Illinois
Fertilizer Conference Proceedings
January 25-27, 1993
| Home | 1993 Index | Search |
| Home | 1993 Index | Search |
Allan S. Felsot1, J. Kent Mitchell2, T.J. Bicki3, J.F. Frank4*
Soils at agrichemical retail facilities throughout the United States are frequently contaminated by pesticides during loading and rinsing operations, even after containment facilities have been installed (Habecker, 1989). In some cases, historical standard operating practices have left a legacy of high pesticide concentrations in soil and ground water that have affected nearby residential water supplies (Long 1989). Sometimes contamination is serious enough to warrant a cleanup order from a state or federal regulatory agency; in other cases, the cleanup may be voluntary as a result of the desire to install new structures or to transfer the property. In either case, the soil must somehow be remediated, but the technical options may be limited and prohibitively expensive.
Current options for cleanup of soil include in situ treatment or excavation followed by treatment (displacement) (Enlow 1990; Norwood and Randolph 1990). In situ methods like vitrification and bioremediation are not presently practical nor commercially available for pesticide-contaminated soils. Currently used displacement techniques include landfilling and incineration. Unfortunately landfilling is becoming a less viable option because of space limitations, and incineration of large quantities of soil is too expensive and not easily accessible for small agrichemical facilities. Treatment of excavated soil by bioremediation may be more suitable to rural facilities because it is an easily transported technology that is cheaper than incineration.
Bioremediation methods under research and development that seem suitable for soil cleanup include composting, bioreactors, and landfarming. Of these three methods, landfarming may be the closest to immediate implementation. Landfarming, also known as land treatment, land application, or land spreading, is a "managed treatment and ultimate disposal process that involves the controlled application of a waste to a soil or soil-vegetation system (Loehr and Overcash 1985)." The waste is spread on agricultural or noncropped land to stimulate degradation, transformation, and/or immobilization of contaminants. The technique is particularly suitable for pesticide-contaminated soils because the practice involves a technology that has been used for many years to dispose of municipal wastewater, sludges, and petroleum refinery wastes (Soil Science Society of America 1986); furthermore, the pesticides in contaminated soils are usually registered by the U.S. EPA for application to soil.
Because of the need for easily implemented methods to dispose of pesticide-contaminated soil during cleanups of agrichemical facilities and the relative ease of excavating and spreading soil, the Illinois Legislature authorized the Illinois Department of Agriculture (IDOA) to permit the land application of pesticide contaminated soil at agronomic rates. The authorization allowed IDOA to prescribe operational control practices to protect the site of application. Successful application of landfarming, however, relies on a detailed site assessment to accurately prescribe the nature and extent of the waste, establishment of cleanup objectives, and development of a remedial action strategy (Bicki and Felsot 1993). Although many states are considering regulations allowing landfarming for disposal of pesticide-contaminated soils, the effectiveness of the technology, methods for stimulation of degradation rates, and possible off-site movement of contaminants has hardly been studied.
In the Corn Belt, candidate land to receive contaminated soils will likely
be fields that cannot be taken out of production. Unresolved questions from
previous landfarming experiments included the translocation of pesticides from
the receiving land, prolonged persistence of pesticides in contaminated soils,
and potential for crop phytotoxicity (Felsot et al. 1988, 1990). Because landfarming
is a leading candidate for treatment of pesticide-contaminated soils, these
questions must be accurately assessed to help define appropriate guidelines
and regulations. This paper describes a field experiment to study the degradation,
translocation, and phytotoxicity of pesticide contaminants in landfarmed soil.
A thorough description of the experimental design, sampling procedures, initial
soil residues, and runoff monitoring have been recently published (Felsot et
al. 1992). This paper presents an update of results following a second year
of monitoring. This paper also presents preliminary results from a companion
field experiment (biostimulation experiment) that investigated the feasibility
of using organic amendments to enhance biodegradation of landfarmed herbicide
residues.
Hypothesis
The hypothesis is stated as criteria of feasibility that after completion of the experiments should enable a judgment about the effectiveness and safety of landfarming . Thus, successful remediation of pesticide-contaminated soil is determined by the satisfaction of four criteria:
(1) Pesticide residues in control (untreated plots) and contaminated soil-treated plots (hereafter referred to as landfarmed plots) after two growing seasons do not differ significantly. In the current design, pesticide behavior in landfarmed plots are compared to pesticide behavior in freshly sprayed plots, which serve as positive controls.
