JEQ Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Stearman, G. K.
Right arrow Articles by Lansford, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Stearman, G. K.
Right arrow Articles by Lansford, S.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Stearman, G. K.
Right arrow Articles by Lansford, S.
Related Collections
Right arrow Wetlands and Aquatic Processes
Right arrow Bioremediation and Biodegradation
Right arrow Agricultural Pesticides
Journal of Environmental Quality 32:1548-1556 (2003)
© 2003 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America

TECHNICAL REPORTS
Wetlands and Aquatic Processes

Pesticide Removal from Container Nursery Runoff in Constructed Wetland Cells

G. Kim Stearman*,a, Dennis B. Georgea, Kris Carlsonb and Stacey Lansfordc

a Center for the Management, Utilization, and Protection of Water Resources, Tennessee Technological Univ., Dixie Ave., Box 5033, Cookeville, TN 38505-0001
b CSC Engineers, 218C E. Tremont Avenue, Charlotte, NC 28203-5364
c Tennessee Technological Univ., Box 14638, Cookeville, TN 38505-0001

* Corresponding author (gkstearman{at}tntech.edu)

Received for publication June 28, 2002.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUMMARY AND CONCLUSIONS
 REFERENCES
 
The increased use of pesticides by container nurseries demands that practices for removal of these potential contaminants from runoff water be examined. Constructed wetlands may be designed to clean runoff water from agricultural production sites, including container nurseries. This study evaluated 14 constructed wetlands cells (1.2 by 4.9 m or 2.4 by 4.9 m, and 30 or 45 cm deep) that collected pesticide runoff from a 465-m2 gravel bed containerized nursery in Baxter, TN. One-half of the cells were vegetated with bulrush, Scirpus validus. The cells were loaded at three rates or flows of 0.240, 0.120, and 0.060 m3 d-1. Herbicides—simazine (Princep) [2-chloro-4,6-bis(ethylamino)-s-triazine] and metolachlor (Pennant) [2-chloro-N-(2-ethyl-6-methylphenyl)-N-2-methoxy-1-methylethyl-acetamide]—were applied to the gravel portion of the container nursery at rates of 4.78 and 2.39 kg ha-1, respectively, 9 July 1998, and at rates of 2.39 and 1.19 kg ha-1, respectively, 17 May 1999. Pesticides entering the wetland and wetland cell water samples were analyzed daily to determine pesticide removal. At the slower flow rate, which corresponds to lower mass loading and greater hydraulic retention times (HRTs), a greater percentage of pesticides was removed. During the 2-yr period, cells with plants removed 82.4% metolachlor and 77.1% simazine compared with cells without plants, which removed 63.2% metolachlor and 64.3% simazine. At the lowest flow rate and mass loading, wetland cells removed 90.2% metolachlor and 83% simazine. Gravel subsurface flow constructed wetlands removed most of the pesticides in runoff water with the greatest removal occurring at lower flow rates in vegetated cells.

Abbreviations: EIA, enzyme immunoassay analysis • GC, gas chromatography • HRTs, hydraulic retention times • SAS, Statistical Analysis System • SF, subsurface flow • TN, Tennessee


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUMMARY AND CONCLUSIONS
 REFERENCES
 
CONTAINER NURSERIES apply pesticides and nutrients at various times throughout the year. Overhead irrigation systems are commonly used to water the plants daily. As much as 70 to 75% of this irrigation water runs off the packed gravel beds that the container plants rest on (Cabrera, 1997; Beeson and Knox, 1991). This runoff may have significant concentrations of pesticides and nutrients (Gilliam et al., 1992). Several researchers have reported runoff of a variety of pesticides from container nurseries (Keese et al., 1994; Briggs et al., 1998, 1999; Wilson et al., 1995). Mahnken et al. (1999) showed that the herbicides simazine and metolachlor were present in runoff from container nurseries. This may cause pollution of receiving water bodies or damage plants as water is reused. It is important to determine pesticide and nutrient runoff from containerized nurseries and to evaluate treatment methods to contain, reduce, or eliminate these potential contaminants.

