| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Earth Sciences, Dartmouth College, Hanover, NH 03755
b Dep. of Biology, Dartmouth College, Hanover, NH 03755
* Corresponding author (Carl.Renshaw{at}Dartmouth.edu)
Received for publication March 15, 2005.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: ICP-MS, inductively coupled plasma mass spectrometer • RUSLE, Revised Universal Soil Loss Equation • SEM, scanning electron microscopy • XRD, X-ray diffraction
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The use of arsenical pesticides was particularly prevalent in apple (Malus sylvestris Mill.) orchards before the invention of 1,1,1-trichloro-2,2-bis(4-chlorophenyl)ethane (DDT) in 1947 (Welch et al., 2000). Many studies have concluded that the redistribution of surface-applied Pb and As within orchard soils is limited; decades after the application of arsenical pesticides, the highest concentrations of Pb and As generally remain in the top 25 cm of contaminated soils (Aten et al., 1980; Benson, 1976; Jones and Hatch, 1937; Peryea and Creger, 1994; Veneman et al., 1983). Arsenic is somewhat more mobile than Pb (Elfving et al., 1994; Merry et al., 1983; Peryea and Creger, 1994), particularly in sandy soils (Veneman et al., 1983). In some cases the mobility of As may be significantly increased by the application of phosphate-rich fertilizers (Johnson and Barnard, 1979; Murphy and Aucott, 1998). Phosphate and As exhibit similar physiochemical behavior in soils and compete directly for specific adsorption sites on soil particles (Woolson et al., 1973).
The limited redistribution of As and Pb implies that these metals are strongly retained in the soil. However, if the metals are retained on fine particles, then the retention of the As and Pb may be impacted by physical erosion (Cooper and Gillespie, 2001), particularly if the upper soil is physically disturbed such as during tilling or land development (Wu et al., 2004). More than 2500 km2 of U.S. farmland are developed each year (USDA, 2003b), suggesting the possibility that land development might mobilize significant amounts of Pb and As in lands where arsenical pesticides were used.
To assess the potential for land disturbance to mobilize arsenical pesticides, we compared the Pb and As budgets of two orchards in southern New Hampshire having similar historical applications of arsenical pesticides during the early 1900s. Trees more than 80 yr old continue to grow in the first orchard, while the original trees planted in the second orchard were removed when the field was tilled and replanted in 1992. A third orchard where lead arsenate was never applied provided a control for our study. We also determined the masses of As and Pb in sediments within streams that drain each field; much of any soil that eroded from these fields would likely be deposited in these channels. Finally, to asses the bioavailability of any Pb and As mobilized by the tilling, we measured metal body burdens in macroinvertebrates inhabiting the wetland ecosystem at the outlet of the stream draining the tilled field.
Site Description
The field site for this study is a commercial orchard located in southern New Hampshire. The orchard is still in production and has been owned and managed by the same family for >80 yr. Active orcharding at the site is indicated on the 1953 U.S. Geological Survey topographic map and the New Hampshire State Blister Rust survey maps from 1936, 1938, and 1960.
Average monthly precipitation at the site is 8.4 cm mo–1 and is approximately uniformly distributed throughout the year. The site receives 
25% of its precipitation in the form of snow. About half the total annual precipitation in this region is lost to evapotranspiration (Randall, 1996). Orchard soils are well drained, shallow to medium depth, gravelly to fine, sandy loam overlying glacial till. Underlying bedrock consists of Silurian metasedimentary shales and schists. The soil is classified as a Hoosic fine gravelly, sandy loam (sandy-skeletal, mixed, mesic typic Dystrudepts).
Three fields within the orchard were selected for sampling:
83% of the field area based on aerial photographs) is regularly mowed during the growing season but is otherwise undisturbed.
|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Stream Channel Sediments
We sampled sediments in the stream adjacent to the undisturbed field. The 30 cm deep sampling pit contained several distinctly stratified, unconsolidated layers. We took 100-g samples from each layer with acid-cleaned implements.