(2) Contaminant concentrations and toxicity in leachates and runoff water are at levels not significantly different from or are even lower than concentrations and toxicity in the controls;
(3) Crop phytotoxicity is not greater than expected from conventional sprays;
(4) Residues of pesticides in crops should not violate established U.S. EPA
tolerances.
Experimental Design--Landfarming Experiment
The experiment consisted of three main effects treatments arranged in a completely randomized design. The treatments were herbicide-contaminated waste soil (landfarmed plots), herbicide sprays (freshly sprayed plots), and no pesticide application (checks). Contaminated soil and herbicide sprays were applied at three rates of application based on the most prevalent pesticide in the waste soil. Each combination of pesticide treatment and rate was replicated four times. The checks were replicated six tidies; three of the checks were hand-weeded, and three were left unweeded.
Plot Design
Replicates of the pesticide and no-pesticide treatments were randomly assigned to one of 30 experimental plots at the University of Illinois (UI) Cruse Farm. The field encompassed about 0.6 ha on a 3-5% sloped gradient. The soil was classified as a Catlin silt loam (Fine-silty, mixed, mesic Typic Argiudoll). Each plot comprised an area 10 m long x 3 m wide with the length of the plot oriented up-and-down the slope (Figure 1).
The down-gradient end of each of two plots was fitted with metal troughs that directed surface runoff into a 208-L polyethylene barrel housed in a 2.4 m x 2.4 m x 1.2 m pit. A second 208-L barrel sat a slightly lower elevation to the primary collection barrel and received runoff overflow through a 1:9 flow splitter. The system was designed to receive all the runoff from a 15-year frequency storm. Surface runoff from individual plots was contained by delineating each plot with soil berms along the length and by a metal barrier along the up-gradient end.
Two porous-cup soil-water samplers were placed along the midline of each plot at a distance of 3.7 m from each end. The samplers were installed at a 45° angle to the horizontal so that the ceramic sampling cup was at a perpendicular distance of 60 cm from the surface. Pressure-vacuum sampling tubing issuing from the samplers was buried 40 cm below the soil surface and extended beyond the soil berg. Results of soil-water sampling are not discussed further in this paper.
Herbicide-Contaminated Soil
During April 1990, water used to fight a fire at a pesticide warehouse in Lexington, IL, flooded the soil surrounding the building and deposited high concentrations of trifluralin and lesser concentrations of atrazine, alachlor, and metolachlor. The soil was excavated and stored until August 1990 at a farm where it was eventually disposed of by landfarming (Felsot 1991; Bicki and Felsot 1993). Analysis of 18 individual cores (5 cm diam. x 10 cm deep) just prior to landfarming showed that trifluralin was the primary constituent (Bicki and Felsot 1993). Approximately 49 Mg of this waste soil was transported to the UI Cruse Farm during August. The soil was covered with a black plastic sheet and stored through the winter. During March 1991, six cores (5 cm diam. x 10 cm deep) were collected randomly from the pile. The following concentrations were found: 118 ± 58 ppm trifluralin, 18 ± 14 ppm metolachlor, 1 t 1 ppm atrazine, and 1 ± 1 ppm alachlor.
Because we were interested in landfarming high concentrations of a greater diversity of herbicides, a simulated spill of alachlor was created on 7 April 1991 by pouring 9.5-L of Lasso 4E (480 g alachlor/L) in 30-cm deep trenches that were dug into the surface of the pile; the trenches were then filled with soil. On 25 April 1991, 1.4 L of Aatrex 4F (480 g atrazine/L) were spilled on the surface of the pile, and the soil was overturned with a spade. On 26 April 1991, a front loader mixed the pile of contaminated soil by completely overturning it in one direction and then overturning it a second time in a direction perpendicular to the first. The soil was then piled about 0.6-0.9 m high within a 7.6 m x 3.0 m area. On 31 May, 10 cores (5 cm diam. x 15 cm deep) were collected along a diagonal transect laid across the surface of the pile, and 10 cores were also collected from a depth of 30-45 cm. Herbicide concentrations (oven-dry weight basis) determined in individual cores averaged 172 ± 99 ppm alachlor, 99 ± 53 ppm trifluralin, 18 ± 9 ppm metolachlor, and 14 ± 11 ppm. atrazine.
Rates of Application
Target rates of application were based on the concentration of alachlor in the contaminated soil pile. The maximum legal application rate of alachlor is 4.48 kg ai/ha; this rate represented the 1X level. Five and 10 times the maximum rate were also determined. The theoretical rates of application of trifluralin, metolachlor, and atrazine were based on their soil concentrations in proportion to the concentration of alachlor at the 1X, 5X, and IOX levels (Table 1).
