site could have precluded enhanced oil biodegradation. Disk
Island may represent a type of beach (low energy, low slope profile, less porous beach
material) that is not amenable to oil bioremediation because of insufficient oxygen
availability and/or high natural organic matter content (peat deposits).
e) Measurements of oil mineralization, based on total C03 production or 14COa released from
radiolabeled hydrocarbons (phenanthracene and hexadecane), did show effects due to
fertilizer application rate. It was quite evident that considerable oil biodegradation was
occurring in the beach samples. Production of large amounts of radiolabeled CO2 from
phenanthracene and hexadecane strongly suggested that this was due to biodegradation of
oil and not other types of organic matter (humic materials). Adding fertilizer generally
caused significant increases in mineralization rates relative to the untreated control plots,
376
-------
DISK ISLAND
with the greatest stimulation occurring with the 500 g/m2 application (approximately 2 to
5 times greater than the untreated controls). The largest application (1,000 g/m2) actually
seemed to inhibit mineralization to some extent. A calculated dose response indicated that
a six-fold increase in fertilizer granule application produced a two-fold increase in oil
degradation rates. In addition, oil-degrading bacteria increased following fertilizer
application. Concentrations were almost 10-fold greater than the untreated controls on the
plots receiving the two highest fertilizer concentrations.
377
-------
SECTION 9
ELRINGTON ISLAND FIELD RESULTS
VISUAL OBSERVATIONS
Section ER 20 on the north end of Elrington Island represented a beach with typical subsurface
oil contamination. The surface was relatively free of oil but extensive amounts of oil were found
approximately .2 m to .3 m below the surface. Three test beaches were established and implanted
with sampling baskets containing a constructed subsurface oil layer designed to simulate the normal
subsurface condition of the beach. A trench was excavated that was long enough to allow all of the
baskets to be simultaneously placed in a line. During excavation it was obvious that oil was very
heterogeneously distributed in the subsurface. In some areas, oil was coagulated in large globs
containing varying degrees of mixed sand and gravel, in other areas the oil was relatively dispersed,
coating most of the surface of the mixed sand and gravel. The oil consistency was dry rather than
runny, gritty to the touch (because of the presence of glacial till), and it did not produce an oil sheen
when disturbed and covered with water.
Excavated beach material was used to fill in the area around the baskets. This, of course, made
it difficult to reestablish the oiled subsurface layer next to the basket. However, due to the oil
consistency and the considerable heterogeneity of the beach material, it appeared that the filling
operation did not affect the movement of oil into or out of the sampling baskets. Baskets were
covered with clean beach material, giving a visual consistency with the surrounding beach areas. As
the tides moved in following the placement of the baskets, very little sheen was observed on the water
surface.
The presence of bed rock just below the beach surface on the Bath beach required that baskets
be placed higher on the beach than intended. As a result, water did not cover the area containing the
baskets during some tidal cycles.
For beach sampling, a set of baskets was removed from the beaches and transported back to
Valdez. Care was taken in the laboratory to remove subsamples of the constructed oil layer, the
unoiled layers (top and bottom), and the interface between the oiled and unoiled layers (top and
bottom). At all samplings, the oiled layer in the baskets remained distinct, and there was no visual
difference over time in the amount of oil seen in each oiled layer from the different treatments. Also,
378
-------
ELRINCTON ISLAND
the consistency of the oil inside the baskets looked very much like the oil coating the beach material
outside and surrounding the baskets.
NUTRIENT CONCENTRATIONS
Nutrient concentrations in the interstitial water from the monitoring wells of the Sprinkler beach
and the untreated Control beach are shown in Figures 9.1 to 9.3 (see Section 3 for positioning of the
wells on the beach). No nutrient monitoring was conducted on the Bath beach. Ammonia, nitrate,
and phosphate concentrations on the untreated Control beach were all below 0.1 ppm. No cross-
contamination from the Sprinkler beach was apparent. Concentrations of ammonia on the Sprinkler
beach increased substantially as a result of fertilizer application. The particularly high values
following the July 10, 1990 application resulted from the accidental use of fertilizer solution with a
higher concentration of nutrients. In all cases, the spike concentrations were rapidly diluted by the
tidal cycles. However, a residual concentration of ammonia in the interstitial water did occur, with
concentrations of 1 to 2 ppm over a seven day period. Based on the beach dynamic studies during
the summer of 1989, this residual was not surprising.
OIL CHEMISTRY
Oil Residue Weight
Changes in the oil residue weights over time are shown in Figures 9.4 a-c. Data from the middle
of the oiled subsurface layer are shown. Examination of the figures reveals that by using the
sampling basket procedure, variability in the data points was very small—a significant improvement
over the sampling procedure used at Passage Cove. The oil concentration on the Bath and Sprinkler
beaches at t = 0 was estimated because the sampling baskets were placed in the beach at different
times prior to fertilizer application (the result of logistical complications). Therefore, starting oil
concentrations were calculated based on the oil decay rate of the untreated Control beach and
extrapolated from data obtained between the time of placement of the baskets and initiation of
fertilizer application.
The subsequent decay curves for all beaches appeared to be zero-order. This suggested that the
oil biodegradation rate was not limited by substrate availability. However, it is presumed that, with
time, the rate would become first-order, possibly when the concentration reached 1,000 to 2,000
379
-------
6
5
k
4
i
3
2
1
i i
o o
a
n
ii
i i
; i
I I
I I
y i
/ i
y i
i i
t i
\ ' \
A---.
e — o o o
©-e — e-o
02JJ90 UJUI90 26JU90
07Aug90
Plot Identification
Control Plot 4- A -A Sprinkler Plot
Figure 9.1. Change in Ammonia Concentration Through Time for the
Incoming Tide at the Untreated Control Beach and Sprinkler
Beach at Elrington Island.
8
7
6
-.5
3
23
2
1
0'
02 J
f
,i
'
;i
i ' $
! \ ;\
* A ' 'l "
* 7 J i ' ',
> ' \ ' 1 ' '
\ ' \ I ! / i
A 1 . 1 ' II
l',l' II
1 / A ' A-&- ' '
i i v i — ii
1 I x 1 ^--^ 1 \
I ' \ I -•-. i A
i i \i "~-^. i
i i A -ft i
i ' \ A i
'/ ' ^^ \ -A i'
l' ^^Q — W \ ^» *"" ' S *
Q Q ^ — — ^-^^^^ " — • 0 AT "" A.
U90 14JU90 26JU90 07 Aug 90
Svnpto Cotoctton Dote
Plot Identification
Control Plot «r A -A Sprinkler Plot
Rgure 9.2.
Change in Nitrate Concentration Through Time for the
Incoming Tide at the Untreated Control Beach and Sprinkler
Beach at Elrington Island.
380
-------
0.5
0.4
-S-0.3
2
02
0.1
. o-e-
02JUI90
i \
/ \
( \
-6 O O
14JUI90 2BJUI90 07Aug90
Sampto CoHacdon Date
Plot Idintification
Control Plot
Sprinkler Plot
Rgure 9.3. Change in Phosphate Concentration Through Time for the
Incoming Tide at the Untreated Control Beach and Sprinkler
Beach at Elrington Island.
381
-------
15000
,12000
9000
6000
3000
22Jun90 04Jul90 16 Jul 90 28 Jul 90
Sample Collection Date
09Aug90
Legend o o o Basket Sample a a a Oiled Homogeoate
«-•-• Median Value of Basket Sample ••• Median Value of Homogenate
* * * Treatment Date
Figure 9.4a. Change in Oil Residue Weight Through Time for the
Sprinkler Beach Oiled Subsurface Layer at Elrington Island.
15000
,12000
9000
6000
3000
*
22Jun90 04 Jul 90 16 Jul 90 28 JJ 90
Sample Collection Dale
09Aug90
o o o Basket Sample A A A Oted Homogenate
-•-e Medtan Value of Basket Sample •• • Median Value of Homogenate
* * * Treatment Date
Figure 9.4b. Change in Oil Residue Weight Through Time for the
Untreated Control Beach Oiled Subsurface Layer at
Elrington Island.
382
-------
15000
,12000
9000
6000
3000
O-.
- - - -Q.
22Jun90 04JUI90 16Jul90 28 Jul 90
Sample Collection Date
09Aug90
L»9«na. ooo Basket Sample a a a Oied Homogenate
•-«-• Meolan VWue of Basket Sample ••• Median Value of Homogenate
* * * Treatment Date
Figure 9.4c. Change in Oil Residue Weight Through Time for the Bath
Beach Oiled Subsurface Layer at Elrington Island.
383
-------
ELRINGTON ISLAND
rag/kg (Passage Cove results showed good first-order kinetics with starting oil concentrations of
approximately 1,000 mg/kg). Since less than a single half-life of decay for the most active beach had
occurred during the test period, it is also possible that first-order kinetics were not yet
distinguishable.
Using a zero-order model, the difference in decay rates between the treated and untreated
beaches is very pronounced. A statistical analysis of the decay curves (regression line) is shown in
Table 9.1. At the 95% confidence level, the slopes of the decay curves for the treated beaches were
significantly different from the untreated Control, but were not significantly different from each
other. The slope of a linear regression on these decay curves produced rates of 118 mg of oil/kg of
beach material/day for the Bath beach, and 103 mg/kg/day for the Sprinkler beach. These rates were
approximately 6 to 7 times faster than the rate calculated for the untreated Control beach
(approximately 16 mg/kg/day), representing the largest fertilizer-enhanced oil removal yet observed.
TABLE 9.1. RESIDUE DECAY RATES FOR ELRINGTON ISLAND
Beach Plot
Treatment
Bath
Sprinkler
Control
Rate
(mg/kg/day)
-118
-103
-16
Standard
Deviation
8
6
7
Significance
0.0008
0.0003
0.09
For the Bath and Sprinkler beaches, values for the treatment day are extrapolated from pretreatment
values based on rate for the untreated Control beach.
Comparing the information with results from the Summer 1989 study in Passage Cove, the rate
of oil disappearance on the untreated Control beach was 2.3 times the value (7 mg/kg/day) obtained
from a similar untreated test plot in Passage Cove. However, the Passage Cove results were for oil
residue on the cobble subsurface. The decay rate for oil in the mixed sand and gravel under the
cobble was essentially zero during the Summer 1989 test period, thus indicating an even further
difference between the two summer seasons. Possibly the physical mixing of the beach material prior
to placement in the sampling baskets, or greater colonization of the oil by oil-degrading
microorganisms could account for this considerable difference. We tend to believe the former. Any
384
-------
ELRINGTON ISLAND
mechanical mixing of the beach material would almost certainly increase oil biodegradation rates
because of increased availability of oil to the microorganisms. On the other hand, considering that
the rates remained relatively constant throughout the test duration, it suggests that the initial mixing
provided a continuously favorable situation for biodegradation (which would not be expected).
Despite the high background of oil residue decay in the untreated Control beach samples, the
effect of applying fertilizer was considerable. A single pulse application of fertilizer was as successful
as multiple pulse applications, at least over a period of approximately 3 to 4 weeks. This verifies
similar results observed in laboratory studies (see Section 10). The mechanism responsible for the
effective single pulse is not known, but we speculate that perhaps it involves a process of luxury
nutrient uptake by the oil-degrading microorganisms, followed by a recycling of the nutrients within
the matrix of the microbial communities. Whether this phenomena occurred during the summer of
1989 is currently unclear. It may have taken an entire season for the colonization of the oil by
bacteria (and their potential nutrient limitation) (i.e., more oil carbon per bacterial cell, but no
corresponding increase in nutrients), to reach a point where a pulse of nutrients was effective. On
the other hand, much of the success of the bioremediation during the summer of 1989, particularly
in August, could be attributed to this pulse phenomena (i.e., if the fertilizer formulations used on the
beaches released much of their nutrients soon after application, the bioremediation response we
observed may have been more controlled by this initial pulse rather than a sustained nutrient release).
Oil residues in the "unoiled" layers of the sampling baskets (we were aware at the beginning of
the field test that the "unoiled" beach material did contain small amounts of oil) and at the interface
between the oiled and unoiled layers were also examined. The results from the "unoiled" layers are
shown in Figures 9.5 and 9.6. Samples from the layer interfaces were quite variable in oil
concentration due to the difficulty in sampling uniformally at the interface, and are not presented.
Concentrations of oil in the "unoiled" layers (top and bottom) ranged from 400 to 1,000 mg/kg,
or approximately 4 to 8% of the oil concentration in the middle oiled layer. Interestingly, there was
a trend toward increasing concentrations in the bottom layer with time, particularly for the treated
beaches. Concentrations on the last sampling date were very high for the Bath beach. The trend was
not present in the top layer. The possibility of oil entering only the bottom layer baskets from the
surrounding oiled beach material is unlikely. Movement of oil from the oiled layer (middle layer) is
more reasonable, but it is not related to the sprinkling operation, as the effect is most pronounced in
samples from the Bath beach which received only a single initial application of nutrients. Somewhat
385
-------
2500
.2000
: isoo
. 1000
5 500
Q
a
22JW190 04JUOO 16JUI90 28 Jul 90
Svnpto Cotoctfon Date
e
a
ODAugOO
oooGonMPk* oaoSp*MwPW
Rgure 9.5. Change in Oil Residue Weight Through Time for the Bottom
Unoiled Layer for All Beaches at Elrington Island.
2500
A 2000
1
1500
1000
5 500
0
22
o
o
8 o
e ° o
£ § * 8
x D B O
5 1 H a a a
i i i i i
JunSO 04JH80 16JUI80 2BJul90 09Aug90
Svnpto Cotectfon
o o o CgnM Ptot a a a SprtnMv Ptot
Figure 9.6. Change in OH Residue Weight Through Time for the Top
Unoiled Layer for All Beaches at Elrington Island.
386
-------
ELRINGTON ISLAND
smaller sampling baskets were used on the Bath beach, but it is unlikely this affected either
subsampling or physical separation of the layers. However, the effect could be related to
biodegradation of the oil, if it is assumed that oil loses its sticky qualities after extensive degradation
and is then free to migrate in the beach material (such as through the physical action of gravity and
tidal movement). This would mean that the hydrocarbon composition of the oil in the bottom layer
could change significantly, even though oil residues increased. This possibility is currently under
investigation.
The fact that oil residues in the top layer remain relatively constant is also interesting since there
are probably as many oil degraders in this layer as there are in the middle oiled layer, and yet there
is little evidence of oil decay from biodegradation. The oil concentrations, however, may be too low
to see significant changes over the short time period of this field test.
Oil Composition
To verify that loss of oil residue weight from the sampling baskets was accompanied by changes
in oil composition (biodegradation), the loss of selected alkanes over time was examined. Changes
in the concentrations of individual alkanes nC18 to nC27, the sum of alkanes nC18 to nC27, phytane,
and the nC18/phytane ratio, are shown in Figures 9.7 to 9.19 for the three test beaches. Note that the
oil composition at t»0 was estimated for each beach because the sampling baskets were placed in the
beaches at different times prior to fertilizer application (the result of logistical complications).
Therefore, starting oil compositions were based on the hydrocarbon decay rate of the untreated
Control beach and extrapolated from the data obtained between the time of basket placement and
initiation of fertilizer application. The responses are quite interesting. On all beaches large decreases
in the nC18/phytane ratios were observed (Figure 9.19), strongly suggesting effects due to
biodegradation (see Section 8). Table 9.2 gives the mean nC18/phytane ratios at each sampling time
and compares statistical differences between beaches. Initial ratios were not statistically different
from each other. Thus, even though the sampling baskets on each beach started with a different oil
concentration, the oil was equally degraded in all cases.
Changes in the nC18/phytane ratios from treated beaches was quite dramatic—decreasing an
order of magnitude on the Bath beach in the first 7 days, and almost an order of magnitude on the
Sprinkler beach in the first 17 days. The greater change on the Bath beach was statistically
significant, as were the changes on the treated beaches relative to the untreated control. The extent
387
-------
10,000
1,000
100
i
c
10
Figure 9.7.
I 1 , p
22Jun90 04JUI90 16Jul90 28Jul90 OQAugQO
Sampte Cotoctton Do*
o o o CanM Plot a a a Sprtttor Plot
Change in the nC18 Concentration Over Time for the Oiled
Subsurface Layer for All Beaches at Elrington Island
(Dashed Lines Extrapolate Control Plot Decay Rate for
Treated Beaches to Date of Fertilizer Application).
10,000
t 1,000
5
•8
100
10
A
A.
22Jun90 04JU90 16JU90
Swnpto Qotadton
2BJUI90
OOAugW
oooOonMnot a D a SprhMir Ptal
Figure 9.8.
Change in the nC19 Concentration Over Time for the Oiled
Subsurface Layer for All Beaches at Elrington Island
(Dashed Lines Extrapolate Control Plot Decay Rate for
Treated Beaches to Date of Fertilizer Application).
388
-------
10.000
1.000
8
100
10
a..
Figure 9.9.
22Jun90 04JUI90 16Jul90 28 Jul 90
Sample Cotectton Date
OOAuggo
A A Bath Plot ooo Control Plot ODD Sprinter Plot
Change in the nC20 Concentration Over Time for the Oiled
Subsurface Layer for All Beaches at Elrington Island
(Dashed Lines Extrapolate Control Plot Decay Rate for
Treated Beaches to Date of Fertilizer Application).
10,000
1,000
5
T5
8
c
100
10
A
A-
22Jun90 04 Jul 90 16 Jul 90 28 Jul 90 09Aug90
Sample Cotection Date
o o o Control Plot n o a SprtnMer Plot
Figure 9.10. Change in the nC21 Concentration Over Time for the Oiled
Subsurface Layer for All Beaches at Elrington Island
(Dashed Lines Extrapolate Control Plot Decay Rate for
Treated Beaches to Date of Fertilizer Application).
389
-------
10,000
1,000
5
•5
!
100
10
22Jun90 04JU90 WJU90 28Jul90 OOAugW
Sample Cofecdon MB
oooConWPtat a a a SprtnMv Plot
Figure 9.11. Change in the nC22 Concentration Over Time for the Oiled
Subsurface Layer for All Beaches at Elrington Island
(Dashed Lines Extrapolate Control Plot Decay Rate for
Treated Beaches to Date of Fertilizer Application).
10,000
1 1,000
8
•8
8
100
10
22Jun90 04JUI90 16JU90 28JJ90 09Aug90
Sampto Cotectfon Dote
oooOonMPtat a a o SprtrMv Plot
Figure 9.12. Change in the nC23 Concentration Over Time for the Oiled
Subsurface Layer for All Beaches at Elrington Island
(Dashed Lines Extrapolate Control Plot Decay Rate for
Treated Beaches to Date of Fertilizer Application).
390
-------
10,000
1.000
5
•8
f 100
10
22 Jin 90 04JU90 16JU90 28 Jd 90 OOAugQO
Sampto Cotocdon Date
A A Uh Plot
oooOortraiPtot a a a SpiWder Plot
Figure 9.13. Change in the nC24 Concentration Over Time for the Oiled
Subsurface Layer for All Beaches at Elrington Island
(Dashed Lines Extrapolate Control Plot Decay Rate for
Treated Beaches to Date of Fertilizer Application).
10,000
; 1,000
I
3
"§ 100
22Jun90 04JUI90 16JUI90 28Jul90 OOAugOO
Sampto Cotocdon Date
o o o Corftd PW o a a Sprtttar Plot
Figure 9.14. Change in the nC25 Concentration Over Time for the Oiled
Subsurface Layer for All Beaches at Elrington Island
(Dashed Lines Extrapolate Control Plot Decay Rate for
Treated Beaches to Date of Fertilizer Application).
391
-------
10,000
1,000
5
T5
I
c
100
10
a-
22Jun90 04JUI90 16JU90 2BJul90 OOAugQO
Swnpto Cotocdon Mb
oooOonWPM a a o SprinMar Ptot
Rgure 9.15. Change in the nC26 Concentration Over Time for the Oiled
Subsurface Layer for All Beaches at EIrington Island
(Dashed Lines Extrapolate Control Plot Decay Rate for
Treated Beaches to Date of Fertilizer Application).
10,000
1 1.000
S
100
10
I I I I '
22JW90 04JU90 WJU90 2BJU90 OQAugW
SamptoCofecHonDato
o o o CorM Ptal a a a SprtaMv PW
Rgure 9.16. Change in the nC27 Concentration Over Time for the Oiled
Subsurface Layer for All Beaches at EIrington Island
(Dashed Lines Extrapolate Control Plot Decay Rate for
Treated Beaches to Date of Fertilizer Application).
392
-------
100,000
3
c
10,000
c 1,000
ts
100
22Jun90 04JUI90 16JU90 28 Jd 90
Sample Cofccfen Data
OQAugQO
& A A Ban Plot
o o o Control Plot a o a SprtnMv Plot
Figure 9.17. Change in the Sum of the Alkane Concentration for nC18 to
nC27 Over Time for the Oiled Subsurface Layer for All
Beaches at Elrington Island (Dashed Lines Extrapolate
Control Plot Decay Rate for Treated Beaches to Date of
Fertilizer Application).
10,000
1,000
5
•8
100
10
22Jun90 04JUI90 16 JJ 90 28Ji490 09Aug90
Sample Cofecflon Date
A A Bah Plot ooo Control Plot ana SpnnMer Ptat
Figure 9.18. Change in the Phytane Concentration Over Time for the
Oiled Subsurface Layer for All Beaches at Eirington Island
(Dashed Lines Extrapolate Control Plot Decay Rate for
Treated Beaches to Date of Fertilizer Application).
393
-------
10
1
I
c
0.1
0.01
22JUR90 04JUI90 16JU90 2BJUI90
Sampte Cdectton Date
oooOonMPW o a a SprinMv Plot
09Aug90
Figure 9.19. Change in the nC18/phytane Ratio Over Time for the Oiled
Subsurface Layer for All Beaches at Elrington Island
(Dashed Lines Extrapolate Control Plot Decay Rate for
Treated Beaches to Date of Fertilizer Application).
394
-------
ELRINGTON ISLAND
TABLE 9.2. ANALYSIS OF BEACH PLOT DIFFERENCES BY DATE FOR ELRINGTON
ISLAND BASED ON nC18/PHYTANE RATIO
Beach
Mean
Std. Dev.
Comparison
t-value
Sig. level
0.95 0.90
Date: Pretreatment
Bath 1.506 0.068
Sprinkler 1.548 0.039
Control 1.678 0.118
Date: 11-July-1990
Bath 0.180 0.017
Sprinkler 0.477 0.060
Control 0.923 0.096
Date: 21-July-1990
Bath 0.080 0.057
Sprinkler 0.265 0.035
Control 0.650 0.028
Date: 30-July-1990
Bath 0.145 0.092
Sprinkler 0.230 0.071
Control 0.415 0.078
Date: 07-August-1990
Bath 0.210 0.028
Sprinkler 0.255 0.007
Control 0.435 0.078
Degrees of freedom: 8
Spr.-Bath 0.536
Control-Bath 1.263
Control-Spr. 1.046
Degrees of freedom: 4
Spr.-Bath 4.763
Control-Bath 7.621
Control-Spr. 3.940
Degrees of freedom: 2
Spr.-Bath 2.766
Control-Bath 8.976
Control-Spr. 8.590
Degrees of freedom: 2
Spr.-Bath 0.731
Control-Bath 2.239
Control-Spr. 1.754
Degrees of freedom: 2
Spr.-Bath 1.559
Control-Bath 2.715
Control-Spr. 2.298
*
*
*
*
*
*
*
*
*
*
*
Critical values for a one-sided t-test at 95% and 90% confidence levels.
n aloha (0.05) aloha (0.10)
8
4
2
1.860
2.132
2.920
1.397
1.533
1.886
395
-------
ELRINGTON ISLAND
of change in the ratios also seemed to maximize at different times—the Bath beach maximized at the
first sampling, the Sprinkler beach at the second sampling, and the untreated Control beach at the
third sampling. At the third sampling, ratios were statistically indistinguishable at the 95% confidence
level. Thus biodegradation of the alkanes was greatly enhanced by the fertilizers and ultimately a
single pulse of fertilizer at the higher nutrient concentration produced the best effect.
We will assume for further discussion that changes in other hydrocarbons are also caused by
biodegradation. Similar extents of biodegradation are reflected in the data shown in Table 9.3 and
9.4. At the 95% confidence interval, starting concentrations of nC18 and the summed alkanes, nC18
to nC27, were not statistically different, again suggesting similar extents of initial alkane degradation
in oil from sampling baskets. As with the nC18/phytane ratio, hydrocarbon concentrations decreased
very rapidly on treated beaches relative to the untreated control, and the single pulse application was
more effective than multiple pulse applications. However, the effect of the Bath application appeared
to maximize early, thus allowing the effect of sprinkler application to catch up—that is, by the last
sampling, hydrocarbon concentrations on both beaches were not statistically different.
Comparing the biodegradation rate of individual hydrocarbons was complicated, due to large
differences in the decay curve response. Similar types of hydrocarbon decay curves were seen in the
Passage Cove study. Rates of biodegradation from the Elrington Island study appeared first-order
only for the untreated Control beach. Treated beaches were decidedly mixed-order, indicating that
biodegradation rate constants were becoming limited by more than just substrate. The much more
slowly degraded phytane was an exception. It appeared first-order on all beaches. We believe that
the mixed-order response on the treated beaches was probably caused by an extremely rapid
degradation of the initial normal alkane concentration in the first 10 to 15 days following fertilizer
application, with a slower degradation once the hydrocarbon concentrations approached detection
limits. That is, by the first sampling and certainly by the second sampling, hydrocarbon
concentrations had been reduced in most cases approximately ten-fold, bringing them very close to
detection limits. Concentrations reported below 100 fig/g are based on very small peaks within the
gas chromatography profiles and may not necessarily be the hydrocarbon in question. The ability to
see further significant changes in concentration was unlikely, due to the small remaining
concentrations.
396
-------
ELRINGTON ISLAND
TABLE 9.3. ANALYSIS OF BEACH PLOT DIFFERENCES BY DATE FOR ELRINGTON
ISLAND BASED ON nC18 LINEAR ALKANE
Beach
Mean
Std. Dev.
Comparison
t-value
Sig. level
0.95 0.90
Date: Pretreatment
Bath 3.370 0.097
Sprinkler 3.412 0.053
Control 3.268 0.068
Date: ll-July-1990
Bath 2.129 0.042
Sprinkler 2.589 0.043
Control 2.869 0.045
Date: 21-July-1990
Bath 1.486 0.326
Sprinkler 2.185 0.100
Control 2.672 0.026
Date: 30-July-1990
Bath 1.644 0.169
Sprinkler 1.972 0.018
Control 2.399 0.170
Degrees of freedom: 8
Spr.-Bath 0.380
Control-Bath -2.861
Control-Spr. -1.670
Degrees of freedom: 4
Spr.-Bath 7.570
Control-Bath 12.022
Control-Spr. 4.579
Degrees of freedom: 2
Spr.-Bath 2.050
Control-Bath 3.627
Control-Spr. 4.713
Degrees of freedom: 2
Spr.-Bath 1.930
Control-Bath 3.150
Control-Spr. 2.498
Critical values for a one-sided t-test at 95% and 90% confidence levels.
n aloha (0.05) aloha (0.10)
8
4
2
1.860
2.132
2.920
1.397
1.533
1.886
*
*
*
*
*
*
Date: 07 -August- 1990
Bath
Sprinkler
Control
1.757
1.843
2.367
0.040
0.062
0.176
Degrees of freedom: 2
Spr.-Bath
Control-Bath
Control-Spr.
1.166
3.380
2.808
-
* *
*
397
-------
ELRINGTON ISLAND
TABLE 9.4. ANALYSIS OF BEACH PLOT DIFFERENCES BY DATE FOR ELRINGTON
ISLAND BASED ON THE SUM OF THE LINEAR ALKANES FROM nC18 TO nC27
Beach
Mean
Std. Dev.
Comparison
t-value
Sig. level
0.95 0.90
Date: Pretreatment
Bath 4.205 0.097
Sprinkler 4.281 0.068
Control 4.120 0.067
Date: 11-July-1990
Bath 3.118 0.039
Sprinkler 3.476 0.036
Control 3.733 0.050
Date: 21-July-1990
Bath 2.810 0.108
Sprinkler 3.155 0.081
Control 3.585 0.008
Date: 30-July-1990
Bath 2.736 0.070
Sprinkler 3.006 0.005
Control 3.328 0.123
Degrees of freedom: 8
Spr.-Bath -0.642
Control-Bath -0.721
Control-Spr. -1.687
Degrees of freedom: 4
Spr.-Bath 6.745
Control-Bath 9.699
Control-Spr. 4.170
Degrees of freedom: 2
Spr.-Bath 2.556
Control-Bath 7.156
Control-Spr. 5.283
Degrees of freedom: 2
Spr.-Bath 3.847
Control-Bath 4.183
Control-Spr. 2.616
Critical values for a one-sided t-test at 95% and 90% confidence levels.
n aloha (0.05) aloha (0.10)
8
4
2
1.860
2.132
2.920
1.397
1.533
1.886
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
Date: 07 -August- 1990
Bath
Sprinkler
Control
2.729
2.832
3.303
0.007
0.035
0.139
Degrees of freedom: 2
Spr.-Bath
Control-Bath
Control-Spr.
2.886
4.124
3.286
-
*
*
*
*
•
398
-------
ELRINCTON ISLAND
In the untreated Control beach, biodegradation remained first-order throughout the test period,
presumably because degradation was slower. Concentrations of hydrocarbons never reached detection
limits during the duration of the test, and thus decay rates could be calculated based on all the
sampling periods. The same can also be said for phytane.
For comparative purposes, decay rates for all three test beaches were examined only during the
first sampling period (day 0 to 7). Table 9.5 shows the estimated rates for nC18, the summed alkanes
nC18 to nC27, and phytane. The difference between treated and untreated beaches was
approximately a factor of three. This is considerably different from the residue weight results in
which differences were 6 to 7. Interestingly, the data from Passage Cove also showed much less of
a difference in rate of hydrocarbon composition change, compared to changes in residue weight,
between treated and untreated beaches. However, Passage Cove rates were approximately 50 to 100
times slower (compare with Table 7.6).
TABLE 9.5. ESTIMATED DECAY RATES IN MG/KG/DAY FOR SELECTED HYDROCARBONS
BASED ON CONCENTRATION CHANGES BETWEEN t = 0 and t « 7 DAYS
nC18
Summed Alkanes
(nC18 to nC27)
phytane
Untreated
Control Beach
1.3
9.0
0.010
Sprinkler Beach
3.8
28.0
0.037
Bath Beach
3.1
20.3
0.035
The higher absolute rates of compositional change on the Sprinkler beach relative to the Bath
beach differed from the results above. This result was generated by normalization to oil residue
weights on each beach, and shows that percent change in hydrocarbon composition was much greater
at the Sprinkler beach.
399
-------
ELRINGTON ISLAND
In summary, changes in hydrocarbon composition in samples taken from the treated and
untreated beaches at Elrington Island reflected the same trends in biodegradation enhancement as seen
with measures of the residue weight changes. Fertilizer addition enhanced oil biodegradation
considerably. This enhancement was more apparent in measures of oil residue weight than measures
of hydrocarbon compositional change. In addition, the effect of the single pulse application was less
effective than the multiple pulse applications, when gauged with changes in hydrocarbon composition
normalized to oil residue weights. Thus, multiple pulse applications of the fertilizer solutions led to
a greater percent degradation in hydrocarbon composition.
Interestingly, the biodegradation of the normal alkanes was apparently not directly coupled to
decreases in oil residue weight; reduction in oil residue continued long after most of the normal
alkanes had been degraded. Phytane degradation actually provided a better correlation with changes
in oil residue weight.
MICROBIOLOGY
Numbers of Oil-Degrading Bacteria
Numbers of oil degraders were assessed using the MPN sheen screen test developed by
researchers at the University of Alaska, Fairbanks, and the ADEC Oil Spill Response Support
Laboratory in Valdez. Results are shown in Table 9.6. As observed in other studies in Prince William
Sound, the number of oil degraders was approximately 106 to 106 cells per gram of beach material and
were quite variable. There was no consistent effect of the fertilizer application on the degrader
numbers. The large numbers of oil degraders in the oiled layer, however, suggests that the high
concentration of oil in the beach material does not prevent colonization by oil-degrading
microorganisms. Determination of total culturable heterotrophic bacteria was also performed on the
oiled samples, and the oil degraders represented approximately 1 to 10% of this population.
400
-------
ELRINCTON ISLAND
TABLE 9.6. NUMBER OF OIL DEGRADERS (MPN'S X 104) IN SAMPLES
FROM ELRINGTON ISLAND
Beach
Untreated
Control
Bath
Sprinkler
Basket Layers*
Interface (Top)
Oiled
Interface (Bottom)
Interface (Top)
Oiled
Interface (Bottom)
Interface (Top)
Oiled
Interface (Bottom)
July 1 1
18.1
14.6
14.0
11.4
11.5
0.9
N.R.b
15.2
14.1
July 21
7.7
18.5
5.3
11.3
3.4
134.3
17.5
13.9
11.6
July 30
88.3
18.9
182.2
5.2
256.2
163.7
13.7
183.2
316.1
August 7
23.0
357.4
32.3
17.5
52.7
228.6
34.2
35.3
83.1
,"Oiled = Oiled layer; Interface = Material between the oiled and unoiled layer, top and bottom of the
oiled layer.
bN.R. » Not reported.
Microbial Activity
In addition to chemical analysis of the samples taken from the sampling baskets, measures of
microbial activity were also performed using several approaches. In one case, 100 g beach samples
were placed in special biometer flasks, and the total amount of carbon dioxide produced over a three-
day period was measured. No nutrients or fertilizer were added to the flasks. CO2 production
measured in contaminated beach material taken from sampling baskets removed from the beach on
July 11, July 21, July 30, and August 7 (10, 20, 29 and 37 days after fertilizer application
respectively), is shown in Figures 9.20a-d. Cumulative CO3 production was approximately linear over
the 72-hour incubation period. The differences in total CO2 production mimic the oil chemistry
results: the Bath beach showed the highest activity but only slightly greater than the Sprinkler beach.
Both treated beaches were substantially more active than the untreated control.
401
-------
210 ^
180
150 \
120
90
60
30
12 24
36 46
Hours
60 72
84
Plot identification A A n Bath Plot o o o Control Plot a a a Sprinkler Plot
Rgure 9.20a. Change In the CO2 Concentration Over Time for all Beaches
at Elrlngton Island on July 11,1990.
