|
Pesticides vary
in how tightly they are adsorbed to soil particles. Degree of soil
binding is measured by binding coefficients or "K" values.
Koc (K of organic carbon) is a measure of adsorption to the organic
matter content of soils, with higher values indicating more binding.
While pesticides are also bound to clay particles, binding to organic
matter is a useful predictor of chemical behavior and movement.
Koc values can be used to predict whether a specific pesticide will
be carried primarily in the sediment or dissolved phase of field
runoff. Koc values greater than 1000 indicate that pesticides are
highly adsorbed to soil. Eroded soil carries the majority of this
kind of chemical leaving fields in runoff. Thus, if conservation
buffers are effective in trapping sediment, they will be effective
in trapping this type of pesticide.
Pesticides with
lower Koc values (less than 300 to 500) tend to move more with water
than adsorbed on sediment. Concentrations carried on sediment are
higher than concentrations in water, but because water quantities
running off fields are so much greater than eroded soil quantities,
water accounts for the majority of chemical leaving fields. To be
effective in trapping this type of pesticide, buffers need to increase
water infiltration or maximize contact of runoff with soil and vegetation
which may adsorb pesticide.
Nitrate is water
soluble and not adsorbed by soil particles. Usually nitrate quickly
enters the soil, with little nitrate carried in runoff water. Rather,
nitrate may leach to shallow ground water and be carried to streams
by subsurface flow. To be effective in trapping nitrate, roots of
conservation buffer plants need to intercept this subsurface flow.
Conditions for denitrification present in this biologically active
zone also reduce nitrate reaching streams. Similarly, some weakly
adsorbed pesticides may leach to shallow groundwater in small amounts.
Although subsurface flow may carry small quantities of pesticides
to streams, usually quantities present in surface runoff are much
greater. The NRCS maintains a current Pesticide Property Database
and can provide Koc values for specific pesticides.
Study Results
One of the earliest
studies of the impact of buffers on pesticide runoff investigated
pesticide retention by grassed waterways (Asmussen et al., 1977).
Runoff from a small plot was directed into an 80-foot-long waterway.
The weakly adsorbed herbicide, 2,4-D, was trapped at a rate of 70%.
In a later, similar study (Rhode et al., 1980), 96% of the strongly
adsorbed trifluralin was trapped when the waterway was dry before
runoff and 86% when the waterway was already wet when runoff was
produced.
Hall et al. (1983)
studied the impact of strip cropping on atrazine runoff in Pennsylvania.
Seventy-two-foot-long plots were constructed up and down a 14% slope
with a 19.7-foot- wide area at the base of the slope seeded to oats
(Avena sativa L.). Runoff of atrazine applied to corn was measured
throughout the season which included a severe, once-in-100-year
frequency storm in June. The oats strip trapped 91% of atrazine
in runoff when the herbicide was applied at a rate of 2 lb/A.
Results of these
early studies surprised some scientists who assumed that buffers
would have minimal impact on runoff of moderately adsorbed pesticides
like atrazine and 2,4-D. However, significant infiltration of runoff
water into buffers was identified as the primary mechanism of pesticide
removal in these studies.
In Mississippi,
Webster and Shaw (1996) measured runoff of metolachlor and metribuzin
from soybean plots with and without tall fescue (Festuca arundinacea
Schreb.) filter strips. Plots were 13 by 72 ft with 3% slope. Filter
strips were either 13 or 6.5 ft wide. Width of the filter strip
did not affect herbicide trapping efficiency. Over 3 years, metolachlor
loss was reduced 55 to 74% by the filter strips, while metribuzin
loss was reduced by 50 to 76%. Much of the trapping of herbicides
could be attributed to infiltration of runoff into the filter strips.
Using similar techniques in a later study (Rankins et al., 1998),
tall fescue buffers reduced runoff of fluormeturon and norfluorazon
applied to cotton by at least 60% and 65%, respectively.
Several studies
in Iowa have investigated herbicide trapping by smooth bromegrass
(Bromus inermis Leysser) filter strips using either simulated or
natural runoff. In some studies field runoff was simulated by adding
known concentrations of herbicide to water released above filter
strips while rainfall was simulated (Mickelson and Baker, 1993).