Constructed wetlands are a treatment method that contain and remove runoff chemicals by various processes including microbial degradation, plant uptake, sorption, chemical reactions, and volatilization. Constructed wetlands, using a subsurface flow (SF) gravel media, have recently been used to clean wastewater, primarily for N and P (Shannon et al., 2000; Tanner et al., 1998).

Although a limited number of studies examining pesticides have been conducted on SF gravel wetlands, there are several studies reporting pesticide removal in relatively large surface flow wetlands. In a 44-ha vegetated surface flow constructed wetland in South Africa, macrophytes and their associated root dwelling microorganisms were responsible for removing 77 and 93% of influent (0.85 µg L-1) azinphos-methyl (O,OdiethylS-[(4-oxo-1,2,3-benzotriazin-3(4H)-yl) methyl]phosphorodithioate) (an orchard pesticide) and removing most of inlet chlorpyriphos (O,O-diethyl O-(3,5,6-trichloro-2-pyridyl) phosphorothioate) and endosulfan (6,7,8,9,10,10-hexachloro-1,5,5a,6,9,9a-hexahydro-6,9-methano-2,4,3-benzodioxathiepin-3-oxide) at inlet concentrations of 0.02 and 0.2 µg L-1, respectively (Schulz and Peall, 2001). Kadlec and Hey (1994) reported removal of 25 to 95% atrazine (6-chloro-N-ethyl-N'-isopropyl-1,3,5-triazine-2,4-diamine) in reconstructed river wetlands in Des Plains, IL. Detenbeck et al. (1996) reported an atrazine half-life of 8 to 14 d in a 230 m long flow-through wetland.

In a study using a gravel recirculating vertical SF system, McKinlay and Kasperek (1999) concluded that microbial degradation was the dominant process for atrazine decomposition rather than plant uptake. This is the only study (McKinlay and Kasperek, 1999) the authors found that examined pesticide removal in an SF constructed wetland. Tanner et al. (1998) examined maturation of an SF constructed wetland and removal of nitrate and phosphate. As SF gravel constructed wetlands matured, chemical removal changed, especially for phosphate, where sorption was the primary means for removal. As wetland sorption sites were saturated, phosphate removal was reduced (Tanner et al., 1998). Shannon et al. (2000) showed removal of N to increase the second year and removal of phosphate to decrease the second year in an SF constructed wetland system. They attributed increased N removal to increased plant density and decreased phosphate removal to partial saturation of sorption sites.

Several studies have reported atrazine (a triazine herbicide similar in structure to simazine) degradation under anaerobic conditions that may be similar to those found in a gravel constructed wetland. Seybold et al. (2001) reported that in anaerobic soils the half-life was 38 d for atrazine and 62 d for metolachlor. In the aqueous phase above the soil, the half-life was 86 d for atrazine and 40 d for metolachlor. Other researchers have reported differing results for atrazine degradation under anaerobic conditions in wetland soil. Chung et al. (1995) showed that 50% of atrazine degraded in 38 wk under anaerobic conditions. Gu et al. (1992) showed no degradation of atrazine under methanogenic conditions. Delaune et al. (1997) reported slower degradation of atrazine under anaerobic vs. aerobic conditions. Goswami and Green (1971) indicated that atrazine degradation would generally be slower under anaerobic compared with aerobic conditions with sediment, while Kearney et al. (1967) showed that atrazine disappeared more rapidly under anaerobic conditions compared with aerobic conditions. Atrazine fate in sterile and unsterile soil was studied by Kruger et al. (1997), who reported faster atrazine degradation in the surface soil than in the subsoil, but this may be due to increased organic matter rather than increased oxidation. Larsen et al. (2001) showed that atrazine in a wetland soil did not degrade anaerobically after addition of various electron acceptors including nitrate, sulfate, and carbon dioxide.

Metolachlor degradation rates have been reported on a limited basis in anaerobic and strongly reducing conditions (Seybold et al., 2001). Metolachlor was degraded at a greater rate under sulfate reducing conditions than without S present, indicating that possibly sulfate-reducing microorganisms degrade metolachlor (Stamper et al., 1997; Stamper and Tuovinen, 1998).