402
-------
2101
180
150
120
90
GO
30
0
6
o
a
D
12 24 36 48 60 72
Hours
Plot identification A A A Bath PkX o o o Control Plot ODD Sprinkler Plot
84
Figure 9.20b. Change In the CO2 Concentration Over Time for all Beaches
at Elrlngton Island on July 21,1990.
403
-------
210
180
150-
120
90
60
30
0
12 24 36 48 60 72 84
Hours
latntmcition a i a 8*ri Plot ooo Control Plot 003 Sprrtdw Plot
Figure 9.20c. Change In the CO, Concentration Over Time for all Beaches
at Elrington Island on July 30,1990.
210
180
150
120
90
60
30
0
12
24
36 48
Hours
60
72
Btm PW o c 3 Control Plot ooo SprinMor Plot
84
Figure 9.20d. Change In the CO2 Concentration Over Time for all Beaches
at Elrington Island on August 7,1990.
404
-------
ELRINGTON ISLAND
Table 9.7 summarizes the rates of CO2 production (slopes of regression lines fit to the data
points) for the four sampling periods. The substantial CO3 production reflects the vigor with which
biodegradation occurred, and strongly supports biodegradation as the primary factor affecting changes
in oil chemistry. Comparison of the mineralization rates between beaches at t = 0 (data not shown)
indicated that rates were very similar, suggesting that preparation of the beach material for the
sampling baskets on different days resulted in excellent homogeneity in initial microbial activity.
With time, the difference in mineralization rates (CO2/day/100 g of beach material) between the
fertilizer-treated beaches and the untreated control increased, and were statistically significant at the
95% confidence level. Note the mineralization rates on the third sampling (July 30) for all three
beaches appear to be unusually high, and the high mineralization rates on the Sprinkler beach seemed
to be maintained through the final sampling. This was not true for the Bath beach where rates
appeared to drop to untreated Control beach levels. It is possible that the effect of the single pulse
fertilizer application had waned on the Bath beach.
TABLE 9.7. MINERALIZATION RATES FROM TOTAL CO2 PRODUCTION
(H MOLES/100 g/hour) AND CALCULATED OIL DEGRADATION (mg/kg/day)a
IN OILED BEACH MATERIAL TAKEN FROM SAMPLING BASKETS ON ELRINGTON ISLAND
Sampling
Date
7/11
7/21
7/30
8/7
Untreated
Control Beach Sprinkler Beachb
CO2 Oil CO3 Oil
5.69 16.4 5.89 16.9
5.47 15.7 8.73 25.1
7.58 21.8 11.55 33.3
5.41 15.6 7.29 20.9
Bath Beachc
CO2 Oil
6.53 18.8
7.96 22.9
8.31 23.9
5.62 16.1
a Based on oil as hexadecane or 85% carbon.
b Fertilizer application occurred on 7/2, 7/6, 7/11, 7/22, 7/27, and 8/3, 1990.
c Fertilizer application occurred on 7/1/90.
405
-------
ELRINGTON ISLAND
When the CO2 production rates are converted to oil carbon degraded, values for the untreated
Control beach were generally in close agreement with changes in the oil residue weight (Table 9.1).
This suggests that oil biodegradation was largely responsible for the changes in residue weight.
Moreover, the carbon turnover rates were almost twice as fast as that for the degradation of the
summed alkanes (Table 9.5). To account for this additional CO2, one would have to assume that
natural biodegradation of the oil involves more than just metabolism of the straight chain alkanes.
In a similar manner carbon turnover rates, as determined by CO2 mineralization in samples from
the fertilizer-treated beaches, were approximately one-fifth the rates for changes observed for the
oil residue weights. On the other hand, they are in the same range as the initial degradation rates for
the summed alkanes. It is concluded at this time that fertilizers significantly enhanced biodegradation
of the alkane fraction of the oil, but may have had a disproportionate positive effect on the
degradation of total oil residues. Thus, as argued above, extensive biodegradation of the oil leads to
a change in oil consistency and stickiness that then allows degraded residues to be scoured from the
beach material by tidal action.
A second type of measurement involved the mineralization of 14C labeled phenanthrene, a
substrate indicative of the biodegradation of PAHs in the oil. The beach material used for the
measurement of total carbon dioxide production in the biometer flasks was amended with the labeled
substrate (see Section 4, Methods) and the production of radiolabeled CO2 was measured over a four
day period. Figures 9.21a-c show that phenanthrene biodegradation was immediate and extensive,
indicating active PAH biodegradation in the subsurface beach material. CO2 production curves are
hyperbolic, indicating depletion of the substrate with time. In all cases the greatest extent of
phenanthrene mineralization occurred in samples from the Bath beach, followed by the Sprinkler and
untreated Control beaches, respectively. Thus, fertilizer application also significantly enhanced PAH
biodegradation. With time, the difference between samples from treated and untreated beaches
increased. This further supports the effectiveness of a pulse application of fertilizer, and suggests
that enhanced biodegradation at the Bath beach (higher concentration of nutrients) was due to the
amount of fertilizer applied at one time, as compared to more frequent applications at lower
concentrations.
406
-------
35 -
30 -
25 -
05
C, 20 -
15 -
10
5 -
12
24
36
48
60
12
84
96
108
TIME (Hours)
D Sprinkler + Untreated Control O Bath
Figure 9.21 a. Production of Radiolabeled CO2 from Phenanthrene Through
Time in Oiled Beach Samples Taken From the Sampling
Baskets (Layer 3) on 7/11/90, for the Sprinkler, Untreated
Control, and Bath Beaches at Elrington Island.
407
-------
t
S
96
TIME (Hours)
D Sprinkler + Untreated Control O Bath
Figure 9.21 b. Production of Radiolabeled CO2 from Phenanthrene Through
Time in Oiled Beach Samples Taken From the Sampling
Baskets (Layer 3) on 7/21/90, for the Sprinkler, Untreated
Control, and Bath Beaches at Elrington Island.
408
-------
0.
o
60 -
s=- 50 J
40 -
30 -
20 -
DBalh
TIME (Hours)
+ Untreated Control
O Sprinkler
Figure 9.21 c. Production of Radiolabeled CO2 from Phenanthrene Through
Time in Oiled Beach Samples Taken From the Sampling
Baskets (Layer 3) on 7/30/90, for the Sprinkler, Untreated
Control, and Bath Beaches at Elrington Island.
409
-------
ELRINGTON ISLAND
Mineralization of I4C-hexadecane was measured in the same beach samples by the ADEC Oil
Spill Response Support Laboratory in Valdez. The mineralization data are shown in Figure 9.22 (data
for samples taken prior to the fertilizer application unfortunately are not available). Although the
percent hexadecane mineralized was relatively low, the results again show the effect of a single pulse
application with a high concentration of fertilizer. On the first sampling date (10 days after fertilizer
application) mineralization activity was approximately three times greater on the Bath beach than the
other beaches. The mineralization activity on the Sprinkler beach (two fertilizer applications) was
essentially the same as the untreated control. Ten days later, the Sprinkler beach (three fertilizer
applications) was showing increased activity, while activity on the Bath beach had decreased. By the
end of the test period, activity on the Sprinkler beach (six fertilizer applications) was the highest
observed overall, and the Bath beach was essentially the same as the untreated Control beach.
This information for the Sprinkler beach differs with observed changes in the oil composition
(Figures 9.7-9.19). In the latter case, the straight chain alkanes were rapidly degraded during the
initial 20 days of the test on the Sprinkler beach, suggesting that highest mineralization should have
occurred in the initial part of the test. The fact that it did not may be a function of isotope dilution;
that is, initially the bacteria populations utilized more of the unlabeled hexadecane in the oil. On the
Bath beach, hydrocarbon degradation must have occurred so rapid initially that isotope dilution was
attenuated, giving higher initial mineralization rates of 14C-hexadecane. Such an attenuation could
only be explained by very rapid increases in biomass specific activity, giving basically zero-order
degradation kinetics.
Dissolved Oxygen and Nutrient Uptake
Specialized sampling baskets were placed in the Elrington beaches to allow measurements of
oxygen uptake as seawater passed through the oiled layer of beach material in treated and untreated
beaches (see Section 4). Two sampling wells were placed in the baskets; one well trapped incoming
tidal water before it passed through the oiled layer, while the other trapped water after it passed
through the oiled layer. With dissolved oxygen probes in each well, the concentration of oxygen was
monitored as the tidal water moved up and through the bottom of the basket and eventually over the
top of the wells.
410
-------
08/07/90
07/30/90
07/21/90
07/11/90
Legend
Sprinkler
Untreated Control
Bath
Median % CO2
Figure 9.22. Median Percent Total CO2 Production (Triplicate Samples)
Over Time for the Sprinkler, Untreated Control, and Bath
Beaches at Elrlngton Island.
411
-------
ELRINGTON ISLAND
Changes in dissolved oxygen concentrations shortly after fertilizer application are shown in
Figures 9.23 to 9.25, and approximately 30 days later in Figures 9.26 to 9.28. Despite the crudeness
of the in situ test basket, the effect of fertilizer-enhanced biodegradation in the subsurface oil layer
could be observed by the differential oxygen uptake, comparing the change in oxygen levels from the
time water entered the wells to its lowest level after that time. Tidal water passing through the oil
layer had a greater oxygen demand than water isolated from it, and the demand was only apparent
in the fertilizer-treated beaches. The general absence of oxygen uptake in the untreated control as
water passed through the oil layer is interesting, since the oil chemistry results showed that some
biodegradation was occurring in the sampling baskets on this beach. Since measurements of oxygen
uptake were performed 4 to 6 days after initial fertilizer application, the response to the added
nutrients was quite rapid. Furthermore, similar degrees of oxygen uptake also occurred 30 days later,
suggesting the effect of fertilizer application was sustained, particularly in the single pulse application
on the Bath beach. This data provides another strong indication that fertilizer-enhanced changes in
oil chemistry are linked to enhanced microbial activities in the beach subsurface.
Samples of seawater were also taken from the in situ test baskets to determine if nutrient uptake
accompanied the dissolved oxygen uptake. The samples were taken just before the wells were totally
filled by the tide, and the concentrations of ammonia, nitrate, and phosphate were measured. If
nutrients were truly linked to enhanced biodegradation of the oil, water that passed through the oiled
layer should have shown fewer nutrients compared to water that did not pass through the oiled layer.
These data are shown in Table 9.8. Any value with a ratio greater than one indicates that lower
nutrient concentrations (NH4, PO4, NO3, and NO2) were found in the wells receiving water that had
passed through the oiled layer. As a result of passing through the oiled layer, nutrient uptake
occurred on all beaches. However, there were several cases where uptake on the fertilizer-treated
beaches appeared to be greater than the untreated control, particularly for the July 2, 1990 ammonia
analysis on the Bath beach (the Bath beach was only sampled twice because a major storm dislodged
the in situ basket), and the July 26,1990 and August 4,1990 ammonia analysis on the Sprinkler beach.
Some of the same trends were seen with the nitrate analysis. It is concluded that there is some
evidence of a connection between enhanced oil degradation and nutrient uptake. This uptake seems
to be related to the large pulse of fertilizer added with the Bath application, and possibly to the
accumulated benefit of several applications at low nutrient concentrations on the Sprinkler beach.
In several cases high ratios were obtained, and the highest ratios occurred on the treated beaches.
412
-------
13
12
11
^ 10
/
-w
H3 O-O
t/
water enters bottom
of well
O Deep well
• Shallow well
0 10 20 30 40 50 60
TIME (min)
Figure 9.23. Change in Dissolved Oxygen Concentration Over Time for
the Untreated Control Beach at Elrington Island on July 6,
1990.
water enters bottom
of well
O Deep well
• Shallow well
40 50 60 70
TIME (min)
Figure 9.24. Change in Dissolved Oxygen Concentration Over Time for
the Sprinkler Beach at Elrington Island on July 3,1990.
413
-------
O Deep well
• Shallow well
A-H2O enters bottom of deep well
B=H2O enters bottom of shallow well
0 10 20 30 40 50 60 70 80 90
TIME (min)
Figure 9.25. Change in Dissolved Oxygen Concentration Over Time for
the Bath Beach at Elrington Island on July 7,1990.
12
11
10
9|
8
7
6
5
4
3
2
1
O Deep well
• Shallow well
enters bottom of shallow well
enters bottom of deep well
0 10 20 30 40 50 60 70
TIME (min)
Figure 9.26. Change in Dissolved Oxygen Concentration Over Time for
the Untreated Control Beach at Elrington Island on August 3,
1990.
414
-------
13
12
11
10
8
O Deep well
• Shallow well
enters bottom of deep well
enters bottom of shallow well
0 10 20 30 40 50 60 70
TIME (min)
Figure 9.27. Change in Dissolved Oxygen Concentration Over Time for
the Sprinkler Beach at Elrington Island on August 5,1990.
12
11
f 10
a
§ .
8
O Deep well
• Shallow well
A=H2O enters bottom of deep well
B=H2O enters bottom of shallow well
0 10 20 30 40 50 60
TIME (min)
Figure 9.28. Change in Dissolved Oxygen Concentration Over Time for
the Bath Beach at Elrington Island on August 3,1990.
415
-------
ELRINCTON ISLAND
TABLE 9.8. RATIO OF NUTRIENT CONCENTRATIONS BEFORE AND AFTER EXPOSURE
TO OILED BEACH MATERIAL IN SPECIAL SAMPLING BASKETS
FROM ELRINGTON ISLAND BEACHES
Beach
Sampling Date
NH.
PCX,
NO,
NO,
Control
Control
Control
Control
Control
Control
Bath
Bath
Sprinkler
Sprinkler
Sprinkler
Sprinkler
Sprinkler
Sprinkler
Sprinkler
June 30
July 4
JulyS
July 14
July 26
August 4
June 30
July 2
June 30
July3
July 4
JulyS
July 14
July 26
August 4
1.33
1.00
1.50
1.08
0.53
0.78
0.80
11.20
2.00
0.86
1.06
1.06
1.25
7.56
3.11
0.74
1.17
1.43
1.50
0.35
1.22
1.13
3.10
0.86
0.96
1.26
0.88
2.00
0.93
0.84
0.45
1.20
1.00
2.00
1.37
NDA»
0.08
8.50
1.94
0.95
1.81
0.37
84.00
4.35
1.17
2.00
1.00
1.00
0.50
0.80
1.25
0.50
4.50
1.00
1.25
0.67
0.50
7.00
1.38
1.20
aNDA = No data available
MICROCOSM STUDIES
For this project, flow-through aquatic microcosms provided the opportunity to study
bioremediation in a more realistic simulated field setting than shake flasks. Although a shake flask
may indicate maximized effects of nutrient additions by increasing oiled beach material contact with
a homogeneous solution, the flow-through microcosms are better at simulating field conditions by
allowing exposure of a column of beach material to two simulated tidal cycles per day, each slowly
rising from the bottom and falling from the top.
The primary objective of the flow-through microcosm studies was to test the effect of nutrients
on oil biodegradation for comparison with field sprinkler application (Elrington Beach) and laboratory
(shake-flask) test results. The systems were used to determine the difference in oil degradation
between simulated "high" and "low" tide conditions.
416
-------
ELRINGTON ISLAND
TOC analyses indicated no significant products of incomplete degradation or bioemulsification
in the effluent water from either treatment. Figure 9.29 indicates the cumulative carbon dioxide
produced by duplicate columns, after subtracting background values from control columns 1 and 2.
Both the "bath" and the "sprinkled" nutrient treatments show accelerated carbon dioxide production
over time compared to the no nutrient microcosm CO2 production rate.
Tidal cycle phase did not seem to have a significant effect on average daily oil mineralization,
as indicated in Figure 9.30. This figure also indicates that replicate columns generated approximately
the same quantity of carbon dioxide. The "bath" nutrient treatment appeared to enhance oil
mineralization by approximately 2-fold over the control without nutrients, and the "sprinkled"
treatment by about 3-fold (Figure 9.31). The relative difference in the mineralization rates was
greater than that seen in samples taken directly from the beach (Table 9.6). In addition, the total CO2
production in the microcosms was only one fourth of that seen in samples taken directly from the
beaches. These differences could be due to the absence of turbulence in the microcosm; the flask tests
were operated with slow mixing, which would have produced some turbulence.
Oil residues appeared to decline somewhat over the 2-week test period. Samples of oiled beach
material at the beginning of the test contained 6363 mg oil/kg (n = 3, s.d. - 168). At the end of the
test, control columns 3 and 4 contained 6667 mg oil/kg beach material (n = 2, s.d., = 0), "bath" treated
columns 5 and 6 had 6129 mg/kg (n = 2, s.d = 81) and the "sprinkled" treatment columns 7 and 8
contained 5284 mg oil/kg (n = 2, s.d. = 907). This gave a rate of oil residue weight change for the
sprinkled treatment (83 mg/kg/day) that was in the same range as seen in the field (110 mg/kg/day,
Table 9.1). If oil residue weight change is a good indicator of oil degradation, then we would surmise
from these microcosm studies that oil carbon was either being converted to bacterial biomass or to
degraded oil residues (at least initially). Since TOC of microcosm effluents was not elevated, we
would conclude that the biomass and oil residues remained associated with the oiled beach material.
In the field, the turbulence of tidal action could have caused the biomass and oil residues to be
sloughed out of the beach material.
Effect of Oil Concentration
The high rates of oil degradation on the treated Elrington beaches (>100 mg oil residue/kg/day)
compared to the previous summer in Passage Cove (10 to 20 mg oil residue/kg/day) raised the
question of the effect of oil mass on the biodegradation rate; that is, does a direct relationship exist
417
-------
No Nutrients
Nutrltnt Bath
Nutrients Sprinkled
7
DAY
•Arrows: NutrtontApplications
Figure 9.29. Cumulative Mineralization of Oil Carbon from Flow-Through
Column Microcosms.
418
-------
CO
•a
Em
•o
Q)
7 8
Nutrlants
Sprinkled
Treatment
Figure 9.30. Average Dally Oil Carbon Mineralization During High and
Low Tidal Cycles of Replicate Column Microcosms.
419
-------
AVERAGE CO2 PRODUCTION PER DAY
8
6
4
2.149
No
Nutrients
4.828
Nutrient
Bath
6.816
Nutrients
Sprinkled
Treatment
Figure 9.31. Average Dally Oil Carbon Mineralization in Column
Microcosms for Each Nutrient Treatment.
420
-------
ELRINCTON ISLAND
between higher oil concentrations (i.e., Elrington Beach) and faster oil degradation rates. To test this
idea, samples of oiled beach material were randomly collected from Elrington Island, placed in the
biometer test flasks, and total CO2 production measured. The contents of the flasks were flooded and
drained of seawater amended with N and P nutrients every 12 hours to simulate the effect of tides.
By randomly selecting samples from the beach, it was assumed that oil concentrations would be quite
heterogeneous and, thus, the mineralization rate, as a measure of the oil degradation rate, could be
correlated with the different oil residue weights.
Figure 9.32 shows that mineralization was linear and did indeed vary. However, if the
mineralization rate is normalized to the total oil residue (Table 9.9), then there appears to be little
reduction in the variability of the rates. Thus, rates are not affected by initial oil weights. It also
appears that biodegradation of the oil is not first-order with respect to oil mass and that many other
factors limit the rate. This is reasonable if one considers that as a hydrophobia substance, oil will not
be totally available to the oil degrader biomass at any point in time.
TABLE 9.9. RELATIONSHIP BETWEEN OIL MINERALIZATION RATES
AND OIL CONCENTRATION IN BEACH MATERIAL FROM ELRINGTON ISLAND
Sample
Designation
1
2
3
4
5
6
7
Mineralization
Rate (ppm/day)
71.88
59.55
56.00
55.00
54.00
51.88
38.13
Residue Weight
of oil (mg)
0.49
0.65
0.19
0.28
0.46
0.19
0.48
Normalized Rate
(ppm CO2/mg oil/day)
146.7
91.6
294.7
196.4
117.4
273.1
79.4
WINTER SAMPLING
On December 7, 1990, Elrington Island was sampled a final time. We were unaware that Exxon
had applied INIPOL and CUSTOMBLEN to the test beaches at the end of the summer. Thus, the
winter sampling information must be weighed against the effects of the fertilizer application.
421
-------
E
Q.
Q.
-------
ELRINGTON ISLAND
Samples were taken along a transect parallel to the water line on the untreated Control, Bath,
and Sprinkler beaches. Samples were analyzed for oil residue weights, changes in oil chemistry,
microbial activity, and numbers of oil-degrading microorganisms. The chemistry data is not yet
available but initial analyses of the microbiology data are presented. Numbers of oil degraders on the
different test beaches are shown in Figure 9.33. Eleven to twelve samples were analyzed from each
beach. Numbers of degraders are only slightly less than observed on the same beaches during the
summer. Thus, winter storm activities do not appear to have substantially affected the oil-degrading
microbial populations. In general, however, there do not appear to be differences between the
beaches.
On the other hand, measurements of hexadecane mineralization activity indicated that some of
the enhanced activity from the summer was still present (Figure 9.34). In other words, a number of
samples from the beaches treated with the fertilizer solution showed the highest level of activity (no
statistical analysis was performed) (keep in mind that INIPOL and CUSTOMBLEN were applied to
all treated and untreated Control beaches at the end of the summer). In addition, samples from the
treated beaches appeared to show more heterogeneity than those from the untreated control. Thus,
it is possible that the enhanced biodegradation activity that occurred in the summer may have been
carried over into the winter.
SUMMARY AND CONCLUSIONS
The following is concluded from the Elrington Island study conducted during the summer of
1990:
a) Bioremediation of subsurface oil is reasonable if sufficient quantities of nitrogen and
phosphorus nutrients can be supplied. This was accomplished by using fertilizer solution.
b) A single pulse application of fertilizer (4 hours, once at low tide) enhanced oil
biodegradation for as long as 3 to 4 weeks. This application was as effective, if not more
so, than a multiple dose application. These results raise the question of whether the effects
of fertilizer application during the summer of 1989, which also consisted of large initial
pulses followed by a gradual release over time of nutrients at lower concentrations, may have
been related more to the extent of nutrient exposure to the microbial communities at one
time than the length of exposure. We believe microbial communities may concentrate
nutrients and recycle them within the biological matrix of the community.
423
-------
BATH
SPRINKLER
UNTREATED
CONTROL
02460
LOG MEAN MPN/GRAM DRY WEIGHT SEDIMENT
LOG MEAN MPN/GRAM DRY WEIGHT SEDIMENT
BATH
SPRINKLER
UNTREATED
CONTROL
Figure 9.33. Number of OH Degrsders at the Bath, Sprinkler, and
Untreated Control Beaches at Elrington Island.
424
-------
# OF VALUES
35
30
25
20
15
10
0 2 4 6 8 10 12 14 16 18 20 22 24
% HEXADECANE TO CO2
Legend
I | Untreated Control
Sprinkler
•• Bath
Figure 9.34. Hexadecane Mineralization Activity for the Untreated Control,
Sprinkler, and Bath Beaches at Elrington Island.
425
-------
ELRINGTON ISLAND
c) Fertilizer-enhanced oil biodegradation rates on Elrington Island were the highest recorded
in the field demonstrations. These rates were approximately 100 mg/kg of beach material/
day, almost six-fold higher than the untreated Control beach. The enhancement was
statistically significant. This occurred despite the fact that rates on the untreated Control
beach were approximately three-fold higher than rates reported at Passage Cove the previous
summer. This increase in effectiveness of fertilizer application could be due to extensive
colonization of the subsurface oil by oil-degrading microorganisms, and/or increased
availability of oil to the bacteria by virtue of impregnation with glacial till (greater exposed
surface area).
d) Measurement of oil mineralization rates in the field and the laboratory using total CO2
production, oxygen uptake, and nutrient assimilation generally coincided with changes in
oil concentration and composition. This supports the use of oil chemistry as a true measure
of oil biodegradation. Relative to oil chemistry analysis, mineralization measurements could
therefore provide simpler procedures for assessing the effect of fertilizer-enhanced
biodegradation in future field studies.
e) The Elrington Island study clearly validated the use of laboratory oil degradation
information from flask and microcosm studies to predict bioremediation events in the field.
f) The use of sampling baskets containing homogenized beach material proved to be a reliable
method for assessing the biological fate of oil, especially in terms of the relative effects of
different fertilizer application strategies. This method considerably reduced sampling
variability, provided information that would have required a much more intensive and costly
sampling effort using standard beach sampling, and provided information that was
representative of the beach it modeled.
426
-------
SECTION 10
SUPPLEMENTAL LABORATORY STUDIES
SHAKE FLASKS
In a series of experiments carried out by Exxon researchers, it was found that microorganisms
indigenous to Prince William Sound have an ability to degrade weathered crude oil if provided
adequate nutrients. Initial experiments in shake flasks demonstrated that organisms in both Prince
William Sound seawater and water from the Alyeska ballast water treatment facility were able to
substantially degrade artificially weathered crude oil in the presence of high levels of nitrogen and
phosphorus (3.5 and 4.1% with respect to oil, 0.03% N and 0.04% P by weight of water) (Figure 10.1).
There was a substantial decrease in the amount of dichloromethane extractable material, and
substantial degradation of both the resolvable fractions and the unresolvable fractions on GC analysis.
Very little organic carbon remained in the aqueous phase after dichloromethane extraction once the
organisms were allowed to settle out.
Biodegradation proceeded much more effectively at warmer temperatures, but there was
significant biodegradation at 5'C in the presence of water-soluble fertilizers (Figure 10.2).
Oleophilic fertilizer shows a sharper temperature dependency than water-soluble fertilizers, and at
20'C there was very little biodegradation when oleophilic fertilizer was used alone (Figure 10.3).
Oleophilic fertilizer stimulated the biodegradation of crude oil. Flask experiments revealed that
the extent of biodegradation increased with increased fertilizer concentration (Figure 10.4). Water-
soluble and oleophilic fertilizers had at least an additive, and perhaps a synergistic, effect on
biodegradation (Figure 10.5).
Oleophilic fertilizer also stimulated the biodegradation of oil on Prince William Sound beach
material (Figure 10.6). Rocks treated with INIPOL and then incubated at 15'C became significantly
cleaner after 14 days; all the resolvable peaks had disappeared in the GC analysis, and 50% of the total
dichloromethane-extractable material had disappeared. The rocks were also clean to the touch. This
biodegradation was not accompanied by a detectable lowering of interfacial tension between oil and
brine, indicating that the microorganisms were not producing significant amounts of surfactants under
the conditions tested.
427
-------
Bushnell-Haas Broth
(3.5% N, 4.1% P)
15°C, 16 days
Alyeska
Inoculum
1SO M.O 48.0 MO 7S.O *0.0 108.0 1M.O 13SO 1SO.O
MINUTES
UNo
loculum
18.0 MO 4S.O M.O 78.0 90.0 108.0 120.0 1SS.O 1M.O
MINUTES
Seawater
Inoculum
1SO 30.0 48.0 §00 7».0 «0.0 10S.O 120.0 13S.O 1SO.O
MINUTES
Figure 10.1. Gas Chromatographic Profiles Showing the Effect of Different
Inocula on Degradation of Artificially Weathered Prudhoe Bay
Crude OH.
-------
Bushnell-Haas Broth
(3.5% N, 4.1% P)
15°C, 16 days
5°C
45.0 60.0 75.0 90.0 105.0 120.0 135.0 150.0
15.0 30.0
MINUTES
15°C
15.O 30.0 45.0
I ' ' • ' I
60.0 75.0 90.0 105.0
MINUTES
120.0
I • ' ' • 'I
135.0 150.0
Rgure 10.2. Gas Chromatographic Profiles Showing the Effect of
Temperature on the Degradation of Artificially Weathered
Prudhoe Bay Crude Oil.
429
-------
Artificial Seawater, 10% INIPOL, 38 days
15.0 30.0 45.0 60.0 75.0 90.0 105.0 120.0 135.0 150.0
MINUTES
15°C
15.0 30.0 45.0 60.0 75.0 90.0 105.0 120.0 135.0150.0
MINUTES
20°C
15.0 30.0 45.0 60.0 75.0 90.0 105.0 120.0 135.0 150.0
MINUTES
Rgure 10.3. Gas Chromatographic Profiles Showing the Effect of
Temperature on the Degradation of Artificially Weathered
Prudhoe Bay Crude Oil Treated with INIPOL.
430
-------
Artificial Seawater, 15°C, 16 Days
3% INIPOL
0.00 500 10.00 1500 20.00 2500 30.00 35.00 40.00
MINUTES
20% INIPOL
000 5.00 10.00 15.00 20.00 2S.OO 30.00 35.00 40. OO
MINUTES
10% INIPOL
0.00 5 00 10.00 15.00 2000 2500 30.00 3500 40.00
MINUTES
50% INIPOL
000 500 10.00 IS 00 20.00 25.00 3O.OO 35.00 4000
MINUTES
Figure 10.4. Gas Chromatographic Profiles Showing the Effect of Different
Concentrations of INIPOL (% of Oil Concentration) on the
Degradation of Artificially Weathered Prudhoe Bay Crude Oil.
431
-------
Artificial Seawater, Poisoned, 15°C, 16 days
15.0 30.0 45.0
60.0 75.0 90.0
MINUTES
i ' ' ' • r
105.0 120.0
135.0 150.0
Oil & 10% INIPOL
15.0 30.0 45.0
60.0 75.0
MINUTES
r
90.0
1 • i • • l • • • « l
105.0 120.0 135.0 150.0
Rgure 10.5. Gas Chromatographic Profile Showing the Effect of Different
Fertilizers, Under Poisoned and Unpoisoned Conditions, on the
Degradation of Artificially Weathered Prudhoe Bay Crude Oil.
432
-------
Artificial Seawater, Active, 15°C, 16 days
Oil + WOODACE Briquettes
(0.4% N, 0.08% P)
-ft--T-r- i-|— I-I-V-T- |-I
60.0 75.0 80.0 105.0 120.0 135.0 150.0
T-i —j— ri— r-t
15.0 30.0
MINUTES
Oil + 10% INIPOL
(0.7% N, 0.06% P)
rt=rir
15.0 30.0
45.0 60.0
75.0 90.0 105.0 120.0 135.0 150.0
MINUTES
Oil + INIPOL and WOODACE Briquettes
(1.1% N, 0.14% P)
15.0 30.0 45.0 60.0 75.0 90.0 105.0 120.0 135.0 150.0
433
Figure 10.5. (Cont.)
-------
Oiled Beach Material, Artificial Seawater, 15°C, 16 Days
No Nutrients
(O.OSg OH)
15.0 30.0
45.0
• l
60.0
75.0 90.0
MINUTES
105.0 120.0 135.0 150.0
I INIPOL, Poisoned
(O.OSg Oil)
15.0 30.0 45.0
60.0
75.0 90.0
MINUTES
105.0 120.0 135.0 150.0
INIPOL
(0.025Q Oil)
* ' ' i ' » ' * i ' ' ' ' i ' ' ' ' f * ' '"i • • * • i ' * » ' i • i » • i i i i i | < i i . )
15.0 30.0 45.0 60.0 75.0 90.0 105.0 120.0 135.0 150.0
MINUTES
Figure 10.6. Gas Chromatographic Profiles Showing the Effect of INIPOL,
Under Poisoned and Unpoisoned Conditions, on the Degradation
of Oil on Beach Material Taken from Prince William Sound.
434
-------
LABORATORY STUDIES
RESPIROMETRY
In a corollary set of shake flask experiments carried out at EPA Cincinnati, analytical
respirometry was used to assess the effects of fertilizer on oil biodegradation. The results are
summarized in Figure 10.7. Cumulative oxygen uptake as a function of time is shown in the
respirometric vessel containing 1000 mg/L oil and INIPOL (5% by weight of oil) and in the
respirometric vessel containing only oleophilic fertilizer (INIPOL). Oxygen uptake began in both
vessels after only a 1.5 day lag period. Maximum uptake of oleophilic fertilizer occurred by the 9th
day, then leveled-off at approximately ISO mg/L. The oxygen uptake rate on weathered oil with
oleophilic fertilizer added was multi-phasic: the first 10 days exhibited the highest uptake rate,
followed by a slower rate for the next 16 days, a somewhat faster rate for the next 4 days, and a much
slower rate after the 30th day. Endogenous oxygen uptake (vessel without oleophilic fertilizer or oil)
was always close to background (data not shown).
The vessel containing oil, oleophilic fertilizer, and the Alyeska ballast water biomass exhibited
an oxygen uptake curve that was almost superimposable on the curve for oil plus oleophilic fertilizer
(data not shown). Thus, in the closed environment of the respirometric vessel, no enhancement of
oil degradation by an external source of oil-degrading organisms was detected.
Shake flasks run in conjunction with the analytical respirometry flasks were periodically
sacrificed following initial application of nutrients, and the oil was extracted for chemical analysis.
Figure 10.8 summarizes the results of GC/FID scans of the alkane hydrocarbons from three sets of
flasks: control (containing 10,000 mg/L weathered crude oil and no nutrients); oleophilic fertilizer-
treated (containing 10,000 mg/L oil and 500 mg/L oleophilic fertilizer); and defined minimal salts-
treated (containing 10,000 mg/L oil and OECD). In the control some minor changes in the alkane
fractions were evident after 6 weeks incubation. Some of these changes may have been due to
biodegradation resulting from background levels of N and P present in the seawater, oil, or beach
material; adsorption to the flask walls; sampling error; or a combination of the above. Whatever the
cause, the magnitude of the changes was relatively insignificant.
Flasks containing oil plus oleophilic fertilizer exhibited complete removal of all aliphatic
components within six weeks. Even the pristane and phytane fractions were reduced to undetectable
levels. The flask containing the minimal-salts solution also exhibited complete removal of the straight
chain aliphatics. However, there were still measurable amounts of pristane and phytane remaining
435
-------
1600
1400
1200
I* 1000
2 800
Q.
0)
600
400
200
10
20
Oil -i- INIPOL + Seawater
INIPOL
Oik INIPOL Without Seawater
30 40
Time (Days)
50
60
70
Figure 10.7. Cumulative Oxygen Uptake on Weathered Prudhoe Bay Crude Oil.
-------
Concentration (ppb)
Concentration (ppb)
Concentration (ppb)
Figure 10.8. Gas Chromatographic Profiles of Alkanes at 0 and 6 Weeks
After Initiation of Flask Studies.