Amount of solution calculated to represent runoff from a 150 foot-long
area was applied to the top of 15 and 30 foot-long filter strips
(thus representing source area to buffer ratios of 10:1 and 5:1).
The 15 foot strip reduced atrazine runoff by 35%, while the 30 foot
strip reduced runoff by 59.5%.
Other similar studies
with smooth bromegrass buffers (Misra et al., 1996) showed less
difference between buffer sizes. When comparing 15:1 and 30:1 source
area to buffer ratios, atrazine removal was 31.2% and 26.4%, respectively.
Concentrations of herbicides in simulated runoff were applied at
either 0.1 ppm or 1.0 ppm. A higher percentage of herbicide was
trapped by buffers when inflow had higher concentrations. When inflow
had 1.0 ppm herbicide, 15:1 area ratio buffers trapped 50%, 47%,
and 47% of atrazine, metolachlor, and cyanazine, respectively. When
inflow concentrations were 0.1 ppm, these buffers trapped 31%, 32%,
and 30% of atrazine, metolachlor, and cyanazine, respectively. Infiltration
of runoff water into buffers accounted for most herbicide trapping.
In other studies
natural runoff from a treated area was collected and distributed
to replicated smooth bromegrass buffers 66 feet long (Arora et al.,
1996). Runoff was distributed to represent source area to buffer
ratios of 15:1 and 30:1. Efficiency of herbicide trapping was determined
for six runoff events over a 2 year period. Trapping of atrazine,
cyanazine, and metolachlor ranged from a low of 8% to a high of
100%, depending largely on the timing and intensity of rainfall
and antecedent soil moisture conditions. Herbicide trapping was
least efficient when soil was saturated due to previous rains. For
most events there were only small differences in herbicide trapping
efficiency between area ratios. Averaging results for all herbicides
and area ratios over the six events, 62% of herbicide contained
in runoff was trapped by buffers. Infiltration of runoff into buffers
was determined to be the major mechanism of herbicide removal. While
sediment retention ranged from 40 to 100%, only about 5% of total
herbicide retention was due to sediment trapping.
Analysis of soil
within the buffer strips confirmed that herbicides were being trapped
and held by soil within the strips (Fawcett et al, 1995). Concentrations
declined during the growing season, presumably due to degradation.
No phytotoxicity was observed on buffer grasses.
Bermudagrass (Coynodon
dactylon L.) and wheat (Triticum aestivum L.) buffers were studied
in Texas (Hoffman, 1995). Three 30-foot-wide buffers were equally
spaced within a 435 wide watershed planted to corn. Hydrologic data
showed that water runoff was reduced 57% by bermudagrass and 50%
by wheat. Total atrazine loss was reduced 30% by bermudagrass and
57% by wheat in one year, and by 44-50% by all buffers in another.
Patty et al. (1997)
studied ryegrass (Lolium perenne L.) buffers in France under natural
rainfall conditions. Buffer widths were varied from 20 to 59 ft.
Runoff volume was reduced by 43 to 99.9%, suspended solids by 87
to 100%, lindane losses by 72 to 100%, atrazine and metabolite losses
by 44 to 100%, isoproturon losses by 99%, and diflufenican losses
by 97%.
The ability of bermudagrass
buffers to trap runoff of turf pesticides was studied in Oklahoma
(Cole et al., 1997). Buffers 16 feet wide trapped from 90 to 100%
of dicamba, from 89 to 98% of 2,4-D, from 89 to 95% of mecroprop,
and from 62 to 99% of chlorpyrifos in runoff. In most instances
buffer mowing height (3.3 in or 9.7 in), buffer length (7.9 or 16
feet), and tine aeration did not significantly affect pesticide
trapping efficiency.
The effectiveness
of a three zone riparian buffer in trapping herbicide runoff was
studied in Georgia (Lowrance et al, 1997). Total buffer width was
164 feet, with a 26-foot-wide grass buffer adjacent to the crop
field, a managed pine forest down slope from the grass, and a narrow
hardwood forest containing the stream channel. More than 90% of
atrazine and alachlor in field runoff was trapped by the buffer.
Most herbicide trapping was accomplished by the grass strip, with
60 to 70% of herbicide in runoff trapped. Grass was a mixture of
bahaigrass grass (Paspalum notatum Flugge), bermudagrass, and perennial
ryegrass.