It is important to determine the required residence time of wastewater in constructed wetlands to adequately reduce the chemicals to nontoxic levels. Some of the removal processes may require hours or days to occur. Since several pesticide removal processes including microbial degradation, chemical hydrolysis/dechlorination, sorption, and plant uptake may occur simultaneously, it is difficult to isolate a single factor that controls pollutant removal in a wetland. Factors that may be important and can be controlled by the design and management of the wetlands include the size and volume of the wetland, the mass loading or flow to the wetlands, and the plant density and composition.

This study examines 14 wetland cells as to the effect of flow (pesticide loading), HRT, presence of plants, depth, and surface area of wetlands on pesticide removal. The objectives of this study were to: (i) determine removal of metolachlor and simazine from container nursery runoff in constructed wetland cells, and (ii) determine the effect of bulrush vegetation, flow, depth, and surface area of constructed wetlands on simazine and metolachlor removal during a 2-yr period from 1998 to 1999.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 SUMMARY AND CONCLUSIONS
 REFERENCES
 
System Design
Subsurface Flow Wetland Pilot System.
In 1992, a pilot-scale SF wetland system consisting of 14 SF cells made from reinforced fiberglass was constructed adjacent to the municipal wastewater treatment plant (WWTP) at Baxter, TN (Fig. 1) . Twelve cells were 4.9 m long by 1.2 m wide, resulting in a 4:1 aspect ratio. Two cells (G and N) were 2.4 m long by 4.9 m wide, which is a 1:2 aspect ratio. Each of the cells was installed at a bottom slope of approximately 1% from head end to tail end. Rock media was quartz gravel and was divided into two diameter classes. The diameter of the larger rock was <=3.8 cm in diameter with an effective size of 1.91 cm, and the smaller rock was <=2.2 cm in diameter with an effective size of 1.11 cm. Cells A, B, C, D, E, F, and G contained 20 cm of the larger gravel overlaid by 10 cm of the smaller gravel. Cells H, I, J, K, L, M, and N contained 30 cm of the larger size gravel overlaid by 15 cm of the smaller gravel (Fig. 2) . The purpose of the smaller gravel was to provide firmer bedding material for the plants.



View larger version (53K):
[in this window]
[in a new window]
  		Fig. 1. Subsurface flow (SF) wetland system.

 


View larger version (30K):
[in this window]
[in a new window]
  		Fig. 2. Cross-section of subsurface flow (SF) wetland cell.

 
Previous field studies at the Baxter wetland have examined the effects of vegetation, media depth, aspect ratio, temperature, drawdown, recycle, and hydraulic loading rate on the treatment of primary treated municipal wastewater (Kemp and George, 1997). Initially, all cells were planted with the common bulrush, Scirpus validus, which was obtained from a nearby lake and wetland. To better assess plant effects on N transformation, plants and root mass were removed from Cells B, D, I, and K (Kemp and George, 1997). For the current study, bulrush (plants and root mass) was removed from Cells F, G, and M. The roots and rhizomes of stray bulrush and weeds in the nonvegetated cells were removed biweekly. The average plant density in the vegetative cells during the study was 600 stems m-2.

Container Nursery
A pilot container nursery was constructed on a 35 by 20 m site adjacent to the existing SF wetland system described in the previous section (Fig. 1). Two hundred each of 5-L containerized landscape plants (Monkey Grass, Spiraea, and Japanese Holly) spaced approximately 50 cm apart were grown on this container nursery area. Each plant was potted in a pine bark–based media and fertilized with an 18–6–12 (N–P–K) slow-release fertilizer before the study.

To facilitate site drainage, the nursery was sloped approximately 1 to 3%. The soil surface was covered with a 0.5-mm thick impermeable polyethylene liner covered with an approximately 15-cm layer of 0.65-cm limestone gravel.

The container bed was irrigated by six overhead rotary impact sprinklers. Manually cleaned, inline filters with a mesh size of 500 cm-1 were installed on each overhead riser to prevent sprinkler clogging. A 745.7-W Hydromatic submersible pump located in the wastewater plant's dechlorination tank provided water to irrigate the nursery. Sulfur dioxide was used to dechlorinate the wastewater treatment plant effluent. The water was transported through 5-cm polyvinyl chloride (PVC) piping throughout the nursery site.