437
Numbers of Carbons
c
o
:
•
:
:
1
"
t)
35
00
TJ
I
10
fj
o
to
ro
to
ro
CO
8
ro
en
ro
a>
ro
00
w
o
w
ro
s
w
o>
u>
o>
ro .tk o> oo o w ^ oj
oo oooogg
oo oooooc
3OO OOOOOO
II 1 1 1 1 1
_ ^3
1
1
'
_;•; -J
'
__I3
1
i
I
i
1
i
1
i
1
i
1
i
H0«
"j
i
j
^i«
i
i
1
I
i
I
"""I
H
1 A
i 2,
I 3
1 A
I 0,
Z
1 £
j 2.
1 •
O O
O O
O O
O) 00
o o
o o
o o
o
o
o
o
ro 4*. ro
o o o
o o o
o o o
to *>• O) CO
o o o o
o o o o
o o o o o
o
o
o
o
o
-U CD
o o
o o
o o
I
o
I
(D
I
-------
LABORATORY STUDIES
at six weeks, although the levels were significantly reduced from the controls. These results suggest
that the oleophilic fertilizer and minimal-salts solution may have enriched different types of
microbial populations. The oleophilic fertilizer-enriched organisms were not only able to break down
straight chain components at a very rapid rate but also branched-chain components. The organisms
enriched by the minimal-salts solution were also able to degrade the branched chain aliphatics, but
at a reduced rate or after a longer lag period.
The GC/MS traces of the aromatic fractions are presented in Figure 10.9 (tabular representation
of the data is given in Table 10.1, as well as the compound names for the alphanumeric designations
in Figure 10.9). In the control several components were reduced to undetectable levels after six weeks
(note fractions H, Q, S, and T, corresponding respectively to dibenzothiophene, C3-fluorenes,
naphthylene, and Cl-naphthylene). The traces from the oleophilic fertilizer and minimal-salts
solution flasks exhibited virtually complete removal of all aromatic fractions after six weeks
incubation.
Results indicate rapid and virtually complete biodegradation of all aliphatic and aromatic
components of the weathered oil contaminating Alaskan beaches occurred in nutrient enriched
respirometer vessels and flasks. Oxygen uptake started after only a 1.5-day lag period and
disappearance of aliphatic and aromatic components occurred within 6 weeks. Different microbial
populations appear to have been enriched by the two types of nutrient solutions (oleophilic fertilizer
and a minimal-salts solution). This suggests a combination of oleophilic fertilizer and a water-soluble
source of nutrients may ultimately be the best way to stimulate rapid bioremediation of crude oil
contaminating Alaskan beaches. Results from ballast water biomass enrichments suggest that external
sources of microbial populations would not enhance biodegradation, and massive inoculations may
not be warranted, at least in the Alaskan bioremediation effort. The respirometric data will
eventually be quantitatively analyzed to calculate the kinetics of oil biodegradation.
438
-------
(Q
O
o
Concentration (ppb)
I?
o w
ro
o
o o
A o>
8 8
§
o
p o
S o
0
3
D>
O
>
m
o
a
in
•n
0
I
_
*
£
Z
O
T)
O
I
u
-t
C
X
u
1 1 1 1 1 1 T — r —
0
m
T]
p z
B m
j ' O
= ' z
F3 c
3>
r- 2
• * ^
2
i — • m
i 5 f~
•
s=>
Concentration (ppb)
Concentration (ppb)
o § 8 8
8
8
lv> *> O)
§88
Z
-D
O
i
f
14 &
p 5
& o
zo
oo
— I-
m
o
o
m
o
D
I
o
D)
a
o>
-------
TABLE 10.1. SHAKER FLASK STUDIES: BIODEGRADATION OF AROMATICS (PAHs)
I. FLASKS WITH OIL + INIPOL
A. Acenaphthene
B. Acenaphthylene
C. Benzo(a)pyrene
D. Benzo(b)fluoranthene
E. Benzo(g,h,i)perylene
F. Chrysene/Benzo(a)anthracene
G. Cl-Chrysenes
H. Dibenzothiophene
I. Cl-Dibenzothiophenes
J. C2-Dibenzothiophenes
K. C3-Dibenzothiophenes
L. Fluoranthene
M. Cl-Fluoroanthenes/Pyrenes
N. Fluorene
O. Cl-Fluorenes
P. C2-Fluorenes
Q. C3-Fluorenes
R. Indeno(l,2,3-cd)pyrene
S. Naphthalene
T. Cl -Naphthalenes
U. C2-Naphthalenes
V. C3-Naphthalenes
W. C4-Naphthalenes
X. Phenanthrene/ Anthracene
Y. Cl-Phenanthrenes/Anthracenes
Z. C2-Phenanthrenes/Anthracenes
1. C3-Phenanthrenes/ Anthracenes
2. C4-Phenanthrenes/Anthracenes
3. Pyrene
0 wks incubation
S1»,b
19.6
165.0
37.2
18.9
<3.0
49.0
56.2
463.0
993.0
269.0
378.0
55.9
<3.0
412.0
155.0
1,267.0
419.0
<3.0
494.0
1,401.0
<3.0
1,621.0
1,141.0
695.0
34.4
1,066.0
445.0
34.6
25.6
SIR*'b
24.1
120.0
30.3
15.2
<3.0
30.7
35.1
330.0
675.0
179.0
235.0
37.0
<3.0
319.0
128.0
820.0
265.0
<3.0
378.0
965.0
<3.0
1,510.0
743.0
489.0
26.2
676.0
307.0
17.2
28.7
6 wks incubation
Sl*-b
<5.0
<5.0
12.0
6.83
<3.0
15.9
18.9
12.8
11.9
8.83
31.4
3.57
<3.0
<5.0
<5.0
38.4
102.0
<3.0
27.1
<5.0
<5.0
<5.0
11.9
6.79
<5.0
38.0
102.0
14.3
5.41
SlRa'b
<5.0
<5.0
8.99
<5.0
<3.0
14.8
19.0
12.2
8.70
8.72
24.3
3.80
<3.0
<5.0
<5.0
30.3
77.6
<3.0
29.2
<5.0
<5.0
<5.0
15.8
<5.0
<5.0
34.4
90.1
18.6
2.96
* Histogram values were the average values from
b Concentration in /ig/L (ppb)
the above duplicate flasks
440
-------
TABLE 10.1. (CONT.)
II. FLASKS WITH OIL + MINIMAL
SALTS (OECD)
A. Acenaphthene
B. Acenaphthylene
C. Benzo(a)pyrene
D. Benzo(b)fluoranthene
E. Benzo(g,h,i)perylene
F. Chrysene/Benzo(a)anthracene
G. Cl-Chrysenes
H. Dibenzothiophene
I. Cl-Dibenzothiophenes
J. C2-Dibenzothiophenes
K. C3-Dibenzothiophenes
L. Fluoranthene
M. Cl-Fluoroanthenes/Pyrenes
N. Fluorene
O. Cl-Fluorenes
P. C2-Fluorenes
Q. C3-Fluorenes
R. Indeno(l,2,3-cd)pyrene
S. Naphthalene
T. Cl -Naphthalenes
U. C2-Naphthalenes
V. C3-Naphthalenes
W. C4-Naphthalenes
X. Phenanthrene/ Anthracene
Y. Cl-Phenanthrenes/ Anthracenes
Z. C2-Phenanthrenes/Anthracenes
1. C3-Phenanthrenes/Anthracenes
2. C4-Phenanthrenes/Anthracenes
3. Pyrene
0 wks incubation
S9a.b
<3.0
88.2
24.4
11.1
<3.0
50.3
47.8
277.0
538.0
159.0
209.0
32.6
<3.0
255.0
87.1
743.0
388.0
<3.0
279.0
776.0
< 3.0
969.0
674.0
423.0
18.6
581.0
263.0
17.2
25.6
S9R*'b
<3.0
75.5
21.6
10.3
<3.0
37.2
44.5
252.0
491.0
141.0
195.0
30.6
<3.0
221.0
76.1
660.0
346.0
<3.0
230.0
663.0
<3.0
869.0
578.0
587.0
17.0
538.0
259.0
14.7
24.9
6 wks incubation
S9».b
<3.0
<5.0
<5.0
<5.0
<5.0
<5.0
<5.0
12.0
8.36
<5.0
12.6
<5.0
<5.0
<5.0
<5.0
28.6
41.4
<3.0
33.8
6.22
< 3.0
<5.0
16.5
<5.0
<5.0
30.9
37.7
7.97
<5.0
S9R»-b
<3.0
<5.0
6.90
<5.0
<5.0
10.5
10.8
12.6
14.0
<5.0
10.5
<5.0
<5.0
<5.0
<5.0
96.1
97.5
<3.0
50.9
19.3
<3.0
<5.0
20.3
<5.0
<5.0
87.3
90.5
20.3
< 5.0
J Histogram values were the average values
b Concentration in /Jg/L (ppb)
from the above duplicate flasks
441
-------
TABLE 10.1 (CONT.)
III. Control: Oil Only
A. Acenaphthene
B. Acenaphthylene
C. Benzo(a)pyrene
D. Benzo(b)fluoranthene
E. Benzo(g,h,i)perylene
F. Chrysene/Benzo(a)anthracene
G. Cl-Chrysenes
H. Dibenzothiophene
I. Cl-Dibenzothiophenes
J. C2-Dibenzothiophenes
K.. C3-Dibenzothiophenes
L. Fluoranthene
M. Cl-Fluoroanthenes/Pyrenes
N. Fluorene
O. Cl-Fluorenes
P. C2-Fluorenes
Q. C3-Fluorenes
R. Indeno(l,2,3-cd)pyrene
S. Naphthalene
T. Cl -Naphthalenes
U. C2-Naphthalenes
V. C3-Naphthalenes
W. C4-Naphthalenes
X. Phenanthrene/ Anthracene
Y. Cl-Phenanthrenes/Anthracenes
Z. C2-Phenanthrenes/Anthracenes
1. C3- Phenanthrenes/ Anthracenes
2. C4-Phenanthrenes/Anthracenes
3. Pyrene
0 wks incubation
SC27*
<3.0
143.0
28.3
15.2
<3.0
62.3
76.4
398.0
843.0
239.0
317.0
49.0
<3.0
382.0
130.0
1,038.0
618.0
<3.0
431.0
1,206.0
<3.0
1,691.0
876.0
556.0
31.7
896.0
406.0
38.9
41.9
6 wks incubation
SC27a
<3.0
89.4
19.9
12.2
<3.0
100.0
105.0
385.0
788.0
222.0
252.0
43.8
<3.0
327.0
82.8
1,015.0
585.0
<3.0
412.0
1,135.0
<3.0
1,570.0
631.0
477.0
23.8
778.0
367.0
105.0
58.8
a Concentration in /ig/L (ppb)
442
-------
LABORATORY STUDIES
TOXICITY OF OLEOPHILIC FERTILIZER
Laboratory Bioassays
Toxicity tests were conducted with oleophilic fertilizer (INIPOL) in the summer of 1989 to
develop definitive acute toxicity values for fishes, invertebrates, and algae. Tests with silver salmon
smolts, herring fry, juvenile sticklebacks, mussel larvae, oyster larvae, mysids, and pandalid shrimp
plus oleophilic fertilizer in seawater have been completed, as well as tests utilizing fertilizer plus
weathered Prudhoe Bay Crude oil in seawater.
Results show that larvae of mussels, oysters and juvenile mysids were two orders of magnitude
more sensitive than salmon, and more sensitive than herring and sticklebacks by a factor of 2 to 6.
When mixed with oil, the toxicity of oleophilic fertilizer to salmon was reduced four-fold. The
mixture was slightly more toxic to most other test species, although the differences were within the
variability of test repeatability.
The data in Table 10.2 were provided to the Shoreline Committee and advisory groups in Valdez,
Seward, and Homer to assist in the evaluation of potential toxic effects associated with large-scale
application of the fertilizer as a clean-up technique. In addition, a risk assessment procedure was
suggested as a means to establish a benchmark no-acute-effect concentration for comparison with
possible environmental concentrations following shoreline treatment. This method was modified by
the Shoreline Committee in Valdez and Homer, and was used to assist in establishing shoreline
segments approved for fertilizer application.
Laboratory toxicity tests were conducted at ERL/GB to determine the toxicity of ammonia
and laural phosphate to sensitive marine biota. A shrimp-like crustacean, Mysidopsis bahia, and an
estuarine fish, the silverside, Menidia beryllina, were tested using standard acute toxicity test methods
(ASTM, 1988). These standard test animals are representative surrogates for invertebrates and fish
in marine systems. In the course of testing, an article by Miller et al. (1990) was published, providing
definitive data on ammonia toxicity for these two standard test species over a range of temperature,
443
-------
LABORATORY STUDIES
TABLE 10.2. RESULTS OF LABORATORY TOXICITY TESTS WITH
THE OLEOPHILIC FERTILIZER, INIPOL, AND VARIOUS MARINE SPECIES'
(VALUES ARE 96-HOUR LC50 ESTIMATES UNLESS NOTED OTHERWISE)
Oleophilic Fertilizer Fertilizer Plus Oil
Fish
Salmon smolts
Herring
Sticklebacks
1,500
100
100
ppm
ppmb
ppmb
6,000
c
110
ppm
ppm
Invertebrates
Mussel larvae 63 ppm (72 hour) 55 ppm (72 hour)
Oyster larvae 41 ppm (48 hour) 9.0 ppm
Mysids 15 ppm 8.5 ppm
Pandalid shrimp 360 ppm 92 ppm.
* Oil concentration used was below toxic concentrations; oil was layered on the surface water at the
start of the test, sprayed with oleophilic fertilizer, and mixed once.
b Best estimate from non-definitive test
c Data are not available
salinity and pH conditions. Results of both acute toxicity tests (96-hour) and chronic estimator tests
(7-day) were published, and are presented below for temperatures of 25°C, 31 ppt salinity, and pH
of 8.0, test conditions for which both acute and chronic tests were conducted:
Acute LC50 Chronic Toxicity Value
Mysids 1.7 ppm 0.232 ppm
Silversides 1.3 ppm 0.061 ppm
444
-------
LABORATORY STUDIES
This publication also addressed the effect of lower temperatures on ammonia toxicity to fish.
Both silversides and sheepshead minnows (Cyprinodon variegatus) were slightly more sensitive when
tested at temperatures less than 25°C (18°C for silversides and 13°C for sheepshead minnows),
indicating that the acute LCSOs listed above are representative of ammonia toxicity at temperatures
encountered in Alaska.
INIPOL also contains tri-laureth phosphate, a common component in cosmetic products.
Because a source of this material could not be located, it was not tested with aquatic animals.
Results of chronic estimator toxicit tests are presented in the following table:
Estimate of Chronic Toxicity of INIPOL
to Two Species of Estuarine Fish
Laboratory
Contract Lab
ERL/GB
No- Effect Concentrations at 7
Test species 7-Dav LCSO
Menidia beryllina 69 ppm
Menidia beryllina 50 ppm
Cyprinodon variegatus 69 ppm
Days
Survival
50 ppm
30 ppm
48 ppm
Growth
25 ppm
30 ppm
48 ppm
These test results place the potential for chronic effects to fish at exposure concentrations
greater than the 15 to 50 ppm concentrations previously reported to be acutely toxic to invertebrates
(i.e., mussel or oyster larvae, juvenile mysids) in laboratory tests. Exposures to INIPOL at
concentrations >50 ppm in the field for 7 days is extremely unlikely. The results do not change the
previous recommendations and assessments that INIPOL can be safely applied at the recommended
rates to oiled shorelines, after consideration of the potential for the development of acutely toxic
conditions for nearshore invertebrates for very short time periods immediately after application.
445
-------
LABORATORY STUDIES
Under routine application procedures and specified environmental conditions, no acute toxic
effects are expected for marine fishes. The likelihood of toxic effects to marine invertebrates is
minimal and transient. This is based on knowledge of (1) the relative sensitivity of fishes and
invertebrates tested under laboratory conditions; (2) the minimal input of oleophilic fertilizer into
marine waters from unintentional over-spraying of marine waters during application; (3) the minimal
release of oleophilic fertilizer from shoreline into the bay following application; and (4) tidal mixing,
dilution, and transport of bay waters contaminated with oleophilic fertilizer that will continue to
decrease fertilizer concentrations in marine waters.
Toxicity of oleophilic fertilizer to marine biota may be associated with release of fertilizer from
the shoreline into the bay immediately after application. A worst-case example would involve
protected embayments with minimal tidal exchange and maximum shoreline-to-water ratios (long,
narrow bays with constricted openings). INIPOL was applied to oiled shorelines at the rate of 293
g/m2 (0.06 lb/ft2). Applied to 100 m of shoreline in a 10-m swath marked from the low-tide line to
the upper storm berm, a total of 293,000 g would be used.
If all of the oleophilic fertilizer reached the water in a pulse and: 1) a generalized, 100 m stretch
of nearshore environment from the tide line to 10 m offshore, with an average depth of 1 m (1,000
m3) is used; and 2) water is assumed to be completely mixed, a "worst-case" expected environmental
concentration of 293 ppm can be calculated. This value is considerably less than the 96-hour LC50
value for salmon and comparable to the LC50 for herring. Toxicity to marine invertebrates residing
in this nearshore area is possible at these concentrations, should unrealistic application conditions
result.
Any exposure resulting from shoreline applications would be mitigated by tidal mixing, dilution,
and transport out of the system into Prince William Sound. The initial concentrations should decrease
by orders of magnitude within 1 to 2 days, to levels considerably less than concentrations acutely toxic
in 2 to 4 days in laboratory tests. Thus, the prospect of sustaining acutely lethal concentrations for
any biota is very unlikely.
Since there are no proven analytical methods to quantify INIPOL in seawater, no measured
environmental concentrations of INIPOL are available from previous field trials to compare with
worst-case predictions. However, a more likely estimate of daily input into nearshore waters is in the
range of 1% to 10% of the applied material, based on: (1) visual observations of the colored film
446
-------
LABORATORY STUDIES
present after spraying; (2) our ability to maintain nutrient-enriched pore water in the intertidal zone;
and (3) the lack of measured nutrient increases in the nearshore zone. When a more realistic input
estimate (assume 10%) is diluted with a nearshore (10 m) volume of water averaging 2 m (a depth
consistent with the steep slope of most shorelines in Prince William Sound), the estimated
environmental concentration is between 3 and 30 ppm. These values indicate that environmental
concentrations less than the laboratory LC50 values of INIPOL and oil mixtures for invertebrates
could develop immediately after application, subject to subsequent tidal dilution and transport. When
considered in this light, the potential for toxic effects appears to be minimal.
Oleophilic fertilizer and oil mixtures that may leave the treated shoreline should have minimal
ecological impact, based on their propensity to degrade and the dilution potential of surrounding
waters. The enhanced microbial biomass and available nutrients associated with mixtures of
oleophilic fertilizer and oil will result in their rapid degradation. In their mixed form, both oil and
INIPOL are less toxic to marine biota, as demonstrated by the reduced toxicity of INIPOL in
laboratory tests where it was mixed with oil.
Wildlife Toxicity Issues
Decisions to apply oleophilic fertilizer (INIPOL) or other fertilizer formulations to enhance
biodegradation must consider potential for toxicological effects on terrestrial and aquatic life.
Information about toxicity of INIPOL to aquatic organisms has been obtained directly from laboratory
and field tests with invertebrates and fish. Toxicity of water-soluble fertilizer formulations was
assessed using toxicity data available on ammonia in the U.S. Environmental Protection Agency Water
Quality Criteria Document. Ammonia is the principle toxic component for water-soluble fertilizers.
Toxicity of several INIPOL components has been sufficiently quantified for small mammals to
allow some realistic assessments of likely events that could lead to lethal or toxic exposures for
mammalian and avian wildlife. Inhalation of toxic vapors of ammonia or 2-butoxy-ethanol poses a
negligible risk to wildlife. Direct ingestion of contaminants from licking treated surfaces poses some
risk if the animals inhabit these areas within a couple of days of treatment. In addition, the potential
for ingestion of contaminants during cleaning behavior of mammals and birds should be considered
for exposures to 2-butoxy-ethanol, laureth phosphate, or a total dose of INIPOL. Skin irritation and
eye irritation are possibile toxic effects to wildlife resulting from exposures to some INIPOL
constituents. The potential for toxic effects lessens with time after application.
447
-------
LABORATORY STUDIES
This review was conducted to obtain information on the toxicity of INIPOL and its constituents
in order to evaluate more fully the potential for adverse effects on wildlife that might enter oiled
shorelines after bioremediation treatments during 1990 oil spill clean-up activities. Toxicity of
fertilizers to small and large mammals and birds was assessed in a cursory manner during the summer
of 1989 by extrapolating available human health information on chemical toxicity and by using the
incidental knowledge of experts working with the Prince William Sound Interagency Shoreline Clean-
up Committee. The purpose of this review was to supplement that information with a systematic
review of what is known about toxicity to mammals and birds for nutrient amendments used in
bioremediation programs.
The data evaluated were those providing exposure (dose)-response information upon which
direct effects to wildlife survival, growth, or reproductive success could be inferred. A considerable
number of abstracts were found in the databases on the biochemical and physiological effects of
atmospheric or aqueous exposures of test animals to ammonia or urea. Most of these reports discussed
metabolic pathways or physiological responses and did not relate effects to survival, growth, or
reproductive success of the animal. Metabolic or physiological studies do not provide a basis for
meaningful extrapolations to direct ecological consequences of exposure. These metabolic and
physiological publications were not included in the evaluation.
Effects data are summarized in Tables 10.3 to 10.7 that group human, other mammalian, and
avian data separately. Not all chemicals had data for each grouping. For chemicals without data on
toxicity to birds, hazards to avian wildlife must be extrapolated from mammalian data. No toxicity
data were retrieved for oleophilic fertilizer (INIPOL); however the use of INIPOL for the oil spill in
Spitsbergen was listed in the Toxline database. Since there were no toxicity data for laureth
phosphate in the databases, information on lauryl sulfate is presented for comparative purposes.
Atmospheric Ammonia
Concentrations of ammonia in air >10 ppm pose potential toxicity problems for birds, based on
the lowest effect exposure noted in Table 10.3. Lethal concentrations of ammonia in air for short-
term exposures (hours) to small mammals appear to be in the thousands of ppm. Sensitivities of
humans and small mammals appear to be the same order of magnitude based on the data in Table 10.3.
Ammonia has induced mutagenic changes in bacteria at high doses; a similar effect has not been
reported in mammals or birds.
448
-------
LABORATORY STUDIES
TABLE 10.3. SUMMARY OF TOXICITY DATA FOR AMMONIA. MOST LETHAL OR
EFFECTS EXPOSURES WERE REPORTED AS PARTS-PER-MILLION (PPM) IN AIR OR AS MG
NH8 PER CUBIC METER OF AIR (1 MG/MS « 1.414 PPM). SOME RESULTS WERE
CONVERTED TO TOXIC DOSES EXPRESSED AS MG NHS/KG TEST ANIMAL. TCLO-LOWER
LEVEL OF TOXIC EXPOSURE, LCLO=LOWER LEVEL OF LETHAL
CONCENTRATIONS, LC50=CONCENTRATION LETHAL TO 50% OF TEST POPULATION
Human Data
Inhalation
Inhalation
Inhalation
Non-human Mammalian Data
Inhalation for Rat
Inhalation for Rat
Inhalation for Mouse
Inhalation for Cat
Inhalation for Rabbit
Ingestion by Guinea Pig
Bird Data
TCLo 20 ppm (duration unspecified)
LCLo 5000 ppm for 5 minutes
LDLo 132 mg/kg
LC50 37,200 ppm for 1 hour
LC50 2,000 ppm for 4 hours
LC50 4,230 ppm for 1 hour
LC50 7,000 mg/ms for 1 hour
LC50 7000 mg/m3 for 1 hour
LD75 900-1200 mg/kg
Inhalation for Pullets: 200 ppm for 17 days;
Toxic effects noted as decreases in food intake, growth, egg
production and increased mortality.
Inhalation for Broiler Chicks: 75 or 100 ppm for 4 days;
Toxic effects noted as respiratory and circulatory damage.
Inhalation for Turkevs: 10 or 40 ppm (duration not specified). Toxic effects noted as
respiratory damage.
Other toxicity data noted
Mutagenic in bacteria (Escherichia coli) at 1500 ppm for 3 hours in atmosphere over culture
disc.
Cytogenic effects noted in rats exposed to 19.8 mg/m3 for 16 weeks.
449
-------
LABORATORY STUDIES
TABLE 10.4. SUMMARY OF TOXICITY DATA FOR AQUEOUS AMMONIA. MOST LETHAL
OR EFFECTS EXPOSURES WERE REPORTED AS PARTS-PER-MILLION (PPM) NH,
IN WATER OR AS MG NHS PER LITER OF WATER. SOME RESULTS WERE CONVERTED
TO TOXIC DOSES EXPRESSED AS MG NH^/KG TEST ANIMAL. TCLO-LOWER LEVEL OF
TOXIC EXPOSURE, LCLO-LOWER LEVEL OF LETHAL CONCENTRATIONS,
LC50-CONCENTRATION LETHAL TO 50% OF TEST POPULATION, LDLO«LOWER LEVEL
OF LETHAL DOSES, LD50-DOSE LETHAL TO 50% OF TEST POPULATION
Hamas Data
Oral Dose LDLo
Inhalation LCLo
Inhalation TCLo
Non-human Mammalian Data
Oral Dose to Rat LD50
Oral Dose to Cat LDLo
Injected into Mouse LDLo
Injected into Rabbit LDLo
Injected into Mouse LD50
Injected into Rabbit LDLo
Eye irritation to rabbit
Eye irritation to rabbit
Eye irritation to rabbit
No Bird Data Available for Review
43 mg/kg
5000 mg/kg (fumes derived from solution)
408 ppm (fumes derived from solution)
350 mg/kg
750 mg/kg
160 mg/kg (beneath skin)
200 mg/kg (beneath skin)
91 mg/kg (into veins)
10 mg/kg (into veins)
0.75 mg causes severe reaction
0.044 mg causes severe reaction
1.0 mg in 30 second rinse causes severe reaction
Other Texicity Data Noted
Mutation in Bacteria (Escherichia coli) at 10 mg/growth disc
450
-------
LABORATORY STUDIES
TABLE 10.5. SUMMARY OF TOXICITY DATA FOR UREA. TOXIC DOSES ARE
EXPRESSED AS MG UREA/KG TEST ANIMAL. LDLO-LOWER LEVEL OF LETHAL
DOSES, LD50-DOSE LETHAL TO 50% OF TEST POPULATION
Human Data
Skin irritation - 22 mg applied intermittently over 3 days caused mild irritation
No other toxicity data were available for review
Non-Human Mammalian Data
Oral Dose Rat
Oral Dose Mouse
Oral Dose Goat and Sheep
Injection Rat
Injection Mouse
Injection Dog
Injection Rabbit
Injection Rat
Injection Mouse
Injection Dog
Injection Rabbit
Bird Data
Injection Pigeon
Other Toxicity data noted
LD50 8,471 mg/kg
LD50 11,000 mg/kg
LDLo 511 mg/kg
LD50 8,200 mg/kg (beneath skin)
LD50 9,200 mg/kg (beneath skin)
LDLo 3,000 mg/kg (beneath skin)
LDLo 3,000 mg/kg (beneath skin)
LD50 5,300 mg/kg (into vein)
LD50 4,600 mg/kg (into vein)
LDLo 3,000 mg/kg (into vein)
LDLo 4,800 mg/kg (into vein)
LDLo 14,800 mg/kg (beneath skin)
Urea has caused reproductive effects in humans and monkeys when injected into placentas
during the 16th or 18th week of pregnancy. Tumorigenic effects in rats and mice have resulted
with one year exposures. Tests with cell cultures of human or other mammal tissues have shown
positive results for mutation screens.
451
-------
LABORATORY STUDIES
TABLE 10.6. SUMMARY OF TOX1CITY DATA FOR 2-BUTOXY-ETHANOL. TOXIC DOSES
ARE EXPRESSED AS MG 2-BUTOXY-ETHANOL/KG TEST ANIMAL. EXPOSURES BY
INHALATION ARE PPM IN AIR OR MG 2-BUTOXY-ETHANOL/M3. LDLO-LOWER
LEVEL OF LETHAL DOSES, TDLO-LOWER LEVEL OF TOXIC DOSES, LD50«DOSE
LETHAL TO 50% OF TEST POPULATION, LC50-CONCENTRATION
IN AIR LETHAL TO 50% OF TEST POPULATION
Human Data
Oral Dose (one woman) TDLo
Inhalation TCLo
Inhalation TCLo
Non- human Mammalian Data
Oral Dose Rat LD50
Oral Dose Mouse LD50
Oral Dose Rabbit LD50
Oral Dose Guinea Pig LD50
Inhalation for Rat LC50
Inhalation for Mouse LC50
Injection Rat LD50
Injection Mouse LD50
Injection Rabbit LD50
Injection Rat LD50
Injection Mouse LD50
Injection Rabbit LD50
Injection Mouse LDLo
Absorbed by Skin Rabbit LD50
Absorbed by Skin Guinea Pig LD50
Eye Irritant Rabbit
Eye Irritant Rabbit
Eye Irritant Rabbit
Bird Data
600 mg/kg
195 ppm for 8 hours
100 ppm (duration unspecified)
470 mg/kg
1,230 mg/kg
300 mg/kg
1,200 mg/kg
2,900 mg/ms
700 ppm in air for 7 hours
220 mg/kg (into abdomen)
536 mg/kg (into abdomeV)
220 mg/kg (into abdomen)
340 mg/kg (into vein)
1,130 mg/kg (into vein)
280 mg/kg (into vein)
500 mg/kg (beneath skin)
220 mg/kg
230 mg/kg
500 mg caused mild irritation
18 mg caused some irritation
100 mg over 24 hours caused moderate irritation
No Bird data were available for review
Other Toxicity Data noted
This chemical has caused reproductive effects in tests with rats, mice, or rabbits using extended
exposures or exposures at TCLo concentrations reported above.
452
-------
LABORATORY STUDIES
TABLE 10.7. SUMMARY OF TOXICITY DATA FOR SODIUM LAURYL SULFATE.
TOXIC DOSES ARE EXPRESSED AS MG SODIUM LAURYL SULFATE/KG TEST ANIMAL.
LD50-DOSE LETHAL TO 50% OF TEST POPULATION
Human Data
Skin Irritant, 250 mg caused mild irritation in 24 hours
Skin Irritant, 25 mg caused mild irritation in 24 hours
No other toxicity data were available for review.
Non-human Toxicity Data
Oral Dose Rat
Injection Rat
Injection Mouse
Injection Rat
Injection Mouse
Skin Irritant Mouse
Skin Irritant Dog
Skin Irritant Pig
Skin Irritant Guinea Pig
Skin Irritant Rabbit
Skin Irritant Rabbit
Skin Irritant Rabbit
Skin Irritant Rabbit
Skin Irritant Rabbit
LD50 1,288 mg/kg
LD50
LD50
LD50
LD50
210 mg/kg (into abdomen)
250 mg/kg (into abdomen)
118 mg/kg (into vein)
118 mg/kg (into vein)
25 mg, moderate irritation in 24 hours
25 mg, mild irritation in 24 hours
25 mg, mild irritation in 24 hours
25 mg, mild irritation in 24 hours
50 mg, severe irritation in 24 hours
25 mg, moderate irritation in 24 hours
250 mg, moderate irritation in 24 hours
10 mg, no degree reported for 24 hours
50 mg, mild irritation in 24 hours
No Bird Data were Available for Review
Other Toxicity Data Noted
Mutagenic effects noted for bacterial and fungal cultures at very high exposure concentrations
when tested in the early 1900's. Results of modern tests at extremely high exposures conducted
with bacterial cultures were negative for mutagenicity (Hazardous Substances Data Base).
Because application of fertilizers produces ammonia from slow-release solubilization of urea or
controlled dissolution of ammonium-containing reagents, release of ammonia to the atmosphere is
*r
expected to be a very indirect process regulated by dissolution of gas from water into the air. It is
expected that this process will yield atmospheric concentrations much less than 10 ppm in air
immediately above treated shorelines. Dilution by directional breezes and thermal air currents will
quickly diminish any atmospheric concentrations of ammonia to background levels. No short-term
453
-------
LABORATORY STUDIES
or chronic exposure problems are expected for wildlife with respect to atmospheric ammonia for
shorelines treated with fertilizers. Although quantitative monitoring data were not obtained, this
assertion would appear to be supported by the fact that teams applying fertilizers during 1989 did not
report fumes of ammonia in treated areas nor did they appear to suffer from acute exposure to
ammonia.
Aqueous Ammonia
Ammonia dissolved in water is toxic to small mammals if ingested in sufficient quantities to
obtain exposures in the hundreds of mg per kilogram of body weight (Table 10.4). Body burdens
obtained by injecting ammonia into animal bodies produced lethal effects at similar or slightly lower
dose rates. No avian toxicity data were available for review.
Maximum aqueous concentrations of ammonia associated with fertilizer additions from the
bioremediation study were reported for subsurface (interstitial) samples. Peak concentrations in
interstitial water ranged from 100 to 400 pg/L. A small mammal would have to drink over one
hundred liters of subsurface seawater to obtain a toxic dose of ammonia. In actuality, wildlife would
be exposed to nearshore surface waters where concentrations of ammonia would be considerably less
compared to subsurface water. For example, water samples collected immediately offshore from
treated areas within a day of treatment were always at background concentrations of <1 to 3 fig
ammonia/L or less, demonstrating the rapid dilution potential from tidal mixing and nearshore
currents. For wildlife frequenting shorelines treated with fertilizers, the potential hazard of ammonia
poisoning from intentional or incidental ingestion of surface waters is negligible.
Aqueous solutions of ammonia can cause eye irritation. Laboratory tests conducted by adding
drops to the eyes of rabbits (a standard procedure for determining eye irritants) determined that as
little as 0.044 mg could cause severe eye irritation. If this quantity was contained in a mL of water,
it would be the equivalent of a concentration of 44 mg/L. None of the surface water samples
analyzed the summer of 1989 approached such concentrations, although some subsurface water
samples did. Eye irritation from ammonia exposures in water would seem unlikely for shoreline
wildlife.
454
-------
LABORATORY STUDIES
Urea
Urea is lethal to mammals and birds at body burdens in the thousands of mg/kg, even though
exposures vary from ingestion to subcutaneous injections (beneath the skin) to intravenous injection
(into a vein) (Table 10.5). There is a single report where 500 mg/kg was lethal to goats and sheep.
Urea was a mild skin irritant for humans when 22 mg was applied over 3 days.
