Project Title:
Principal Investigators:
Nancy E. Kinner
Thomas P. Ballestero
Environmental Research Group
University of New Hampshire
Stephen H. Jones
Jackson Estuarine Laboratory
University of New Hampshire
Submitted: February 15, 2000
I. Accomplishments
During the period covered by this progress report (August 1, 1999 through January 31, 2000), we had two goals: (1) field evaluation of the effectiveness of the bioremediation technologies, and (2) submission of this progress report. This period included the end of our second field season which began on April 10, 1999 and ended on October 16, 1999.
The first sampling event for the second field season (1999) occurred on April 10 and 11. Sampling is a two day process because four randomly-selected grids must be sampled within each plot due to high in situ and analytical variability of sediment total petroleum hydrocarbons (TPH). Sampling can only be conducted in a 3-4 hour window when the tide is low exposing the marsh sediments. Parameters measured include TPH, pH, salinity, plant density, plant height, abundance of oil-degrading microorganisms, nitrate, ammonia, sulfate and phosphate. Porewater samples, in particular, must be obtained slowly (at low tide, porewater yields are low), requiring as much as 45 minutes each. To prevent trampling the marsh grass and compacting the sediments, portable catwalks must be moved onto the marsh and placed to allow access to each randomly-selected grid sampled , some of which are not accessible from the sides of the plots. Sampling is facilitated by having a large crew at the site. This typically consists of 2 graduate students, 1-3 undergraduates and 2-3 high school students.
Weekly nitrate addition began in April this year and was terminated in October. Continuous air injection also began in April and was terminated in October. Three sampling events have occurred (one per month August through October) during the period covered by this report. Sampling will recommence on a monthly basis beginning in April 2000 and continue through July 2000 (the end of the project).
Three high school students from the Waynflete School (Portland, ME) continued to help with sampling and nitrate additions during the 1999 field season. One of the students is a senior and will not be able to continue working with us during the 2000 sampling season. He has worked both field seasons and will be sorely missed. A new student, who is currently a junior, began to work with us in October and will continue into the 2000 field season. We plan to employ 1-2 more new students identified by the Chair of the Waynflete School’s Science Department (Carol Titterton) for the 2000 sampling season.
The two graduate students (Jennifer Gilbert and Dalia Hildebrand), who began working on the project in June 1999, will continue the field and laboratory work for the 2000 season. Fabio Roldan-Garcia and Geoffrey Grant (the two original graduate students on the project) are now working and will complete their theses in 2000.
During the period February to December 1999, gas flux chambers were installed in the Fore River Creek salt marsh and used to collect N2O gas samples every two weeks over the growing season (March to November). These samples were collected and analyzed by Richard Puk (Ph.D. student in Natural Resources — Advisor: Prof. William McDowell). Money for supplies and equipment for this gas sampling are provided by the CICEET grant.
A Tekmar 7000 headspace autosampler interfaced to an HP5290 GC is used for the N2O analysis. The autosampler controls injection of samples and re-starts the integrator for each injection. Gas samples are analyzed with an electron capture detector (ECD) at 394° C, using a 2.5 x 4 mm ID stainless steel packed column (HayeSep Q, 80-100 mesh). An Ar/CH4 mixture is used as the carrier gas. The run time is 20 min at a column temperature of 70° C and a carrier flow rate of 30 mL/min.
Two chambers (25" x 25" x 14.5" each) and twenty collars were designed and constructed of aluminum. Each chamber is equipped with two fans for continuous mixing of the headspace and has ports for sampling and thermocouple installation. The collars, which are pressed into the sediments and support the chambers, have 6 cm holes on each side to allow filling and drainage of water during a tidal cycle. Movement of water on the marsh is mostly lateral rather than vertical and the holes are sufficient to drain or fill the collar quickly during an ebb and flood tide, respectively. A total of 12 collars were installed during late February 1999 in the three experimental (nitrate, nutrients, air) plots and four collars are in the control plot. The collars are placed 11 cm into the sediment. A knife was used to cut a thin groove to allow each collar to be placed between the roots of the Spartina. Since the collars were installed, there has been no dislocation during the tidal cycles. The collars were removed for the winter.
Samples for N2O, NH3, plant height, and temperature were taken every two weeks beginning March 14, 1999. Sampling is usually a one or two day process because there are 16 collars to sample. Sediment samples are collected using a 60 mL syringe, dried, and CO2 is removed before they are placed into evacuated 22 mL autosampler vials. Samples are taken every 5 minutes over a 20 minute period. During sampling, the holes in the collars are closed with rubber stoppers. Sampling is conducted in a 3-4 hour window when the tide is low and the marsh is exposed.