Soil berms were placed around the nursery to prevent intrusion of runoff from outside the nursery area and to ensure that all runoff from the nursery site was collected. Perforated piping was distributed along the lowest end of the site and covered with gravel to transport percolated water to a partially buried 1800-L polyethylene tank. This collected water was pumped to an 8.5-m3 holding tank by a float-controlled, 1491.4-W Hydromatic submersible sump pump. A Shur-Dri 1491.4-W timer-controlled pump subsequently transferred water from the holding tank to the wetland system. The nursery runoff water was distributed by PVC pipe either across the width of the cell or by a single orifice inlet. Lines were flushed for 5 to 10 s each day to prevent clogging. The water level in each cell was controlled by the discharge standpipe elevation (Fig. 2).

Herbicide Application
To simulate a typical weed-management procedure, simazine (Princep) and metolachlor (Pennant) were applied to the gravel surface using an air-pressurized backpack sprayer on 9 July 1998 and 19 May 1999. The liquid Princep was formulated by the Novartis Crop Protection Corporation at 89% pure simazine, 1% related triazines, and 10% inert ingredients. The liquid Pennant product was formulated by the Ciba-Geigy Corporation at 85.1% pure metolachlor and 14.9% inert ingredients. The liquid pesticide formulations were measured using a graduated cylinder and thoroughly mixed with water in a 20-L bucket before pouring the pesticide solution in the backpack sprayer. Princep was applied at 4.78 kg ha-1 (220 g) and Pennant was applied at 2.39 kg ha-1 (110 g) in 1998 and at one-half these rates in 1999.

System Operation
The container nursery was irrigated from 0700 to 0900 h each morning. Irrigation was automatically controlled by an Intermatic 24-h dial time switch (Model T104), which activated the submersible pump in the dechlorination tank of the water treatment plant. Once the water reached the container nursery, a Toro Vision II Plus Series Controller (Model V2-PO6) operated the solenoid valves, allowing flow to specific sprinklers. The average irrigation rate was 0.7 cm h-1, producing approximately 6.05 m3 of nursery runoff each day.

Nursery runoff was applied to all 14 wetland cells using peristaltic pumps with microprocessor pump controllers. The pumps were started manually and operated for 2 h at the same time each day (0900–1100 h) at flow rates of 0.120, 0.06, and 0.03 m3 h-1. Total water entering cells daily was 0.240, 0.120, and 0.06 m3. The pumps were calibrated weekly during pesticide sampling by volumetrically measuring the quantity of influent into a cell in a 1-min period and comparing this measurement to the flow rate setting on the pump controllers. If necessary, the pump controllers were adjusted accordingly. The flow rates, physical dimensions, and theoretical retention time for each SF cell are listed in Table 1 . Evapotranspiration of cells was measured several times over the course of the experiments and did not vary significantly over the 1-d sampling time. Therefore, evapotranspiration of cells was considered constant and not significantly variable over sampling dates.


View this table:
[in this window]
[in a new window]
  		Table 1. Subsurface flow (SF) wetland cell characteristics.

 
Water Quality Measurements
Herbicide Analysis.
Effluent water samples were collected from each cell at the effluent standpipe using a 60-mL plastic syringe with a 1-mm orifice at the same time each day (0900 h). The syringes were prerinsed with cell effluent before taking samples and a separate syringe was used for each of the 14 cells. The samples for the enzyme immunoassay analysis (EIA) of simazine and metolachlor (Stearman and Adams, 1992) were immediately transferred from the syringe to individual amber-colored 15-mL glass vials with Teflon lids. Starting on 9 July 1998 and 19 May 1999 these samples were collected every 24 h for 10 d after application, then approximately every 48 to 72 h until the herbicide concentrations were near preapplication levels (18 Sept. 1998 and 25 June 1999).

Effluent water samples were analyzed by EIA for simazine and metolachlor. Enzyme immunoassay analysis of selected pesticides and herbicides has been shown to be a practical and cost-effective method (Stearman et al., 1997). Goh et al. (1992) and Leavitt et al. (1991) reported that the EIA of atrazine corresponded analogously with gas or liquid chromatography analysis.