The application rate of oleophilic fertilizer (INIPOL) during the summer of 1989 was
approximately 300 g/m2, equivalent to 48 g of urea/m2 because INIPOL is 16% urea. A small animal
(1 kg) would have to eat all the INIPOL applied to an area 0.25 m by 0.25 m (0.0625 square meters,
or an area approximately .3 m by .3 m) in order to obtain a lethal dose of urea (using 3,000 mg/kg
as an LD50). Effects of skin irritation could be possibile for animals active on recently treated
shorelines.
2 - bu toxy-ethanol
Information on the toxicity of 2-butoxy-ethanol to mammals is shown in Table 10.6. Body
burdens of 200 to 500 mg/kg from ingestion, injection, or absorption through the skin are lethal.
Mild eye irritation in rabbits occurs at doses of 18 to 100 mg. Reproductive effects in mammals have
been reported for a range of doses administered to pregnant females. A human exposure limit of 25
ppm in air has been established to protect workers. No avian toxicity data were available.
Monitoring by Exxon industrial hygiene personnel during initial oleophilic fertilizer (INIPOL)
applications during the summer of 1989 indicated that exposure limits of 25 ppm were not exceeded.
As a result, workers in the field were not required to wear personal protection equipment during
applications. Inhalation exposures for wildlife would likewise seem to pose no toxic threat.
The 2-butoxy-ethanol comprises 11% of the formulated INIPOL, providing an application rate
of 30 g/m2 for this chemical. Acute toxicity could be expected for 1 kg animals that ingest freshly
applied INIPOL (licking treated substrate, chewing coated sticks, ingesting coated gravel) from an
area approximating 0.01 m2 (0.1 m by 0.1 m, or 4 inches by 4 inches) to obtain a 300 mg/kg dose.
Ingestion during cleaning behavior also should be considered, because wildlife could contaminate their
feet, fur, or feathers with 2-butoxy-ethanol if the animals venture onto a treated shoreline
immediately after application. Transfer of mg quantities of butoxy-ethanol from feet or fur into the
455
-------
LABORATORY STUDIES
eyes could occur during cleaning or sleep, causing eye irritation. As butoxy-ethanol is volatile and
should quickly evaporate from beach surfaces, substantial body contamination would be likely only
for animals utilizing shorelines treated with INIPOL within hours of application.
Lauryl sulfate (surrogate for latireth phosphate)
A considerable amount of data exists on the skin irritant capabilities of sodium lauryl sulfate
(Table 10.7) because it is a common ingredient in soap. Laureth phosphates could be expected to act
similarly. Skin exposure to >25 mg of lauryl sulfate can cause irritation. No effects were seen with
an application of 10 mg to rabbits. Body doses >100 mg/kg by injection or ingestion can be lethal
to small mammals. No avian data were available for review.
The prospects of adverse effects to wildlife attributable to laureth phosphate following oleophilic
fertilzier (INIPOL) applications are intermediate of those posed by urea and 2-butoxy-ethanol. This
constituent is neither volatile nor readily absorbed across the skin, so toxicity risks are considerably
less than those of 2-butoxy-ethanol for body burdens. Ingestion appears to be the only likely route
of exposure. If the rat oral LD50 of 1288 mg/kg for lauryl sulfate is used for calculation purposes,
an acutely toxic dose of laureth phosphate for a 1 kg animal would be contained in the INIPOL
applied to approximately 0.02 m2 (or an area .1 m by .1 m), as this component comprises
approximately 25% of the INIPOL. Potential for toxic effects from ingestion of treated surfaces or
from cleaning contaminated feet, fur, of feathers should be considered. Incidental eye irritation is
possible for animals visiting treated shorelines shortly after application.
If necessary, some of the uncertainty about the availability of applied INIPOL could be
quantified or characterized qualitatively during monitoring programs. Collection of samples from
shoreline surfaces could be scheduled on a daily basis for quantification of grams of nitrogen or pg
of butoxy-ethanol per square meter through time after treatment to measure residual concentrations
of these components. Such data would assist in quantifying exposure potential for animals visiting
the shoreline over several days after fertilizer application.
As with other environmental risks associated with clean-up activities, the potential for wildlife
effects following fertilizer application must be compared to and balanced with risks posed to wildlife
by the residual oil along the shoreline.
456
-------
LABORATORY STUDIES
TOXICITY OF CUSTOMBLEN PELLETS TO BIRDS
The EPA Environmental Research Laboratory at Corvallis conducted a screening-level acute
toxicity test with CUSTOMBLEN and adult (9 month) bobwhite quail. Quail were dosed with
CUSTOMBLEN pellets using a dosing gun to push the pellets down their throat. Birds were dosed
at the rate of 5 g, 1 g or 0.2 g per bird (birds weighed approximately 200 g). The 5 g/bird dose (20
g/kg) caused acute mortality within 8 hours; 1 g/bird (4 g/kg) dose caused mortality within 36 hours
for half the test population; 0.2 g/bird (1 g/kg) caused no mortality among the test population. When
quail were offered CUSTOMBLEN pellets mixed with their food, they did not discriminate between
the two; the pellets were not preferentially ingested, nor were they avoided. The acute LD50
calculated from these screening tests converts to the consumption of 40 to 60 pellets for a 200 g adult
quail. Subsequent feeding tests with black-legged stilts, a heron-like bird that normally feeds on live
invertebrates inhabiting the shoreline, showed that they did not ingest the CUSTOMBLEN pellets and
preferred live food.
No definitive toxicity tests with birds were conducted because of limited bird availability. This
was a low priority issue in light of these data, reviewed literature, and the lack of reports of bird
mortality or consumption of CUSTOMBLEN pellets in the field. The screening level tests indicate
that seed-eating birds are at greater risk than live-feeders. Since shoreline and aquatic birds are not
seed eaters, their risks of mortality from consumption of CUSTOMBLEN pellets are minimal.
BIOMETER STUDIES
During the winter of 1989/1990 and early spring 1990, further laboratory experimentation was
conducted in an attempt to fully optimize oil degradation in the field using refined and improved
methods for fertilizer application.
Fertilizer Specific Activity
When treated with a single application of inorganic nutrient solutions containing varying
amounts of nitrogen and phosphorus followed by subsequent additions of 3% NaCl, microorganisms
associated with oiled beach material from Disk Island responded with increased respiratory activities.
These increases in activity were directly related to the amount of N and P added to each system, with
10-fold dilutions of these solutions producing proportional decreases in the microbial activities
457
-------
LABORATORY STUDIES
observed (Figure 10.10). Over the 3-day incubation period, material treated with the level 1 nutrient
solution produced approximately 400 M CO2 (cumulative) while consuming close to the same amount
of oxygen (460 M) for an average respiratory coefficient of about 1.1. Since crude oil represented
the primary source of carbon in these systems, this activity was presumably due to the stimulation of
oil-degrading microorganisms.
Material treated with various dilutions of the level 1 solution exhibited proportional decreases
in activity. Both the lowest level of nutrient addition (level 4) and unamended Prince William Sound
water exhibited similar responses, suggesting that the level of available N and P in natural waters was
roughly equivalent to 35 and 0.07 g/L, respectively (nutrient analysis of this water determined N and
P concentrations of 35 and 9 g/L, respectively). These responses are compared to those obtained with
the 3% NaCl control. Preliminary studies demonstrated that this treatment was equivalent to a
poisoned control (formalin), but did not produce some of the undesirable effects associated with
acidification and the addition of formaldehyde, which precipitated during chemical extraction
procedures (data not shown). Therefore, 3% NaCl was used as the control in subsequent studies.
The stair-step response in CO2 production data, most notably evident with the level 1 treatment,
was a result of intermittent sampling during both high- and low-tide periods. During high-tide
periods (0 to 12, 24 to 36, and 48 to 60 hour), the rate of CO3 production appeared to be much slower
due to the ability of the tide waters (pH=8.1) to adsorb evolved CO2. The presence of tide waters did
not affect oxygen this way.
Following 3 days of measuring microbial respiration rates, radiolabeled phenanthrene was used
as a reporter chemical to further demonstrate the ability to stimulate the activity of indigenous
oil-degrading microorganisms through the addition of inorganic nutrients (Figures 10.11 and 10.12).
Radiolabeled substrate was added during the first high-tide period, which lasted for 6 hours, and
systems were monitored for 60 hours. During the first low-tide period (6 hours of incubation), there
was a dramatic increase in the amount of 14CO2 produced in all systems except the 3% NaCl control.
With continued incubation, specific treatment effects were pronounced: level l>level 2>level 3>level
4=Prince William Sound. These radiorespirometric data correlate well with data presented in Figure
10.10.
458
-------
•<£>
•n
I <
r
0>
o
to
i 8 8
.1.1
to
8
8
o 1
• A • *
ro
8
•
W
8
• .
8
i
Ul
8
//
.' /
/ /
/
• t
-------
80 n
60-
40-
a.
O
S 20 H
3
E
3
O
24 36
Incubation Time (hrs)
48
leveM
level 4
level 2
PWS water
-w-
60
level 3
sterile NaCI
Legend
Level 1 =350 ppm N/70ppm P
Level 2=35 ppm N/7ppm P
Level 3-3.5 ppm N/0.7 ppm P
Level 4-0.35 ppm N/0.07 ppm P
PWS = Prince William Sound
Figure 10.11. Mineralization of Radiolabeled Phenanthrene Over Time As
Influenced by Fertilizer Application.
460
-------
c
0
c
a
c
a>
*!
o
c *"
o
T»
N
0
c
12 24 36 46 60 72 84 96 108
Incubation Time (hrs)
Level 1
350 ppm N
Level 2
70 ppm P
PWS
-•— 3% NaCI
(Control)
Legend
Level 1 =350 ppm N/70ppm P
Level 2=35 ppm N/7ppm P
PWS =Prince William Sound
Figure 10.12. Stimulation of Oil-Degrading Microflora with Inorganic Nutrient
Supplementation Over Time.
461
-------
LABORATORY STUDIES
The observed relationship between N concentration and the activity of oil-degrading
microorganisms is summarized in Figure 10.13. Based on these data, the optimum concentration of
N, the primary factor limiting the biodegradation of crude oil, was determined to be approximately
35 ppm.
Corresponding changes in the amounts of oil present in each system further demonstrated the
ability to stimulate the activity of oil-degrading microorganisms through the addition of inorganic
nutrients (Table 10.8). Following 6 days of incubation after initial treatment, beach material treated
with level 1 nutrients contained 202 mg hexane-soluble and 212 mg hexane-insoluble oil constituents.
TABLE 10.8. CHANGES IN OIL CONCENTRATION AND DISTRIBUTION
DURING BIODEGRADATION WITH VARIOUS
NUTRIENT AMENDMENTS
Treatment
Level 1
Level 2
Level 3
Level 4
PWS
water
Control
Hexane-
Soluble
mg
202
224
204
213
227
273
Oiled
Hexane-
insoluble
mg
212
223
214
228
244
298
Rocks
Total
peak
area
Io8io
5.7
5.8
5.9
5.9
6.0
6.2
Hexane-
Soluble
mg
47
32
42
32
36
75
Tide
Hexane-
insoluble
mg
73
47
62
50
62
52
Waters
Total
peak
area
loSio
6.0
5.8
5.7
6.1
6.0
6.5
Oiled rocks extracted after 6 days incubation;
Tide waters combined for extraction;
Fertilizer treatments applied on day 1 only (initial high-tide);
Control = level 1 plus 3.7% formaldehyde added at each high tide.
462
-------
LABORATORY STUDIES
.1
"o
Q. C~
Si
o
«
cc
3-
2-
1 -
•50
50
150
250
350
Nitrogen Concentration (ppm)
Figure 10.13. Specific Relationship Between N Concentration and the Activity of
Oil-Degrading Mlcroflora.
When compared to the values of 273 mg hexane-solubles and 298 mg asphaltenes for the control,
the ability to enhance the extent of oil biodegradation through nutrient additions was clearly
demonstrated. Moreover, total peak area (calculated from gas chromatographic analyses) was less with
the level 1 treatment than with any other treatment. These further support the ranking of treatments
for their ability to accelerate the rate and extent of oil biodegradation as described above.
Changes in the amount and composition of oil recovered from tide waters was less definitive.
For all parameters measured (hexane-soluble weights, asphaltene weights and total peak area) there
were no noticeable differences between treatments.
463
-------
LABORATORY STUDIES
Using CO2 evolution data as a representative measurement of relevant microbial activities, the
importance of nitrogen concentration on the activity of oil-degrading microorganisms was
demonstrated. This relationship is shown in Figure 10.13 where the rate of CO2 production is plotted
as a function of N concentration. From these data, it was determined that the most effective
concentration of soluble nitrogen was between 350 and 35 ppm N. Therefore, these two
concentrations were used in subsequent studies.
Agitation Rates
The effects of physical agitation on the removal of crude oil from contaminated beach material
were determined using radiolabeled phenanthrene as a reporter chemical. Generally, increased
agitation stimulated the activity of oil-degrading microorganisms (Figure 10.14). However, when
agitation was excessive (125 rpm), the amount of labeled substrate physically removed from the
system was high (Figures 10.15 and 10.16; Table 10.9). Since most of the oil was washed away under
these conditions, the amount of labeled phenanthrene available for biodegradation was reduced.
Fertilizer Application Strategies
Oleophilic fertilizer application to beach material in biometer flasks was based on field
remediation activities where fertilizer was applied at a rate of 20% (weight) of the oil present.
Preliminary studies showed that beach material from Passage Cove was contaminated with 4.5%
(weight) partially weathered, Prudhoe Bay crude oil (450 mg oil per 100 g beach material). Therefore,
a single application of 80 L oleophilic fertilizer was applied to 100 g of oiled beach material, resulting
in the addition of 19 mg laureth phosphate and 12.6 mg urea per 100 g oiled beach material.
When compared with level 1 (350 ppm N, 70 ppm P) and level 2 (35 ppm N, 7 ppm P) inorganic
nutrient solutions, the effect of oleophilic fertilizer application was proportional to its N
concentration effect (126 ppm N) (Figure 10.17). However, the stimulatory effect of oleophilic
fertilizer on the activity of oil-degrading microorganisms appeared to decrease 4 days after fertilizer
application.
After 6 days of incubation, radiolabeled phenanthrene or oleic acid was added to each flask and
the production of 14CO2 was monitored for 24 hours (Figures 10.18 and 10.19). Within the 12 hour
high-tide incubation period, only 10 to 14% of the added phenanthrene was removed from the system
464
-------
o
I
i
0 12 24 36 48 60 72 84 96 108120132
,/- Sample Intervals (hr)
Figure 10.14. Cumulative Percent Mineralization of 14C Phenanthrene Over Time.
465
-------
a.
o
200000
180000
160000 -
140000 -
120000 -
100000 -
D 125 rpm
T 100 rpm
V 75 rpm
• 50 rpm
O 0 rpm
24 36 48 60 72 84 96 108
Sample Intervals (hr)
Figure 10.15. Tldalnates1* Counts per Sample Interval.
200000
180000
160000
140000
120000
100000
80000
60000
40000
20000
—
_
.
" \
't
s
1
/
•^
' \
S
/ \
c
- /$
X -V
/ S
y N
^
1
™
—
™
™
~
—
~
""
—
~
~
E
= i
•7,
y
X
\
^
^
^
m
1
•
a
E
—
=
1
o
^
o Is
i ^ :<
i i i i
E=j 125 rpm
[mrm 100 rpm
^^^ 75 rpm
^^ 50 rpm
s
| |^ 0 rpm
j
i
-
i
i
i
i
i
i
I _> ^ ^
j ^dl§ ^u
i 1 !!M ilf
i = i Z$t = i ^i
12 24 36 48 60 72 84
Sample Intervals (hr)
96 108
Figure 10.16. Tldalnatas" Counts per Sample Interval.
466
-------
LABORATORY STUDIES
TABLE 10.9. TOTAL PEAK AREAS: GC ANALYSIS OF OIL RESIDUE
FROM SHAKER EXPERIMENT
Condition
0 rpm
50 rpm
75 rpm
100 rpm
125 rpm
TO Control
(no "tidal" cycle)
Condition
0 rpm
50 rpm
75 rpm
100 rpm
125 rpm
Rocks
Total Peak Area
3.05xl06
2.65xl06
2.34x1 O6
2.40xl06
2.35xl06
3.32xl06
"Tides"
Total Peak Area
6.69x1 O6
1.16xl06
2.22x1 06
22.06x1 06
2.31xl06
% 0 rpm Control
100.0
87.0
76.8
78.6
77.1
109.0
% 0 rpm Control
100.0
173.6
331.8
308.4
345.5
TO Control: 100 g oiled beach material that was not subjected to simulated "tidal" cycling or shaking.
467
-------
1600
J»
o
C
O
O
3
•o
o
CM
O
U
0
3
3
O
control
Incubation Time (days)
Legend
IN = INIPOL
Lev 1 = 350 ppm N/
70ppmP
Lev 2= 35 ppm N/
7 ppm P
Figure 10.17. Effect of Various Treatments on the Activity of Indigenous,
Oil-Degrading Microorganisms.
468
-------
0
c
e
c
* ^
81
*~ 3
«i
o
c —
o
o
c
INt-lev! daily
tevl daily
Iev2 daily
Iev1 bath
3% NaCI
(Control)
12 24 36 48 60 72 84 96 108
Time (hrs)
Legend
IN =INIPOL
Lev 1 = 350 ppm N/
70 ppm P
Lev 2= 35 ppm N/
7 ppm P
Figure 10.18. Effect of Various Treatments on the Activity of Indigenous,
Oil-Degrading Microorganisms.
469
-------
s
15
u
i
o [14C] Oleic Acid
• [14C] Phenanthrene
0 10 20 30 40 50 60 70 80 90 100 110 120
Time (hr)
Figure 10.19. Effects of INIPOL and Soluble Nutrients on 14C Oleic Acid and 14C
Phenanthrene Mineralization Over Time.
470
-------
LABORATORY STUDIES
with the "out-going tide". The extent of phenanthrene mineralization also depended on the amount
of inorganic nutrients present. With the level 1 and 2 nutrient solutions, 41 and 21% of the radiolabel
was mineralized, respectively. Treatment with oleophilic fertilizer alone, 6 days prior to the addition
of radiolabel, resulted in 17% phenanthrene mineralization. When treated with level 1 and oleophilic
fertilizer together, the effect was equivalent to level 1 alone. These values were much higher than
the 3% NaCl control (<3% mineralization over 24 hours).
Considering these data, it was postulated that some of the CO2 production data presented in
Figure 10.17 may have been due to the biodegradation of oleophilic fertilizer constituents (i.e., oleic
acid, laureth phosphate). Utilizing radiolabeled oleic acid as a reporter chemical, oleophilic fertilizer
biodegradation was indeed shown to be a potential contributing factor (Figure 10.19). Within the 24
hour incubation period that followed the 6 day period during which microbial respiration was
recorded, approximately 23% of the added 14C-oleic acid was mineralized when pretreated with
oleophilic fertilizer. However, approximately the same amount of oleic acid biodegradation was
observed when beach materials were treated solely with inorganic nutrients. Regardless of the
treatment applied, the amount of either 14C-oleic acid and 14C-phenanthrene mineralization was
greater than that observed with the untreated control.
Corresponding changes in the amount of oil in each system and its physical distribution further
demonstrated the ability to accelerate the rate and extent of oil biodegradation through the addition
of inorganic nutrients (Table 10.10). Since previous studies showed little difference in the amount
of oil in removed tide water, these data are not presented. In terms of the amount of oil biodegraded
within the 7 day incubation period, the same trend was again observed: level 1 inorganic nutrient
solution was most effective followed by level 1 + INIPOL, level 2, and INIPOL alone; all were more
effective than no treatment (3% NaCl control). Chemical profiles from gas chromatographic analysis
of hexane extracts clearly reflect this response.
When treated daily with a solution containing only nitrogen at a concentration of 350 ppm N,
the response of oil-degrading microorganisms (as determined by phenanthrene mineralization) was
equal to those treated with level 1 or level 2 inorganic nutrient solutions (Figures 10.20 and 10.21).
Conversely, the effect of phosphorus alone (70 ppm P) was much less. Therefore, the stimulatory
effect of inorganic nutrient amendments was primarily a function of nitrogen addition.
471
-------
LABORATORY STUDIES
TABLE 10.10. CHANGES IN OIL CONCENTRATION AND OIL DISTRIBUTION
AFTER 6 DAYS INCUBATION AND TREATMENT WITH INIPOL
AND/OR INORGANIC NUTRIENT SOLUTIONS
Oiled Rocks
Treatment
INIPOL + high-level
soluble nutrients
High-level
soluble nutrients
Pulse high-
level soluble nutrients
INIPOL
Low-level
soluble nutrients
3% NaCl
(control)
Hexane-
soluble
mg
434
423
429
434
430
435
Hexane-
insoluble
mg
31
39
38
34
39
38
Total
peak
area
Io8io
5.8
5.6
5.7
6.1
6.1
6.3
Gravimetric measurements obtained following 7 days of incubation to record respirometric responses
to treatments;
INIPOL applied on day 1 only;
Soluble nutrients added daily with each high-tide.
When oil contaminated beach material was treated with a single application of level 1 nutrient
solution (pulse application), response in terms of oil-degrading ability was equivalent to daily
treatment with the same solution. The initial stimulatory response persisted for 3 to 4 days. Field
data collected over a 30-day time period demonstrated that the stimulatory effects of the pulse
application strategy persisted for 10 to 14 days (see Elrington Island results in Section 9). Therefore,
the requirement for inorganic nitrogen by indigenous, oil-degrading microflora may be fulfilled
relatively quickly. Once this requirement is satisfied, these microorganisms appear to be active for
extended periods of time before nitrogen again becomes an essential limiting factor for oil
biodegradation.
472
-------
100-1
%adsorption
% in tidewater
%mineralization
HLSN+IN HLSN IN LLSN
Treatment
3% NaCI
(Control)
Legend
HLSN=High-l_evel Soluble Nutrients
IN =INIPOL
LLSN =Low-Level Soluble Nutrients
Figure 10.20. Distribution of 14C from Radiolabeled Phenanthrene.
473
-------
c
o
HLSN
Treatment
IN
oleic acid
Legend
HLSN * High-level
Soluble Nutrients
IN = INIPOL
Figure 10.21. Accelerated Biodegradation of 14C-Phenanthrene and 14C-Ole
Acid.
474
-------
LABORATORY STUDIES
Bioaugmentation Studies
The possibility of enhancing the rate and extent of oil biodegradation through the addition of
oil-degrading microorganisms was evaluated. Microbial inoculum of specially-selected, oil-degrading
marine isolates indigenous to Prince William Sound are described in Table 10.11. Of the various
organisms available, a mixed-culture community (DI/EI l-5a) and a pure bacterial culture (strain
El 2V) were selected for further study based on their catabolic abilities and observed growth rates
in the presence of Prudhoe Bay crude oil.
When oiled beach materials were exposed to the action of these organisms, the amount of oil
biodegradation was enhanced (Figures 10.22 and 10.23, and Table 10.12). Therefore, under controlled
conditions, the potential of enhancing the bioremediation process through bioaugmentation was
demonstrated. Nutrient supplementation was again effective in stimulating the activity of indigenous,
oil-degrading microorganisms. By no means, however, do these data state that a similar response
would be observed if these organisms were applied in situ.
Oleophilic Fertilizer Mode of Action
Laboratory studies designed to discern the mode of action of oleophilic fertilizer were basically
inconclusive. While the fertilizer increased microbial activities above background levels, the most
significant response was related to the addition of soluble nutrients (Figure 10.24). Moreover, the
effect of oleophilic fertilizer did not appear to be directly related to the presence of laureth phosphate
or oleic acid (Figure 10.24). Conversely, the addition of nitrogen, as urea in the case of INIPOL or
NH4NO3 for soluble nutrient additions, showed the most significant response.
Mineralization of radiolabeled phenanthrene (Table 10.13) and oil chemistry data (Figure 10.26)
supported microbial respiration data. The addition of soybean oil, an alternative source of
triglycerides, was ineffective (Figure 10.25; Table 10.14). Therefore, the unique effectiveness of
INIPOL appeared to be due to a synergistic effect of constituents resulting from its novel formulation
(i.e., synergistic interaction between INIPOL constituents).
475
-------
TABLE 10.11. ORGANISMS ISOLATED FROM OILED BEACH MATERIAL IN PRINCE WILLIAM SOUND, ALASKA
Organism
Designation
DI/EI #1
DI/EI #2
DI/EI #3
DI/EI #5a
Hexadecane #2
Hexacosane #1
Hexacosane #2
Pristane #2
AK-Naph-B
AK-Phen-6
El 2V
Enriching
Substrate
crude oil
crude oil
crude oil
crude oil
nC16
nC26
nC26
pristane
PAHs+ naph
crude oil + phen
crude oil
Genus*
Rhodococcus sp.
Rhodococeus sp.
Pseudomonas sp.
Rhodoozus sp.
Rhodoezus sp.
Acinetibacter sp.
Rhodococcus sp.
Rhadoceccus sp.
Pseudomonas sp.
Pseudomonas sp.
Rhodococcus sp.
Carbon source tested as growth substrates
C8 C16 C26 Pristane Naph Phen
t t I t II
- +
~- ++ 1 - II
- +
- - - + -
* +
• Bacterial identification was determined by analysis of fatty acid profiles (Microbial ID, Inc., Newark, DE)
-------
2
I
Sample Intervals (hr)
Legend
• Oil Subculture/SN
V EI-2/SN
V No Inoculum/SN
D NoSN
SN Soluble Nutrients
Figure 10.22. Cumulative Total CO2 Production Over Time.
477
-------
DU
g
84 96
Sample Intervals (hr)
Legend
O Oil Subculture
• EI-2 Subculture
V No Inoculum/SN
V No Inoculum/ No SN (Control)
SN Soluble Nutrients
Figure 10.23. Cumulative Percent Mineralization of 14C-Phenanthrene Over Time.
478
-------
3
c
o
o
3
13
2
0,
CM
O
O
0>
"3
"5
E
3
O
1600-
1400-
1200-
'
1000-
(
800-
'
600-
400-
t
200-
n *
-• HLSN+IN
••— HLSN
•* IN
LLSN
-D-
3% NaCI
(Control)
Incubation Time (days)
Legend
HLSN=High-Level Soluble Nutrients
IN -INIPOL
LLSN =Low-Level Soluble Nutrients
Figure 10.24. Effect of INIPOL and Inorganic Nutrients on the Activity of Oil-
Degrading Mlcroflora Over Time.
479
-------
2500
_ 2000-
O
I
s
o
2
I
"5
1500-
1000-
500-
SN + IN
SN + LP
IN
12
60
84
108
132
156
Incubation Time (hrs)
Legend
SN ^Soluble Nutrients
IN =INIPOL
LP -Laureth Phosphate
OA =Oleic Acid
SO =Soybean Oil
Figure 10.25. Effect of INIPOL and its Constituents on Mlcrobial Activities Over
Time.
480
-------
INIPOL Alone
TimeO
Soluble Nutrients
+ INIPOL
Soluble Nutrients
(Daily)
Soybean Oil
Alone (T=0)
Figure 10.26. Oil Chemistry Data.
481
-------
LABORATORY STUDIES
TABLE 10.12. COMPARISON OF GC PROFILE TOTAL PEAK AREAS FROM MICROBIAL
INOCULATION BIOMETER STUDY
Condition
Mean Peak Area % Control % TO
Oil Subculture
with SN
EI-2, Mixed Isolate
Culture with SN
No Inoculum
with SN
Control, No
Inoculum, No SN
TO Extraction
1.59xl06
1.55xl06
I.6U106
1.75xl06
2.31xl06
90.9
88.4
92.0
100.0
132.2
68.7
66.8
69.6
75.6
100.0
SN: Soluble Nutrients, 35.7 mMol N/8.07 mMol P
Filtered Prince William Sound, AK "tide"
TO Extraction: 100 g oiled beach material that was not subjected to "tidal" cycling
TABLE 10.13. MINERALIZATION OF 9-14C-PHENANTHRENE (24 HOUR INCUBATION)
FOLLOWING 7 DAYS OF TREATMENT EXPOSURE
Mass Balance of Radioactivity
Treatment
Counts in
"tidalnate"
Counts as
C02
Counts absorbed
or assimilated
(calculated)
Sterile SN
3% NaCl
Soybean Oil
INIPOL
SN + Soybean Oil
SN + INIPOL
SN + OA
SN
SN + LP
SN = Soluble Nutrients
OA = Oleic Acid
LP - Lauryl Phosphate
9.6
15.0
19.1
8.8
15.9
14.1
9.3
13.2
14.6
<0.01
4.7
5.5
15.1
18.6
36.5
38.1
40.7
42.1
90.4
80.3
75.4
76.1
65.5
49.4
52.6
46.1
43.3
482
-------
TABLE 10.14. CHANGES IN OIL CONCENTRATION AND OIL DISTRIBUTION
AFTER 7 DAYS INCUBATION
Oil residue analysis
Treatment
CH2C12-
soluble
Hexane-
soluble
Total peak
area
mg
mg
log
10
Sterile SN
3% NaCl
Soybean Oil
INIPOL
SN + Soybean Oil
SN + INIPOL
SN + OA
SN
SN + LP
SN = Soluble Nutrients
OA - Oleic Acid
LP = Lauryl Phosphate
421.3
451.3
413.7
511.0
377.7
428.7
488.4
429.4
435.7
397.8
425.6
418.7
447.2
375.8
434.3
427.5
422.2
436.5
6.34
6.25
6.34
6.48
6.34
6.23
6.23
5.81
6.32
SUMMARY AND CONCLUSIONS
Treatment with inorganic nutrient solutions increased the rate and extent of oil biodegradation
2 to 3 fold. This stimulatory effect was primarily due to nitrogen addition with phosphorus
amendment exhibiting little effect on the microbial activities measured. Moreover, a pulse
application strategy appears to offer an efficient means of stimulating the rate and extent of oil
biodegradation by microorganisms indigenous to Prince William Sound. Application of these findings
may enhance the efficiency of future oil spill bioremediation efforts.
483
-------
SECTION 11
STABLE ISOTOPES
Stable isotopes were used in the oil spill bioremediation project to evaluate the assimilation of
carbon from oil carbon into microbial food chains, and trace nitrogen fertilizer used for
bioremediation into bacteria and other organisms on the beach. Studies were conducted on beaches
at Snug Harbor, Passage Cove, Elrington Island, and Disk Island. At these sites algae, seagrasses, and
heterotrophic consumers were sampled to examine the assimilation of nitrogen fertilizer into
organisms other than bacteria. Microcosms were used to study bacteria assimilation of nitrogen
fertili/er and the effect of the nitrogen on bacteria assimilation of oil.
FOOD CHAIN STUDIES
Food Chain Studies at Snug Harbor and Passage Cove, Summer 1989
Potential damage (other than from oil) to the ecological balance of intertidal and shallow subtidal
communities in Prince William Sound through addition of large quantities of fertilizer to contaminated
beaches was possible. Stable carbon and nitrogen isotopes in algae and higher consumers were used
in the summer of 1989 to trace assimilation of nitrogen in bioremediation treatments on beaches in
Prince William Sound.
The stable carbon and nitrogen isotope ratios of primary producers and consumers are located
in Table 11.1. The £16N for Laminana spp only ranged from +6.3 to +9.9 %o. The 613C values for
Uruspora spp., F. distichus and Laminaria spp. were consistent with previous measurements of marine
C3 plants (Fry and Sherr, 1984). The £13C values for Z. marina were as expected for C4 plants (Fry
and Sherr, 1984). Seston, which in seawater environments contains primarily algal material and
detritus, exhibited the most negative carbon isotope measured in these samples (51SC - -23.7), with
the exception of an unidentified terrestrial plant spp, (51SC « -28.4 %o).
The range of 51SC and 516N for different primary producers and consumers observed in Snug
Harbor and Passage Cove are shown in Figure 11.1. Although not all species are represented at both
sites, there wasn't a significant difference among the ranges within the two environments. There was
a progressive enrichment in i16N among limpets, periwinkles, M. edilus, Balanus spp, and whelks.
The values for consumers were equal to or more positive than those for the primary producers. Only
seston contained fi16N as enriched in 516N as the consumers. In contrast, consumers were depleted in
484
-------
TABLE 11.1. ISOTOPIC DATA FOR SAMPLES TAKEN FROM SELECTED BEACHES
OF PRINCE WILLIAM SOUND, AK BETWEEN JUNE AND AUGUST, 1989
Sample
Urospora spp
Urospora spp
Urospora spp
Urospora spp
Urospora spp
Urospora spp
Urospora spp
Urospora spp
Urospora spp
Fucus distichus
Fucus distichus
Fucus distichus
Fucus distichus
Fucus distichus
Fucus distichus
Fucus distichus
Fucus distichus
Fucus distichus
Fucus distichus
Laminaria spp
Laminaria spp
Laminaria spp
Laminaria spp
Laminaria spp
Laminaria spp
Zostera marina
Zostera marina
Zostera marina
Zostera marina
Zostera marina
Zostera marina
Zostera marina
Zostera marina
Zostera marina
Zostera marina
Zostera marina
Date
6/13/89
6/17/89
8/10/89
8/10/89
8/10/89
8/13/89
8/20/89
8/13/89
8/20/89
6/13/89
6/17/89
8/10/89
8/10/89
8/10/89
8/10/89
8/13/89
8/20/89
8/13/89
8/20/89
6/13/89
6/17/89
8/10/89
8/10/89
8/20/89
8/20/89
6/13/89
6/17/89
6/17/89
8/10/89
8/10/89
8/10/89
8/10/89
8/13/89
8/20/89
8/13/89
8/20/89
Location
Tatitalek Is.
Snug Harbor
Snug Harbor
Snug Harbor
Snug Harbor2
Passage Cove1
Passage Cove1
Passage Cove2
Passage Cove2
Tatitalek Is.
Snug Harbor
Snug Harbor
Snug Harbor
Snug Harbor2
Snug Harbor2
Passage Cove1
Passage Cove1
Passage Cove2
Passage Cove2
Tatitalek Is.
Snug Harbor
Snug Harbor1
Snug Harbor2
Passage Cove1
Passage Cove2
Tatitalek Is.