We sampled 10 different locations in the ephemeral stream channel adjacent to the tilled and control fields (Fig. 1). In three of these pits we collected samples (200 g) with acid-cleaned implements in 5-cm increments down to a depth of 30 cm. At the remaining sites we dug 30 cm deep pits and took samples only where the texture or composition of the sediments appeared to change. All samples to be analyzed for Pb and As concentrations were refrigerated, while all bulk density samples were weighed and then placed on clean plastic plates and set aside for air-drying.
Wetland Macroinvertebrates
Two common macroinvertebrate taxa, midge fly larvae (Diptera chironomidae) and dragon fly larvae (Odonata libellulidae) were collected from four sites along a transect across the wetland at the outlet of the ephemeral channel draining the tilled and control fields. As a control, the same taxa were collected from five uncontaminated lakes in New Hampshire and Vermont. At each of the sites, trace metal clean techniques with a 250-µm d-frame net was used to collect organisms by sweeping macrophytes and by dragging along the top 5 cm of sediments (Back et al., 2003; Chen et al., 2000). Net contents were sieved through 250-µm mesh and pooled together into trace metal clean polyethylene buckets with filtered wetland water for transport.
Within 48 h of collection, organisms were sorted live in the lab before digestion for metals analysis. Five replicate samples of libellulids, and two to three replicate samples of chironomids were taken per site. Samples were composites of individual organisms (100 mg dry wt.), consisting of two to three individuals for libellulids and approximately 20 individuals for chironomids, to ensure detectable metal levels.
Analytical Methods
After air-drying, bulk sediment density samples were oven-dried at 100°C for 24 h, allowed to cool, and then weighed. Approximately 100 g of all samples from each pit dug in the stream channel sediments adjacent to each field was then sieved through a 2-mm sieve, weighed, and sealed into round plastic tins for 137Cs analysis by 
-counting on a Canberra Broad-Energy Intrinsic-Ge Detector. Each sediment sample was counted for 10 to 12 h to obtain sufficient counting statistics. The morphology and mineralogy of selected air-dried samples was examined using scanning electron microscopy (SEM) and X-ray diffraction (XRD). These samples were first sieved (<63 µm) to isolate the silt and clay-sized fraction. Samples for SEM analyses were C-coated before analysis.
Selective sequential extractions are commonly used to provide information about the chemical speciation of associated trace metals (e.g., Shuman, 1985; Tessier et al., 1979). Accordingly, sequential extractions were used to identify the phases that retain As and Pb in these soils. For these extractions, all reagents were trace metal grade or better. We note that sequential extractions depend on the specific and complete extraction of a target phase, and are at best operationally defined mineralogical fractions. Despite limitations caused by inefficient or nonspecific extractions, selective sequential extractions provide a tested, albeit approximate, method of quantifying trace metal speciation in soils.
Sequential extractions were performed on samples from pits in the two sites most likely to be impacted by land disturbance; three pits from the tilled field and three pits from the ephemeral stream channel down gradient from the tilled field. In each pit samples were taken from three representative depths (surface, middle, and deepest depths). We used a five-step sequential extraction procedure designed to extract exchangeable ions, carbonate phases, organic matter, acid-soluble oxides and sulfides, and crystalline oxides. Exchangeable ions were first extracted by mixing 4.5 g of soil with 20 mL of 1 M magnesium chloride (MgCl2) at a pH of 7. Samples were shaken for 1 h and then centrifuged for 15 min and 8 mL of supernatant fluids extracted, filtered with 0.2-µm filters, acidified with two drops concentrated HCl, and then refrigerated until analyzed for Pb and As using inductively coupled plasma mass spectrometer (ICP-MS). The remaining soil was then cleaned by adding 5 mL of deionized water, vortexing, centrifuging, and decanting.