The impact of untreated
setbacks around tile inlets in tile-outlet terrace systems was studied
in Iowa, Nebraska, and Missouri (Mickelson et al.; Franti et al.).
In all studies, setbacks provided no reductions in herbicide runoff
into inlets beyond what would be expected due to reduced area treated.
This result is not surprising, since these terraces are designed
to pond water around inlets, causing sediment to settle and increasing
water infiltration. Much of the untreated setback area would be
under water during runoff events and could not be expected to increase
infiltration or sediment trapping beyond that caused by normal functioning
of the terrace system. The studies' authors concluded that alternative
BMPs such as herbicide incorporation and no-till production were
more effective in reducing herbicide runoff into inlets. Based on
this research, USEPA allowed label changes on atrazine and cyanazine-containing
products. Three alternative BMPs are now described on these labels
for use in tile-outlet terrace systems: 1) 66 foot untreated setback
around inlet; or 2) incorporation of herbicide in areas draining
to inlet; or 3) no-till production with high crop residue levels
in areas draining to inlet.
The importance of
water infiltration as a mechanism of trapping moderately adsorbed
pesticides is illustrated by some studies which have shown that
buffers do little to reduce concentrations of this type of pesticide
in runoff. In Nebraska (Yonts et al., 1996) smooth bromegrass or
intermediate wheatgrass buffers did not significantly reduce concentrations
of alachlor, cyanazine, 2,4-D, or atrazine present in furrow irrigation
runoff water, although concentrations of the strongly adsorbed chlorpyrifos
were reduced.
In contrast, some
studies have shown significant reductions in concentrations of moderately
adsorbed pesticides caused by buffers. In Mississippi (Tingle et
al., 1998), tall fescue buffers as narrow as 1.6 feet at the base
of 72 foot plots reduced concentrations of metolachlor and metribuzin
in runoff by almost 50%.
Biological and physical
conditions, which develop in buffers planted to grasses and/or trees,
favor increased water infiltration and subsequent attenuation of
nutrients and attenuation and degradation of pesticides. Bharati
(1997) compared infiltration and soil properties under a multi-species
riparian buffer in Iowa to adjacent cultivated fields and a grazed
pasture. Average cumulative water infiltration was five times greater
under the buffer than the cultivated field and pasture sites. Soil
bulk density was also consistently lower under the buffer. Wood
(1977) compared hydrologic characteristics of forested land to adjacent
sugarcane, pineapple, or pasture land of the same soil series at
15 sites in the Hawaiian islands. Infiltration rates were higher
under forest cover at 14 of the 15 sites. Mean weight diameters
of the surface soil aggregates were larger for forested soils.
The extensive root
growth in buffers and superior soil structure likely explain observed
water infiltration increases. This root growth also impacts biological
activity, supplying an organic carbon energy source to soil microorganisms.
These microorganisms in turn are responsible for degrading pesticides
and denitrifying nitrate. Such untilled areas also attenuate atmospheric
carbon dioxide in tree and grass vegetation and soil organic matter.
Tillage and crop production often deplete soil organic matter (Reicosky
et al., 1995). Increased soil organic matter in buffers serves to
better adsorb pesticides in runoff. Nutrients are taken up by vegetation
and stored in living tissue. Periodic harvest may be desirable to
prevent later rerelease. Pesticides are also taken up by roots and
may be metabolized in plants. In addition, vegetation at the soil
surface adsorbs pesticides during runoff events. In Iowa (Fawcett
et al., 1995), atrazine concentrations in plant residue collected
in buffers ranged from 80 ppb to 740 ppb depending on collection
date, and were similar to concentrations found in surface buffer
soil.
Few studies have
investigated pesticide trapping by constructed wetlands. Some pesticides
are relatively short-lived in water. Thus degradation occurring
while runoff or tile effluent is sequestered in wetlands reduces
contaminants reaching streams. Wetlands can also serve to ameliorate
pulses of concentrated runoff before it enters streams. Some pesticides
are relatively persistent once they reach water. However, the high
organic matter content of wetlands sediments binds these pesticides,
removing them from water. Matter (1993) used intact freshwater wetland
sediment microcosms to study the behavior of atrazine. Atrazine
was removed from the overlying water column at a rate of 15.8%/day
for 3 days after introduction. After 10 months, 88% applied atrazine
was unextractable from sediment and none was recovered from the
overlying water.