Snug Harbor
Snug Harbor
Snug Harbor1
Snug Harbor1
Snug Harbor2
Snug Harbor2
Passage Cove1
Passage Cove1
Passage Cove2
Passage Cove2
51BN
+8.5
-
+4.9
+8.3
+5.6
+4.6
+6.5
+2.3
+3.3
+8.5
+5.9
+7.6
+7.9
+5.8
+7.4
+6.3
+7.0
+6.7
+4.6
+9.9
+7.5
+8.4
+7.6
+7.5
+6.3
+ 1.8
+4.5
+4.5
+7.4
+7.3
+6.9
+6.7
H-S.l
+6.7
+8.1
+7.3
*13C
-19.1
-
-21.7
-19.4
-16.0
-14.6
-18.9
-18.1
-18.4
-15.4
-20.0
-16.2
-18.9
-18.8
-16.6
-19.5
-15.2
-17.4
-16.0
-
-18.4
-21.2
-19.3
-
-22.4
-10.6
-9.5
-9.7
-11.0
-10.4
-11.6
-11.2
-11.3
-9.3
-13.9
-8.9
1Samples collected directly from untreated control plots
2Samples collected directly from fertilized plots
485
-------
STABLE ISOTOPES
TABLE 11.1. CONTINUED
Sample
Seston
Seston
Seston
Seston
Seston
Collisella pelta
Collisella pelta
Notoacmea scutum
Notoacmea scutum
Littorina spp
Littorina sitkana
Littorina spp
Littorina sitkana
Littorina sitkana
Mytilus edilus
Mytilus edilus
Mytilus edilus
Mytilus edilus
Mytilus edilus
Mytilus edilus
Mytilus edilus
Mytilus edilus
Mytilus edilus
Mytilus edilus
Mytilus edilus
Mytilus edilus
Mytilus edilus
Balanus glandula
Balanus glandula
Balanus glandula
Balanus rastratus
Balanus glandula
Balanus glandula
Balanus glandula
Balanus glandula
Balanus glandula
Balanus glandula
Date
7/16/89
8/13/89
8/22/89
8/13/89
8/22/89
8/13/89
8/20/89
8/13/89
8/20/89
8/10/89
8/13/89
8/20/89
8/13/89
8/20/89
6/8/89
6/17/89
7/11/89
7/11/89
7/11/89 '
8/10/89
8/10/89
8/10/89
8/10/89
8/13/89
8/20/89
8/13/89
8/20/89
6/13/89
6/17/89
8/10/89
8/10/89
8/10/89
8/10/89
8/13/89
8/20/89
8/13/89
8/20/89
Location
Snug Harbor
Passage Cove1
Passage Cove1
Passage Cove2
Passage Cove2
Passage Cove1
Passage Cove1
Passage Cove2
Passage Cove2
Snug Harbor2
Passage Cove1
Passage Cove1
Passage Cove2
Passage Cove2
Tatitalek Is.
Snug Harbor
Snug Harbor3
Snug Harbor8
Snug Harbor*
Snug Harbor1
Snug Harbor1
Snug Harbor2
Snug Harbor2
Passage Cove1
Passage Cove1
Passage Cove2
Passage Cove2
Tatitalek Is.
Snug Harbor
Snug Harbor1
Snug Harbor1
Snug Harbor2
Snug Harbor2
Passage Cove1
Passage Cove1
Passage Cove2
Passage Cove2
«16N
+8.2
+9.1
+5.7
+6.6
+7.7
+7.8
+7.4
+7.3
+7.6
+7.9
+8.4
+8.3
+7.7
+9.1
+7.8
-
+8.5
-
+7.8
+8.3
+8.4
+7.9
+8.2
+8.3
+8.5
+8.7
+9.9
-
+8.9
+9.9
+9.9
+9.2
+9.9
+9.5
+9.9
+ 10.1
51SC
-22.9
-22.2
-23.2
-21.6
-23.7
-20.1
-15.5
-22.6
-19.4
-21.0
-18.3
-18.0
-17.2
-17.8
_
-21.5
-18.9
-21.4
-20.3
-20.3
-20.4
-20.2
-21.4
-20.1
-20.9
-20.1
-20.7
-19.0
-20.2
-19.0
-19.8
-20.9
-20.8
-19.5
-20.9
-17.7
-19.4
1Samples collected directly from untreated control plots
2Samples collected directly from fertilized plots
'Samples collected from experimental cages
-------
TABLE 11.1. CONTINUED
Sample
Nucella lima
Searlesia dira
Nucella lima
Nucella lima
Root feeder parti
Root feeder parti
Z. marina epiphytes
Z. marina epiphytes
Brown algal spp
Mollusk spp
Terr plant spp
Brown algal spp
Red algal spp
Ammonia, Root Feeder
Date
8/13/89
8/20/89
8/13/89
8/20/89
8/13/89
8/20/89
6/13/89
6/17/89
6/13/89
6/13/89
6/13/89
8/10/89
8/10/89
7/16/89
Location
Passage Cove1
Passage Cove1
Passage Cove2
Passage Cove2
Passage Cove2
Passage Cove2
Tatitalek Is.
Snug Harbor
Tatitalek Is.
Tatitalek Is.
Tatitalek Is.
Snug Harbor1
Snug Harbor1
Snug Harbor
515N
+10.4
+11.4
+ 10.3
410.0
+3.4
+6.0
_
-
+6.8
+ 30.2
-
+6.5
+8.4
-
61SC
-18.1
-16.2
-20.3
-17.4
-28.4
-23.9
-10.8
-12.4
-21.4
-15.3
-28.4
-20.0
-18.6
N.A.
1Samples collected directly from untreated control plots
2Samples collected directly from fertilized plots
the heavier isotope of carbon, £1SC, compared with primary producers, but were more positive than
the seston. Limpets had the largest carbon isotope variability among consumers. Finally, a
progressive enrichment pattern was observed for 51SC among M. edilus, Balanus sp, and whelks.
Fertilizer Nitrogen Assimilation by Food Chains on Beaches, Summer 1989
At Snug Harbor (August 10) and Passage Cove (August 13; August 20; August 22), biological
samples and organisms were taken directly from either untreated control or fertilized plots within the
beach. In both cases, the area was contaminated by oil. INIPOL, the oleophilic fertilizer used in the
test sites, had an 616N of +2.0 %o. The 513C of INIPOL was -29.9 %o, while cruds oil from the Alyeska
Terminal used to fill the Exxon Valdez had a 513C of -30.1 %». Both the nitrogen and carbon isotope
ratios of the fertilizer and oil-contaminant are isotopically lighter than valises measured for biological
samples and seston (Table 11.1). Mean 513C and 515N for different species ar»d seston from untreated
control and fertilized plots within the beach are shown in Figure 11,2 (In most cases, n>3, refer to
487
-------
CO
00
Whelk -
Balanus spp -
M. Edilus -
Periwinkle -
Limpet -
Seston -
Z. marina -
Laminaria spp -
F. distichus -
Urospora spp
Whelk -
Balanus spp -
M. Edilus -
Periwinkle -
Limpet -
Seston -
Z marina-
Laminaria spp -
F. distichus -
Urospora spp
Snug Harbor
24 -22 -20 -18 -16 -14 -12 -10
813C
Passage Cove
-24 -22 -20 -18 -16 -14 -12 -10
513C
Whelk -
Balanus spp -
M. Edilus -
Periwinkle -
Limpet -
Seston -
Z. marina -
Laminaria spp -
F. distichus -
Urospora spp -
Snug Harbor
Whelk -
Balanus spp -
M. Edilus -
Periwinkle -
Limpet -
Seston -
Z. marina -
Laminaria spp -
F. distichus -
Urospora spp -
4 6
815N
8 10 12
4 6
615N
8 1012
Figure 11.1. The 513C and 519N of Plants, Animals, and Seston Collected
from Snug Harbor (6/13/89, 7/16/89, 8/10/89) and from
Passage Cove (8/13/89, 8/20/89, 8/22/89). The Bar indicates
the Range of Measured Vaiues (Refer to Table 11.1).
-------
815N
813C
IVJ
oo
Uruspora spp
F. distichus
'/////////////{/////,
^'S^&2£dr&%*Z
Uruspora spp
F. distichus
Laminaria spp
Z. marina
Seston
Limpet
Periwinkle
M. edilus
Balanus spp
Whelk
Y//////////////////A
Figure 11.2. Mean 813C and 815N of Plants, Animals, and Seston Collected
Directly from Untreated Control and Fertilized Plots.
-------
STABLE ISOTOPES
Table 11.1). Lack of a major difference among samples from untreated control and fertilized plots
for within the beach either carbon or nitrogen isotope ratios suggested that in the short period after
application of the fertilizer, utilization of fertilizer carbon or nitrogen and transfer up the food chain
was not significant. There was, however, a small decrease in the £16N in seston, F. distichus, and
Laminaria spp, which might indicate that fertilizer nitrogen was beginning to be incorporated into
algae.
Limpets feed mostly on microscopic algae. Periwinkles feed on films of diatoms, blue green
algae, filamentous algae, and occasionally on small pieces of macroscopic plants. M. edilus and
Balanus spp are filter feeders. Whelks are carnivores, and feed primarily on snails and limpets. With
a trophic level transfer, there generally is a small positive shift in the carbon isotope ratio (up to 1
%o), and a slightly larger shift in the nitrogen isotope ratio (up to 3 %o) (Fry and Sherr, 1984; Peterson
and Fry, 1987). Thus, the carbon isotope ratio of these consumers should be similar to that of its food
source, and the nitrogen isotope ratio should be slightly enriched in 515N (more positive).
The isotopic data shows that herbivores were primarily feeding on algae and not on seagrasses
(Figure 11.3). Similarly, the filter-feeders M. edilus and Balanus spp were utilizing suspended
organic matter derived from marine algal material. The isotopic difference between the filter-feeders
and seston demonstrated that these organisms were selectively removing material from suspension.
In addition, there was a difference between the food source ingested by M. edilus and Balanus spp.
Finally, the similarity among carbon isotope ratios and the clear trophic level shift in the nitrogen
isotope ratio between whelks and the limpets and periwinkles indicated the latter were a likely food
source for the carnivore.
The following preliminary conclusions can therefore be drawn: 1) there is no clear evidence for
the immediate incorporation of fertilizer carbon or nitrogen into the food-web of affected intertidal
beaches, and 2) the isotopic evidence does not indicate a shift from the expected trophic level
interactions among the organisms that were sampled in this study. In addition, it was demonstrated
that there was no immediate (weeks to months) effect of the addition of fertilizers on the food-web
structure in Prince William Sound.
Preliminary analysis from the summer of 1989 data indicated some assimilation of nitrogen from
the fertilizer. The sampling frequency through the summer, however, was not great enough to
establish a relationship between stable nitrogen isotope ratios in organisms and the fertilizer. The
summer 1990 ecological monitoring study was designed to confirm the trends observed during the
summer of 1989.
490
-------
14
12
10
3 Q ~
in
*> 6 -
4 -
2 -
Limpet
Periwinkle
M. edilus
Balanus spp
Whelk
Seston
Seagrass
i i I r^
-28 -24 -20
-16
-1 2
- 8
813C
Figure 11.3. Limpet, Periwinkle, M. edilus, Balanus spp, and Whelk Data
Showing 615N Plotted as a Function of 80. The Boxes
Indicate the Range of Values Measured for These Respective
Consumers. The Crosses are Averages and Ranges of 615N
and 513C for Algae, Seagrasses, and Seston (Refer to Table
11.1; Values from Tatitalek Island not included)
491
-------
STABLE ISOTOPES
The position of organisms in the food chain may be established by plotting the range of stable
nitrogen or carbon isotope values for each trophic level in an ascending trophic status. It is important
to establish these relationships to understand the assimilation of fertilizer nitrogen into algae and
consumers. Fractionation of isotopically heavy carbon isotopes is not significant between trophic
levels; therefore, a direct correspondence between the trophic levels indicates a dietary dependence
by the higher tropic level.
Food Chain Studies at EIrington and Disk Island, Summer 1990
On Disk Island stable carbon isotope values (518C) indicated that the brown and/or green algae
plus seston were important carbon sources for barnacles (Figure 11.4A). Limpets and periwinkles
relied on the same carbon sources; however, red algae was another possible food source. In addition,
the wide range in 513C of limpets and periwinkles suggested a greater number of carbon sources,
probably relating to the feeding strategy. Sources of carbon for eel blennys and whelks appeared to
be organisms that feed on brown and/or green algae. With some interesting exceptions, similar results
were noted on EIrington Island.
On EIrington, red algae did not appear to be a significant source of carbon for higher trophic
levels (Figure 11.4B). In fact, the 51SC of the Odonthalia on the two islands was significantly
different, suggesting different species were analyzed. Another interesting difference was a
substantially narrower range in £1SC of limpets and periwinkles on EIrington. The difference in
feeding exposure for these organisms on the islands may be related to the physical structure of the
beaches, since the slope of the beach on Disk Island was considerably shallower than EIrington. The
steeper slope may create more distinct ecological niches for organisms resulting in a smaller number
of species of algae that provide organic matter for the limpets and periwinkles.
Food chain studies have demonstrated that stable nitrogen isotopes are generally fractionated by
3%ofor each trophic step. Considering this fractionation, data from EIrington and Disk Islands suggest
that eel blennys may prey on all of the lower trophic consumers (Figure 11.5A and B). Also, whelks,
periwinkles, limpets and barnacles may depend on any of the primary producers as food sources.
Stable carbon isotopes provided an indication of the transfer of carbon sources from lower to higher
trophic levels. Nitrogen isotopes were used to place the organisms that were studied in the food chain
in an appropriate order. Furthermore, it is clear from the nitrogen data that the normal status of the
food chains were not altered by the addition of nitrogen.
492
-------
Disk Island 1990
Eel Blenny -
Whelk -
Periwinkle -
Limpet
Barnacle -
Seston
Red Algae
Brown Algae
Green Algae
Eel Blenny
Whelk
Periwinkle
Limpet
Barnacle
Seston
Red Algae
Brown Algae
Green Algae
-35 -30 -25 -20
13
8 C
-15
-10
Elrington Island 1990 B
-35 -30 -25 -20 -15
613C
-10
Figures 11.4A, 4B.
Stable Carbon Isotope Values for Ecological Samples
Collected at Disk Island and Elrington Island.
493
-------
Disk island 1990
Eel Blenny
Whelk
Periwinkle
Limpet
Barnacle
Seston
Red Algae
Brown Algae
Green Algae -
Eel Blenny
Whelk
Periwinkle
Limpet
Barnacle
Seston
Red Algae
Brown Algae
Green Algae
2 4 6 8 10 12 14
815N
Elrington Island 1990
6 8
15
6 N
10 12 14
Figures 11.5 A, 5B.
Stable Nitrogen Isotope Values for Ecological Samples
Collected at Disk Island and Elrington Island.
494
-------
Fertilizer Nitrogen Assimilation by Food Chains on Beaches, Summer 1990
The stable nitrogen isotope ratio £1BN of the fertilizer was approximately 0.0 %o, while isotope
ratios of organisms on the beach were substantially higher, ranging from approximately 8.0 %o to
12.0 %o. In contrast, 51BN of NH4+ and NOS" measured in pore waters from Disk Island were -1.4 %o
and -2.3 %o, respectively. Therefore, declining 516N values indicated that organisms on the beach
assimilated nitrogen from the fertilizer. Figure 11.6 presents 516N for algae and consumers from
Elrington Island from June to August 1990. Only brown algae appeared to assimilate fertilizer
nitrogen; 516N started to decline soon after fertilizer application was initiated. It appears that the
uptake of fertilizer nitrogen is related to the ecological niche of the algae in the rocky intertidal zone.
Brown algae, especially Fucus, are resistant to drying and thus would be found closest to the
fertilizer-treated beaches. Red and green algae are not as resistant to drying and would be found in
the lower intertidal and subtidal zones.
The transfer of fertilizer nitrogen from algae to consumers depends on the feeding strategies of
the organisms. Trends in the stable nitrogen isotope values in the consumer organisms suggested that
fertilizer nitrogen was transferred to barnacles, whelks, and eel blenny biomass (Figure 11.6B).
Barnacle £16N values appeared to respond dramatically to fertilizer nitrogen while more gradual
changes were evident in whelk and eel blenny isotope ratios. Barnacle stable nitrogen isotope values
declined rapidly because these organisms are immediately dependent on algae as a food source. Stable
nitrogen isotope values in the eel blennys and whelks declined more gradually because the isotope
ratios of organisms in higher trophic levels are more buffered from rapid changes than in the lower
trophic levels.
On Disk Island brown and red algae became isotopically lighter through June and July,
suggesting the uptake of nitrogen in the fertilizer applied to the beaches (Figure 11.7A). It is possible
that red algae consumed fertilizer nitrogen on Disk and not on Elrington because different species of
Odonthalia were examined and each was located in a different area of the intertidal zone. Another
interesting result was that £15N in seston samples taken from waters a couple of hundred meters of
the beach from Disk Island did not vary through the summer (Figure 11.7A). It was expected that
seston and the algae on the beaches would have the same nitrogen source. The data, however, suggest
two nitrogen sources. This supports our findings that organisms on the beaches assimilated fertilizer
nitrogen. In contrast to beaches on Elrington Island, consumer organisms did not appear to assimilate
nitrogen on Disk Island (Figure 11.7B). This may also be related to the physical differences between
the beaches. If ecological niches are more distinct on steeper beaches, then the higher trophic levels
on the shallower beaches of Disk Island would have more sources of substrate and, as a result, 615N
signals from fertilizer nitrogen could be diluted.
495
-------
10-r
8--
15
6"N
7--
-A- - Brown Algae
••• Gre«n Algae
-*— Red Algae
4. , 4. * ,
01
14 T
B
12 --
10 --
8 -
6 -
A -
r —-- -x — ••..
[" A
, * , * * , *, * *,
— 4- • Barnacle
••-•>• Periwinkle
— *— Limpet
- -x- Whelk
-»••»- Eel Blenny
^ i
<»
~
1C
(O
CU
Date
Figures 11.6 A, 6B.
Stable Nitrogen Isotope Values Over Time for Ecological
Samples Collected at Elrington Island.
496
-------
•—*— Brown Algae
-B- Green Algae
-•«•• Red Algae
-x- Seston
Date
B
513N
14-
13-
12-
10-
9-
8-
7 _
CD
' -f
^"^X^ "* _--•-
' ' -* —
*-
-------
STABLE ISOTOPES
Stable nitrogen isotope values indicated that fertilizer nitrogen was assimilated into intertidal
food chains on Elrington and Disk Islands. However, both stable carbon and nitrogen isotope values
indicated that the structure of the food chains were not altered by the application of fertilizer. It is
hyphothesized that the slope of the beaches on each island controls spatial distribution of the
organisms and, therefore, the distance from fertilizer application. Isotope ratios indicated that on
Disk Island (the beach with a shallower slope), limpets and periwinkles had a greater diversity of food
sources. As a result, nitrogen appeared to be assimilated only by the organisms located close to
fertilizer treated areas. Furthermore, assimilation df nitrogen into the heterotrophic food chain was
not observed on Disk Island, and may be due to dilution. These results support the use of stable
isotopes as tracers of elements in aquatic food chains.
ANCILLARY FIELD DATA, SUMMER 1990
The stable carbon (S13C) and nitrogen isotope (£16N) measurements of samples taken on Disk
Island are shown in Table 11.2. The nitrogen isotope data from Disk Island shows that although
suspended particulate matter (SPM) in the water removed from the well in the CUSTOMBLEN
fertilized plot within the beach (100 g/m2) was more positive (+7.5 %o), it was similar to SPM from
intertidal and cove waters, +5.8 %o and +5.7 %*, respectively. Finally, the «16N of bacterial assays and
nucleic acid extracts were all greater than 6 %o. Table 11.3 presents stable carbon and nitrogen isotope
values of carbon and nitrogen added to microcosms.
Carbon isotope values were similar to SPM from fertilized (-23.7 %o) and unfertilized (-24.3 %o)
wells, but more negative than SPM from intertidal and cove waters (-21.7 %o). The «1SC data of
bacteria from bacterial bioassays conducted in intertidal and cove waters (-22.0) was similar to the
SPM recovered from these waters. In contrast, the £1SC data from bacterial bioassays conducted with
well water (-26.5 %o) were more negative compared to the SPM, and, in addition, were significantly
more negative than bacteria incubated in intertidal and cove water. However, the 81SC of nucleic acid
extracted from intertidal waters (-20 %«) was similar to the $18C extracted from well water (-19.9 %o).
MICROCOSM EXPERIMENT
The microcosm study was designed to determine if bacteria growing on different carbon and
nitrogen sources could be distinguished isOtopically. The experimental design consisted of four
treatments of oiled gravel: I) fertilizer; 2) fertilizer and seagrass detritus; 3) seagrass detritus; and 4)
an untreated control. Table 11.3 presents stable carbon and nitrogen isotope values of carbon and
nitrogen added to the microcosm.
498
-------
TABLE 11.2. STABLE CARBON AND NITROGEN ISOTOPE DATA I ROM DISK ISLAND
AND RM ATII) BIOASSAY EXPFRIMFNTS
Location
Disk Island
SPM
Disk Island
SPM
Disk Island
SPM
Disk Island
Bioassay
Disk Island
N no Ir- ic Acid
Disk Island
Bioassay
Disk Island
Nucleic Acid
Disk Island
Bioassay
SPM = Suspended
Location
Intertidal and
cove water
Untreated control
plot
Fertilized plot
Intertidal and
cove water
Intertidal and
oovr walri
Porewater
Porewater
Porewater +
fertilizer
particulate matter
513C(%o)
-21.7+1.0(5)
-24.3±0.9(5)
-23.7+0.7(4)
-22.0+2.0(2)
-20.0(1)
-26.5(1)
-19.9(1)
-27.4(1)
516N(%o)
+5.8+0.3(6)
+7.5+1.7(6)
+5.7+1.4(6)
+8.4+0.1(2)
+6.8, 14.8
+8.0(1)
+9.1(1)
-7.4(1)
TABLE 1 1.3. STABLE CARBON AND NITROGEN ISOTOPE DATA OF CARBON AND
NITROGEN SOURCES IN THE MICROCOSM EXPERIMENT
Source
A'-Y(%«>)
Alf'N(%o)
Oil -30.1
Seagrass Detritus -16.8 +13.5
Fertilizer Nitrate - -2.4
Fertilizer Ammonium - +0.5
499
-------
STABLE ISOTOPES
The acridine orange direct counts (AODC) of bacteria, organic carbon content (PC), elemental
carbon to nitrogen ratio (C:N)wt, and isotopic data are shown in Figures 11.8-11.12. Cubic splines
have been drawn through the data to facilitate comparisons among treatments, and are not meant to
interpolate between data points. The solid dark line indicates the untreated control (oiled gravel
without fertilizer or seagrass).
In all treatments a maximum in bacterial abundance was observed within the first 5 days of the
experiment, followed by a sharp decline (Figure 11.8). Both fertilizer treatments had a secondary,
although smaller, maximum occurring between days 7 and 11. Organic carbon concentrations
decreased uniformly during the first 5 days of the experiment, and then for the unfertilized
treatments remained relatively constant (Figure 11.9). In contrast, PC in the fertilized treatments
increased after day 9. Except for the untreated control, elemental carbon to nitrogen ratios by weight
((C:N)wt) were in the range expected for bacteria at the beginning of the experiment (4 to 6; Figure
11.10). As the experiment continued, however, (C:N)wt increased in all treatments. The (C:N)wt
variation among treatments was also greater at the end of the experiment compared to the beginning
of the experiment.
The 61BN of the seagrass was +13.5, while the «16N of NO3' and NH4+ from the CUSTOMBLEN
fertilizer was -2.4 and +0.5, respectively (Table 11.3). Throughout the experiment, the nitrogen
isotope ratio of the untreated control remained relatively constant (+8.1+1.4 %o), while all the
treatments decreased in comparison to the untreated control (Figure 11.11). This decrease was much
more evident in the fertilized treatments where values at the end of the experiment were similar to
the fertilizer.
The 51SC of the oil and seagrass were -30.1 and -16.8, respectively (Table 11.3). As observed
with the 516N of the untreated control, the £13C of the untreated control was also relatively constant
(-25.8+0.8 %o). In contrast, the values for the treated tanks changed significantly during the course
of the experiment (Figure 11.12). Both fertilizer treatments attained the lowest values at the end of
the experiment. The seagrass treatment without fertilizer, however, was higher than the untreated
control at the end of the experiment.
A comparison of the carbon isotope data measured on GF/F filter concentrates of bacteria with
carbon isotope data of nucleic acid extracts is given in Table 11.4. The latter data are less likely to
include carbon or nitrogen from non-bacterial sources. The £1SC of nucleic acids decreased during
the experiment in all treatments. For the untreated control, the initial 513C value of nucleic acid was
500
-------
rr 4 -
2 -
0
1 I 1 I I
O Oil,Pert
D Oil,Seagrass,Pert
O Oil,Seagrass
A Oil
'
468
DAYS OF INCUBATION
10
12
Figure 11.8. Acridine Orange Direct Counts (AODC) of Bacteria for the
Four Microcosm Treatments Over Time.
_ 4
2 -
MICROCOSM II
O Oil,Pert
O Oil,Seagrass.Pert
o Oil,Seagrass
A Oil
O-
468
DAYS OF INCUBATION
10
12
Figure 11.9. Organic Carbon Content (PC) for the Four Microcosm
Treatments Over Time.
-------
MICROCOSM II
14
O Oil,Pert
O Oil,Seagrass,Pert
O Oil,Seagrass
A Oil
T 1 1 1 T
468
DAYS OF INCUBATION
10
12
Figure 11.10. Elemental Carbon to Nitrogen Ratio (CrN)^ for the Four
Microcosm Treatments Over Time.
MICROCOSM II
20 -
15 -
10 -
5
0 -
-5 -
I
O Oil,Pert
D Oil,Seagratt,Pert
O Oil,Seagrass
A Oil
I
I
468
DAYS OF INCUBATION
10
12
Figure 11.11. 615N for the Four Microcosm Treatments Over Time.
502
-------
MICROCOSM II
A
o
n
'Co
-20
-22
-24
-26
-28
-30
0
~Q \
O Oil,Pert
D Oil,Seagrass,Pert
o Oil,Seagrass
A Oil
8
DAYS OF INCUBATION
10
12
Figure 11.12. 813C for the Four Microcosm Treatments Over Time.
503
-------
STABLE ISOTOPES
TABLE 11.4. STABLE CARBON ISOTOPE DATA OF MATERIAL COLLECTED ON GF/F
FILTERS, NOMINALLY CONSIDERED TO BE BACTERIA, AND NUCLEIC ACID
EXTRACTS FROM THE MICROCOSM EXPERIMENT
Sample
Oil, Fertilizer
Oil, Fertilizer
Oil, Fertilizer
Oil, Fertilizer, Seagrass
Oil, Fertilizer, Seagrass
Oil, Fertilizer, Seagrass
Oil, Seagrass
Oil, Seagrass
Oil, Seagrass
Oil, Seagrass
Oil
Oil
Oil
Date
8/13/90
8/16/90
8/20/90
8/13/90
8/16/90
8/20/90
8/13/90
8/16/90
8/20/90
8/25/90
8/13/90
8/16/90
8/20/90
513C (%o)
GF/F
-26.1
-26.2
-27.7
-24.8
-23.5
-25.9
-23.5
-25.9
-23.9
-26.4
-25.1
-26.4
513C (%o)
Nucleic Acid
-29.4
-29.7
-33.3
-18.9
-24.2
-32.7
-18.6
-24.2
-32.7
-31.6
-26.8
-30.9
-33.4
similar to the filter, but as the experiment continued the 61SC of nucleic acids became more negative
than the filter concentrates. Initially, both seagrass treatments had heavy 61SC of nucleic acids. These
values subsequently decreased, becoming more negative than the filter concentrates. This can be
contrasted to the oil and fertilizer treatment (more negative 51SC of nucleic acids) compared to filter
concentrates throughout the experiment.
The isotope ratios of organic matter less than 1 /* that was recovered from the microcosms and
a GF/F filter (termed filter concentrates of bacteria), varied between treatments. Although AODC
and C:N data were consistent with a bacterial source of organic matter at the beginning of the
experiment, another source with a higher C:N was present at the end of the experiment on these filter
concentrates. In general, the nucleic acid carbon isotope data suggested the other organic sources
were isotopically heavier (see Table 11.4). The most positive 613C of nucleic acids was observed at
the start of the experiment in the seagrass treatments. This suggests the bacteria were using seagrass
carbon as a substrate source (compare Tables 11.3 & 11.4). All treatments reached light 61SC values
at the end of the experiment. Petroleum hydrocarbons, which have more negative carbon isotope
values (-28 to -30 %o), could be the source of carbon resulting in the lighter values.
504
-------
The nitrogen isotope ratio at the end of the experiment was lighter in all treatments compared
to the untreated control (oiled-gravel only) and was similar to the CUSTOMBLEN in the fertilizer
treatment. This suggests fertilizer nitrogen was utilized during the metabolism of available carbon
substrates.
The data from the field do not resemble the microcosms data. The fertilized plot within the
beach on Disk Island did not reach the lighter nitrogen or carbon isotope ratios. Values similar to
those observed in the microcosm were found only when CUSTOMBLEN was added to a bioassay,
where high fertilizer concentrations were available to bacteria (see Table 11.2). This result suggests
that the bacteria in the fertilized plot within the beach were not subjected to the same concentration
of nutrients as the bacteria in the microcosm.
505
-------
SECTION 12
MODELING OIL BIOREMEDIATION
APPROACH AND ANALYSIS
A structured protocol was formulated to devise a predictive methodology to determine the time
needed to clean up the oil contamination in Prince William Sound beaches using bioremediation. A
corollary objective was to quantify the impact on clean-up time that various beach material
amendments or treatments (the use of fertilizers) have on shortening the period required for oil
decontamination using bioremediation.
The overall strategy was to formulate a predictive algorithm calibrated using data collected
generated using bench-scale laboratory microcosm systems. This particular strategy was more
economical and less time consuming for evaluating the effectiveness of various treatments on
enhancing bioremediation rates. Ideally, it is desirable to quantify the impact of the rate enhancement
on lessening clean-up time.
To develop the methodology, the following five tasks were implemented:
• A model was devised utilizing accepted procedures for biological systems that can predict
clean-up times.
• Key modeling parameters were identified.
• Existing laboratory and field data were analyzed using modeling (i.e., attempt to use the
laboratory data to predict field results).
• What would be needed to improve the model was determined.
• Additional field and laboratory efforts for improving predictive capability and streamlining
batch data collection and analysis were suggested.
The four major areas relevant to development of a methodology for predicting bioremediation
rates at the Prince William Sound beaches are as follows:
Review and synopsis of existing predictive modeling procedures for biological waste
treatment systems.
Proposed modeling approach for the Prince William Sound bioremediation systems.
506
-------
• Review of laboratory and field data collected to date.
• Methodology to calibrate the bioremediation model.
It was deemed appropriate to develop the predictive methodology based on analogous work
which involved the predictive modeling of biological wastewater systems. For this work,
respirometric techniques provided the means to calibrate the model cost-effectively.
Modeling Procedures for Biological Waste Treatment Systems
The effluent quality for a full-scale biological wastewater treatment plant was predicted by
calibrating a model by using kinetic coefficients that were obtained via analysis of respirometric
measurements of plant biomass (Colvin et al., 1990).
Modeling procedures for biological waste treatment systems are usually based on being able to
quantify and hopefully control or influence the specific growth rate of microorganisms that are
degrading target waste components. As an example, consider the modeling of a simple chemostat
system depicted in Figure 12.1. A steady-state mass balance for cell concentration, X, shows that the
growth rate is controlled by the dilution rate, which is a product of the influent substrate flow rate
divided by the reactor volume (F/V). Since cell growth rate is also a function of substrate
concentration, S, then specific output substrate concentrations can be selected by adjusting the flow
rate, which in turn controls the growth rate.
This basic procedure was used to formulate a predictive modeling approach for activated sludge
systems which are used for treating municipal and industrial wastewaters (Gaudy and Gaudy, 1980).
In this case, predictive equations are derived by writing mass balances for both substrate (waste) and
biomass concentrations around the reactor. An appropriate equation is used to relate growth rate to
substrate concentration and the equations are solved to obtain predictors for substrate and biomass
concentrations. This approach is enhanced by using respirometric (oxygen uptake) measurements of
the biomass on the target waste to quantify the relationship between growth rate and waste
concentration; in other words, the respirometric technique affords a relatively rapid and
cost-effective calibration of the process model for the activated sludge system. Such an approach was
successfully employed to obtain accurate predictions of a full-scale activated sludge system treating
a municipal waste that had received a heavy industrial contribution. A field study showed that
507
-------
MODELING
Flow
Substrate
Flow
Substrate,
Biomass
Flow rate
reactor volume
Figure 12.1. Chemostat Reactor.
predicted and measured values for effluent quality were in agreement and that a process model was
calibrated using separate bench-scale respirometric measurements is a reliable predictor for activated
sludge systems.
Modeling Approach for Bioremediatioo
Both biological waste treatment systems and bioremediation techniques are dependent on
microbial growth and subsequent utilization of target waste components. This feature enables one to
assume with a degree of certainty that a reasonable modeling approach for bioremediation systems
should be based on the growth rate characteristics of the biomass. Consequently, a similar mass
balance approach can be employed to obtain a predictive model.
508
-------
The bioremediation model for the Prince William Sound beaches was derived by using the
diagram shown in Figure 12.2. The mass balances for substrate and biomass are written around the
control volume of beach material.
M(dX/dt) - /iXM - F, (1)
M(dS/dt) - - /iXM/Ys (2)
X represents biomass concentration in the beach material in mg/kg; S is the oil concentration in
mg/kg; M is the mass of beach material in kg; /* is the specific growth rate in days'1; F8 is the
sloughing rate of biomass in mg/day; Yt is the cell yield in mg/mg; and t is time in days.
It is assumed that biomass does not accumulate and is relatively constant because of sloughing
that occurs due to tidal washing of the beach material. Using this assumption, dX/dt equals 0 and
X is assumed to be at a relatively constant value.
fiXM = F. (3)
dS/dt = - pX/Yt (4)
The substrate mass balance equation can now be solved using integration and the initial
condition that at t=0, S=S; which is the initial oil concentration in the control volume. This procedure
provides a predictive equation for S, the oil concentration in the beach material as a function of time.
S = ^ - MXt/Yt (5)
It is appropriate at this point to compare and contrast this equation with the predictive equations
employed in biological wastewater treatment systems (e.g., the chemo;tat). With the sloughing
assumption, it is also assumed that the growth rate of the cells is eficcu^ely constant. This is
analogous to a chemostat at steady-state, since these reactors attain steady growth rates when influent
flow rates are constant.