The carbonate extraction began by adding 30 mL of 1 M sodium acetate (NaOAc) at a pH of 5 to the cleaned soil samples. The samples were then shaken for 5 h, centrifuged, and the supernatant sampled, acidified, and refrigerated for later analyses using ICP-MS. The remaining soil was then cleaned as described above. Acid-volatile sulfides and amorphous oxides were next extracted using 30 mL of hydrochloric acid (HCl) and shaken for 10 h. The soil was the then sampled and cleaned as before.
The organic matter extraction began by adding 10 mL of 5% sodium hypochlorite (NaOCl) to the soil samples. These samples were vortexed and heated in water baths for 30 min and then centrifuged. After centrifuging, the supernatants were decanted and the above process repeated twice. The resulting supernatant was sampled and analyzed and the soil cleaned as before. Finally, crystalline oxides were extracted using 30 mL of 1 M hydroxylamine–hydrochloride–acetic acid. No residual or silicate extraction was performed as these extractions, while efficient, often result in the erroneous classification of inefficiently extracted phases (La Force and Fendorf, 2000). Each sample was vortexed and heated for 6 h and then centrifuged, decanted, filtered, and refrigerated until analyzed.
Soil As and Pb concentrations were determined by ICP-MS in a clean laboratory environment. The ICP-MS had typical detection limits of 0.1 µg L–1 and analytical uncertainties of ±2% or better. Total digestions of soils were also determined to establish the efficiency of sequential extrations. Soil samples for total digestions were digested in aqua regia (3:1 concentrated HCl/HNO3), evaporated to dryness at 70°C, resuspended in 1 M HCl, and stored in acid-cleaned plastic bottles before analysis by ICP-MS.
Macroinvertebrate samples preparation methods were similar to those described by Quinn et al. (2003). Samples were frozen in Teflon vials, lyophilized, weighed, then acidified and homogenized with trace metal grade 70% HNO3 (Seastar grade) and H2SO4. Separate sulfuric acid digestions were also performed to determine total phosphorus (APHA, 1995). Samples were microwave digested for 2 h, diluted, and analyzed using ICP-MS.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
30 cm. Assuming all pesticide was applied at the soil surface, it appears that As and Pb have migrated vertically about 0.2 cm yr–1 since they were applied. Consistent with what has been observed elsewhere (Elfving et al., 1994; Merry et al., 1983; Peryea and Creger, 1994), in the untilled field the As appears to be slightly more mobile than Pb, particularly under the canopy where concentrations are highest.
|
17% of the untilled field underlies canopy, the total masses of As and Pb in the upper 35 cm of the untilled field are estimated as 295.8 ± 1.5 kg ha–1 and 907.8 ± 3. 0 kg ha–1, respectively (error reported as SD from mean total mass). In the tilled field, the total As and Pb masses in the upper 35 cm are slightly less (292.4 ± 1.5 kg ha–1 As and 897.9 ± 3.0 kg ha–1 Pb). The total As and Pb masses in the upper 35 cm of the control field are much less (45.6 ± 1.5 kg ha–1 As and 27.4 ± 3 kg ha–1 Pb). The masses of applied As and Pb (determined by difference between the orchard and control sites) are consistent with the stoichiometry (Pb/As of 3.5 by mass) of the most commonly used lead arsenate pesticides (Frank et al., 1976).
The variation of As with depth in pits in the perennial stream channel adjacent to the untilled field and in the ephemeral stream channel adjacent to the tilled field are plotted along with 137Cs concentrations in Fig. 3
. Both profiles have a peak in As concentration at a depth of 30 cm. At this same depth the concentration of 137Cs < < 1 Bq kg–1. Radioactive Cs is present in soils largely as a result of fallout from atmospheric nuclear weapons during the 1950s and 1960s. Thus, the absence of 137Cs is consistent with sediment deposition before 1950. A secondary As peak occurs in the sediments below the tilled field at a depth of 10 to 15 cm. This peak is associated with elevated concentrations of 137Cs, consistent with deposition after 1950, and is absent in the sediments below the untilled field. We therefore postulate that a secondary mobilization event occurred in the tilled field that deposited As and Pb into the stream after 1950.
