Using Plants to Remediate Petroleum-Contaminated Soil

Period Covered by Report:
Date of Report:
EPA Grant Number: R827015-01-0
Title: Using Plants to Remediate Petroleum-Contaminated Soil
Investigators: Greg Thoma, Duane Wolf, Craig Beyrouty
Institutions: University of Arkansas
EPA Project Officer: Bala Krishnan
Project Period: September 1, 1999 to August 31, 2000 (N/C Ext. to June 3 0, 2001)
Project Amount: $134,949
Research Category: Phytoremediation

Description:

Objective(s) of the Research Project:

1. To conduct greenhouse studies to screen plants for their ability to germinate and grow in weathered crude oil-contaminated soil with or without amendments.

2. To survey and collect plant species currently growing on contaminated sites and screen the plants and rhizosphere microorganisms for their ability to enhance biodegradation of petroleum contaminants.

3. To conduct an on-site field study to evaluate likely combinations of plants and management systems to enhance phytoremediation of weathered crude oil -contaminated sites.

4. To develop a model that can be used to summarize and aid in the interpretation of experimental data collected in both the laboratory and field during the first experimental season.

Progress Summary/ Accomplishments:

This report covers the September 1 to November 30, 2000 period and summarizes our current IPEC phytoremediation studies that consist of an on-site field project in southern Arkansas, greenhouse studies, and a mathematical modeling project.

Field Study

The field site in El Dorado, AR is located in a bermed area that is the site of an intentional spill by vandals approximately three years ago. The experimental plots consist of four replicates of the following treatments: (1) nonvegetated-nonfertilized control, (2) fescue-ryegrass-alfalfa + fertilizer, and (3) fescue-bermudagrass + fertilizer. Each field plot has 12 microplots ('soil socks') that contain homogenized soil that allow monitoring of the field treatments, on a smaller scale, with less effect of field variability of the contaminant levels.

On 10-12 July 2000, 6 months after establishment of vegetation at the site, soil and plant samples were collected from the plots. Plant shoot biomass and root biomass, length, surface area, and volume for each of the treatments are summarized in Table 1. All plant species appeared to be exhibiting adequate plant growth.

The Total Petroleum Hydrocarbon (TPH) and biomarker (hopane) analyses of the soil samples collected 6 months after plot establishment (t=6) are currently being conducted.

The sampling that was scheduled for December 2000 was canceled due to ice storms that made travel nearly impossible. We propose to reschedule the sampling for May 2001 to allow for sampling of actively growing plant material.

Table 1. Shoot biomass and root biomass, length, surface area, and volume for samples collected 12 July 2000 or 6 months after initiation of the field study in El Dorado, AR.

Greenhouse Study

Because oil-contaminated soils have different nutrient requirements than conventional agronomic recommendations suggest, a nutrient rate study using warm-season plant species was conducted. Soil analyses for the petroleum contaminated soils amended with organic materials are summarized in Table 2.

Table 2. Soil Chemical Data from Greenhouse Study.

Mathematical Model

We have consulted with a mathematician regarding the solution to the system of differential equations because we found certain input parameters cause the numerical solution to fail. We have identified the cause of this failure, and are currently working to correct it. For parameter values that do not cause the failure, we have demonstrated, through grid independence testing and solutions computed with different numerical algorithms (4th order explicit Runge-Kutta, Gear's method, and the modified Rosenbrock method) that the numerical solutions are valid. Specifically, this means that the solution has converged, within a pre-specified error level, to the true solution of the system of equations.

Simulations to evaluate the significance of the bulk and rhizosphere kinetic rate constants, and the rhizosphere volume are presented in Figure 1. Several comparisons can be drawn regarding the phytoremediation reaction system. A comparison of curves 1 and 3 indicates that the bulk soil degradation rate constant, here taken to be 10% of the maximum rhizosphere rate constant, still significantly contributes to the overall hydrocarbon dissipation. Thus, even when phytoremediation is used, some management strategy focusing on increasing the microbial activity in the bulk soil should be considered. Comparison of curves 2 and 3 shows the effect of root turnover rate. Thus for the conditions simulated here, an annual species would be recommended. Comparison of curves 3 and 5 shows the effect of increasing the rhizosphere zone to 1.5 mm from the root surface. This increase is simulated for the same total root biomass. It would represent the effect of a management strategy that enhanced the microbial biomass extent around the roots (perhaps through additional aeration, irrigation, fertilization, or plant species selection). Finally, comparison of curves 3 and 4 shows the impact of a management strategy that increased the degradation rate constant, k, through manipulations focused on increasing the microbial activity by, for example, increasing the number of degrader organisms in the rhizosphere through selective enhancement such as adjusting soil pH, moisture, or nutrient availability. It is clear from the model predictions that this model can be very useful in guiding the phytoremediation research effort.

Publications/Presentations:

Thoma, G.J., T.B. Lam, E.N. Dempsey, D.C. Wolf, and C.A. Beyrouty. 2000. A mathematical model of phytoremediation of oil contaminated soil. In K.L. Sublette (ed.) Proc. 7th International Petroleum Environmental Conf., Albuquerque, NM. 7-10 Nov. 2000. Integrated Petroleum Environmental Consortium, Tulsa, OK.