509
-------
Oil-contaminated beach material
of mass "M"
F8, Biomass
Sloughing Rate
Beach Material
Figure 12.2. Schematic for Modeling the Beach Bloremediatlon System
510
-------
MODELING
The question of model calibration must be addressed. For a biological wastewater treatment
system, this is accomplished by quantifying the relationship between specific growth rate and
substrate concentration. This approach is unlikely to be applicable for an oil bioremediation
application because substrate concentration will not be the limiting factor. In other words, the rate
at which the target organics solubilize will control the availability of the carbon source to the
microbial populations. Consequently, an alternative calibration strategy was warranted. This strategy
consisted of utilizing bench-scale laboratory methods to determine the growth rates the biomass will
achieve under various conditions. For example, flasks can be set-up to determine growth rate
potential for control conditions, for various amendments or soil additives which are employed, or for
extremes in weather conditions (e.g., cold temperatures). The laboratory approach simulated the
target microcosm and then measured the resulting microbial growth rate, presumably using a surrogate
indicator such as oxygen uptake or carbon dioxide evolution to calculate a cell growth rate.
This modeling approach affords separate consideration for the influence of microbial growth
rate and biomass concentration. If one merely uses laboratory bench-scale carbon dioxide evolution
or oxygen uptake data to compute predicted oil degradation rates, an aggregate oil degradation rate
which is the resultant product of the interactive effect of biomass growth rate and biomass
concentration can not be predicted. If the predictive methodology is incapable of meeting these
criteria, it will be difficult to extrapolate results and identify design or operational modifications
which will enable engineers and scientists to optimize and refine bioremediation systems.
Specifically, if growth rate is optimized and the cells are achieving maximum rates, the model
indicates that an increase in biomass concentration will enhance oil degradation rates. The question
then is how to engineer an increase in biomass concentration in a bioremediation system. This aspect
is discussed in greater detail in another part of this section.
Review of Field and Laboratory Data Collected the Summer of 1990
Existing data were reviewed and assessed for applicability to modeling. Both field and
laboratory data were collected by EPA (see Sections 8 and 9). The pi^arry field data collected
consisted of:
• Weight of oil in beach material
• Biomass in beach material using a modified most probable number (MPN) technique for
hydrocarbon degraders
511
-------
MODELING
• Hydrocarbon degradation activity in beach material using radiolabeled phenanthrene as a
reporter chemical
The bulk of the bench-scale laboratory data in the summer of 1990 consisted of microcosm
studies. In addition to residual oil and biomass data, oxygen uptake and carbon dioxide evolution
measurements were taken as a means to indicate microbial activity.
Microcosm studies were initially performed in a batch mode. Beach material was added to a
flask and the various parameters were monitored with time. These studies were later modified to
reflect the impact of tidal action on the bioremediation rates. This modification changed the reactor
configuration of the bench-scale system to a mode that ameliorated biomass retention and prevented
biomass accumulation.
Methodology to Calibrate Bioremediation Model
The most direct way to calibrate the bioremediation model is to calculate microbial growth rates
by measuring increases in cell numbers. This approach was utilized in shake flask batch methods
where a relatively easily measured parameter such as absorbance was employed as a surrogate
indicator of microbial numbers. The most practical way to determine microbial numbers in oil
bioremediation systems is the modified MPN technique that measures the number of hydrocarbon
degraders. This methodology is relatively labor intensive when considering the number of data points
needed to compute a growth rate. Alternatively, oxygen uptake or carbon dioxide evolution data can
be used as a surrogate indicator of microbial growth and "translated" into cell numbers or mass. The
assumption is that cell replication is occurring simultaneously with oxygen uptake or carbon dioxide
evolution. As discussed previously, an analogous approach was successfully employed to calibrate
biological wastewater treatment models. It is recommended that the bench-scale systems which are
currently used can be employed for generating the data needed for calibrating a bioremediation
model. Modifications may be necessary to streamline or enhance the testing methodology.
The field data were analyzed using the proposed bioremediation model as represented by
equation 5. To apply the model the values of a number of parameters had to be assumed. The value
of cell yield was assumed to be 0.5 mg cell/mg oil COD (chemical oxygen demand) and the mass of
oil was converted into an oxygen demand equivalent using a conversion factor of 2.8 mg COD/mg
oil. Biomass concentration estimates represented a problem since most of the data were collected
512
-------
MODELING
using an MPN method and the results were reported in logarithms of numbers of organisms (e.g., 106,
107, etc.). This means that the smallest increment between biomass concentration readings is one
logarithm. In terms of the MPN test this is reasonable, but in terms of modeling this is problematic.
Consequently, when both field- and bench-scale data were analyzed, different values of biomass
concentration were assumed for calculating growth rates or for predicting oil degradation rates. MPN
values were based on the results of measurements made by the University of Alaska (Fairbanks), and
they also provided input to assumptions for biomass concentrations ranges for both the bench and
field work.
For the bench-scale work, equation 5 was employed to calculate microbial growth rates based
on oxygen uptake or carbon dioxide evolution data. For analysis of the field data, the growth rates
that were measured in the batch tests were used in equation 5 to predict oil removal rates in the field.
To calculate batch growth rates and to apply the model for predicting field results, a number of
assumptions had to be made and are detailed below.
Modeling Assumptions
The first major assumption is that it is reasonable to use carbon dioxide evolution data or oxygen
utilization data as a surrogate indicator of cell growth; in other words, it is assumed that cell
replication or proliferation is occurring simultaneously with substrate or oil degradation. The next
step is to "transform" these data into growth data. This is accomplished by using the mass balance
principle as given in the equation below:
Substrate + O2 = CO2 + Cells (6)
The substrate in this case is oil. It is assumed that the oil has a chemical oxygen demand (COD) value
of 2.8 mg COD/mg oil; it is also assumed that the oil contains 0.8 mg C (carboa)/mg oil. Carbon
dioxide contains 0.273 mg C/ mg CO2 (12/44).
The cell yield value, Yt, is taken as 0.3 mg X/ mg COD or 1.05 mg X/ mg C ((0.3)(2.8)/0.8).
COD and C values for biomass are computed assuming a cell stoichiometry of C5H7O2N; these are
1.42 mg COD/ mg X and 0.57 mg C/ mg X.
513
-------
MODELING
Equation 5 is manipulated to compute growth rate from the bench-scale data:
fi = Yt(Si - S)/(Xt) (7)
In equation 7, the term "Sj - S" represents the amount of oil degraded; this will be calculated or
estimated using oxygen uptake or carbon dioxide evolution data. The oxygen uptake or CO2 evolution
data is converted into oil degradation information by using the aerobic balance principle for microbial
systems:
ACOD = O2 uptake + ACOD cells (8)
OR
ATOC - CO2 evolved + ATOC cells (9)
These equations merely state that the degraded substrate (ACOD or ATOC) is accounted for in new
cells (ACOD cells or ATOC cells) and oxygen uptake or carbon dioxide evolution (provided that there
are insignificant physical removal or chemical oxidation mechanisms operative). Our model
assumption is that with both the beach and bench-scale reactors, the biomass concentration is at
steady state due to the effects of sloughing. This means that the ACOD cells and ATOC cells terms
are equal to the mass of COD or TOC which is incorporated into biomass production. Equation 7 can
then be rewritten to allow computation of growth rates directly from the bench-scale oxygen uptake
or carbon dioxide evolution data.
The first step is to write expressions for the ACOD cells and ATOC cells terms in terms of
biomass production:
ACOD cells « 1.42/iXt (10)
ATOC cells - 0.53jiXt (11)
These terms represent the mass of COD or TOC which is incorporated into cell material; these are
used in conjunction with the COD and TOC balance equations to provide an expression for the SrS
term in equation 7.
514
-------
MODELING
ACOD - (S; - S) - O2 uptake + 1.42/iXt (12)
ATOC = (S; - S) = CO2 evolved + 0.53/iXt (13)
These equations are now inserted into equation 7:
H = (Yt/Xt)(O2 uptake + 1.42/iXt) (14)
H = (Yt/Xt)(CO2 evolved + 0.53^X1) (15)
These equations can now be solve to provide expressions for /i:
H = O2 uptake/((l/Yt - 1.42)Xt) (16)
H = CO2 evolved/((l/Yt - 0.53)Xt) (17)
Regarding other model parameters, X represents the biomass concentration on the beach
material; X is calculated by using an MPN value and by assuming that one cell weighs 10~12 grams.
Two or three MPN values will be assumed when calculating growth rates from the bench-scale data.
Time is represented by "t".
For predicting field results, equation 5 is used as a predictor for oil concentration in the
beach material:
S - Ss - /*Xt/Yt (5)
S; represents the initial oil concentration in the beach material. This equation predicts the time it
takes to achieve oil clean-up levels in the beach material.
It is appropriate at this point to discuss the utility of equation 5 and the modeling approach in
general. The overall strategy is to use the bench-scale tests to quantify the growth rates which the
cells can achieve while degrading the oil. If a particular amendment or additive enhances the growth
rate, and consequently the oil degradation rate, equation 5 will predict the impact on clean-up time.
515
-------
MODELING
In other words it will quantify the reduction in clean-up time that can be expected by using various
additives to enhance bioremediation rates.
The biomass term in equation 5 demonstrates that the overall rate of oil degradation is a
combined result of biomass growth rate and biomass concentration. This means that merely increasing
microbial growth rates represents a relatively limited technique for increasing oil degradation rates.
Specifically, much faster oil degradation rates can be achieved if the biomass concentration on the
beach material is increased. Practically speaking, this is an engineering question. For example,
consider the extended aeration activated sludge process where biomass is deliberately retained to
achieve relatively high reactor degradation rates by maintaining high biomass concentrations. For
bioremediation projects such as the one on the Prince William Sound beaches, one way to increase
biomass concentration is to add an additional carbon source that enables the beach to support a higher
biomass concentration that may be more resistant to sloughing.
From an analytical perspective, the significance of the biomass term in equation 5 is somewhat
negative, since it points out that accurate bench estimates of growth rate are contingent on the ability
to quantify biomass concentrations in the bench-scale systems. Similarly, accurate beach material
biomass concentrations are needed to apply the current version of the model for predicting oil
degradation rates. In view of the accuracy of the existing biomass concentration data (+ or - one
logarithm), it was decided to generate growth rate information by analyzing the bench-scale data by
assuming two or three different biomass concentrations representative of current MPN data. This
results in producing two or three different growth rates which are predicted by each set of batch data.
These growth rates are, in turn, used to predict oil degradation rates at the various test plots.
Different growth rate and biomass concentration combinations are employed in the field data
predictive modeling effort and the predictive results can be compared with actual field data.
Bench-Scale Data Analysis
To evaluate the efficacy of using the bench-scale data to calibrate the predictive model, it is
appropriate to use equation 16 or 17 (for O2 and CO2 data, respectively) to compute growth rates
using the Micro-Oxymax data. Growth rate calculations were made for different biomass, x,
concentrations.
516
-------
MODELING
Results of bench-scale evaluation of the microbial activity on the oil adhering to the beach
material were analyzed using the current modeling approach. Typical computational results are
presented in Tables 12.1 through 12.4. These data present calculated growth rates for oxygen uptake
data that were reported elsewhere (USEPA, 1990). As expected, the impact of varying the values of
the biomass concentration by one logarithm produces a tenfold variation in the values of the
calculated growth rates. For most biokinetic work, the quantification of biomass concentration or the
time course variation in biomass or cell number concentration is an essential component of the data
collection effort. This does not, however, diminish the significance of the results presented in these
tables. This is stated because the relatively rough method of analysis (i.e., lack of more refined
biomass concentration data, cell yield values, data on COD and TOC content of oil and cells, etc.) of
the bench-scale data produced computational results for growth rate which are reasonable; calculated
p values range from 0.005 to 0.10 hours"1 for Tables 12.1 and 12.2. It should be noted that this
variation does contain a comparison of a control and a fertilizer amended sample of beach material.
A comparison of the kinetic calculations in Tables 12.1 and 12.2 makes a good point regarding
the potential significance and effectiveness of fertilizers and other bioremediation amendments. The
comparison of growth rates shows that the populations which were growing on the amended beach
material achieved growth rates that were approximately double those of the untreated control systems.
This can be concluded because it is reasonable to assume that both systems had approximately the
same biomass concentration. As shown by equations 14 and 15, an observed doubling of the rate of
oxygen uptake or carbon dioxide evolution realizes a doubling of the cell growth rate; knowledge of
the biomass concentration is required in order to compute the value of the growth rate in each
bench-scale reactor system. However, it is safe to conclude that the comparative bench evaluations
provide a quantitative relative comparison (e.g., 10%, 50%, etc.) of the impact of fertilizers and other
amendments on cell growth rates on the beach material. For quantitatively extrapolating the bench
data for predicting field results, more refined biomass estimates are needed for both beach and field
activities.
Field Data Analysis
Two sets of corresponding bench and field data collected the summer of 1990 for studies at
Elrington Island were analyzed and the results are given in Tables 12.3 and 12.4. The average of the
computed bench-scale growth rates were used in conjunction with equation 5 to predict oil
degradation rates in the test plots and these analytical results are given in Tables 12.5 and 12.6.
517
-------
MODELING
TABLE 12.1. GROWTH RATE CALCULATIONS FOR 350 PPM NITROGEN FLASK
Time
(hours)
12
24
36
48
60
72
O2 Uptake'
per 100 g Beach
Material in Moles
65
155
225
300
375
445
Calculated /
X-107 MPN
perg
Beach Material
0.09150
0.109
0.1050
0.1050
0.1050
0.1040
*b (h'1)
X=108 MPN
perg
Beach Material
0.0092
0.00109
0.0105
0.0105
0.0105
0.0109
a Data obtained from Figure 6b (USEPA, 1990)
O2 uptake
1
(Equation 16)
xt
TABLE 12.2. GROWTH RATE CALCULATIONS FOR 0.35 PPM NITROGEN FLASK
Time
(hours)
12
24
36
48
60
72
O2 Uptake'
per 100 g Beach
Material \i Moles
50
85
120
155
200
240
Calculated fi
X«107 MPN
perg
Beach Material
0.0690
0.060
0.0560
0.0543
0.0560
0.0560
:b (h-1)
X=108 MPN
perg
Beach Material
0.0069
0.0060
0.0056
0.0054
0.0056
0.0056
a Data obtained from Figure 6b (USEPA, 1990)
O2 uptake
1
(Equation 16)
xt
518
-------
MODELING
TABLE 12.3. GROWTH RATE CALCULATIONS FOR ELRINGTON ISLAND CO-
EVOLUTION DATA, FERTILIZER-TREATED BEACH MATERIAL"
Calculated /ib (IT1)
Time
(hours)
MC02
per 100 g
Beach Material
X-107 MPN
per g
Beach Material
X=108 MPN
per g
Beach Material
12
24
36
48
60
150
270
370
480
580
0.356
0.320
0.292
0.284
0.275
0.0356
0.0320
0.0292
0.0284
0.0275
" Data obtained from Figure 4a (USEPA letter, August 7, 1990)
CO, evolved
1
Yt-0'53 Ixt
TABLE 12.4. GROWTH RATE CALCULATIONS FOR ELRINGTON ISLAND CO,
EVOLUTION DATA, UNTREATED CONTROL BEACH MATERIAL*
Calculated Mb (h"1)
Time
(hours)
12
24
36
48
60
/iC02
per 100 g
Beach Material
70
155
220
280
325
X=107 MPN
Per g
Beach Material
0.165
0.185
0.174
0.166
0.154
X=108 MPN
per g
Beach Materal
0.0165
0.0183
0.0174
0.0166
0.0154
a Data obtained from Figure 4b (USEPA letter, August 7, 1990)
CO2 evolved
~j (Equation 17)
Yt-0'53 Ixt
519
-------
MODELING
TABLE 12.5. PREDICTED S (RESIDUAL OIL) VALUES, MG OIL/KG FOR
ELRINGTON ISLAND UNTREATED CONTROL8
FOR DIFFERENT ^ AND MPN VALUESb
M = 0.170 h'1
Assumed X, MPN
Time
Date (days)
6/29/90 0
7/12/90 13
7/19/90 22
106
--
7,937
7,893
107
--
7,368
6,930
10*
--
1,679
0
H = 0.017 h'1
X, MPN
106
--
7,994
7,989
107
--
7,937
7,893
108
--
7,368
5,930
Actual
8,000
8,000
7,700
a Actual data obtained from Figure 3b (USEPA Letter, August 7, 1990)
b Predictions made using equation (5)
TABLE 12.6. PREDICTED S (RESIDUAL OIL) VALUES, MG OIL/KG FOR
ELRINGTON ISLAND FERTILIZER TREATED BEACH MATERIAL'
FOR DIFFERENT n AND MPN VALUESb
H = 0.31 h"1
X, MPN
Time
Date (days)
6/29/90 0
7/12/90 13
7/19/90 22
106
—
11,884
11,805
107
--
10,847
10,049
108
--
470
0
H = 0.031 h'1
X, MPN
106
--
11,988
11,980
107
--
11,885
11,805
108
--
10,847
10,050
Actual
12,000
10,500
10,000
a Actual data obtained from Figure 3c (USEPA Letter, August 7, 1990)
b Predictions made using equation (5)
520
-------
MODELING
The results in Tables 12.5 and 12.6 are noteworthy because several of the growth rate and
biomass combinations resulted in yielding reasonable predictions using equation 5. Based on previous
knowledge of microbial kinetics and other supporting field data, there would be a tendency to believe
that the combination of the lower growth rate values and the MPN values of 108 tend to be more
realistic. Growth rates in excess of 0.10 hours"1 are somewhat high for organisms in natural
environments, especially organisms that must contend with difficult -to-degrade substrates such as
petroleum hydrocarbons. In addition, results of field analyses indicate that MPN numbers of 108are
reasonable for hydrocarbon degraders. These facts and the favorable results of the predictive
modeling exercise suggest that the approach advocated herein merits further consideration.
These results are especially positive considering that the predictions of oil degradation rates were
largely performed using data collected from bench-scale systems. In other words, the estimates of
growth rate potential, as quantified by the measured p values, for the control and fertilizer
remediation conditions were used to calibrate the predictive model.
SUMMARY AND CONCLUSIONS
The following recommendations are suggested for future work regarding predictive model
development
• To devise a rigorous field test of the modeling approach, it is not necessary to totally
simulate conditions at the beaches to obtain verification of, or the data to modify, the
modeling approach presented. For example, it is suggested that a modified field and
associated bench-scale testing regimen be implemented which eliminates the impact of the
tidal washing effect. This program can be performed by isolating a substantial portion of
sieved beach material and confining it to a location where the tidal effect is nullified.
Separate samples can be removed from this test area for respiro; netric (O2 uptake or CO2
evolution) testing to obtain kinetic parameters (i.e., model calibration). Different test areas
can be devised to evaluate and model the impact of different trea;n;ents or amendments on
bioremediation rates. The advantages of this approach are that it simplifies the system model
and permits focus on evaluating and/or modifying the modeling approach. Additionally, the
"biomass retention hypothesis" (i.e., the idea that enhancing biomass retention in the beach
bioremediation systems will accelerate overall bioremediation rates), can be tested in a more
structured and controlled manner.
521
-------
MODELING
• Additional work for improving and/or modifying the modeling approach should target issues
such as the refinement of system biomass estimates (+ or - one logarithm is simply not
acceptable for model testing purposes) and defining other parameters such as the cell yield
and cell carbon content. Consideration should also be given to the use of a substrate
measurement parameter which is easily convertible to COD or TOC readings. This will
enable analysts to check mass balance calculations during batch respirometric tests.
• Grazing of bacteria may play a significant role in the modeling procedure. If data are
available which enables computation of grazing rates (which are presumed to be
commensurate with a decay rate), then it should be relatively straightforward to compute the
aggregate effect of grazing and ascertain whether or not it warrants inclusion in the
modeling algorithm. For example, if the number of protozoa per gram of beach material and
the associated "grazing rate" are known, a biomass attrition rate attributable to grazing can
be computed. This rate can be compared to current estimated values for growth rates and
the associated biomass concentrations to determine the significance of the potential impact
of grazing.
522
-------
SECTION 13
COMMERCIAL PRODUCTS TESTING
THE FOLLOWING EXPERIMENTS WERE NOT FUNDED OR SUPPORTED BY
THE EXXON CORPORATION. THIS WORK WAS CONDUCTED BY EPA'S
RISK REDUCTION ENGINEERING LABORATORY IN CINCINNATI, OHIO
WITH SUPPORT FROM THE NATIONAL ENVIRONMENTAL TECHNOLOGY
APPLICATIONS CORPORATION (NETAC), FOR PROPOSAL SELECTION
AND HELD ANALYTICAL SUPPORT. IT HAS BEEN INCLUDED IN THIS
REPORT FOR INFORMATIONAL PURPOSES ONLY.
The commercial products testing effort conducted by EPA's Risk Reduction Engineering
Laboratory (RREL) in Cincinnati, Ohio, with support from NETAC, involved: 1) laboratory
screening of ten commercial products; and 2) field testing of those commercial products in Prince
William Sound that met EPA's three criteria for field testing. The results of these two phases are
described below.
LABORATORY STUDIES
After the EPA oil spill bioremediation project showed in the summer of 1989 that
bioremediation of oil-polluted beaches was enhanced by fertilizer addition, the question then arose
whether further enhancement was possible with the addition of microbial inocula prepared from oil-
degrading populations not indigenous to Alaska. Seeding experiments have been conducted in
previous studies with mixed results (Leahy et al., 1990). In a recent study, Dott et al., 1989, compared
fuel oil degradation rates of activated sludge microorganisms with nine different commercial bacterial
cultures in separate laboratory flasks. They found that the rate and extent of n-alkane and total
hydrocarbon degradation by the diverse populations in activated sludge were significantly higher than
any of the highly adapted commercially available cultures. Most success with biodegradation
enhancement by allochthonous microbial cultures has been achieved when chemostats or fermentors
were used to control conditions or reduce competition from indigenous microflora (Wong et al., 1988).
In February, 1990, EPA issued a public solicitation for proposals to the bioremediation industry
to test the efficacy of commercial microbial products for enhancing degradation of weathered Alaskan
crude oil. The Agency commissioned NETAC, a non-profit organization dedicated to the
commercialization of environmental technologies, to convene a panel of experts to review the
proposals and choose those that offered the most promise for success in th; field. Forty proposals
523
-------
COMMERCIAL PRODUCTS
were submitted, and 11 were selected for the first phase of a two-tiered testing protocol (only 10 were
tested because one company did not participate). The laboratory testing consisted of electrolytic
respirometers set up to measure oxygen uptake over time, and shake flasks to measure oil degradation
and microbial growth. If one or more products were found effective, the second tier would take
place, consisting of small field plots on an actual contaminated beach in Prince William Sound in the
summer of 1990. The first phase of testing, the laboratory batch flask and respirometric evaluations,
are discussed below.
The objective of the laboratory protocol was to determine if commercial bioremediation products
could enhance the biodegradation of weathered crude oil to a significantly better degree than that
achievable by simple fertilizer application. Testing was conducted in a controlled and closed
environment designed to give quick results under ideal conditions. It was not meant to simulate the
open environment of the oiled beaches of Prince William Sound, where conditions are in a constant
state of flux with respect to tidal cycles and washout, temperature variation, climatic changes,
freshwater/saltwater interactions, etc. The organisms inside the respirometer vessels were in
*
continuous contact with the oil, seawater, and nutrients that were added initially. The seawater was
not replenished every 12 hours as is the case in nature. The test was merely a screening procedure
designed to determine if there was sufficient enhancement by the commercial additives to justify
proceeding to the next tier of testing. To proceed to the field phase, three lines of evidence were used
for decision-making: rapid onset and high rate of oxygen uptake, substantial growth of oil degraders,
and significant degradation of the aliphatic and aromatic hydrocarbon fractions of the weathered
Prudhoe Bay crude oil.
The 10 companies participating in the laboratory testing phase were (in alphabetical order):
Alpha Environmental, Bioversal, Elf Aquitaine, ERI-Microbe Masters, Imbach, Microlife Technics,
Polybac, Sybron, WasteMicrobes, and Woodward Clyde. Of the 10 products tested, 8 were microbial
and 2 were non-microbial formulations (i.e., fertilizers and/or dispersants).
METHODS
Electrolytic Respirometry
The studies were conducted using four automated continuous oxygen-uptake measuring Voith
Sapromats (Model B-12), which consist of a temperature-controlled water bath containing measuring
524
-------
COMMERCIAL PRODUCTS
units, and a recorder for digital indication and direct plotting of the oxygen uptake curves. The
measuring units comprise 12 reaction vessels each with a carbon dioxide absorber mounted inside, 12
oxygen generators each connected to its own reaction vessel by tubing, and 12 pressure indicators
connected electronically to the reaction vessels. The measuring units are interconnected by tubing,
forming an air-sealed system, so that the atmospheric pressure fluctuations do not adversely affect
the results.
Depletion of oxygen by microbiological activity creates a vacuum, which is sensed by the
pressure indicator. The oxygen generator is triggered to produce just enough oxygen to
counterbalance the negative pressure. The electrical current used to generate the oxygen is measured
by the digital recorder, and the data are converted directly into mg/L oxygen uptake. The CO2
produced by microbial activity is absorbed by soda lime. The oxygen generators of the individual
measuring units are electrolytic cells that supply the required amount of oxygen by electrolytic
decomposition of copper sulfate/sulfuric acid solution.
A recorder/plotter constructs an oxygen uptake curve as a function of time and displays it on
the computer screen while digitally saving the data on disc. For frequent recording and storage of
oxygen uptake data, the Sapromat B-12 recorders are interfaced to an IBM-AT personal computer
via the Metrabyte interface system. A software package, Labtech Notebook version 2.8 (Laboratory
Technologies Corp., Wilmington, MA) allows the collection of data at 15 minute intervals.
Experimental Design
All commercial products were tested in duplicate at the concentration recommended by the
manufacturer. Each 500 mL respirometer flask was charged with the following materials in the order
listed: 250 mg weathered crude oil, 250 mL seawater from Prince William Sound, and commercial
product. Seawater was prepared as follows: 25 g of oiled rocks from a contaminated beach in Prince
William Sound were placed in a 4-L flask; 2 L of seawater were added; and the mixture was shaken
for approximately 30 minutes to wash off all microbial inoculum from the rocks. The flask contents
were allowed to settle, and the supernatant was mixed with more seawater for use in the respirometer
vessels. Table 13.1 presents the summarized experimental design showing all control and experimental
flasks.
525
-------
COMMERCIAL PRODUCTS
TABLE 13.1. EXPERIMENTAL DESIGN FOR RESPIROMETRIC STUDIES
Reaction Vessel Nutrients Weathered Commercial Seawater TOTAL
Oil Product
TEST FLASKS:
Tpn + + + + 20
Fj, F2 + + _ + 2
CONTROL FLASKS:
CPn +
PF PF 4-
v_*r 1 , v^r* T
Cj-inoculum _ _
C2-no nutrients _ +
+ + 20
_ + 2
_ + 2
_ + 2
TOTAL 48
TPn = duplicate commercial product flasks (n = 10)
Fj, F2 = fertilizer flasks (mineral N and P nutrients)
CFn, CFj, CF2 = no-oil controls for products and fertilizer.respectively
Cj, C2 = inoculum and no-nutrient controls
Flasks F1 and F2 represented simple inorganic fertilizer application and contained the following
ingredients
-------
COMMERCIAL PRODUCTS
All respirometer flasks were incubated at 15' C in the dark and continuously stirred at 300 rpm
by magnetic stirrers. The first set of control flasks (CPn, CF1, CF2) represented background oxygen
uptake of the product and nutrient-supplemented seawater without oil. Results from these flasks
were subtracted from the appropriate test flasks to obtain the net oxygen uptake on the weathered oil.
The inoculum control represented the endogenous oxygen uptake of the organisms from the washed
beach material and the seawater without oil or nutrients. The no-nutrient control represented the
oxygen uptake of the organisms from the washed beach material and seawater on weathered oil
without any external source of nutrient addition (i.e., background nutrient levels from Prince William
Sound).
Shake Flasks
Shaker flasks duplicating the respirometer flasks were used to assess the quantitative changes
in oil composition by chromatographic separation of the individual components. Although it was
possible to remove samples from the respirometer flasks, it was deemed more prudent not to disturb
the respirometric runs, but to instead use the shake flasks with proportionately higher levels of oil,
commercial products, etc., to facilitate sampling and precision/accuracy of the analytical chemistry.
Table 13.2 summarizes the shaker flask experimental design.
The test flasks corresponded exactly to the 22 test flasks listed in Table 13.1 but with the
following modifications: flask size, 250 mL; seawater, 100 mL; weathered oil and commercial
products, 10 times the final concentrations used in the respirometer flasks; and mineral nutrients,
same final concentration used in the respirometer flasks. The higher concentration of weathered oil
was used to improve the final sensitivity of the chemical analyses.
In addition to the 22 test flasks, 18 supplemental flasks were set up. These reactors represented
9 sterile product controls, which determined whether the enhancement was due to the microorganisms
or to the nutrients or metabolites in the product, and 9 sterile background controls (i.e., sterile oil and
seawater, but non-sterile product) to evaluate, by comparison to non-sterile flasks, the effect of
competition from naturally occurring organisms (one of the 10 products did not receive these sterile
treatments). Sterilization of materials was accomplished by autoclaving at 121* C for 15 minutes.
527
-------
COMMERCIAL PRODUCTS
TABLE 13.2. EXPERIMENTAL DESIGN FOR THE SHAKER FLASK STUDIES
Reaction Vessel
Nutrients Weathered Commercial Seawater TOTAL
Oil Product
TEST FLASKS:
Tpn + + +
SPn + + sterile +
TpnSb + sterile + sterile
Ft.F, + + _
20s
9"
9m
2"
CONTROL FLASKS:
cpn + _ + +
CFI + _ _ +
C1-inoculum _ _ _ +
C2-no nutrients _ + _ +
TOTAL
TPn = duplicate commercial products (n - 10), non-sterile system
SFn = sterile products in non-sterile seawater/oil, non -duplicated
TpnSb = non-sterile products in sterile seawater/oil, non-duplicated
Fj, F2 = fertilizer (mineral N and P nutrients) in non-sterile system
Cpn, CF1 = no-oil controls for products and fertilizer, respectively
Cj, C2 = inoculum and no-nutrient controls
a - microbiological and chemical analysis
b = microbiological analysis only
10b
lb
lb
2*
54
528
-------
COMMERCIAL PRODUCTS
Sampling
There were three sampling events for analytical chemistry and microbiology, days 0,11, and 20.
These events were determined by the shape of the oxygen uptake curves from the respirometry
experiments. Each shaker flask was sacrificed at the indicated sampling time by mixing the contents
with methylene chloride and performing the extraction on the entire mixture. Before sacrificing a
flask, a small aliquot was removed for determination of microbial density changes.
Nutrient Analysis
The nitrogen species NH3-N, NO2-N, and NO3-N were determined by U.S. EPA Methods [EPA
600/4-79-020]. The NHS-N method was No. 350.1 and the NO2-N/NOS-N method was No. 353.1.
Oil Chemistry
The oil constituents were analyzed by measuring the aliphatic and aromatic fractions of the
methylene chloride extracts. The extracts were concentrated and passed through a silica gel
fractionation column to separate the alkanes and the polycyclic aromatic hydrocarbons (PAHs). The
column was first eluted with hexane to collect the alkane fraction and then with a 1:1 mixture of
hexane and benzene to collect the aromatic fraction. Any polar compounds remaining in the extract
stayed bound to the silica gel column. Aliphatic fractions were measured by gas chromatography
using a flame ionization detector. The PAH fractions were characterized by gas chromatography/mass
spectrometry (GS/MS).
Microbiology
Growth of oil-degraders was measured by spread plates on oil agar (Bushnell-Haas medium
supplemented with Prudhoe Bay crude oil as the carbon source). Plates were incubated at 15' C for
21 days prior to counting.
529
-------
COMMERCIAL PRODUCTS
RESULTS
Respirometry
The net oxygen uptake curves (oxygen uptake in product flasks with oil minus oxygen uptake
in flasks without oil) for all 10 products (curves with symbols) compared with the curve for mineral
nutrients (curve with no symbols) are summarized in Figures 13.la and 13.Ib. In Figure 13.la
products F and G showed significantly higher alkane degradation and also exhibited higher net
oxygen consumption than mineral nutrients. The final plateau in total oxygen uptake was slightly less
than 500 mg/L for both Products E and G compared with about 340 mg/L for the mineral nutrient
flasks. The acclimation lag period for products E and G was approximately 2 and 4 days,
respectively, compared with 5 days for mineral nutrients. Product A gave the highest maximum net
uptake (630 mg/L compared with 340 for mineral nutrients) but the lag period was almost 10 days.
Products B and D exhibited O2 uptake characteristics that were not different from the nutrient
control.
In Figure 13.1b only products J and C gave higher overall net O2 consumption than mineral
nutrients, although product F exceeded the control after 27 days. The lag period for both products
J and C was only 1 day. The shape of the product F curve was multi-phasic, suggesting the organisms
were consuming different substrates at different rates and at different times (diauxie). Very little net
oxygen consumption was observed with product I.
Nutrient Concentrations
Product flasks requiring nutrient addition, as specified by the product manufacturer, received
the same level of mineral nutrients as the fertilizer flasks. The concentrations of ammonia nitrogen
measured in each product flask at day 0 are summarized in Table 13.3.