Application of Contaminated Soil and Sprays
The weight of soil needed per plot to give an equivalent alachlor application rate of 4.5, 22.4, and 44.8 kg ai/ha was based on the average concentration (172 ppm) determined in May. Thus, on an oven-dry weight equivalent basis, 78, 391, and 782 kg of contaminated soil were needed to produce proportional application rates of 1X, 5X, and 10X. On 6 June 1991 the appropriate amount of soil was loaded into a 1.3-m wide manure spreader from a front-end loader that had been calibrated by weighing empty and full. The comparatively small plot size necessitated delivery of the soil from the manure spreader without the use of the beater blades; the soil dropped off the back of the spreader as the drive chains moved. The soil was applied in one pass and then raked evenly across the entire 3-m width of the plot.
For the freshly sprayed treatments, enough alachlor (Lasso 4E, 480 g/L), trifluralin (Treflan EC, 480 g/L), metolachlor (Dual 4E, 960 g/L), and atrazine (Aatrex 4L, 480 g/L) were mixed together with tap water to give theoretical application rates equivalent to the proportional rates calculated for the landfarmed plots (Table I). On 5 June 1991 the spray was delivered from a tractor-mounted boom calibrated to deliver 336 L of spray per ha.
Plot Preparation and Planting
After application of contaminated soil and herbicide sprays, all plots were disked to a depth of 10 cm in an up-and-down slope direction; the soil surface was smoothed with a rolling bar cultivator. On 7 June 1991 four rows of soybeans (Glycine max L.) were planted in each plot. Pesticidetreated plots were not cultivated throughout the growing season; three of the six check plots were hand-weeded. On 28 May 1992, plots were disked and replanted in soybeans. Once again, three of the six check plots were hand-weeded, and the other plots were not cultivated further.
Sampling of Applied Soil and Sprays
During application of contaminated soil and herbicide sprays, three aluminum pans (29.2 x 19.3 cm) were placed along the vertical midline of each plot to intercept the applied material. The pans were packaged in individual polyethylene bags and returned to the laboratory for analysis. Contaminated soil intercepted on the landfarmed plots was weighed, sieved through an 8-mesh screen (3-mm openings), and stored at -20°C before analysis. A subsample was oven-dried to determine the percentage moisture content. Bags containing the pans with intercepted spray material were frozen immediately.
Immediately after application, six cores (5 cm diam. x 10 cm deep) were randomly collected from each plot (landfarmed and sprayed). The cores were placed in individual polyethylene bags and returned to the laboratory for analysis. Each core was sieved and stored at -20°C.
Water Sampling
Prior to sampling, the depth of water to the bottom of the barrel was measured within 24-48 hours after a runoff event to determine the volume of runoff. The water and sediment were stirred with a paddle, and water was collected for pesticide analysis by submersing two 500-rnL glass bottles into a barrel. An additional two samples of mixed sediment and water was collected by submersing a hlalgene bottle into the barrel; this sample was used to determine sediment concentrations necessary to calculate the weight of eroded soil. Samples for pesticide analysis were returned to the laboratory shortly after collection and stored at 4°C for 48-72 hours before analysis. Prior to extraction, water was separated from sediment by filtration through a glass microfiber filter.
Toxicity Assays
Phytotoxicity of herbicide residues in soil was determined by counting total emerged soybean plants in each plot during July 1991 and 1992. Total number of weeds in each plot were also counted later in July. Toxicity of herbicide residues in runoff water were determined by the algal photosynthetic inhibition bioassay (Ross et al. 1988); because toxicity was not observed, results of algal assays are not present in this paper.
Analytical Methods
Methods for extraction of soil and water and quantitation of herbicide residues by gas chromatography were detailed in Felsot et al. (1992). Limits of detection were 150 ppb in soil and 1 ppb in water. Extraction efficiencies from soil and water were determined by use of fortified blanks, which consisted of soil and runoff water collected from untreated plots. Recovery of all herbicides from soil were greater than 80%. From water fortified at 1 ppb, recoveries were 64% for trifluralin and >85% for atrazine, alachlor, and metolachlor. Data were not corrected for extraction efficiencies.