|
10 g kg–1). Concentrations increase to 30 to 70 g kg–1 at depths between 10 and 20 cm, and then return to background levels at depths >25 cm. All samples analyzed have 137Cs concentrations >1 Bq kg–1, consistent with deposition after 1950. In Fig. 4 we plot the total excess As and Pb mass in each pit in the ephemeral channel as a function of distance downstream from the uppermost reach of the channel draining the tilled field. Excess metal mass is calculated by subtracting the average background metal concentration measured in the control field soils (14.2 g kg–1 As and 8.8 g kg–1 Pb) from the measured concentrations at each depth, converting concentrations to mass using the measured bulk densities, and then integrating to a depth of 30 cm. There is no correlation between metal mass and distance downstream, so we calculate the total excess metal mass in the ephemeral stream channel as the average excess metal mass (44.3 ± 6 kg ha–1 As, 57.9 ± 12 kg ha–1 Pb) times the area of the stream channel (450 ± 150 m2), yielding 2 ± 1.0 kg As and 2.6 ± 1.6 kg Pb.
|
|
Macroinvertebrate body metal burdens (Table 1) are more variable in the wetlands than in the survey of regional lakes. Mean Pb burdens in the wetland macroinvertebrates are significantly lower than those measured in the control sites (lakes). There were no significant differences in As burden in macroinvertebrates between the wetland and the lakes.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
4.4 kg ha–1 yr–1 of the original lead arsenate applied to the orchard soils could be dissolved each year. Thus, in the 70 yr since the lead arsenate was applied, 300 kg ha–1 of the As could have dissolved. This estimate is similar to the total mass of As measured in the untilled field (296 kg ha–1). We therefore conclude that dissolution alone may be sufficient to explain the absence of lead arsenate in its original mineral phase in the orchard soils. Although lead arsenate appears to dissolve, both lead and arsenate are retained in the soil on secondary phases. The results from the XRD and SEM analyses and sequential extractions suggest that most (70–80%) of the dissolved Pb and As adsorbs onto fine particles of amorphous oxides or crystalline oxides. This is consistent with the known behavior of As and Pb, both of which strongly adsorb to oxide minerals. Lesser quantities of the As and Pb were extracted by NaOCl (15–30%), an extraction that targets organic matter. The suggested organic matter association could indeed be observed in the case of Pb, which forms strong organic complexes, but probably reflects the extraction of nontarget As-bearing phases since As is not generally associated with organic matter. Additional work to determine the speciation of the As would provide additional insight into the specific mechanisms by which the As is retained, particularly in the channel sediments.
For both contaminated orchards our results are similar to those from previous studies in that the highest concentrations of Pb and As are restricted to the top 25 cm of soil, indicating only limited vertical redistribution of metals (Aten et al., 1980; Benson, 1976; Jones and Hatch, 1937; Peryea and Creger, 1994; Veneman et al., 1983). A comparison of Pb and As inventories in the tilled and untilled fields suggests the loss of these metals from the tilled field. Both the tilled and untilled fields were originally planted at the same time and managed by the same orchard owners using the same application rates of pesticides. If we assume that arsenical pesticides were applied approximately equally to both fields, then the tilled field appears to be missing 3.4 ± 3 kg As and 9.9 ± 6 kg Pb. The assumption of identical applications of the arsenical pesticides is tenuous, but additional support for the loss of As and Pb from the tilled field comes from three lines of evidence.