There is some very preliminary evidence that both the nutrient, and to a lesser extent, the nitrate plots, may be exhibiting greater production of N2O with time compared to the controls. Flux data, blade height of Spartina and water samples still need to be evaluated. Porewater samples will be evaluated to determine nitrate, phosphate, ammonium and dissolved N2O. Mr. Puk and Professor McDowell will calculate gas fluxes, and determine the effects of experimental treatment, Spartina growth, temperature, and humidity on them.
Dr. Stephen Jones and John Sowles (Maine Department of Environmental Protection, ME DEP) discussed sampling mussel tissue in the Fore River to determine if the marsh is chronically exposed to some background level of petroleum contaminants from commercial/industrial activities in Portland Harbor. Mussels from the Fore River have been used as indicators of chronic oil exposure in the Gulfwatch Program. Dr. Stephen Jones and Charles Penny (MEDEP) collected mussels from four different mussel beds in the Fore River/Fore River Creek area on October 15, 1999. Two tissue samples from each of the four sites were processed and analyzed in the laboratory using the same Massachusetts DEP TPH method that is used for the marsh sediment samples. The mussel tissue is also being processed and analyzed by Gulfwatch. Results from these analyses are not yet available.
The software for the gas chromatograph used in the TPH analysis was upgraded to handle Y2K problems (Version 6.50 upgraded to Version 6.51). The new version was installed in December 1999 and no problems have been encountered in 2000.
We will recommence sampling in April 2000. Sampling will occur monthly through July 2000. Drs. Nancy Kinner, Ballestero and Jones submitted a pre-proposal to CICEET to continue the project for additional years. CICEET has requested that a full proposal be submitted in March.
Percent recoveries for the surrogate in the TPH analysis have been lower than the range accepted by the Massachusetts DEP Method (40-140%). This is mainly attributed to the matrix of the samples (marsh sediments vs. the sand samples the method was developed to analyze). Using standard QC protocols, we have generated our own percent recovery control charts for the surrogate based on the sediment matrix. As per standard QC procedures, any sample with a percent recovery out of this range is not included in the data set. Sand blanks have been in the acceptable surrogate recovery range (40-140%) for the Massachusetts DEP method.
Our calibration standards have produced response factors that are within the Massachusetts DEP Method’s specified range (+/- 25% relative percent difference) for short and long chain aliphatics and aromatics. Mean values for the internal standard area counts for aliphatics and aromatics are calculated for each sample. Any sample with an internal area count that is not within two standard deviations from the mean (accepted QC criterion) is considered out of control and excluded from the data set. Generally, our internal standard area counts are within the acceptable range.
A statistical analysis of Year I and Year II data is being conducted by modeling the TPH degradation within the plots using multivariate regression. Log transformations of the data are necessary because the TPH concentrations are not normally distributed, which is typical for organic contaminants in sediments. The two control plots are being treated as one entity (n = 16 samples/event). The general multivariate regression can be simplified as the following linear equation:
log TPH = at + b
where:
t = time
b = the cumulative effect of the individual parameters
a = the cumulative effect of the interaction between parameters and time.
The detailed linear expression for our experiment is:
log Y = b 0 + b 1· air addition + b 2· nitrate addition + b 3· nutrient addition +
b 4· control + b 5· time + time(b 6 · air addition + b 7· nitrate addition + b 8· nutrient
addition + b 9· control)
where:
Y = TPH, short chain aliphatics, long chain aliphatics, or aromatics
b 0 - b 5 = coefficient for parameters in the b term
b 6 - b 9 = coefficient for parameters in the a term.
Four detailed linear equations can be produced — one for each fraction of the hydrocarbons (short chain aliphatics, long chain aliphatics, and aromatics and one for TPH (the sum of the three fractions). Table 1 contains the values for each of the b coefficients.