Experimental Design - Biostimulation Experiment
In separate plots (1 m2) adjacent to the runoff plots, contaminated soil and
sprays were applied at 1X, 5X, and 10X equivalent alachlor rates, similar to
rates of application in the landfarming experiment. These plots were designated
"biostimulation plots." Sewage sludge (25 g/kg soil based on a 7.5
cm depth) and corn meal (50 g/kg soil) were incorporated into the top 5 cm of
selected plots to stimulate the degradation of herbicide residues. Unamended,
pesticide-treated plots were used as positive controls, and amended, no-pesticide
plots were used as controls for analytical background and microbial activity.
Three soil cores (7.5 cm deep x 5 cm diam.) were collected from each plot on
0, 30, 60, and 100 days after application of soil and sprays. Individual cores
were sieved and extracted as described by Felsot et al. (1992). Also, microbial
activity (soil dehydrogenase activity and microbial enumeration) was determined
on triplicate subsamples of bulked cores from each plot (Dzantor and Felsot
1991).
Our previous studies (Felsot et al. 1988, 1990) indicated that landfarming required at least two years to remediate soil containing elevated concentrations of alachlor and metolachlor; concentrations of these contaminants from a theoretical 1X application rate were not significantly different after two years than concentrations of herbicides in untreated soil. Thus, the present landfarming experiment was designed to last a minimum of two years.
Landfarming Experiment
Herbicide Residues Following Application of Contaminated Soil and Spray Mixture
Because a major criterion of feasibility of landfarming was to be determined by comparing herbicide residue behavior in landfarmed and sprayed plots, an accurate assessment of actual material applied and initial concentrations in soil was imperative. The actual rates of herbicides applied were determined by analysis of spray and soil intercepted by pans placed in three locations in the plots and compared to the target rates of application (Table 2). In each case, residues intercepted by pans in sprayed plots were approximately 2-3 times lower than the theoretical application rate. Average residues recovered from soil cores yielded rates of application similar to those calculated from interceptor pans and confirmed the lower than expected application rate of the sprays. An analysis of the actual spray solution indicated that the amount of pesticides added to the tanks were near theoretical amounts; thus, it was concluded that the application of finished spray at a targeted rate of 336 L/ha was not achieved. Indeed, an excessive volume of _spray solution remaining after application supported this hypothesis and suggested that either the sprayer was improperly calibrated and/or the tractor speed was excessive.
After application of contaminated soil, soil collected in the interceptor pans
was weighed and scaled up to a plot area basis (Table 3).
For the 1X and 1OX treatment the actual weight of soil applied per plot was
within 8% of the target rate; the weight of soil applied in the 5X plot was
26%
greater than the target weight. In contrast to residues recovered after spraying,
residues recovered after application of contaminated soil tended to be higher
than the expected amounts (Table 2). The largest discrepancies
(approximately 150-160% of target rates) were observed in the 5X plots as expected
from the weights of soil actually applied.
Comparative Degradation of Herbicide Residues in Landfarmed and Sprayed Plots
Because average initial residues of the herbicides in sprayed plots were significantly below the average residues in the landfarmed plots, degradation was compared by normalizing herbicide residues to a percentage of actual amounts applied and plotting the percentages recovered relative to time after application. Degradation curves for alachlor and trifluralin at the 1X and IOX application levels are discussed in this report (Figure 2). At the 1X rate of application, alachlor had an estimated half-life of 30 days in landfarmed and sprayed plots; however, by the end of the experiment (425 days after application), recovery of alachlor residues in landfarmed plots were slightly greater than in sprayed plots. Similar results were observed for alachlor at the 1OX application level; however, initial rates of degradation in both landfarmed and sprayed plots were much slower than at the 1X rate.
Trifluralin at the 1X rate initially degraded faster in the landfarmed plots (half-life ~30 days) than in the sprayed plots (half-life -125 days). After 425 days, however, percentage recoveries were similar. Trifluralin degraded more slowly when applied at a IOX rate; initial rates of degradation and final percentage recoveries were slightly greater in landfarmed plots than in sprayed plots.
Based on the degradation curves for alachlor and trifluralin, 1X rates of application seem safe for landfarming because compared to sprayed residues, there is little difference in proportion of residues remaining following two growing seasons. At higher rates of application, however, initial rates of degradation are very slow regardless of application method, and residues of trifluralin can be high enough to cause phytotoxicity even after two growing seasons.