First, As concentrations in the drainage down gradient from the tilled field have two distinct peaks (Fig. 3). The lower peak is not associated with significant 137Cs concentrations, consistent with deposition before 1950. A similar peak is found in the sediments below the untilled field. In fact, below 20 cm the As concentration profiles in both drainages are remarkably similar, consistent with their assumed similar histories. We suggest that these lower peaks in As concentrations are associated with the initial application of the arsenical pesticides and are probably due to both the drifting of the pesticide spray as it was applied and the subsequent erosion of the residual pesticide from the land surface soon after it was applied. The upper peak in the tilled field drainage is associated with elevated concentrations of 137Cs, suggesting deposition after 1950. No similar peak appears in the As concentration profile in the untilled field drainage. That this upper peak in the tilled drainage does not represent the remobilization of Pb and As originally deposited during the application of the pesticides is supported by the distinct bimodal concentration profile; we would expect that the reworking of sediments during remobilization would mix the sediments and smooth the concentration profile and result in changes in the phases that retain the As and Pb. We suggest that the disturbance that remobilized the Pb and As from the tilled field and ultimately deposited these metals into the drainage after 1950 was the tilling and replanting of the field in 1992.
Second, if the As and Pb apparently missing from the tilled field was lost due to tilling-related physical erosion of the fine particles of amorphous oxides and organic matter, then we should expect to find similar masses of Pb and As in the accumulating sediments of the drainage adjacent to the tilled field. In fact, within the uncertainty of our measurements, the mass of Pb and As in the 137Cs-rich sediments in the ephemeral stream channel draining the tilled field is the same as the masses of these metals missing from the tilled field.
And finally, we developed an independent estimate of the soil loss expected due to tilling a field using the Revised Universal Soil Loss Equation (RUSLE) (Renard et al., 1991; Wischmeier, 1976). Use of parametric values appropriate for southern New Hampshire and the appropriate field geometry (USDA, 2003a) yields an expected soil loss due to tilling of 18 to 36 Mg ha–1. Using the average bulk density of the upper 20 cm of the tilled field (1.3 g cm–3), this is equivalent to 0.14 to 0.28 cm yr–1 soil loss due to tilling, well within the range of commonly encountered values.
The above results can be summarized by converting the masses of As and Pb missing in the tilled field and present in the ephemeral stream channel draining the tilled field into an equivalent depth of soil eroded. These eroded soil depths, along with that predicted using the RUSLE, are shown in Fig. 6 . The agreement between the different estimates is remarkable and provides support for our suggestion that soil erosion associated with tilling mobilized the Pb and As.
|
The fact that the masses of As and Pb "missing" in the tilled field are approximately equivalent to the masses of these metals present in the ephemeral stream channel argues against the tilled field being a major source of As and Pb to the wetland at the outlet of the ephemeral channel at the current time; most of the As and Pb eroded from the tilled field appears to remain in the channel sediments and not to have reached the wetland. This is consistent with the lack of elevated As and Pb body burdens in wetland macroinvertebrates. Mean As and Pb body burden in both chironomids and libellulids are similar (As) or lower (Pb) in the wetland than in regional lakes. However, it is also possible that any metals that did reach the wetlands are biologically unavailable. Alternatively, high primary productivity typical in wetlands could result in lower metal body burdens through bloom dilution (Pickhardt et al., 2002).
Finally, while this work only considers the effect of tilling on the mobilization of residual arsenical pesticides, our work shows that the Pb and As are bound to small and presumably highly mobile particles. It is therefore likely that other types of land disturbances will also mobilize significant amounts of Pb and As in lands where arsenical pesticides were used, particularly over longer timescales. In southern New Hampshire, for example, former orchard land is currently being rapidly developed and urbanized. Our results suggest that as this land is developed, attention should be given to the possibility of mobilizing previously immobile reservoirs of Pb and As.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  G. R. Robinson Jr., P. Larkins, C. J. Boughton, B. W. Reed, and P. L. Sibrell Assessment of Contamination from Arsenical Pesticide Use on Orchards in the Great Valley region, Virginia and West Virginia, USA J. Environ. Qual., April 5, 2007; 36(3): 654 - 663. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. T. Jones A Broad View of Arsenic Poult. Sci., January 1, 2007; 86(1): 2 - 14. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Soil Science Society of America Journal | Journal of Plant Registrations | The Plant Genome | |||