Statistical Analysis of Alkane Degradation Data
The percent reductions of the resolved aliphatic constituents of the weathered oil (nC12 through
nC34 plus the isoprenoid hydrocarbons pristane and phytane) were computed at day 11 for each
product flask and the results compared with the percent reduction computed for the mineral nutrient
flasks. Table 13.4 summarizes the statistical differences observed using Tukey's Studentized Range
530
-------
COMMERCIAL PRODUCTS
TABLE 13.3. NHS-N LEVELS IN EACH PRODUCT FLASK
AT THE START OF THE EXPERIMENT
PRODUCT NH3-N NUTRIENTS
mg/L ADDED
A 8.0 YES
B 2.1 NO
C 1080.0 NO
D 11.8 YES
E 11.3 YES
F 10.0 YES
G 24.9 YES
H 426.0 NO
I 0.5 NO
J 1.5 NO
FR* 6.9 YES
FR = mineral fertilizer
TABLE 13.4. TUKEY*S STUDENTIZED RANGE TEST FOR DETECTING DIFFERENCES IN
MEAN PERCENT REMOVAL OF ALKANES BY PRODUCTS IN 11 DAYS
PRODUCT % REMOVAL SIGNIFICANTLY DIFFERENT
FROM INORGANIC NUTRIENTS*
E
G
B
A
D
FR
C
J
H
F
I
94.5
93.6
87.9
75.9
74.2
68.4
67.8
59.9
49.5
33.3
27.9
YES
YES
NO
NO
NO
NO
NO
NO
NO
YES
YES
Minimum Detection Difference = 21.3% at 5% Significance Level
531
-------
(a)
700
10
(b)
Time, days
Legend
Product
— FR
• E
V G
• B
A A
• D
700
Legend
Product
— FR
• C
T J
• H
A F
Time, days
Figure 13.1. Net Oxygen Uptake Curves for Products and Mineral
Nutrients: a) Products E, G, B, A, and D; (b) Products C, J,
H, F, and I.
532
-------
COMMERCIAL PRODUCTS
Test (Tukey, 1953). The products are arranged in descending order of significance. Only Products
E and G gave significantly higher removals (p < 0.05) than inorganic fertilizer after 11 days. Six of
the other products gave results that were not different from mineral nutrients, while two actually gave
significantly lower removals. The latter results suggest that the products may have been inhibitory
to the biomass at the levels used in the closed flasks.
Total Alkane Reduction
The total alkane degradation data from the product flasks and the corresponding sterile controls
at days 11 and 20 are summarized in Figure 13.2. The products are arranged on the x-axis in the
order determined by the statistical analysis (this same ordering has been made on all figures). At day
11 (top half of Figure 13.2), better degradation was observed in every case when the commercial
products were first sterilized, suggesting that the indigenous Alaskan populations were doing most,
if not all, of the bioremediation. In contrast, less degradation occurred in every case except Product
I when the background (seawater and oil) was first sterilized. This suggests that when left alone, the
product organisms were less able to degrade the alkane fraction than the indigenous organisms. In
the non-sterile treatments, enhancement was observed for Products E and G compared with mineral
nutrients, suggesting that the products exhibiting the enhancement were providing metabolites or
some other form of nutritional benefit that was lacking in the mineral nutrient flask. By day 20
(bottom half of Figure 13.2), all products except Products F and I caught up, giving greater than 85%
reduction in the total alkane levels in the flasks. However, most of the flasks containing oil and
seawater that were first sterilized still significantly lagged behind the non-sterile systems.
Typical Chromatographic Profiles
Typical day 11 Chromatographic profiles of the individual alkane components for three of the
products are shown in Figure 13.3. The three profiles shown were selected to represent products
giving significantly better (Product E), equivalent (Product A), and lower (Product I) alkane removals
than mineral nutrients after 11 days. The solid bars in the figure are the non-sterile systems (i.e., as
they would be used in the field). The small-hatched bars are the flasks containing non-sterile oil and
seawater inoculated with sterile product, while the large-hatched bars are those containing sterile
seawater and oil inoculated with non-sterile products. The no-nutrient controls are represented by
solid lines with symbols to facilitate comparisons.
533
-------
(a) DAY 11
en
tu
z
o
p
u
Q
LU
CC
NON- .
STERILE
STERILE
PRODUCT
STERILE
BACKGROUND
0 FR C
PRODUCT
H
00
(b) DAY 20
100
<
_J
<
O
U
3
Q
lit
OC
EGBADFRCJ
00
Figure 13.2. Total Alkane Reduction in the Product Flasks: (a) day 11,
(b) day 20.
534
-------
CONCENTRATION, ppm
3
to
3
(*}
•
o ?
Q. O
||
m«a
** "^
,>-o
i o
"I
o
—&
^*
zr
>
3
©
O
a
o
s>
0)
Sooobot aaooooofaiocg^gg! Qaoooa
I
i!
I
II
OX
-------
COMMERCIAL PRODUCTS
In the case of Product E, virtually all of the normal alkanes in both the non-sterile and sterile
product flasks were at or below detectable limits. Even the branched chain aliphatics were
substantially reduced. In the flasks containing sterile seawater and oil with non-sterile product,
changes were not observed in the normal alkanes nC15 and above, while some decreases occurred in
the lower molecular weight fractions. In the case of Product A, the concentrations of normal alkanes
in the non-sterile flasks were higher than the corresponding Product E flasks. Levels of the branched
chain alkanes (pristane and phytane) were unchanged. In the flasks containing sterile product, most
of the normal alkanes were nearly undetectable, while the pristane and phytane remained close to
starting levels. In the sterile background flasks, levels of all constituents were similar to the
corresponding Product E flasks. Finally, in the case of Product I, most of the alkanes were unchanged
in all flasks, regardless of the sterile nature of the controls. The exceptions were the normal alkanes
nC16 through nC22, which averaged approximately 53% lower than starting levels.
Total PAH Reduction
A summary of the total PAH reduction data at days 11 and 20 is presented in Figure 13.4.
Differences are less clear among the products, although Products C, F, H, and I gave total reductions
considerably less than mineral nutrients. By day 20, PAH reduction by Product C was somewhat
closer to the others, while Products H, F, and I substantially lagged. Excellent removal of aromatics
was observed in all other flasks.
Typical Mass Spectral Profiles
Typical day 11 mass spectral profiles of the individual PAHs for the same three products are
shown in Figure 13.5. Mass spectral analyses of the fractionated extracts from the sterile controls
were not performed. Day 0 and day 11 PAH levels for each of the non-sterile product flasks are
shown on the figure. Differences in the PAH levels among the three product flasks are clearly
evident. Higher concentrations of PAHs, especially in the substituted naphthalene group, occurred
in the Product A flasks and even higher in the Product I flasks. The dibenzothiophene group was also
more resistant to degradation in the Product I flasks than the other two.
536
-------
DAY 11
DAY 20
B
D FR C
PRODUCT
H
Figure 13.4. Total Aromatic Reduction in the Product Flasks at Days 11
and 20.
537
-------
DAY O
DAY 11
E
a
a
z"
O
UJ
O
z
O
O
PRODUCT e
PRODUCT A
•"
PRODUCTI
i —B-.
.H.
LI
ISA
A • c D e
IJKLMNOPQRttTUVWXYZl9 » 4 •
AROMATICS
A Acenaphthene
B Acenaphthytene
C Benzo(a)anthracene
D Benzo(a)pyrene
E Benzo(b)fluoranth
F Benzo(g,h,i)perylene
G Chrysene
H C1-Chrysenes
I C2-Chrysenes
J C3-Chrysanes
K Dibenzo(a,h)anthr.
L Dlbcnzothiophena
M C1-0ibenzothiophs
N C2-Dlbenzothlophs
O C3-Oibenzothiophs
P Fluoranthen*
Q Fluorene
R C1-Fluorenes
S C2-Fluorenes
T C3-Fluoranes
U lndeno(1,2,3-cd)pyr
V Naphthalene
W C1-Naphthalenes
X C2-Naphthalenes
Y C3-Naphthalenes
Z C4-Naphthalenes
1 Phenanth./anthra.
2 C1-Phenanth./anths
3 C2-Phenanth./anths
4 C3-Phenanth./anths
5 C4-Phenanth./anths
6 Pyrene
Figure 13.5. GC/MS Profile of the Aromatic Data at Day 11 for Products
E, A, I.
538
-------
COMMERCIAL PRODUCTS
Microbiology
Virtually all changes in oil-degrader densities occurred by day 11. The populations leveled off
in all flasks thereafter. Consequently, the growth of oil-degraders has been summarized for days 0
and 11 only, and the results from all flasks, including sterile controls, are presented in Figures 13.6
and 13.7. Figure 13.6 depicts the yield of oil-degraders for all products at day 11, and Figure 13.7
shows the Iog10 change in oil-degraders in 11 days. Data absent from the figures were caused by
missed dilutions.
Products E and G, which gave the best alkane degradation of all the products (Table 13.4 and
Figure 13.2) and displayed net oxygen uptake characteristics superior to most (Figure 13.la), also
exhibited excellent yield and growth of oil-degraders in 11 days. Products C, J, and F yielded high
levels of oil-degraders and good oxygen uptake curves, but alkane degradation was not better than
the populations growing in simple mineral nutrients. Oil-degrader populations actually declined in
the Product B and A flasks, and the increase in oil-degraders in the flasks containing Products D and
I was minimal.
Flasks containing sterile seawater and oil inoculated with non-sterile products (large-hatched
bars in Figure 13.6) gave higher oil-degrader counts in all cases except Products C and F. The
organisms in the latter two products were less able to grow on the weathered crude oil than the
indigenous populations. If Products C and F were first sterilized, however, the indigenous
populations grew better than either the non-sterile systems or the sterile background control flasks.
This suggests that either there was an antagonism between the indigenous populations and product
organisms, or the products contained one or more inhibitors that prevented the indigenous bacteria
from achieving their ultimate biodegradative potential.
Flasks containing Products B and A also gave better growth when the products were first
sterilized (Figure 13.7, small-hatched bars), again suggesting either a microbial antagonism or the
presence of a heat-labile inhibitory substance. Product D gave the best growth when the indigenous
populations were first sterilized. Products E, G, and J gave good final yields and density increases,
but Product J did not perform well with respect to oil constituent degradation.
539
-------
QC
HI
Q
<
cc
a
tu
Q
o
u.
O
Non-
Sterile
Sterile
product
Sterile
background
B
D FR C
PRODUCT
H
Figure 13.6. Yield of Oil-Degraders for All Products at Day 11.
540
-------
Non-
Sterile
Sterile
product
Sterile
background
oc
LU
Q
<
OC
C5
UJ
Q
UJ
O
O
o
B
D FR C
PRODUCT
H
Figure 13.7. Log Increase in Oil-Degraders for All Products in 11 Days.
541
-------
COMMERCIAL PRODUCTS
SUMMARY AND CONCLUSIONS
The objective of this laboratory experiment was to determine if weathered crude oil could be
degraded faster when natural microbial populations were supplemented with exogenous oil-degraders
and excess nutrients than excess nutrients alone. Oil degradation chemistry, oxygen uptake in
respirometer flasks, and microbial density changes were used to decide which product(s) would
proceed to field testing.
Of all the products tested, the two that provided the most consistent results in all three tests were
products £ and G. Both gave higher oxygen uptake with more rapid onset, greater growth of oil-
degraders, and superior alkane degradation than mineral nutrients. Products C and J showed good
growth of oil-degrader populations and gave excellent net oxygen uptake curves, but were not better
oil-degraders than indigenous populations supplied with simple mineral nutrients. Product F yielded
the highest oil-degrading population of all, yet it's oxygen uptake curve was not better than the
mineral nutrient curve until after day 27, and alkane degradation was relatively poor. Product A gave
the best overall net oxygen consumption, but the change in oil-degraders and the relative alkane
degradative capability at day 11 were unsubstantial, perhaps because of the extended acclimation lag
period as noted in the oxygen uptake curve for Product A. Flasks containing products B, D, H, and
I also produced only minor changes in oil-degrading populations and unenhanced oil-degradative
capability.
The sterile controls revealed that the indigenous Alaskan oil-degrading populations were
performing most, if not all, of the biodegradative activity. The organisms present in products £ and
G did not appear to contribute significantly to such activity. This suggests that a co-metabolite, a
nutrient, or some other unknown factor exists in these two products and stimulates the indigenous
microorganisms to degrade the crude oil constituents at faster rates than is possible with simple
nutrient addition. Further work needs to be done to define the enhancement factor(s) in these
products.
Correlations have not yet been made between weathered crude oil degradation and oxygen
uptake, nor have carbon balances been performed. Work is being planned to measure carbonaceous
metabolic end products, CO2 production, and total biomass yield, and then to correlate this
information with the oxygen consumption data. If such correlations can be established, then use of
oxygen consumption data for estimating biodegradation efficacy as part of a screening protocol will
542
-------
COMMERCIAL PRODUCTS
be made possible. The respirometric technique requires much less effort than conventional shake
flask studies because data gathering is automated and computerized, and it is not necessary to collect
samples manually during the course of a biodegradation experiment. All that is required, assuming
the proper correlations have been established, is the careful measurement of initial substrate and
biomass values followed by the measurement of the residual soluble product value at the plateau of
the uptake curve (Grady et al., 1989). From the analysis of this information, treatment decisions can
be facilitated.
It appears from all the available evidence that the indigenous Alaskan microorganisms were
primarily responsible for the biodegradation in the closed flasks and respirometer vessels, and that
any enhancement provided by products £ and G might have been due simply to metabolites,
nutrients, or co-substrates fortuitously present in the product. The NETAC panel reviewed the
results of the tests and agreed with the recommendation for further testing of two products that
exceeded the performance of inorganic nutrient addition. Results from all three lines of evidence
(respirometry, microbiology, and oil chemistry), supported the decision to only field test products E
and G. The field testing of these two products at Disk Island in Prince William Sound is described
below.
FIELD STUDIES
In addition to application of nutrients, bioremediation in the field may be enhanced by
inoculation with allochthonous microorganisms. Cultures and cultural products have been added to
different environments to stimulate biological removal of contaminants. Some of the investigations
have demonstrated enhancement, while others have not (Leahy and Colwell, 1990). Lehtomaki and
Niemela, 1975, found that addition of brewers' yeast to oil-contaminated soil enhanced oil removal
by factors of 2- to 10-fold. This was most likely due to the supply of critical nutrients, vitamins, or
cofactors naturally lacking in the soil. Christiansen and Spraker (1982), reported a series of case
histories of refinery wastewater treatment plants using commercial cultures to overcome various
specific problems, such as foaming, toxic loads, low biomass, etc. Most success with biodegradation
enhancement by allochthonous microbial cultures has been achieved when chemostats or fermentors
were used to control conditions or reduce competition from indigenous microflora (Wong and
Goldsmith, 1988).
543
-------
COMMERCIAL PRODUCTS
The results of the field testing of the two selected products (E and G) are presented below. The
objective was to determine if commercial microbiological products were able to enhance
bioremediation of an oil-contaminated beach in Prince William Sound to a greater extent than that
achievable by simple fertilizer application. The two companies that participated in the testing were
Sybron, Inc. and ERI-Waste Microbes, Inc.
METHODS
Plot Description
A schematic representation of the experimental layout is depicted in Figure 13.S. The
experiment was a randomized complete block design. Four beach segments ("blocks"), each 20 m wide
(labeled 1 through 4) were staked out in the intertidal zone on Disk Island (DI-67a). Within each
block were 4 treatment plots, labeled A through D, 2 m wide by 5 m long (top-to-bottom). The plots
were separated by a 3 m buffer zone. Plot A was the no-treatment control; plot B was the
nutrient-only treatment; plot C received nutrients plus Sybron's product; and plot D received nutrients
plus ERI's product. The treatment plots within each block were randomly distributed.
Each plot was subdivided horizontally into three equal segments 2 m wide by 1.67 m long, as
shown schematically in Figure 13.9. In each of the three segments, four bags made of fiberglass
screening material were filled with approximately 750 to 1000 g of oiled gravel, buried approximately
5 to 10 cm below the surface, and covered with mixed sand and gravel. To obtain the gravel for each
bag, oiled gravel from Disk Island was first sieved through a 25 mm coarse screen to remove large
stones and then through a 4.75 mm sieve to remove the small sand granules that compact the beach
material. The gravel was mixed manually with shovels and hoes in a large wooden container to
achieve reasonable homogeneity with respect to oil contamination and rock size. The four bags
corresponded to the four sampling events planned for the experiment. A surveyor's ribbon was
attached to each bag for easy identification. The 12 samples within each block were numbered 1
through 4 in the top third, 5 through 8 in the middle third, and 9 through 12 in the bottom third.
544
-------
I""! A: No Nutrients
C: Sybron * Nutrients
B: Mineral Nutrients
D: ERI * Nutrients
BLOCK 1
I
B
1
A D
BLOCK 2
I
1
C B D A
BLOCK 3 BLOCK 4
1
I
A D C B
I
A B
Figure 13.8. Schematic Diagram of the Experimental Plot Layout on
Disk Island.
545
-------
2 m
9
v—x
(ft)
6
s—-•
/• -.
8
5 m
Figure 13.9. Schematic Diagram of a Typical Beach Plot Showing
Dimensions and Location of Sampling Bags.
546
-------
COMMERCIAL. PRODUCTS
Sampling
On a given sampling day, triplicate samples from each plot within a block were collected
according to a random schedule. One sample was randomly taken from each of the three identical
sectors of each plot. Some of the gravel was transferred into 500 mL I-Chem jars, labeled, and placed
in a cooler to be carried back to Valdez for freezing and shipment via Federal Express to the
analytical chemistry laboratory located in Pittsburgh, PA. The rest of the gravel was archived in
aluminum foil and frozen. Thus, 48 samples were collected on each of the four sampling days, giving
a total of 192 samples for the entire experiment.
Nutrients
To track the fate of the added nutrients with time, wells were installed in some of the plots for
collecting subsurface water samples. Well points extending approximately 60 cm below the surface
were driven into the center of the four control plots and the four plots receiving only nutrients.
Subsurface water from these 8 wells, which served as samples for nutrient analysis, were collected
according to a pre-determined sampling schedule (see below).
Nutrient Application
The source of nitrogen was ammonium nitrate. Each 2 m x 5 m plot received 200 g of N (20
g/m2). At 35% N, the amount of NH4NOS containing 200 g of N was 570 g or 1.25 Ib per plot. This
amount, less approximately 40 g to account for the N in the product containing the phosphate salt (see
next paragraph), was added to 6 gallons of seawater and the contents stirred until dissolved. A
2-gallon plastic sprinkling can was filled with the solution and the entire contents poured onto the
top third of a plot earmarked for nutrients. The sprinkling can was again filled and the contents
poured onto the middle third. The procedure was repeated for the bottom third.
The source of phosphorus was an Ortho product named "Upstart," which had an N-P-K. analysis
of 3-10-3. At 10% P2OB, the amount of Upstart used was 450 g (1 Ib) per plot. This corresponded
to a phosphorus loading of 20 g P per plot (2 g P/m2). The 450 g of Upstart was added to the 6
gallons of seawater above (after the NH4NOS had been dissolved) before applying to each plot. Note
that this product contained 3% N in the form of NH4NO3. The amount of N in Upstart had already
been accounted for in the above 530 g computation of NH4NOS needs.
547
-------
COMMERCIAL PRODUCTS
Schedule
The experiment lasted only 27 days because severe Alaskan winter weather precludes field
activities beyond the month of August. Day 0 occurred on Sunday, July 29, 1990. Nutrients and
commercial products were applied on days 0, 4, 8, 12, 16, 20, and 24. One extra application (day 2),
was used for an additional commercial product application, as specified by the two vendors. After
nutrients and products had been delivered to the appropriate plots, randomly assigned triplicate
sampling bags were removed from the plots for time 0 sediment chemistry and microbiology analysis.
The other triplicate sampling bags were collected on days 9,18, and 27. Nutrient sampling took place
on days 1, 2, 3, 4, 17, 18, 19, and 20. This allowed determination of nutrient concentrations
throughout the four-day interval between applications at two different times in the experiment.
Sediment Chemistry
All samples were analyzed for oil residue weight by methylene chloride extraction followed by
evaporation to dryness and weighing on an analytical balance. After weighing, each sample was
reconstituted with methylene chloride, passed through a silica gel fractionation column, and analyzed
for the normal alkanes nC12 through nC34, plus the isoprenoid hydrocarbons pristane and phytane
by gas chromatography using a flame ionization detector. The aliphatic fraction was eluted from the
silica gel column with hexane prior to GC injection.
Microbiology
Sub-samples from the 8 plots of blocks 2 and 3 were analyzed for oil-degrading bacteria by
standard plate count, using Bushnell-Haas medium supplemented with Prudhoe Bay crude oil as the
carbon source. Only one of the three triplicates from those 8 plots was analyzed for microbial
numbers. The plates were incubated at 15* C for 21 days and the colonies counted.
Data Analysis
The data were analyzed by analysis of variance and individual contrast methodology using SAS
Software Release 6.06.
548
-------
COMMERCIAL PRODUCTS
RESULTS
Persistence of Nutrients
Figures 13.10-13.12 summarize the average changes in nutrient levels over time in each block
on Disk Island. Figure 13.10 shows the ammonia-N data, Figure 13.11 the nitrate-N data, and Figure
13.12 the phosphate-P data.
Persistence of ammonia-N was the most erratic. In block 1 the levels of NH3-N in the
nutrient-treated plot were measured at 1.1 and 4.0 mg/L one and two days after application,
respectively, and in block 2 the NH3-N was 1.7 mg/L one day after application. Little NH3-N was
measured in any of the control plots at any time except in block 4, where 0.1 mg/L was measured
after one day and almost 1.0 mg/L after four days. The source of the high NH3-N spike in the
control plot of the fourth block may have been caused by carry-over of nutrients from the
nutrient-treated plot onto the control plot. The nutrient-treated plot had to be placed above the
control plot (see Figure 13.8) because of the presence of compacted peat on the extreme right end of
the beach. There was a surface flow of water from a saltwater lagoon located approximately 50 m
above the test area that flowed across the nutrient-treated plot onto the control plot. This stream was
not noticed when the plots were first delineated. Although this explains the higher levels of NH3-N
measured one day after application, it does not explain why such a high spike was observed on the
fourth day.
The nitrate and phosphate data indicate significant but decreasing levels of nutrients in the
nutrient-treated plots as time progressed to four days after application (Figures 13.11 and 13.12).
Again, high levels of NO3-N and measurable levels of PO4-P appeared in the control plot of the
fourth block four days after application.
Numbers of Oil-Degrading Microorganisms
Oil-degrader counts in all plots of blocks 2 and 3 are shown in Figure 13.13. Although the levels
of oil degraders were high in each of the plots, there were no significant changes or differences in
any of the plots after 27 days of field testing.
549
-------
6.0
4.0
^»
E 3.0
m
z
i 2-°
z
1.0
fk l\
mM No Nutrient
Control
BLOCK 1
-
-
-
•
•
|
i
BLOCK 2
po
]
nm
Mineral
Nutrients
BLOCK 3
BLOCK 4
I
1234 1234 1234
TIME, DAYS
1234
Figure 13.10. Average Changes in Ammonia-N Levels in the Four Days
Between Fertilizer Applications.
550
-------
No Nutrient
Control
Mineral
Nutrients
1.5
o» 1.0
»
i
TO
o °-«
o.o
BLOCK 1
K la •_ >3
BL
.<
DCK 2
B I_£
BLOCK 3
K _S _ra »^
BLOCK 4
bB. I
il B^ k
1234 1234 1234
TIME, DAYS
1234
Figure 13.11. Average Changes in Nitrate-N Levels in the Four Days
Between Fertilizer Applications.
551
-------
0.3
_J
O 0.2
oT
i
*
O 0-1
Q.
A A
•1 No Nutrient km
Control
BLOCK 1
•
~
!
BLOCK 2
F
•
JjJ
Mineral
Nutrients
BLOCK 3
I
I
J-i
BLOCK 4
jj
1234 1234 1234 1234
TIME, DAYS
Figure 13.12. Average Changes in Phosphorus Levels in the Four Days
Between Fertilizer Applications.
552
-------
I-
CO
oc
IU
o
IL
O
d
10'
104
10'
10'
101
10°
No
Nutrients
Mineral
Nutrients
Sybron +
Nutrients
ERI +
Nutrients
5 10 15 20 25 30
TIME, DAYS
Figure 13.13. Oil-Degrader Counts in All Plots of Blocks 2 and 3 as a
Function of Time.
553
-------
COMMERCIAL PRODUCTS
Oil Residue Weight
Changes in oil residue weight, averaged over all four blocks, are summarized in Figure 13.14
as a function of time. The points on the connected curves are the mean residue weights for each of
the four treatments, and the error bars depict one standard deviation unit above and below the means.
These error bars represent the variation in oil residue weight among the four blocks and are indicative
of the overall experimental error.
Visual inspection of the data from the plots treated with mineral nutrients alone and mineral
nutrients supplemented with Sybron's product indicates a decrease in oil residue weight of
approximately 33% at the end of the experimental period, compared with no net change in the
no-nutrient control plot and a slight increase in the ERI plot. When the data were subjected to
analysis of variance, however, there were no statistically significant differences among any of the
four treatments at the 5% significance level. This was true even after the data were log transformed
to stabilize the variance.
Note the broad error bars on Figure 13.14 at the day 0 sampling time compared with the other
three sampling times. Despite the effort to control the heterogeneity of rock size and contamination
by the sieving and mixing techniques, there was still substantial variation in oil residue weight
between plots and between blocks at day 0. To ascertain the source of this variation, a breakout of
plot oil residue weights by block was conducted. Results are shown in Figure 13.15.
Examination of these data reveals the differences in the distribution of oil between plots. Note
that very little change took place in any of the treatments in blocks 2 and 4. Oil residue weights in
the no-nutrient control plot and the nutrient-only plot of block 1 and the Sybron plot of block 2
appeared to decline markedly within the first 9 days and then level-off for the remainder of the
experimental period. The oil residue weights in the ERI plots of blocks 1 and 3 actually showed an
increase between days 18 and 27.
The error bars shown on this figure are the standard deviations of the triplicate samples within
each plot and are indicative of the total sampling and analysis error. At day 0 the agreement of the
triplicate samples averaged within each plot (Figure 13.15) was better than the agreement of identical
plots averaged over blocks (Figure 13.14). This suggests that the cause of the variation among plots
was consistent within each of the plots.
554
-------
4.0
2 3-0
..0
9
S
1.0
0.0
NO NUTRENTS
SYBRON * NUTREKT3
4.0
1.0
1.0
•.0
1JJ
OjO
NUTRENTS ALONE
ERI
o « io it 10 at 30 -• o 5 10 11 20 as ao
TIME, DAYS
Figure 13.14. Changes in Oil Residue Weight Averaged Over All Four
Blocks as a Function of Time.
555
-------
* a
CM
o
o
CD
O
O
_l
CD
o
c»
O
04
oo
ooooooooo
CO Z
_ a
II
I i
8
o
m
o
44
255
«««
s :'
CO
u
1AA
HO
Figure 13.15. Changes in Oil Residue Weight in Each Block as a
Function of Time.
556
-------
COMMERCIAL PRODUCTS
Total Resolvable Alkanes
All samples were subjected to GC analysis to determine the changes in the aliphatic profiles of
the oil among the various treatments. The concentrations of all the normal alkanes and the isoprenoid
alkanes pristane and phytane resolvable by GC/FID were summed together for each treatment,
averaged over all four blocks, and plotted as a function of time. The data with associated error bars
are shown on Figure 13.16.
Except for the day 0 data, the error bars in Figure 13.16 were generally higher than the
corresponding residue weight error bars (Figure 13.14). Although a downward trend in resolvable
alkane concentrations was perceptible in all of the treatments after 27 days, the analysis of variance
revealed no significant differences among the treatments (p<0.05). This agrees with the findings of
no significance among treatments in the oil residue weight data.
Figure 13.17 was constructed to examine the behavior of the GC data in the individual plots
within each block. A downward trend in alkane hydrocarbon levels is generally perceptible in the
control, nutrient-only, and Sybron plots of blocks 1 to 3 and the ERI plot of block 1. Temporal
changes in the alkane levels from the ERI plot of block 2 are highly variable, showing an increase at
day 9 followed by decreases at days 18 and 27, while in block 3 increases are observed successively
after day 9.
The error bars represent the sampling error associated with the triplicate samples in each plot.
The error bars are higher overall than the corresponding oil residue weight data (Figure 13.15). The
sampling errors associated with the GC data appear to be no better than the overall experimental
error, which contrasts somewhat with the residue weight data.
Total Resolvable Alkanes as a Percent of the Residue Weight
The previous two figures depicted observed temporal changes in total resolvable alkanes
normalized to sediment weight but not to the weight of the oil. Since there might have been
significant differences in the extent of sediment contamination among samples, the measured
resolvable alkanes were normalized to oil residue weight within each sample and the results plotted
in Figures 13.18 and 13.19. Figure 13.18 shows the changes in the total resolvable alkanes as a percent
of the residue weight, averaged over all four blocks for each treatment, and plotted as a function of
557
-------
01
Jt
I
w
tr
u
14
12
1O
4
t
1«
It
10
NO NUTRIENTS
8YBRON * NUTRIENTS
10
to is »o -•
10 18 20 tS 30
TIME, DAYS
Rgure 13.16. Changes in Total Resolvable Alkanes Averaged Over All
Four Blocks as a Function of Time.
558
-------
+ a
CO Z
•
_ a
a
cc
Ul
a
i
i
CO
u
anavAiosaa ivioi
Figure 13.17. Changes in Total Resolvable Alkanes In Each Block as a
Function of Time.
559
-------
#
2
Q
(0
u
5
03
0.7
0.«
O.S
0.4
0.9
0.2
0.1
0.9
O.t
0.8
0.4
0.9
0.2
0.1
0.0
NO NUTRIENTS
SYBRON * NUTRIENTS
O.T
0.4
OJ
0.4
0.1
0.»
0.4
OJt
OU
OJ
OJ
0.1
0.0
NUTRIENTS ALONE
ERI * NUTRIENTS
o • io
ao
»0 -4 0 • 10 1t 20 26 «0
TIME, DAYS
Figure 13.18. Changes in Total Resolvable Alkanes as a Percent of OH
Residue Weight Averaged Over All Four Blocks as a
Function of Time.
560
-------
UJ Z
Si
CO
UJ
% M.M 3naisau/S3Nv>nv nvioi
Figure 13.19. Changes in Total Resolvable Alkanes as a Percent of OH
Residue Weight in Each Block as a Function of Time.
561
-------
COMMERCIAL PRODUCTS
time. Figure 13.19 shows the changes within each block. Again, there were no significant differences
among the four treatments. Note that the general behavior of the residue weight-normalized curves
in Figures 13.18 and 13.19 is similar to the corresponding behavior of the sediment weight-normalized
curves in Figures 13.16 and 13.17. This suggests that the contamination of the sediment samples was
moderately homogeneous.
An important observation from Figure 13.18 is the magnitude of the total alkane/residue weight
ratio. The total alkane hydrocarbons resolvable by GC/FID are less than 0.5% of the total oil residue
weight. In other words, over 99.5% of the oil remaining on Disk Island 1.5 years after the spill is not
resolvable by conventional gas chromatography. The compounds comprising this persistent fraction
are likely the tars and asphaltines that degrade slowly with time.
SUMMARY AND CONCLUSIONS
The conclusions reached in this field study were based on three sources of information: nutrient
persistence, microbiology, and sediment chemistry. The nutrient data clearly demonstrated that
nitrogen and phosphorus persisted at measurably higher levels in the treated plots compared with the
control plots throughout the four days between applications. These measurements were taken
approximately 60 cm below the surface of the beach, suggesting that nutrients were in constant
contact with the subsurface sediment layers for relatively long periods of time.
The microbiology data clearly demonstrated no net increase in oil-degrader populations in any
of the plots after 27 days, and no differences among the four treatments at any time during the 27
day period. The oil-degrader populations were initially high and were maintained with or without
the presence of excess nutrients. Either the oil-degraders were dormant or, more likely, they were
sufficiently able to sustain their activity with the oligotrophic levels of nutrients present in the
ambient environment.
Sediment chemistry revealed the most definitive information because it was the basis of the
statistical analyses conducted. No significant differences were found among the four treatments at
the 5% significance level either from the standpoint of oil residue weight, total resolvable alkane
hydrocarbons, or total resolvable alkanes as a percent of oil residue weight. The relatively high
variation observed in the day-0 residue weight data clearly indicates the necessity of replicating
treatments when conducting field experiments.
562
-------
COMMERCIAL PRODUCTS
One possible explanation for the apparent downward trend in residue weights and alkane
hydrocarbons in both the control and treated plots is the perturbation of the beach material as a result
of the sieving and mixing techniques prior to experimental initiation. Although day 0 occurred 2 days
after the rocks were prepared and buried, that might not have been enough time for stabilization of
the perturbed oil on the bagged rocks. Thus, the apparent disappearance of oil might have simply
been mobilization due to tidal action.
Most of the readily biodegradable compounds in the aliphatic fraction of the oil had disappeared
in the 1.5 years since the spill took place. This is the most likely explanation for the lack of any
significant enhancement by either nutrient addition alone or nutrient addition supplemented with
commercial microbial cultures. Further evidence supporting this conclusion derives from examining
the n-alkane/isoprenoid alkane ratios. These ratios have been used in past literature to indicate extent
of biodegradation; the lower the ratio, the more extensive the biodegradation. The average
nC17/pristane and nC18/phytane ratios on day 0 for all the plots on Disk Island were 0.18 and 0.27,
respectively. This compares to approximately 1.5 to 1.8 for unweathered Prudhoe Bay crude oil.
Thus, the remaining oil present on Disk Island will likely degrade very slowly from now on because
of the recalcitrant nature of the substrate. If either nutrient application or commercial inoculation
can accelerate this rate, the time period must extend significantly beyond the 27 days allotted for this
study or the trial must be conducted on beaches with fresher oil contamination.
563
-------
COMMERCIAL PRODUCTS
LITERATURE CITED
Atlas, R.M. 1981. Microbial degradation of petroleum hydrocarbons: an environmental perspective.
Microbiol. Rev. 45:180-209.
Christiansen, J. A. and P. W. Spraker. 1982. Improving effluent quality of petrochemical wastewaters
with mutant bacterial cultures. Proc. Purdue Ind. Wastes Conf. 37:567-576.
Cooney, J. J. 1984. The fate of petroleum pollutants in freshwater ecosystems. In R. M. Atlas (ed.),
Petroleum Microbiology, Macmillan Publishing Co., New York.
Dott, W., D. Feidieker, P. Kampfer, H. Schleibinger, and S. Strechel. 1989. Comparison of
autochthonous bacteria and commercially available cultures with respect to their effectiveness in fuel
oil degradation. J. Ind. Microbiol. 4:365-374.
Floodgate, G. D. 1984. The fate of petroleum in marine ecosystems." In R. M. Atlas (ed.), Petroleum
Microbiology, Macmillan Publishing Co., New York.
Grady, C. P., Jr., J. S. Dang, D. M. Harvey, A. Jobbagy, and X. -L. Wang. 1989. Determination of
biodegradation kinetics through use of electrolytic respirometry. Water Science Technol. 21:957-968.
Leahy, J. G. and R. R. Colwell. 1990. Microbial degradation of hydrocarbons in the environment.