Results of Toxicity Assays
During 1991, total number of soybean plants in 1X landfarmed plots was significantly less (p<0.05) than the number of plants in check plots (Table 4). The number of soybean plants in 1X sprayed plots was also lower but not significantly. Phytotoxicity was attributed to the presence of atrazine which is not registered for use on soybeans. The greater inhibition of soybean growth in the 1X landfarmed plots than in the sprayed plots probably resulted from the greater initial concentrations of atrazine (0.39 vs. 0.17 gg/g soil). Inhibition of soybean germination was greatest for 5X and 10X sprayed plots, which did not differ from one another. Activity against weeds was evidenced by the significantly lower plant numbers in all treatments (except 1X landfarmed plots) when compared to the check.
During 1992, soybean plant stand counts were similar in sprayed and landfarmed plots at the 1X application rate, which suggested that plant emergence was not affected by the relatively higher herbicide residues in the landfarmed plots (Figure 3). However, at the higher rates of application, significant phytotoxicity was noted, which was probably due to the relatively high concentrations of atrazine in the plots receiving 5X and IOX rates of application. Herbicidal activity against weeds was evident in all plots but most pronounced in the 5X and IOX plots (note that the handweeded plots were not weeded until after weed counts were made). The extremely low weed counts in the 1OX plots supported the observation that residue levels in these plots degraded much slower than in the 1X plots.
Cumulative Loss of Herbicides in Runoff Water
Losses of alachlor and trifluralin in runoff water during the two growing seasons were cumulated and expressed as a percentage of the initial application rate that was experimentally determined for landfarmed and sprayed plots (Figure 4). Percentage loss of both pesticides was greater from 1X sprayed plots than from 1X landfarmed plots. No significant differences in cumulative percentage runoff was observed for 5X and lox plots. Based on these observations, landfarming does not pose an increased risk of herbicide runoff.
Biostimulation Experiment
In laboratory experiments, corn meal and sewage sludge have been shown to enhance the biodegradation of high concentrations of alachlor (Dzantor and Felsot 1992). To determine the feasibility of using these organic amendments in the field to augment the degradation of landfarmed herbicide residues, corn meal and sewage sludge were added to small plots containing landfarmed and sprayed herbicide residues. Alachlor and trifluralin residues were expressed as a percentage of residues initially recovered immediately after application of contaminated soil, sprays, and organic amendments.
Corn meal and municipal sewage sludge stimulated degradation of alachlor residues derived from contaminated soil (Figure 5, 1X & lox Landfarmed) or sprays (Figure 5, 1X & lox Sprayed). In the 1X plots at 100 days after application, percentages of alachlor recovered from unamended plots were similar to percentages recovered in amended plots; this similarity in recovery suggested that the stimulatory effects of amendments did not significantly affect alachlor degradation beyond two months. Residues recovered from amended lox plots were still lower at 100 days post application than residues recovered from unamended plots.
In contrast to alachlor, trifluralin degradation was generally stimulated only by corn meal (with the exception of the 1X sprayed plots); sewage sludge did not significantly affect the loss of trifluralin (Figure 6). In the 1X landfarmed plot, sludge may have even inhibited trifluralin degradation, perhaps through providing increased sorption regions; this tendency was not as pronounced in the lox landfarmed and sprayed treatments.
In addition to following herbicide concentrations, soil dehydrogenase was measured
in soil collected from each plot. Compared to unamended soil, dehydrogenase
activity was significantly elevated for at least 60 days following corn meal
and sewage sludge amendments (data not shown). Increased dehydrogenase activity
suggested that enhanced degradation of alachlor and trifluralin was associated
with increases in general microbial activity.
Four criteria of feasibility were presented as a measure of the efficacy and safety of landfarming herbicide-contaminated soils. Three of these criteria have been preliminarily assessed in this report. Degradation of the most prevalent herbicide constituents, alachlor and trifluralin, did not significantly differ between landfarmed and sprayed plots treated at a 1X rate of application. However, at the IOX rate of application, degradation was very slow and did not result in nonphytotoxic herbicide residues even after two growing seasons. Results from the biostimulation experiment suggested that degradation of landfarmed herbicide residues can be enhanced by amendment of soil with corn meal. Sewage sludge should not be used because of the possibility of prolonging the persistence of certain herbicides.
Soybean plant stand counts indicated that initial rates of application must be carefully controlled and made as low as feasible to avoid crop phytotoxicity, especially when a mixture of herbicide contaminants are present. Evidence of weed control confirmed that the herbicide residues in landfarmed soil were biologically active. Total herbicide runoff as a percentage of initial application seemed independent of increasing herbicide loads, and sprayed herbicide residues at the 1X treatment rate actually ran off in proportionately greater amounts than residues from landfarmed soil. Herbicide concentrations in runoff water were probably not high enough to cause significant toxicity in the algal photosynthetic inhibition bioassay. Work that remains to be completed for this project includes analysis of herbicide residues in soybeans. Past research, however, has shown that residues of herbicides are infrequently detected and not different at the 1X rate of application (Felsot et al. 1988).