The model produced for short chain aliphatics, long chain aliphatics, aromatics, and TPH resulted in r2 values of 0.453, 0.436, 0.241, and 0.406 respectively. The models indicate that all plots are showing a decline in petroleum concentration over time. As expected, time is a significant variable for the decline in all petroleum components. Because of the time effect, only b 6-b 9 terms (which account for the time effect) can be used to determine the effects of each of the treatments on the petroleum hydrocarbon concentrations. The air and nitrate plots are the only ones that shows any significant decline in TPH over time (b 6 air treatment for TPH > 90% significance, b 7 nitrate treatment for TPH > 75% significance). The other two (nutrients and controls) plots show a decline, but it is much less than 75% significant. When the hydrocarbon fractures are evaluated individually as short chain aliphatics, long chain aliphatics, and aromatics, the model exhibits somewhat different results for the decline in petroleum over time. For short chain aliphatics, only the nitrate plot and nutrient plots are showing a decline (greater than 95% and 90% significant, respectively). All plots are showing a 99.9% significant decline in the long chain aliphatics and only the nitrate plot is declining with a significance greater than 80% for the aromatic compounds. [N.B., Because the model is multidimensional showing the results graphically is difficult, so no figures are included.]
The data set used in the model only includes data that falls within the QC requirements outlined above. However, a couple of the data sets for the first two sampling events have only a few data points (n = 2-4 vs. n = 8) and this may be affecting the model and masking results. We are currently working with Prof. Philip J. Ramsey (Applied Mathematics) and Prof. L David Meeker (Applied Mathematics) to address these concerns and better understand the results of the model.
Overall, the data are generally showing that the nitrate treatment is significantly improving the rate of hydrocarbon degradation. The results for the air and nutrient treatments are also promising. These results are especially encouraging considering that we have not optimized the delivery system for the treatments. We suspect that the nitrate and nutrient solutions are not being totally distributed to the marsh sediments. To check this and optimize delivery, we are proposing to continue the CICEET project. In anticipation of possible funding, we installed three horizontal injection wells (similar to the existing ones) at a site adjacent to the nutrient plot. These need to "age" (which will take several months) to be similar to those currently in the marsh. If we receive the additional CICEET funding, we will use these wells and adjacent lysimeters to monitor fluid distribution in the sediments.
There has been no significant change in the population of oil-degrading bacteria with time (Figure 1). Initial MPN counts were higher and more variable due to some initial problems with the analysis, but that problem is no longer occurring. We did not expect large increases in these data over time because the amount of hydrocarbons per unit microorganism is relatively low.
The complete analysis of the blade height and abundance of Spartina alterniflora in the marsh will be available for the final report. The plant height for each treatment is shown graphically and as expected, follows a seasonal trend (Figure 2).
II. Tasks and Activities for Next Reporting Period
During the period February 2000 to July 2000, we will continue field sampling, off-gas measurements, adding air and nitrate to the plots (starting in April 2000), adding nutrients twice in the spring (after plant growth starts), and analyzing porewater samples that have been frozen for nutrient analysis. The TPH results for the mussel analysis will be compared to those obtained from Gulfwatch.
We will continue using the same sampling procedures and analytical protocols we used during the current reporting period, as outlined in the original proposal and the previous progress reports.
At the present time, we have no concerns or difficulties.
III. Expenditures
The project is within budget at this time. There have been no unusual expenditures.
Table 1: Coefficients for multivariate regression linear equations.
|
Coefficient
|
TPH
|
Short
Chain Aliphatic Fraction
|
Long
Chain Aliphatic Fraction
|
Aromatic Fraction
|
| b 0 |
3.416
|
2.940
|
2.806
|
2.868
|
| b 1 |
1.83x10-1
|
1.715x10-1
|
4.137x10-1
|
1.561x10-1
|
| b 2 |
5.35x10-2
|
1.426x10-2
|
3.419x10-1
|
-2.422x10-3
|
| b 3 |
1.738x10-1
|
-1.085x10-1
|
-5.241x10-1
|
-1.477x10-1
|
| b 4 |
-6.27x10-2
|
-7.724x10-2
|
-2.351x10-1
|
-5.959x10-3
|
| b 5 |
-1.6x10-3
|
-1.883x10-3
|
-2.025x10-3
|
-1.272x10-3
|
| b 6 |
-3.46x10-4 **
|
-1.29x10-4
|
-1.282x10-3 ***
|
-1.49x10-4
|
| b 7 |
8.15x10-5 *
|
4.375x10-4***
|
-1.329x10-3***
|
3.393x10-4*
|
| b 8 |
-5.55x10-5
|
-3.909x10-4**
|
1.547x10-3***
|
-1.713x10-4
|
| b 9 |
2.09x10-4
|
8.24x10-5
|
1.064x10-3***
|
-1.9x10-5
|
* = >75% confidence
** = >90% confidence
*** = >95% confidence