Each of the simazine and metolachlor EIA kits consisted of 96 antibody-coated wells in a microtiter plate and solutions or reagents and standards (Strategic Diagnostics, Newark, DE). Eight standards including a blank were prepared in the same matrix as the collected water samples and were analyzed on the microtiter plate in duplicate. Simazine standard solutions were made over a range of 0.20 to 20 µg L-1 and metolachlor standards ranged from 0.25 to 20 µg L-1.

To provide an accuracy check for the EIA procedure, selected cell effluent samples were analyzed for simazine and metolachlor using gas chromatography (GC) (Method 505, Method 507, U.S. Army Corps of Engineers, 1989). A single 1000-mL sample was obtained from a specific cell during each EIA sampling day starting with Cell A. The sample was collected from the cell effluent standpipe using a prerinsed 60-mL plastic syringe with a 1-mm orifice and transferred immediately to a 1000-mL amber-colored glass jar with a Teflon lid and labeled as previously described. Figures 3 and 4 show the best fit line for the comparison between EIA and GC analysis. For both simazine and metolachlor, EIA and GC were highly correlated. In all cases, EIA results were used to compute pesticide input and pesticide remaining in the wetland cells. Once all cells were sampled (i.e., the end of the 14th day of sampling), the rotation repeated back to Cell A and was continued until the study ceased.



View larger version (13K):
[in this window]
[in a new window]
  		Fig. 3. Comparison of enzyme immunoassay analysis (EIA) and gas chromatography (GC) analysis of simazine for the years 1998 and 1999.

 


View larger version (10K):
[in this window]
[in a new window]
  		Fig. 4. Comparison of enzyme immunoassay analysis (EIA) and gas chromatography (GC) analysis of metolachlor for the years 1998 and 1999.

 
In addition to EIA samples taken from the effluent of each cell, daily water samples were obtained from the container nursery at the sprinkler heads, from the nursery effluent collection gutter, and in the nursery sump. Daily EIA water samples were also collected from the holding tank discharge, which was used to compute pesticide concentration entering the wetland cells. The water samples from the container nursery locations were obtained by the grab-method using a 15-mL vial. A Teflon bailer was used to sample water from the holding tank. Samples were transferred from the bailer to amber-colored 15-mL vials and labeled as previously specified. The samples were placed on ice and transported to the Environmental Quality Laboratory at the Center for the Management, Utilization, and Protection of Water Resources in Cookeville, TN, for EIA analysis. All samples were obtained and analyzed at the same frequency as those from the cell effluents. Irrigation water was analyzed by EIA during the first week of sampling to check the Baxter Wastewater Treatment Plant effluent for possible background levels of simazine and metolachlor.

Organic Carbon Measurement
Organic C from each wetland cell was measured gravimetrically by dry combustion (Davies, 1974). Each wetland cell had eight removable black plastic mesh containers that were removed for sampling the gravel media. After sampling, new mesh containers were filled with gravel and immediately installed to replace the sampled mesh containers in the wetland cells. The sediment was removed from the gravel by scraping and washing the gravel. The resulting sediment solution was dried at 105°C, weighed, and then combusted at 400°C for 4 h and weighed to compute (by difference) the mass of CO2 combusted.

Field Measurements
Probe well clusters, located at one-third intervals in each cell, were installed during the initial wetland construction (Fig. 2). The wells were constructed of PVC pipe and allowed internal sampling at various depths in each wetland cell. Rubber inflatable bladders were inserted in each well to prevent atmospheric oxygen from re-aerating the cell and to allow fresh water to be available for each probe measurement. The bladders were kept inflated until immediately before obtaining measurements, when they were deflated and removed from the well clusters. Dissolved oxygen (DO), water temperature, and pH sensor probes were lowered into the PVC wells until partially submerged in the cell bulk liquid. The meters were allowed to stabilize for approximately 30 to 45 s before a reading was recorded. Probes were rinsed with distilled water before each measurement.

The DO and water temperature measurements were obtained using a Yellow Springs Instrument (YSI) (Yellow Springs, OH) Model 59 Dissolved Oxygen Meter with sealed sampling probe. The pH was measured by an Orion pH meter. Meters were calibrated both before and after sampling. The DO meter was calibrated using the Winkler-Azide method (Method 360.2, USEPA, 1983) while the pH meter was calibrated using premade buffer solutions (Method 150.1, USEPA, 1983).