Considering buffer
research to date, under controlled conditions, buffers have been
effective in trapping both highly adsorbed and moderately adsorbed
pesticides.
Table 1 is a summary
of buffer studies showing trapping efficiency for specific pesticides
and pesticide Koc values. Highly adsorbed pesticides were trapped
at rates of from 62 to 100%. Trapping of moderately adsorbed pesticides
was more variable and ranged from 8 to 100%. Lowest percent pesticide
retention by buffers occurred when buffer soil was saturated due
to previous rains. Many studies found pesticide trapping efficiencies
of 50% or more.
Table 1. Summary
of buffer studies measuring trapping efficiencies for specific pesticides.
Koc values listed for each pesticide are from the NRCS Field Office
Technical Guide Section II Pesticide Property Database.
| Pesticide |
Koc |
Study
Reference |
Range
% Pesticide Trapped |
| |
Highly Adsorbed Pesticides |
|
| Chlorpyrifos |
6070 |
Cole
et al. 1997 |
62-99 |
| Diflufenican |
1990 |
Patty
et al. 1997 |
97 |
| Lindane |
1100 |
Patty
et al. 1997 |
72-100 |
| Trifluralin |
8000 |
Rhode
et al. 1980 |
86-96 |
| |
Moderately Adsorbed Pesticides |
| Alachlor |
170 |
Lowrance
et al. 1997 |
91 |
| Atrazine |
100 |
Arora
et al. 1996
Hall et al. 1983
Hoffman 1995
Lowrance et al. 1997
Mickelson & Baker 1993
Misra et al. 1996
Patty et al. 1997 |
11-100
91
30-57
97
35-60
26-50
44-100 |
| Cyanazine |
190 |
Arora
et al. 1996
Misra et al. 1996 |
8-100
30-47 |
| 2,4-D |
20 |
Asmussen
et al. 1977
Cole et al. 1997 |
70
89-98 |
| Dicamba |
2 |
Cole
et al. 1997 |
90-100 |
| Fluormeturon |
100 |
Rankins
et al. 1998 |
60 |
| Isoproturon |
120 |
Patty
et al. 1997 |
99 |
| Mecoprop |
20 |
Cole
et al. 1997 |
89-95 |
| Metolachlor |
200 |
Arora
et al. 1996
Misra et al. 1996
Webster & Shaw 1996
Tingle et al. 1998 |
16-100
32-47
55-74
67-97 |
| Metribuzin |
60 |
Webster
& Shaw 1996
Tingle et al. 1998 |
50-76
73-97 |
| Norflurazon |
600 |
Rankins
et al. 1998 |
65 |
Do results of these
controlled studies predict what will happen in the real world? Nearly
all of these studies (with the exception of early grassed waterway
studies) were designed to encourage sheet flow across buffers. Therefore,
they represent the maximum trapping expected. In the real world,
concentrated flow often occurs across buffers, reducing their effectiveness.
In order to maximize trapping of both sediment-adsorbed and dissolved
pesticides, sheet flow needs to be encouraged through proper buffer
design and maintenance.
Dillaha et al. (1989)
analyzed 33 existing buffers in Virginia for sediment trapping efficiency.
They found sediment trapping was often poor either due to concentrated
flow where topography was hilly, or due to sediment that accumulated
in the buffer, causing runoff to flow parallel to the buffer until
a low point was reached, where concentrated flow occurred.
Excessive sediment
load in runoff may not only change flow patterns due to accumulation
in buffers, but may also reduce water infiltration, making buffers
less effective in trapping dissolved pesticides. In an Iowa simulation
study (Misra, 1994), runoff with and without suspended sediment
was introduced into buffers. In absence of sediment, buffers removed
over 80% of atrazine, cyanazine, and metolachlor. When sediment
at 10,000 mg/L was included in runoff, trapping of the three herbicides
fell to about 50%. Accumulation of sediment apparently caused surface
sealing, reducing total water infiltration from 83% in absence of
sediment to 30% with sediment. It is thus critical that soil conservation
methods be used above conservation buffers to reduce amounts of
sediment entering buffers.
|