Figure 1. Basic plot design for landfarming experiments at the University of Illinois
Table 1: Theoretical Application Rates (kg ai/ha) of Herbicide Contaminants
Table 2: Targeted and Calculated Rates of Application of Herbicides
Table 3: Dry Weight (kg) of Contaminated Soil Applied Per Plot
Table 4: Phytotoxicity of Sprayed Herbicides and Landfarmed Soil
*1Professional Scientist, Center For Economic Entomology, Illinois Natural History Survey; 2Professor, Department Of Agricultural Engineering, Univ. Of Illinois; 3Environmental Scientist, Ocean Spray Inc., Massachusetts; 4Vice President, Andrews Environmental Engineering, Springfield.
Bicki, T. J.; Felsot, A.S. 1993. Remediation of pesticide-contaminated soil at agrochemical facilities. In Advances in Understanding Mechanisms of Movement of Pesticides Into Groundwater. R. C. Honeycutt, and D. Schabacker, eds. Lewis Publishers, Chelsea, MI (in press).
Dzantor, E. K.; Felsot, A.S. 1991. Microbial responses to large concentrations of herbicides in soil. Environmental Toxicology and Chemistry 10:649-655.
Dzantor, E. K.; Felsot, A.S. 1992. Landfarming and biostimulation for decontaminating herbicide wastes in soil. In Proc. International Workshop on Research in Pesticide Treatment/Disposal/Waste Minimization. T. D. Ferguson (ed.). USEPA Doc. no. 600/S9-91/047 (NTIS PB92-119940/AS). pp. 46-67.
Enlow, P. D. 1990. A Literature Review of Nonbiological Remediation Technologies Whiich May be Applicable to FertilizerlAgrichemical Dealer Sites; TVA Bulletin Y-213; Tennessee Valley Authority National Fertilizer & Environmental Research Center: Muscle Shoals, AL.
Felsot, A. S. 1991. Use of landfarming for remediation of pesticide contamination at agrochemical facilities. Proc. 9th Ann. Hazardous Materials Management ConferencelInternationall; Tower Conference Management Co.: Glen Ellyn, IL. pp 506-515.
Felsot, A. S.; Liebl, R.; Bicki, T; 1988. Feasibility of land application of soils contaminated with pesticide waste as a remediation practice. Final Project Report HWRIC RR-021; Hazardous Waste Research and Information Center: Savoy, IL. 55 pp.
Felsot, A.; Dzantor, E. K.; Case,L.; Liebl, R. 1990. Assessment of problems associated with landfilling or land application of pesticide waste and feasibility of cleanup by microbiologial degradation. Final Project Report HWRIC RR-053; IL Hazardous Waste Research & Information Center: Champaign, IL. 68 pp.
Habecker, M. A. 1989. Environmental contamination at Wisconsin pesticide mixing/loading facilites: case study, investigation and remedial action evaluation. Wisconsin Department of Agriculture, Trade, and Consumer Protection Agricultural Resource Division, Madison, WI. 80 pp.
Loehr, R., M.; Overcash, M. R. 1985. Land treatment of wastes: concepts and general design. J. Environ. Engineering 111:141-159.
Long, T. 1989. Groundwater contamination in the vicinity of agrichemical mixing and loading facilities.Proc. Illinois Agricultural Pesticides Conf. '89. Coop. Ext. Serv.Univ. of Ill., Urbana-Champaign, IL. pp. 139-149.
Norwood, V. M.; Randolph, M. E. 1990. A Literature Review of Biological Treatment and Bioremediation Technologies Which May Be Applicable at Fertilizer/Agrichemical Dealer Sites.; TVA Bulletin Y-215; Tennessee Valley Authority National Fertilizer & Environmental Research Center: Muscle Shoals, AL. 34 pp.
Ross, P. E.; Jarry, V.; Sloterdijk, H. 1988. A rapid bioassay using the green alga Selenastrum capricornutum to screen for toxicity in St. Lawrence River sediment elutriates.In ASTM STP 988; Cairns, Jr.; Pratt, J. R., Eds.; American Society for Testing and Materials: Philadelphia, PA. pp 68-73.
Soil Science Society of America. 1986. Utilization, treatment, and disposal of waste on land. Soil Science Society of America, Inc., Madison, WI.318 pp.