Data Analyses
Total influent and effluent simazine and metolachlor mass for each cell was computed using the following equation:

[1]
where TMj is total mass for herbicide j; Cj,i is concentration of herbicide j at time i; Qj,i is hydraulic flow rate transporting herbicide j at time i; and {Delta}tj,i is 1-d time increment. A general linear model (GLM) (Neter et al., 1990) of the means of the dependent variables simazine and metolachlor was conducted to determine which independent variables (i.e., media depth, presence of vegetation, hydraulic loading or flow, and surface area or aspect ratio) significantly affected herbicide total effluent mass. This method was used to take into account the unbalanced design and unequal subclass distributions within the SF system. The p values were considered significant when {alpha} < 0.05.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 SUMMARY AND CONCLUSIONS
 REFERENCES
 
Field Measurements
The mean water temperature during the study sampling periods was 23.7 (±0.55) °C. Although mean DO concentrations measured in the bulk solution were normally <0.5 mg L-1, oxygen was transferred from the roots of bulrush to the rhizosphere (Reddy et al., 1989). Consequently, the DO in a zone extending only a few millimeters from the root may be higher than measured in the bulk solution (Allen et al., 2002).

The pH in the SF cells ranged from 6.41 to 7.45, which should be favorable for microbial activity. The cells with plants had pH values that were generally 0.3 units lower than cells without plants.

The percentage of the total influent simazine and metolachlor removed by each constructed wetland cell for the 2 yr is shown in Table 2 . Simazine removal was above 90% of that applied only in vegetated cells with HRTs at or above 13.3 d. Metolachlor removal was above 86% in cells with HRTs at or above 5.1 d.


View this table:
[in this window]
[in a new window]
  		Table 2. Percentage metolachlor/simazine removed in constructed wetland cells.

 
All of the nonvegetated cells removed <80% of the simazine applied while exactly half of the vegetated cells removed >80% of the simazine applied. Cell F removed 78% of applied simazine in both years, which was the highest percent removal of the nonvegetated cells. Cell F also removed the highest percentage of applied metolachlor (mean 91.5%) of the nonvegetated cells. Cell F had an 8.3 d retention time. Metolachlor removal was greater than 80% in two of the nonvegetated cells (F and M) and in all but two of the vegetated cells (A and H). Organic C ranged from 0.682 g L-1 in Cell A to 0.003 g L-1 in Cell I (Table 3) . The source of organic C was primarily from root residues and dead microbiota.


View this table:
[in this window]
[in a new window]
  		Table 3. Organic C sediment and herbicide loading and removal in constructed wetland cells.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
SUMMARY AND CONCLUSIONS
 REFERENCES
 
All of the cells removed at least 50% of the simazine applied while all but two of the cells, B and I (nonvegetated cells at 2 and 3 HRTs, respectively), removed at least 50% of the metolachlor. Sorption of simazine and metolachlor is a relatively rapid process, requiring only minutes to hours to occur (Talbert and Fletchall, 1965; Kookana et al., 1992; Bouchard et al., 1982; Pusino et al., 1992). The 50% simazine removed by all cells may likely be that simazine readily adsorbed onto the organic sediment and biofilms of the gravel media. This sorption process would be expected to occur in the first 48 h (the shortest HRT) after simazine entered the wetland cell (Kookana et al., 1992; Talbert and Fletchall, 1965). Some of the herbicide removal processes, such as biodegradation, may require more than several days to remove most of the herbicides. Once herbicides are sorbed, desorption may occur. The wetland cells would have some herbicides desorbed although this was thought to be consistent and relatively low. There has been no known research on release of pesticides previously sorbed in submerged wetlands. The ratio of adsorbed atrazine and metolachlor on aged samples from fields 2 to 15 mo after their last application to the normal adsorption constant (Kd) was 2.3 to 42 times higher, indicating that these pesticides would be less likely to desorb as they aged (Pignatello and Huang, 1991).

The nonvegetated cells were more erratic (based on difference between years) in the percent herbicide removed while the vegetated cells had a more uniform removal of herbicides. Possibly, the plant roots contribute to a more uniform environment for microbial degradation. In general, as HRT increased the percent herbicide removed increased (Fig. 5 and 6) . At retention times of 5.1 d, all vegetated cells removed at least 76% of the herbicides. At the high flow rates (high mass loadings) and low retention times, about 60% of the simazine and metolachlor was removed except in Cells B and I for metolachlor, which had 17.9 and 38.8% removal of metolachlor, respectively (Table 2).