Microbiol. Rev. 54:305-315.
Lehtomaki, M. and S. Niemela. 1975. Improving microbial degradation of oil in soil. Ambio
4:126-129.
OECD. OECD guidelines for testing of chemicals. Section 3, Degradation and accumulation, Method
301C, ready biodegradability: modified MITI test (I) and Method 302C, inherent biodegradability:
modified MITI test (II). Director of Information, OECD, Paris, France.
Pritchard, P. H. and C. F. Costa. 1991. EPA's Alaska oil spill bioremediation project. Environ. Sci.
Tech. 25:372-379.
564
-------
COMMERCIAL PRODUCTS
Tukey, J. W. 1953. The problem of multiple comparisons. 396 pp. Princeton University,
Princeton, NJ.
U.S. Environmental Protection Agency. 1991. Methods for chemical analysis of water and wastes.
EPA 600/4-79-020, U.S. Environmental Protection Agency, Washington, D.C.
Venosa, A. D., J. R. Haines, W. Nisamaneepong, R. Govind, S. Pradhan, and B. Siddique. 1991.
Efficacy of commercial inocula in enhancing oil biodegradation in closed laboratory reactors. J. Ind.
Microbiol.
Vestal, J. R., J. J. Cooney, S. Crow, and J. Berger. 1984. In R. M. Atlas (ed.), Petroleum
Microbiology, Macmillan Publishing Co., New York.
Wong, A. D. and C. D. Goldsmith. 1988. The impact of a chemostat discharge containing oil-
degrading bacteria on the biological kinetics of a refinery activated sludge process. Water Sci.
Technol. 20:131-136.
565
-------
REFERENCES
Antrim, L.D., and J.Q. Word. 1989. Acute toxicity of Inipol and weathered petroleum to salmon,
herring, and mussel larvae. Final Report to EPA Gulf Breeze. Battelle contract #68-03-3479.
Battelle Marine Science Laboratory, Sequim, WA.
APHA, AWWA, WPCF. 1989. Standard Methods for the Examination of Water and Wastewater. 17th
ed. Washington, DC. APHA.
ASTM. 1980. Standard practice for conducting static acute toxicity tests with larvae of four species
of bivalve molluscs. In: Annual Book of ASTM Standards, Water and Environmental Technology, vol.
11.04, E724-80. American Society for Testing and Materials, Philadelphia, PA.
ASTM. 1986. Standard practice for conducting static acute toxicity tests with larvae of four species
of bivalve molluscs. In: Annual Book of ASTM Standards, Water and Environmental
Technology, vol. 11.04. American Society for Testing and Materials, Philadelphia, PA, pp. 368-384.
ASTM. 1988. Proposed revision of standard E724 guide for conducting static acute toxicity tests with
embryos of four species of saltwater bivalve molluscs. In: Annual Book of ASTM Standards, Water
and Environmental Technology, E724-88. American Society for Testing and Materials, Philadelphia,
PA.
ASTM. 1988. Standard practice for conducting acute toxicity tests with fishes, macroinvertebrates,
and amphibians. In: Annual Book of ASTM Standards, Water and Environmental Technology, vol.
11.04, E729-80. American Society for Testing and Materials, Philadelphia, PA, pp. 304-323.
Atlas, R.M. 1981. Microbiol. Rev. 45:180-209.
Atlas, R.M., ed. 1984. Petroleum Microbiology. Macmillian, New York.
Atlas, R., and M. Busdosh. 1976. Microbial degradation of petroleum in the Arctic, la: Proceedings
of the Third International Biodegradation Symposium, Kingston, RI, August 1975; edited by J.M.
Sharpley and A.M. Kaplan. Applied Science Publishers, Ltd., London, pp. 79-86.
566
-------
REFERENCE
Atlas, R., et al. 1978. /. Fish. Res. Board Can. 35:585-590.
Bartha, R. 1986. Microbiol. Ecol. 12:155-172.
Brown, E.J., and J.F. Braddock. 1990. Sheen Screen, a Miniaturized Most-Probable-Number Method
for Enumeration of Oil-Degrading Microorganisms. Applied Env. Microbiol. Vol. 56.
Brown, R., et al. 1985. Pollut. Eng. 25:234-239.
Caron, D.A. 1983. Technique for enumeration of heterotrophic and phototrophic nanoplankton using
epifluorescence microscopy and comparison with other products. Appl. Environ. Microbiol. 46:491-
498.
Chaloud, D.J., D.T. Heggem, M.L. Papp, T.C. Chiang, A.C. Neale, R.W. Gerlach, and J.M. Pollard.
1990. Alaska Bioremediation Project Quality Assurance Plan Summer 1990. Draft. EPA/600/x-xx-
xxx. U.S. Environmental Protection Agency, Las Vegas, NV.
Cifuentes, L.A., M.L. Fogel, J.R. Pennock, and J.H. Sharp. 1989. Biogeochemical factors that
influence the stable nitrogen isotope ratio of ammonium in the Delaware estuary. Geochim.
Cosmochim. Acta 53:2713-2721.
Clark, J., et al. 1990. Results of environmental monitoring in Prince William Sound following
fertilizer additions to enhance oil degradation. In: Proceedings of the 83rd Annual meeting and
Exhibition, Air and Waste Management Association, Pittsburgh, PA, June 24-29, 1990. 90-22.4.
Claxton, L.D., J. Allen, A. Auletta, K. Mortelmans, E. Nestman, and E. Zeiger. 1987. Guide for the
Salmonella typhimurium/mammalian microsome tests for bacterial mutagenicity. 189:83-91.
Coffin, R.B., B. Fry, B.J. Peterson, and R.T. Wright. 1989. Identification of bacterial carbon sources
with stable isotope analysis. Limnol. Oceanogr. 34:1305-1310.
Coffin, R.B., D. Velinsky, R. Devereux, W.A. Price, and L.A. Cifuentes. 1990. Stable carbon isotope
analyses of nucleic acids to trace sources of dissolved substrate used by estuarine bacteria. Appl.
Environ. Microbiol. 56:2012-2020.
567
-------
REFERENCE
Colvin, R.J., A.F. Rozich, F.F. Gaudy, Jr., and J. Martin. 1990. Application of a process model
calibrated with respirometry to predict full-scale activated sludge performance. Presented at 45th
Purdue Industrial Waste Conference. West Lafayette, IN, May 1990.
Cook, F., and D. Westlake. 1974. Task Force on Northern Oil Development. Report. No. 74-1.
#R72-12774.
Dibble, J., and R. Earth. 1979. Soil Sci. 125:56-60.
EVS. 1990. Measurement of acute toxicity of the oleophilic fertilizer, Inipol EAP22. Final Report
to EPA Gulf Breeze. EVS Project # 2/294-05. EVS, Seattle, WA.
Fedorak, P., and D. Westlake. 1981. Can. J. Microbiol. 27:432-443.
Fry, B., and E.B. Sherr. 1984. «1SC measurements as indicators of carbon flow in marine and
freshwater ecosystems. Contrib. Mar. Sci. 27:13-47.
Fuhrman, J.A., and F. Azam. 1982. Thymidine incorporation as a measure of heterotrophic
bacterioplankton production in marine surface waters: Evaluation and field results. Mar. Biol.
66:109.
Gaudy, A.F., Jr., and E.T. Gaudy. 1980. Microbiology for Environmental Scientists and Engineers.
McGraw Hill Co., New York.
Halmo, G. 1985. Enhanced biodegradation of oil. In: Proceedings 1985 Oil Spill Conference,
February 25-28, Los Angeles, CA. American Petroleum Institute, Washington, DC, Publication
#4385, pp. 531-537.
Hildebrand, R. 1989. Draft fucus protocols. United States Coast Guard Research and Development,
Mobile Laboratory.
Hobbie, J.E., R.J. Daley, and S. Jasper. 1977. Use of nuclepore filters for counting bacteria by
fluorescence microscopy. Appl. Environ. Microbiol. 33:1225-1228.
568
-------
REFERENCE
Horowitz, A. and R. Atlas. 1977. Appl. Environ. Microbiol. 33:647-653.
Houk, V.S., S. Schalkowsky, and L.D. Claxton. 1989. Development and validation of the spiral
Salmonella assay: An automated approach to bacterial mutagenicity testing. Mutal. Res. 223:49-64.
Houk, V.S., L.D. Claxton, and G. Early. 1991. Use of the spiral Salmonella assay to detect the
mutagenicity of complex environmental mixtures. Environ. Molec. Mutagen. 17:112-121.
Kreitler, C.W. 1975. Determining the source of nitrate in ground water by nitrogen isotope studies.
Report of Investigations No. 83, Bureau of Economic Geology. The University of Texas at Austin,
Austin, TX.
Kremer, F., et al. 1990. Implementation of the field demonstration for the bioremediation of the
Alaskan oil spill. In: Proceedings of the 83rd Annual Meeting and Exhibition, Air and Waste
Management Association, Pittsburgh, PA, June 24-29, 1990. 90-22.1.
Leahy, J.G., and R.R. Colwell. 1990. Microbiol. Rev. 54:305-315.
Lee, K., and E.M. Levy. 1989. In: Aquatic Toxicology and Water Quality Management, edited by J.A.
Nriagu. John Wiley and Sons, New York, pp. 217-243.
Lee, M.D., et al. 1988. CRC Crit. Rev. Environ. Control 18:29-89.
Lindstrom, J., J. Clark, and R. Prince. 1991. Bioremediation Monitoring Program. Draft Final
Report.
Liu, K-K. and I.R. Kaplan. 1989. The eastern tropical Pacific as a source of 15N-enriched nitrate
in seawater of southern California. Limnol. Oceanogr. 34:820-830.
Macko, S.A. 1981. Stable nitrogen isotope ratios as tracers of organic geochemical processes. Ph.D.
thesis, University of Texas at Austin, Austin, TX. 181 pp.
Maniatus, T., E.F. Fritsch, and J. Sambrook. 1982. Molecular Cloning. A Laboratory Manual. Cold
Spring Harbor Laboratory, Cold Spring Harbor, NY, pp. 109-112.
569
-------
REFERENCE
Maron, D.M., and B.N. Ames. 1983. Revised methods for the Salmonella mutagenicity test. Mutat.
Res. 113:173-215.
MEC. 1989. Results of bivalve larvae bioassays for the EPA Bioremediation Demonstration Project,
Valdez, Alaska. Final Report to U.S. EPA. MEC Analytical, Tiburon, CA.
Miller, D.C., S. Poucher, J.A. Cardin, and D. Hansen. 1990. The acute and chronic toxicity of
ammonia to marine fish and a mysid. Arch. Environ. Contam. Toxicol. 19:40-48.
Minugh, E., et al. 1983. A case history: Cleanup of a subsurface leak of refined product. In:
Proceedings 1983 Oil Spill Conference, February 28 - March 3, San Antonio, TX. American
Petroleum Institute, Washington, DC, Publication #4356, pp. 397-401.
Morgan, P., and R.J. Watkinson. 1989. CRC Crit. Rev. Biotech. 8:305-333.
National Academy of Sciences. 1985. Oil in the Sea: Inputs, Fates and Effects. National Academy
of Sciences, Washington DC.
Nelson, M.J.K., et al. 1987. Appl. Environ. Microbiol. 53:949-954.
Office of Research and Development. 1989. Alaskan Oil Spill Bioremediation Project. EPA/600/8-
89/073. U.S. Environmental Protection Agency, Washington, DC.
Office of Research and Development. 1990. Biosystem Technology Development Program.
EPA/600/9-90/041. U.S. Environmental Protection Agency.
Papp, M.L., D.T. Heggem, R.D. Van Remortel, R.W. Gerlach, J.M. Pollard, and L.D. Stetzenbach.
1989. Alaskan Oil Spill Bioremediation Quality Assurance Quality Assurance Plan and Standard
Operation Procedures. Draft. EPA/600/x-xx/xxx. U.S. Environmental Protection Agency, Las
Vegas, NV.
Parsons, T.R., Y. Maita, and C.M. Lalli. 1984. A Manual of Chemical and Biological Methods for
Seawater Analysis. Pergamon Press, Inc., Maxwell House, Elmsford, NY.
570
-------
REFERENCE
Peltier, W.H., and C.I. Weber. 1985. Methods for measuring the acute toxicity of effluents to
freshwater and marine organisms. 3rd ed. EPA 600/4-85-013. U.S. Environmental Protection
Agency, Cincinnati, OH.
Peterson, B.J., and B. Fry. 1987. Stable isotopes in ecosystems studies. Anna. Rev. Ecol. Syst. 18:
293-320.
Peterson, B.J., R.W. Howarth, and R.H. Garritt. 1985. Multiple stable isotopes used to trace the flow
of organic matter in estuarine food webs. Science 227:1361-1363.
Pritchard, P. H. and C. F. Costa. 1991. EPA's Alaska oil spill bioremediation project. Environ. Sci.
Tech. 25:372-379.
Pritchard, H., and C. Costa. 1990. U.S. EPA. Letter to R. Mastracchio, Exxon, August 7.
Raymond, R.L., V.W. Jamison, and J.O. Hudson. 1976. AIChE Svmp. Ser. 73:390-402.
Raymond, R. et al. 1978. American Petroleum Institute, Washington, DC. Project No. 307-77.
Rittmann, B.E., and N.M. Johnson. 1989. Water Sci. Tech. 21:209-219.
Rogers, J. et al. 1990. The role of microorganisms in the bioremediation of the oil spill in Prince
William Sound, Alaska. In: Proceedings of the 83rd Annual Meeting and Exhibition, Air and Waste
Management Association, Pittsburgh, PA, June 24-29, 1990. 90-22.2.
Spj-ent, P. 1989. Applied Nonparametric Statistical Methods. London. Chapman and Hall.
Stanley, T.W., and S.S. Verner. 1985. The U.S. Environmental Protection Agency's Quality Assurance
Program. In: Quality Assurance for Environmental Measurements. ASTMSTP867. American Society
for Testing and Materials, Philadelphia, PA., pp. 12-19.
Steel, R.G.D., and J.H. Torrie. 1980. Principles and Procedures of Statistics a Biometrical Approach.
2nd ed. McGraw-Hill.
571
-------
REFERENCE
Svem, P. 1987. Accidentally Spilled Gas-Oil in a Shoreline Sediment on Spitzbergen: Natural Fate
and Enhancement of Biodegradation. Sintef, Applied Chemistry Division, N-7034 Trondheim,
Norway. 16 pp.
Taylor, J.K. 1987. Quality Assurance of Chemical Measurements. Lewis Publishers, Chelsea, MI.
328 pp.
USEPA. 1984. National oil and hazardous substances pollution contingency plan; Final Rule 18.
Federal Register 40 CFR Part 300, July.
USEPA. 1987. Methods for estimating the chronic toxicity of effluents and receiving waters to
marine and estuarine organisms. EPA 600/4-87-028. U.S. Environmental Protection Agency,
Cincinnati, OH.
USEPA. 1990. Significance of bioremediation in Prince William Sound. Internal USEPA Report.
USEPA ORD. 1990. Alaska Oil Spill Bioremediation Project Status Report as of June 12, 1990.
Valdez.
USEPA ORD. 1990. Alaska Oil Spill Bioremediation Project Status Report as of July 16, 1990.
Valdez.
Velinsky, D.J., L.A. Cifuentes, J.R. Pennock, J.H. Sharp, and M.L. Fogel. 1989. Determination of
the isotopic composition of ammonium-nitrogen from estuarine waters at the natural abundance level.
Mar. Chem. 26:351-361.
Williams, R., T. Pasley, S. Warren, R. Zweidinger, R. Watts, A. Stead, and L. Claxton. 1988.
Selection of a suitable extraction method for mutagenic activity from woodsmoke-impacted air
particles. Int. J. Environ. Anal. Chem. 34:137-154.
Yaninga, P., et al. 1985. Restoration of water quality in a multiaquifer system via biodegradation
of the organic contaminants. In: Proceedings 5th National Symposium on Aquifer Restoration and
Ground Water Monitoring, National Water Well Association, Worthington, OH, pp. 510-515.
Zar, J. 1984. Biostatistical Analysis, 2nd edition. Prentice-Hall, Englewood Cliffs, NJ.
572
-------
APPENDIX A
PROJECT ORGANIZATION AND RESPONSIBILITIES
The Oil Spill Bioremediation Project was directed by a team of scientists from EPA, with
support from EPA laboratory contractors, the State of Alaska, the University of Alaska, and Exxon.
The project was a team effort and personnel assisted where needed. C. Costa, Division Director at
EPA/EMSL Las Vegas, and P.H. Pritchard, Branch Chief at EPA/ERL Gulf Breeze, shared overall
responsibility for the project. C. Costa handled logistics, administration, and other nontechnical areas
of project management; H. Pritchard served as the Principal Investigator with responsibility for the
scientific aspects of the program. Visits from Headquarters personnel provided oversight of program
operations, objectives, and management. Project staff were drawn from the following organizations:
• EPA ORD Headquarters in Washington, D.C.
• EPA Environmental Research Laboratories - Gulf Breeze, Florida; Athens, Georgia; and
Ada, Oklahoma
• EPA Risk Reduction Engineering Laboratory - Cincinnati, Ohio
• EPA Environmental Monitoring Systems Laboratories - Las Vegas, Nevada, and Cincinnati,
Ohio
• EPA Health Effects Research Laboratory - Research Triangle Park, North Carolina
• EPA Center for Environmental Research Information - Cincinnati, Ohio
• University of Alaska
• Alaska Department of Environmental Conservation
• Exxon Research and Engineering - Annandale, New Jersey
• Exxon Production Research - Houston, Texas
• Exxon Biomedical Services - East Millstone, New Jersey
The organizational structure for the Oil Spill Bioremediation Field Project is shown in Figures
A.I and A.2. Principal scientists and support staff for the project are listed in Table A.I.
A-l
-------
Investigation Areas
Support Areas
EPA
EXXON
Pro|*ct ftlamgw
Costa
Principal hiv
QMC
Administration
AdmlnttlraUv* Autotanl
8p*dal
Figure A.1 Project Organization Chart for Summer 1989
-------
EPA
Haadquartara
EIIMI
CooparaNng
Protect Managar
Coala
Principal ktvMMgator
Pritdiard
QM3C
InvesUgatlon Areas
AddMoH
>
Ubontory
Support Areas
Admlnklnttlon
AdmlntetraHv* AMtotonI
SpMtalProtacto
Octa ManagMMnt/
An«ly«to
Figure A.2 Project Organization Chart for Summer 1990
-------
APPENDIX A
TABLE A.I. EPA BIOREMEDIATION PROJECT STAFF
Responsibility
Project Management
EPA Headquarters
Special Projects
and Administration
Nutrient Addition
Ecological
Monitoring
Name
H. Pritchard1'2'3
C. Costa1-2-3
J. Skinner1-2'3
D. Valentinetti1
T. Baugh2'3
E. Sullivan1
E. Gray1
K. Brown'
S. Doherty3
D. Rosenblatt3
L. Suit3
V. Furlong1'2-3
R. Shoemaker1
E. Clay1
W. Barlow1
T. Morton1
A. Venosa1-2
J. Glaser1
F. Kremer1-2
J. Haines1
E. Opatken1
S. Safferman1
A. Horowitz1
J. Clark1'2-3
L. Claxton1
R. Parrish1'3
R. Coffin1'2-3
L. Cifuentes1'2'3
J. Macauley1
J. Hoff1
J. Patrick1
R. Stanley1
G. Primrose1
T. Heitmuller1
B. Dorn3
Affiliation
EPA/ERL
EPA/EMSL
EPA/ORD
EPA/ORD
EPA/ORD
TRI
TRI
TRI
TRI
TRI
TRI
EPA/EMSL
EPA/EMSL
EPA/EMSL
EPA/EMSL
EPA/EMSL
EPA/RREL
EPA/RREL
EPA/CERI
EPA/RREL
EPA/RREL
EPA/RREL
Indep.
Consultant
EPA/ERL
EPA/HERL
EPA/ERL
TRI
TRI/Texas A&M
EPA/ERL
TRI
EPA/ERL
EPA/ERL
TRI
TRI
TRI
Location
Gulf Breeze, FL
Las Vegas, NV
Washington, DC
Washington, DC
Washington, DC
Rockville, MD
Seattle, WA
Rockville, MD
Gulf Breeze, FL
Rockville, MD
Rockville, MD
Las Vegas, NV
Las Vegas, NV
Las Vegas, NV
Las Vegas, NV
Las Vegas, NV
Cincinnati, OH
Cincinnati, OH
Cincinnati, OH
Cincinnati, OH
Cincinnati, OH
Cincinnati, OH
Shaker Heights,
OH
Gulf Breeze, FL
Res. Tri. Park, NC
Gulf Breeze, FL
Gulf Breeze, FL
College Sta., TX
Gulf Breeze, FL
Gulf Breeze, FL
Gulf Breeze, FL
Gulf Breeze, FL
Gulf Breeze, FL
Gulf Breeze, FL
Gulf Breeze, FL
Participated during the Summer of 1989.
Participated during the Winter of 1989/1990.
'Participated during the Summer of 1990.
A-4
-------
APPENDIX A
TABLE A.I (CONTD)
Responsibility
QA/QC
Data Management/
Analysis
Field Support
Logistics
Microbiology
Analysis
Name
D. Heggem1'2-3
M. Papp1-8
L. Stetzenbach1
J. Pollard1-'
W. Beckert1
D. Chaloud8
T. Chiang3
W. Horn1
B. Damiata1
B. Gerlach2-3
J. Pollard1
A. Neale2'3
A. Borders8
J. Baker1
W. Kinney1
F. Kremer1
D. Miller1
R. Wright1
M. Dillon1'2'3
J. Wilson1
K. Cabbie3
E. Eschner8
R. Dushek3
G. Merritt3
N. Halsell8
B. Shokes1
K. Schmidt1-2'3
D. Peres3
M. Sweeney3
J. Rogers1
R. Araujo1-2'3
S. Montgomery1
C. Robinson1
M. Shields1
L. Bosworth1
J. Mueller1-2'3
S. Baraket1
S. Resnick3
Affiliation
EPA/EMSL
Lockheed
EPA/UNLV
Lockheed
EPA/EMSL
Lockheed
Lockheed
SAIC
SAIC
Lockheed
Lockheed
Lockheed
EPA/EMSL
Lockheed
EPA/EMSL
EPA/CERI
EPA/ERL
SAIC
SAIC
EPA/ERL
Lockheed
Lockheed
Lockheed
Lockheed
Lockheed
SAIC
SAIC
Lockheed
Lockheed
EPA/ERL
EPA/ERL
TRI
TRI
TRI
TRI
So. Bio.
TRI
TRI
Location
Las Vegas, NV
Las Vegas, NV
Las Vegas, NV
Las Vegas, NV
Las Vegas, NV
Las Vegas, NV
Las Vegas, NV
San Diego, CA
Golden, CO
Las Vegas, NV
Las Vegas, NV
Las Vegas, NV
Las Vegas, NV
Las Vegas, NV
Las Vegas, NV
Cincinnati, OH
Ada, OK
San Diego, CA
Valdez, AK
Ada, OK
Las Vegas, NV
Las Vegas, NV
Las Vegas, NV
Las Vegas, NV
Las Vegas, NV
San Diego, CA
San Diego, CA
Las Vegas, NV
Las Vegas, NV
Athens, GA
Athens, GA
Gulf Breeze, FL
Gulf Breeze, FL
Gulf Breeze, FL
Gulf Breeze, FL
Gulf Breeze, FL
Gulf Breeze, FL
Gulf Breeze, FL
Participated during the Summer of 1989.
Participated during the Winter of 1989/1990.
Participated during the Summer of 1990.
A-5
-------
APPENDIX A
TABLE A.l (CONTD)
Responsibility
Microbiology
Analysis
(Cont.)
Microcosms
Hydrocarbon/
Nutrient Analysis
Exxon
Cooperating
Scientists
Name
E. Brown3''
G. Winter3
E. Pritchard3
L. Pritchard3
A. Rozich3
J. Kittigawa3
R. Cripe1'2-3''
D. Dalton3
S. Friedman3
T. Mandeville3
M. Shelton3
J. Morin3
J. Rogers1-3
M. Dillon1-3
J. Payne1
D. McNabb1
R. Sims1'3
M. McCabe1
J. Evans1
J. Stiefvater1-8
E. Smiley1
J. Lopez1
T. Fogg1
M. Hart1-3
M. Peters3
C. Timm3
J. Clayton1'3
G. Salata1'3
W. Horn1-8
R. Prince3-"
R. Atlas3
D. Elmendorf3'8
R. Bare3''
M. Grossman8
J. O'Bara3
R. Chianelli1'2'3
S. Hinton1-'
J. Wilkinson2-3
Affiliation
U. of AK
ADEC
U. of AK
U. of AK
ERM
ADEC
EPA/ERL
TRI
TRI
TRI
TRI
TRI
EPA/ERL
Lockheed
SAIC
SAIC
SAIC
SAIC
SAIC
SAIC
SAIC
SAIC
SAIC
SAIC
SAIC
SAIC
SAIC
SAIC
SAIC
Exxon
U. Ky.
Exxon
Exxon
Exxon
Exxon
Exxon
Exxon
Exxon
Location
Fairbanks, AK
Valdez, AK
Fairbanks, AK
Fairbanks, AK
Exton, PA
Valdez, AK
Gulf Breeze, FL
Gulf Breeze, FL
Gulf Breeze, FL
Gulf Breeze, FL
Gulf Breeze, FL
Gulf Breeze, FL
Athens, GA
Las Vegas, NV
San Diego, CA
San Diego, CA
San Diego, CA
San Diego, CA
San Diego, CA
San Diego, CA
Anchorage, AK
San Diego, CA
San Diego, CA
San Diego, CA
San Diego, CA
San Diego, CA
San Diego, CA
San Diego, CA
San Diego, CA
Annandale, NJ
Lexington, KY
Annandale, NJ
Annandale, NJ
Annandale, NJ
Florham Pk, NY
Annandale, NJ
Annandale, NJ
Anchorage, AK
Participated during the Summer of 1989.
Participated during the Winter of 1989/1990.
Participated during the Summer of 1990.
'Oil biodegradation studies.
A-6
-------
APPENDIX A
ON-SITE ORGANIZATION AND RESPONSIBILITIES
On-site personnel provided the following support:
• The special projects staff handled helicopter and boat schedules, housing and transportation
issues, travel, EPA timecards, weekly reports, office management and support, and
equipment deposition.
• The nutrient addition staff was responsible for nutrient application and monitoring for Snug
Harbor in 1989. This involved staking plots, loading of briquettes into bags, excavation of
monitoring wells, and monitoring of well samples.
• The ecological monitoring staff collected water and biological samples for ecological effects
testing and prepared the samples for shipment. Water samples were collected from the field
experiment beaches and Prince William Sound. Biological samples, including aquatic
vegetation and fauna, were collected from the experimental beaches, identified, and frozen
for analyses.
• The QA/QC staff designed the QA/QC program, prepared the QA Plans, monitored QA/QC
data, resolved data quality issues, and prepared an assessment of overall program data quality.
The QA Plans provided the written documentation of the QA program. QA staff monitored
data quality and performed informal, internal audits of program operations.
• The data management/analysis staff was responsible for software development and testing,
development and input of field data forms, sample tracking and archiving, and data analysis.
Plots and statistical outputs of the data were generated by the staff as requested by the
Principal Investigator, QA staff, and program scientists.
• The field crew set up beaches for field experiments, collected field samples and data, and
removed all installed equipment following completion of an experiment. Activities included
staking plots, preparing oiled homogenates and sample baskets, installing monitoring wells,
and applying fertilizer.
• The logistics staff was responsible for ordering and tracking all equipment and supplies for
the program staff. An inventory of all equipment and supplies was maintained, and the staff
oversaw removal and proper distribution of materials from Valdez at project completion in
accordance with EPA regulations and guidelines.
• The microbiology staff performed field and laboratory-based experiments to measure
numbers of oil degraders, and used biometers to monitor micrubial activity. Analyses
included MPN viable count, carbon dioxide, TOC, and radiolabsled carbon analyses. The
staff also input data and prepared plots for data analysis and interpretation.
• The microcosm staff conducted experiments using three types of microcosms: jar, tank, and
column. These systems used oiled beach material from Prince William Sound. The jar and
column microcosm studies were performed in a laboratory in Valdez, while the tank
microcosm study was performed aboard the motor vessel AUGUSTINE in 1989.
A-7
-------
APPENDIX A
• The analytical laboratory conducted nutrient and oil chemistry analyses, performed QC
checks, and prepared data for transfer to the data base management system. Additional
responsibilities included transfer of samples to the sample archive, and set-up and removal
of the laboratory.
• The cooperating scientists from Exxon provided a linkage between EPA and Exxon dealing
with scientific issues, experimental designs, data analysis and interpretation, and logistics.
Individuals were also located in Valdez to perform their experiments and provide scientific
expertise and assistance to the Exxon/EPA/ADEC monitoring program.
A-8
-------
APPENDIX B
CHRONOLOGY OF EVENTS
An overview of the events that took place since the initiation of the oil spill project is
presented.
MAY 1989
15
16
19
EPA command center for the Bioremediation Project established in Valdez.
Investigation of potential fertilizers for field tests initiated.
Snug Harbor selected as the initial test site.
Efforts initiated to establish test plots within the beaches and set up a staging operation.
Fishing vessel AUGUSTINE available for use.
Studies on the movement of ground water and nutrients in the beaches initiated.
Methods for adequately sampling contaminated beach material developed, a sampling
20
23
design established, and methods for fertilizer application to the beaches finalized.
25 • Snug Harbor background data on the extent of oil contamination on the test beaches
collected.
JUNE 1989
2 • Collection of background ecological monitoring data from Snug Harbor initiated.
4 • Microcosm design and construction initiated the first week of June (placed aboard the
AUGUSTINE).
• Due to the length of time to establish a microbiology laboratory, only one type of method
established to measure the number of oil degraders.
8 • Following delays due to complications with logistics and weather, nutrient field
application began. Stable isotope study initiated.
12 • A major rainstorm washed much of the oleophilic fertilizer off the beaches, so it was
reapplied.
15 • A workshop on "Beach and Nearshore Hydraulics" was held in Seattle, Washington. (See
Appendix B)
20 • Visual loss of oil from rock surfaces of the cobblestone beaches treated with oleophilic
fertilizer apparent.
JULY 1989
1 • First status report on the field demonstration project submitted to Exxon.
18 • A recommendation for large-scale application of fertilizers to beaches in PWS submitted
to Exxon.
24 • Passage Cove selected for additional nutrient and beach hydraulic investigations.
25 • Application of INIPOL and slow-release CUSTOMBLEN granules to Tern Beach at
Passage Cove.
27 • Sampling wells for beach hydraulic study installed on Kittiwake Beach.
B-l
-------
APPENDIX B
AUGUST 1989
2 • Began operation of sprinkling system on Kittiwake Beach at Passage Cove.
10 • Passage Cove beaches treated with oleophilic and slow-release fertilizer visually showed
marked disappearance of oil relative to the reference beaches.
18 • Disappearance of oil evident from the Passage Cove beach treated with fertilizer solution
from a sprinkler system.
SEPTEMBER 1989
5 • Final sampling at Passage Cove.
12 • Final sampling at Snug Harbor.
15 • Addendum to the July 1 Status Report submitted to Exxon.
NOVEMBER 1989
8 • Workshop held in Gulf Breeze, FL, to discuss data analysis, interpretations, and
Significant findings.
12 • Winter sampling on Passage Cove test beaches.
DECEMBER 1989
13 • Workshop held in Washington, D.C. to develop winter research plan for oil
bioremediation.
APRIL 1990
• EPA, Exxon, NOAA, UAF, and ADEC scientists met to develop strategies for the
summer research.
• Joint monitoring program conceived and developed.
• Research plan for testing the effects of fertilizer concentrations and fertilizer
applications developed by EPA scientists.
MAY 1990
22 • Peer scientists met in Anchorage to review research plans; plan revised and submitted to
Exxon.
JUNE 1990
• Sampling baskets constructed and tested on Disk Island.
• Test plots on Disk Island for fertilizer-specific activity established.
• A small test plot with monitoring wells established at Snug Harbor to assess scaling
effects of fertilizer granule application.
11 • Elrington Island beach selected for testing fertilizer
to solution application through a sprinkler system.
IS • Oil chemistry analysis laboratory established at Prince William Sound Community
College; microbiology laboratory established at ADEC lab.
18 • Experimental plots and well points installed at Disk Island, and also at Snug Harbor for
to scaling experiment.
22 • Motor vessel INSPECTOR made available by Exxon for research support.
B-2
-------
APPENDIX B
20 • Sampling baskets prepared and installed at Disk Island.
23 • CUSTOMBLEN fertilizer applied at Disk Island and monitoring of nutrient
concentrations initiated.
30 • Sampling baskets prepared and installed on the Bath beach at Elrington Island.
JULY 1990
1 • Sampling baskets prepared and installed on the Sprinkler beach at Elrington Island.
• Fertilizer applied on the Bath beach.
2 • Sampling baskets prepared and installed on the Control beach at Elrington Island.
• Initial run of the sprinkler system conducted at the Sprinkler beach.
9 • QA/QC audits of field and laboratory operation conducted.
15 • Alaska Department of Natural Resources barred beach access to Elrington Island pending
resolution of the site permit.
20 • Site permit problem for Elrington Island resolved.
21 • Fertilizer application on the Sprinkler beach resumed. Experiment to detect nutrient
enrichment initiated.
23 • EPA Administrator William Reilly visited various beach sites in Prince William Sound.
27 • Field activities completed for the Disk Island fertilizer application rate study.
30 • Ecological monitoring initiated on Elrington Island.
AUGUST 1990
6 • External QA audit conducted.
to • Meeting held with representatives of Exxon to discuss results of the Disk Island and
10 Elrington Island experiments.
7 • Final sampling (basket removal) and removal of all remaining field items at Elrington
Island.
13 • Samples upon which to base winter sampling collected at Elrington Island along a transect
from the three treatment areas.
DECEMBER 1990
12 • Winter sampling at Elrington.
FEBRUARY 1991
19 • Meeting with EPA, Exxon, ADEC, U. of Alaska held in Las Vegas, NV to discuss results
to of the oil spill project and plan for the summer of 1991.
21
U.S. Environmental Protection Agency
Region 5, Library (PL-12J)
77 West Jackson Boulevard, 12th Floor
Chicago, IL 60604-3590
B-3
------- |