View larger version (10K):
[in this window]
[in a new window]
  		Fig. 5. Metolachlor removal in constructed wetland cells with plants.

 


View larger version (10K):
[in this window]
[in a new window]
  		Fig. 6. Simazine removal in constructed wetland cells with plants.

 
There was no correlation between total organic C and pesticide removal, although the vegetated cells usually had more organic C and more percent herbicide removed. Organic C has been shown to be correlated with microbial biomass, which influences biodegradation of herbicides. Also, organic C provides sorption sites for the herbicide removal (Stevenson, 1982). Table 3 also shows side by side the paired vegetated and nonvegetated cells that have the same dimensions and flow rates. Comparing Cells A and B, then Cells C and D, and so forth, cells with the same depth and flow rate can be compared with and without vegetation (Table 3). In Table 3, the percent of applied herbicides removed by each wetland cell (combining simazine and metolachlor) can be examined and compared for each cell treated with the same flow and depth. In general, the vegetated cells removed 10 to almost 30% more herbicide than the nonvegetated cells at the same depth and flow rate.

Tables 4 and 5 show the statistical results for years 1998 and 1999 for metolachlor and simazine, respectively. Herbicides were removed at greater percentages at low flow rates compared with high flow rates. As flow increased, mass loading increased, retention time decreased, and percent herbicide removed decreased. Conversely, at the lowest flow rates (highest HRTs) a higher percentage of herbicide was removed, and the HRT was increased allowing more time for removal processes including microbial degradation to occur. On average, constructed wetland cells with plants removed 20.6% more simazine than cells without plants. Planted cells removed 19.2% more metolachlor than nonvegetated cells. Cells with 45 cm depth removed more metolachlor than cells with 30 cm depth, but there was no significant difference for simazine removal with depth of cell. Since flow and depth determine hydraulic residence time, metolachlor removal was highly dependent on residence time while simazine was also to a lesser degree dependent on residence time since flow was significant and depth was not significant. Cells with larger surface area (aspect) were not different in the percent of herbicide removed compared with the cells with less surface area.


View this table:
[in this window]
[in a new window]
  		Table 4. Metolachlor % removal in constructed wetland cells and significant factors (1998, 1999).

 

View this table:
[in this window]
[in a new window]
  		Table 5. Simazine % removal in constructed wetland cells and significant factors (1998, 1999).

 
Figures 5 and 6 show results from the vegetated cells and show the percentage of metolachlor and simazine that was removed in the cells as related to HRT. In Fig. 5, percent removal of metolachlor by the wetland cells increases as the HRTs increased from 2 to 5 d and then remained relatively steady through HRTs of 20 d. In Fig. 6, percent removal of simazine by the wetland cells steadily increased as the HRTs increased from 2 to 13 d. In the wetland cells that have HRTs from 5 to 20 d, metolachlor removal was relatively unchanged. Therefore, there was no benefit to increasing HRT beyond 5 d for metolachlor or 13 d for simazine with the vegetated wetland cells under the conditions studied.

The mass of herbicide applied in 1998 was twice that applied in 1999. Therefore, if mass loading determined pesticide removal, then there should have been differences in percent removal from 1998 to 1999 when examining percent removal for an individual cell. The percent removals shown in Table 2 were not usually different from year to year except in a few cells. Also, there was no difference for pesticide removed between years (Tables 4 and 5). Therefore, mass loading did not appear to be as important as HRT in determining pesticide removal. The cells that had the highest flow rates also had the highest mass loadings and removed the most mass of herbicides but the lowest percent of herbicides. Results show that time is more important than mass loading to achieve removals of 79.3% or more, which were achieved at HRTs above 7.9 d for vegetated cells for both herbicides (Table 2). Also when comparing herbicide removal during the 2-yr sampling period, the pesticide mass entering the cells was much greater initially (immediately after application) and then decreased each week. There was little difference in the daily amounts removed by the cells regardless of high or low mass loading (daily pesticide removal is not shown). This also supports the hypothesis that time was more important than mass loading when examining the percent of pesticide removed. In other words a 2-d retention time may not be enough to remove much more than 60% of the herbicides, no matter what the mass loading.

Flow and volume of cell determine HRT. The HRTs of 2 or 3 d were not long enough for some pesticide removal processes to significantly occur, such as microbial degradation (McCormick and Hiltbold, 1966; Erickson and Lee, 1989; Delaune et al., 1997; Stamper and Tuovinen, 1998). Flow was also equal to mass loading for a particular year, but not between years since pesticide application rates were reduced to half of the first year application rates. The pesticide concentration, combined with flow (i.e., mass loading), should indicate what the system can handle. In this study, even at high mass loadings (1998) the pesticide removals were similar to the subsequent year's removal. By comparing pesticide removal in the vegetated Cells A and H (2 and 3 d HRTs, respectively), which had the same flow (mass loading) and only differed in depth, effects of both depth and HRT could be determined. There was no significant difference between percent removal of herbicides for Cells A and H.

Depth was significant for metolachlor, but not for simazine (Tables 4 and 5). Comparing the pesticide removals in Table 3 shows that the extra 15 cm depth did not improve removals except for a few cells; notably Cell B had only 40.8% removal while the deeper Cell I had 51.6% removal for the combined pesticides. Cells B and I were nonvegetated cells with high flows.

Wetland cells with plants have higher pesticide removals for a number of reasons, including plant uptake, oxygenated rhizosphere zones that enhance aerobic degradation, and increased microbial activity from plant root exudate stimulation.


   SUMMARY AND CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUMMARY AND CONCLUSIONS
REFERENCES
 
The wetland cells are complex systems where many variables including flow, HRT, depth, surface area, organic matter, and presence of plants may influence and control pesticide removal. The presence of plants and higher HRTs (determined by flow, depth, and surface area of cells) significantly affected pesticide removal in wetland cells in this study. Although microbial processes were not studied in this project, future research concentrating on frequency of pesticide applications would be beneficial since removal of pesticides by microbial action is reported to be partially affected by repeated pesticide applications (McKinlay and Kasperek, 1999). Microbial degradation of pesticides may be the rate-limiting step compared with pesticide sorption.

This 2-yr study found that metolachlor and simazine removal in gravel SF wetlands was significantly improved at lower flow rates with vegetated cells, compared with higher flow rates and nonvegetated cells. The vegetated cell (Cell A) with 2.3-d HRTs removed 62.0% of herbicide applied. The nonvegetated paired Cell B at 2.1-d HRTs removed 40.8% of herbicide applied. The vegetated cell (Cell C) with 5.1-d HRTs removed 82.2% of herbicide applied. The nonvegetated paired Cell D at 4.1-d HRTs removed 64.0% of herbicide applied. Subsurface flow vegetated gravel constructed wetlands can be designed to remove more than 60% of pesticides, metolachlor and simazine.


   ACKNOWLEDGMENTS
 
Funding for this research was provided in part by the Tennessee Department of Agriculture through the USEPA Region 4 (Contract ID no. 98-06661-00), 319 Nonpoint Source Pollution Program. These results have not been subjected to the agency's peer and administrative review. Special thanks to Amy Knox for editing and formatting the final manuscript.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUMMARY AND CONCLUSIONS
 REFERENCES
 


Related articles in JEQ:

This Issue in Journal of Environmental Quality

JEQ 2003 32: 1167-1172. [Full Text]  



This article has been cited by other articles:


Home page
 J. Environ. Qual.Home page
S. Xu and P. R. Jaffe
Effects of Plants on the Removal of Hexavalent Chromium in Wetland Sediments
J. Environ. Qual., January 5, 2006; 35(1): 334 - 341.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Stearman, G. K.
Right arrow Articles by Lansford, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Stearman, G. K.
Right arrow Articles by Lansford, S.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Stearman, G. K.
Right arrow Articles by Lansford, S.
Related Collections
Right arrow Wetlands and Aquatic Processes
Right arrow Bioremediation and Biodegradation
Right arrow Agricultural Pesticides

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Vadose Zone Journal Journal of Plant Registrations
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal