| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
School of Natural Resources, Univ. of Missouri, Columbia, MO 65211
* Corresponding author (udawattar{at}missouri.edu)
Received for publication August 24, 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Abbreviations: NPS, nonpoint source • TN, total nitrogen • TP, total phosphorus
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Agroforestry farming practices provide multiple benefits including high productivity and additional income while maintaining soil health (Kang et al., 1984). Vegetative filter strips established in the form of agroforestry or contour grass buffer strips have the potential to improve water quality, wildlife abundance, biodiversity, and aesthetic value. Filter strips of permanent vegetation that reduce runoff and trap sediment can be used to greatly reduce nonpoint-source (NPS) pollution (Robinson et al., 1996; Cooper and Lipe, 1992). Nonpoint-source pollution is a landscape-scale phenomenon and its diffuse nature complicates mitigation (Verchot et al., 1998), but vegetative filter strips have distinct advantages over other erosion control technologies (Robinson et al., 1996). Normally, interest in the use of agroforestry practices and contour grass strips for various environmental benefits relates to their potential to increase infiltration, reduce runoff, and reduce NPS pollution.
Limited research suggests that properly established filter strips can reduce runoff, sediment load, and NPS pollution (Dillaha et al., 1989) although the design factors of such practices have not been clearly defined. The increased infiltration found under natural forests also occurs in agroforestry and with other vegetative filter practices. Multistrata systems, combined with litter cover and dense root systems, hold runoff when it first reaches the surface and subsequently promote infiltration. The vegetation in an agroforestry practice serves two major purposes: (i) the fine root system holds soil in place, reducing susceptibility to erosion, and (ii) plant stems decrease the flow velocity, enhancing sedimentation. Tree roots can also take up nutrients that would otherwise be lost by leaching (van Noordwijk et al., 1996). Moreover, the addition of organic matter from trees in agroforestry improves soil physical properties, chemical properties, and infiltration, thus reducing runoff, NPS pollution, and sediment loss (Young, 1997). Nonpoint-source pollution removal capability of vegetative strips depends on the nature of the pollutant, hydrology of the area, soil properties, and nature of the trees. Site-specific interrelationships of these factors contribute to large variability in pollution control effectiveness.
Schmitt et al. (1999) compared grass versus grass–shrub–tree buffers to test their effectiveness in protecting waterways from contaminants. The grass–shrub–tree combination produced significantly lower sediment and total nitrogen runoff than the grass buffer strips. Numerous other studies have shown that forest vegetation removes significant quantities of NPS pollution from agricultural runoff (Cooper and Gilliam, 1987; Lowrance et al., 1984; Peterjohn and Correll, 1984). It might therefore be assumed that incorporation of trees and grass in upland buffer designs would enhance water quality. However, sufficient quantitative information from direct experimental studies designed to evaluate the relative filtering performance of agroforestry and grass strips is lacking to provide guidance in designing effective agroforestry buffer strip practices.
The watershed practices in our study are complex, so to investigate their influences we had to minimize extraneous effects. No reported studies on the effects of agroforestry and contour grass strips on water quality that included a calibration and a treatment period were found in the literature. The research we are reporting is unique as it employs paired watersheds to examine the effect of agroforestry while eliminating extraneous effects. In a paired watershed study, watersheds need not be identical (Hewlett and Pienaar, 1973). However, areas should be small enough to obtain uniform treatment over the entire watershed; watersheds should be similar in size, slope, location, and land cover; and watersheds should be in a steady state prior to the study (USEPA, 1993). Since climatic and hydrologic differences are statistically described using calibration data, water quality changes can be attributed to treatments.
Although several benefits of agroforestry have been claimed, in fact, there has been little research to demonstrate these benefits, thereby restricting the systematic selection of best management practices suitable for the region. This is especially true with highly erodible landscapes dominated by corn–soybean agriculture as is practiced in northern Missouri and much of the U.S. Midwest. The primary objective of this study was to examine whether agroforestry is a viable and effective land management strategy when protection of water sources from sediment and nutrients is a goal. Specific objectives were to (i) quantify runoff, sediment loss, and nutrient loss from typical corn–soybean rotations; (ii) determine the effectiveness of agroforestry when employed in conjunction with corn–soybean agriculture in reducing runoff, sediment, and nutrient losses; and (iii) examine the specific mechanisms of agroforestry and contour grass filter strips in reducing sediment and nutrient loss from watersheds.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
|
Thirty-year mean annual precipitation in the region is 920 mm yr-1, of which more than 66% falls from April through September (Owenby and Ezell, 1992). Mean annual air temperature is approximately 11.7°C with an average monthly low of -6.6°C in February and an average monthly high of 31.4°C in July (Owenby and Ezell, 1992). Snowfall averages about 590 mm per year and can stay on the ground for extended periods.
Agricultural activities since 1991 are summarized in Table 1 . Prior to 1991, the field containing these watersheds was in a corn–soybean rotation with cultivation along straight rows conforming to the field boundaries. Since 1992, cultivation has been on the contour, unless otherwise noted.
|
Chemical and physical analyses of composite water samples were performed in the Forest Hydrology Laboratory at the University of Missouri. A known volume of a well-mixed sample was filtered through a preweighed 934-AH glass microfiber filter (Whatman, Maidstone, UK) using a vacuum pump to estimate sediment concentration (American Public Health Association, American Water Works Association, and Water Environment Federation, 1992, p. 2–56). These filters were dried at 105°C to a constant weight and their dry weights were recorded. The difference between two dry weights and the filtered sample volume were used to estimate total suspended sediment concentration.
Unprocessed samples were refrigerated at 4°C until analysis. From 1991 to 1998, total nitrogen (TN), nitrate, and ammonium were determined with a Technicon (Terrytown, NY) autoanalyzer. This method determined total Kjeldhal nitrogen in ammonium form. Nitrate N concentrations of those samples were added to total Kjeldhal nitrogen to estimate total nitrogen of samples analyzed by the Technicon autoanalyzer. A Lachat (Milwaukee, WI) Quick-Chem 8000 Analyzer was purchased in 1998 and subsequent TN, nitrate, and ammonium analyses were performed on this new instrument. Total nitrogen was determined using cadmium reduction on unfiltered samples following potassium perusulfate digestion. Total phosphorus (TP) was determined by ascorbic acid–molybdate procedure on unfiltered samples following ammonium peroxidisulfate digestion. Total phosphorus, TN, nitrate N (cadmium reduction), and ammonium (phenolate) were determined as outlined by Lachat Quickchem methods 10-115-01-1-F, 10-107-04-1-C, 10-107-04-1-B, and 10-107-6-1-A, respectively. The detection limit for the four methods was 
0.002 mg L-1. Quality control for the Lachat analyzer was maintained by randomly positioning three control standards with differing concentrations, four duplicate samples, and one quality control sample in each tray (90 samples). All samples with suspect concentrations and trays with unacceptable concentrations were reanalyzed.
Calibration relationships between control and treatment watersheds were developed using runoff and chemistry data. Runoff calibration was developed using 110 data pairs between the control and contour strip and control and agroforestry watersheds. For nutrients and sediment, 64 and 71 data pairs were used to develop calibration regressions. A few (maximum three) suspect or out of range data pairs were not used when developing these relationships. The treatment period started in June 1997. Treatment effects on runoff were examined using 44 events. However, only 17 runoff events produced sufficient runoff to collect samples for sediment and nutrient analysis, which allows us to compare sediment and nutrient losses due to treatments.
Approximately 9% of the runoff data collected during the calibration period on the control and agroforestry watersheds was lost due to sampler and flow meter failures caused by lightening damage, electronic malfunction, missed samples, and pump tube loosening or presence of debris in water samples. The contour strip watershed suffered a similar loss of 12% of the sampling events. Runoff and nutrient data for the control and one remaining treatment (when only one was available) were used to develop calibration for that treatment even though the other treatment data were lost. However, if a flowmeter or a sampler recorded any malfunction at any time during the treatment period, runoff and nutrient data for the entire event were discarded.
Statistical analyses of the data were performed using Statistical Analysis Systems (SAS Institute, 1999). Statistical relationships between the control and treatments were developed using regression. Our study contained a sufficient number of observations to determine a 10% change in discharge (USEPA, 1993). The procedure also explains testing the significance of the regression, testing the significance of overall regression, testing for a significant worthwhile difference, testing residual error, and evaluating the range of values obtained.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
|
Calibration Parameters
Highly significant (R2 = 0.97 and p = 0.0001) relationships for discharge (m3 ha-1) exist between the control and two treatments for the calibration period (Table 3)
. Regression relationships for sediment and nutrients (kg ha-1) from treatment watersheds and the control were significant during the calibration period (p = 0.0001). Total nitrogen, nitrate, and ammonium calibrations had an R2 greater than 0.92 for both treatments. Agroforestry possessed a better relationship for sediment (R2 = 0.88) than the contour strip treatment (R2 = 0.77). The lowest regression coefficients for calibration were found for total phosphorus (0.74 and 0.69 for contour strip and agroforestry, respectively). Calibration slopes for each relationship of the contour strip treatment were always greater than the slopes of the agroforestry treatment (Table 3). This could be due to watershed morphology differences, inherent soil–site characteristics, relative effectiveness of the grass waterways, or other unknown factors.
|
While Fig. 2 shows the runoff reduction trends on both treatments through the three-year time period, the initial 11 runoff events in 1998 on both agroforestry and contour strip treatments produced more discharge than predicted. As trees started to transpire, and vegetation buffer strips became better established, watersheds began showing discharge reductions (Fig. 2). The first reductions on both treatments appeared on 3 July 1998 (event 25), about one year after the grass–legume strips were established. The agroforestry and contour strip treatments resulted in 23 and 10% runoff reductions based on calibrations for this single event. Of the subsequent eight events (26th to 33rd events), only three generated runoff on the control. In October 1998, when fall precipitation commenced, both treatment watersheds consistently produced less runoff than predicted. Based on calibration relationships, agroforestry and contour strip treatments reduced runoff by 0.2 and 6% during the 1998 sampling period.
|
The largest decreases in runoff from both watersheds were associated with the largest runoff events. The largest deviations occurred on 15 Apr. 1999. This indicates the effect of grass strips before tree leaf out and before crop planting. During this period, the ground had the maximum exposure, which resulted in more runoff, but the grass strips reduced runoff by 20 and 29% on agroforestry and the contour strip treatments, respectively. Studying runoff control mechanisms, Robinson et al. (1996) reported that vegetative filter strips encourage infiltration of water into the soil. They found that the first 3 m of the vegetative filter strip was the most effective in decreasing the runoff volume. However, other studies have indicated that 4.6-m grass strips were more effective in controlling runoff and sediment loss than more narrow ones (Schmitt et al., 1999). Grass strips in this study are 4.56 m wide and reduced runoff significantly. In spite of the differences, this study confirms that agroforestry and contour grass strips markedly reduce runoff in corn–soybean rotation watersheds in the Midwest.
The largest runoff event during the treatment period occurred on 4 Oct. 1998 and accounted for 16, 14, and 17% of the total runoff on the control, agroforestry, and contour strip watersheds, respectively. The second and third largest events occurred on 3 July 1998 and 29 Oct. 1998. Averaged over the three watersheds, the largest two and three runoff events accounted for 26 (24.6 to 28.1) and 36 (33 to 37.2)%, respectively, of the total runoff during the treatment period. Some major precipitation events generate greater runoff in severe storms than in a more gradual rainfall event. Studies suggest that control of runoff and erosion caused by such catastrophic events is more important for water quality improvement than regulation of smaller events (Edwards and Owens, 1991). During the largest three runoff events, the contour strip and agroforestry treatments reduced runoff by 10 and 11%, respectively, on the two watersheds compared with the predicted runoff losses.
Sediment Loss
During the treatment period, the control, agroforestry, and contour strip treatments lost 200, 264, and 242 kg ha-1 sediment, respectively. The predicted losses, based on calibrations on the agroforestry and contour strip treatments, were 195 and 206 kg ha-1. The agroforestry and contour strip treatments lost respectively 35 and 17% more than the predicted amounts. The first two runoff events during the treatment period on average caused 45 and 36 kg ha-1 sediment loss on agroforestry and contour strip treatments, respectively. Soon after treatments were established, both treatments lost more sediment. This could be due to soil disturbance during tree planting and grass strip establishment.
However, Fig. 3 shows that sediment loss has declined over the treatment period. As the grass strip and permanent vegetation became established and started to transpire and infiltration improved, runoff loss was reduced and associated losses declined. During the first two years (1997 and 1998) of sampling, the treatments did not result in reductions of sediment loss. However, more recently, both the agroforestry and contour strip watersheds have shown trends toward declining differences. At this time we cannot speculate whether these treatment differences will continue to trend lower.
|
Figure 3 also shows that soil loss was largely determined by individual precipitation events and the ground condition. The greatest sediment loss by an individual runoff event on the three watersheds was caused by different precipitation events. The precipitation event on 27 Mar. 1998 caused the greatest loss on agroforestry and contour strip treatments while the precipitation event on 15 Apr. 1999 caused the greatest loss on the control watershed. The second and third largest sediment losses on the agroforestry and contour strip watersheds, respectively, occurred on 2 June 1998. The large sediment losses on the treatment watersheds in the beginning of the treatment period indicate that grass and tree components of these two watersheds were not well enough established to be effective in controlling sediment loss. As time progressed and the buffers became better established, treatment effects significantly improved. Results from these initial three years of the study show that grass strips without trees were more effective than grass strips with trees (agroforestry).
Historical data shows that grass buffer strips improve infiltration and trap sediment. In France, 5.7- and 11.1-m-wide grass filter strips reduced suspended sediment in runoff by 69 to 90% and 69 to 97%, respectively (Patty et al., 1997). Studying soil erosion on corn, wheat, and meadow watersheds, Edwards and Owens (1991) observed 92, 8, and 0% soil loss on their watersheds, respectively. During the third year of this study soil loss was reduced by 19% on the contour strip watershed. We attribute these reductions mainly to the grass strips. However, as much as 87 and 93% reductions of sediment in runoff on grass strips and grass–shrub–tree strips, respectively, have been reported by Schmitt et al. (1999). Research has also shown that most of the sediment and nutrients are trapped within the first 4 to 7.5 m of the strip and, thereafter, increasing width results in marginal improvements in retention (Schmitt et al., 1999). Our study design consists of several grass strips and trees within grass strips and, therefore, we anticipate higher percentage reductions in sediment as runoff travels through these buffers.
Total Phosphorus (TP) Loss
During the three-year treatment period, the control, agroforestry, and contour strip treatments lost 3.11, 2.41, and 3.26 kg TP ha-1. The predicted losses based on calibration relationships for the agroforestry and contour strip were 2.89 and 3.53 kg TP ha-1, respectively. Treatments have reduced TP loss by 17 and 8% on the agroforestry and contour strip watersheds, respectively.
Annual total phosphorus loss from the control, agroforestry, and contour strip watersheds averaged 1.0, 0.8, and 1.1 kg TP ha-1 during the three-year treatment period. However, TP reductions generally did not begin to occur until fall 1998 (Fig. 4) . No significant reductions in TP loss were noticed for either treatment watershed in 1997. The agroforestry watershed showed greater reductions (18%) than the contour strip treatment (3.7%) in 1998. In fact, most of the reductions in TP loss occurred in 1998 when watersheds produced the greatest number of runoff events during the treatment period. During 1999, the agroforestry and contour strip treatments produced 14 and 26% reductions, respectively, in TP loss based on calibration relationships. The contour strip watershed indicated a marked improvement in reduction of P in runoff from 1998 to 1999. In contrast, the agroforestry watershed did not indicate a similar improvement. However, it appears that loss of P in runoff varies from year to year and from watershed to watershed.
|
The largest TP losses on each watershed were associated with heavier precipitation events and open ground conditions. The greatest losses occurred on 4 and 29 Oct. and 9 Nov. 1998 on the three watersheds. These losses occurred after crops had been harvested and above-average fall precipitation occurred. The largest single TP loss accounted for 19 to 22% of the total P loss from the watersheds during the three-year treatment period. Since the heaviest rainfall events result in the greater P losses, grass buffers with or without trees are especially important in reducing TP losses from agricultural fields and their subsequent effect on water quality. Permanent vegetation such as grass and trees actively transpires water, increases infiltration, and traps sediment, thereby reducing TP loss.
In findings reported by Dillaha et al. (1989), orchard grass filter strips of 9.1 and 4.6 m in width removed 79 and 61% P in runoff, respectively. Strips of varying widths have been shown to remove 22 to 89% TP (Sharpley et al., 1992; Patty et al., 1997; Schmitt et al., 1999). Wider strips compared with narrower strips and more established buffers compared with newer ones are normally regarded as being more effective. The reduced TP loss from treatment watersheds in our study may have resulted from greater infiltration and less interaction of runoff with surface soil. Surface coverage in the filter strips would logically reduce runoff and enhance settling of material from the runoff. Also, roots of the vegetation should effectively remove portions of the inorganic P in the soil solution through uptake, causing increased P adsorption capacity of soils and increased P retention (Lyons et al., 1998). Subsurface losses of agricultural P in some soils are important (Gburek and Sharpley, 1998) and, therefore, management strategies should consider minimization of such losses (Sims et al., 1998). However, in the northern Missouri soybean–corn areas, most soils have a restrictive argillic horizon, which inhibits deep subsurface flow and promotes copious runoff. Our study demonstrates that contour grass strips and agroforestry treatments can be used to effectively control TP in runoff from row-cropped fields.
Total Nitrogen (TN) Loss
During the treatment period, the control, agroforestry, and contour strip watersheds lost 11.3, 10.1, and 9.7 kg TN ha-1, respectively. On average, control, agroforestry, and contour strip watersheds lost 3.77, 3.37, and 3.24 kg TN ha-1 per year. Compared with predicted losses based on calibrations, TN loss was not reduced in 1997 or 1998 but small reductions occurred in 1999. During the 1998 and 1999 sampling years, TN loss was reduced for only three storms on the agroforestry watershed. However, on the contour strip watershed, TN loss was generally reduced in October 1998 and thereafter (Fig. 5)
. The storm on 4 Apr. 1999 had the greatest reduction in TN loss associated with the treatments. On this day, the agroforestry and contour strip treatments reduced loss by 21 and 20%, respectively, compared with their predicted discharges.
|
The largest two and three TN losses accounted for 42 and 50%, respectively, of the total loss averaged over the three watersheds. This finding is consistent with those from other studies in which a few large storms have been shown to account for most of the annual loss (Edwards and Owens, 1991). These losses occurred early in the second year of treatments, with the largest loss being associated with the first precipitation event in the second year and the second largest associated with the second event. While additional data are required to determine if TN losses resulting from large precipitation events can be effectively reduced by contour strip and agroforestry treatments, our data suggest a trend toward reduced discharges following the initial establishment period for the strips and trees.
The presence of stem, thatch, and roots effectively slows runoff, promotes sediment removal, and increases uptake of N and other nutrients. Other studies have also shown decreased nitrogen loss due to grass strips and agroforestry (Dillaha et al., 1989; Lowrance et al., 1984; Schmitt et al., 1999). Plant uptake is the single most important process that removes N during the growing season (Lowrance et al., 1984), while dentrification is the dominant process in the winter months (Gilliam, 1994). On our treatment watersheds, contour grass strips were established in June 1997. Treatment effects during the first year were not observed. However, during the second year, especially after October, trends were observed that suggest both treatments have the potential to reduce TN loss from the site (Fig. 5).
Burwell et al. (1976) reported that in northwestern Missouri, subsurface discharge of NO3–N accounted for up to 84% of the total annual stream discharge. Comparing flow components on two watersheds, one with riparian buffer strips and the second with grass filter strips (in northeastern Missouri), Schmitt (1999) found that shallow subsurface and surface flow were the primary storm flow processes for each. He attributed the differences to vegetation, infiltration, flow resistance, antecedent moisture, and season. The location of the grass strip, number of strips, and the width of the strip appear to influence proportion of surface and subsurface losses. In our study, each treatment watershed has several grass strips parallel to the contours. Therefore, we would expect greater reductions in nutrient losses due to flow resistance, interception, infiltration, and vegetation uptake.
In riparian studies, subsurface water appears to be the dominant pathway of nitrate flux between croplands and riparian forests (Peterjohn and Correll, 1984). In a 208-ha forested watershed in Pennsylvania, subsurface flow accounted for more than 95% of storm flow (DeWalle et al., 1988). In an area where intact permanent vegetation with widely distributed root systems exists, one might expect it to be capable of taking up most of the subsurface nutrients before they leave the watershed. Reduction in losses results from a combination of reduction in runoff and utilization of nutrients by the vegetation. Although our watersheds are dominated by surface flow discharge, we anticipate that the subsurface flow component on the agroforestry watershed may account for a greater portion of storm flow as the trees grow (their current age is three years) and exert their influence on filtration and uptake.
Nitrate Loss
The measured loss from the control, agroforestry, and contour strip watersheds averaged 1.9, 1.8, and 1.5 kg ha-1 yr-1 nitrate N, respectively, during the first three years of treatment. Only one runoff event occurred in 1997, after treatments were established. However, during the 1998 sampling year, 11 runoff events occurred on the three watersheds that produced samples (Fig. 6)
. The control, agroforestry, and contour strip treatments lost 4.35, 4.43, and 3.17 kg ha-1 yr-1 nitrate N, respectively, in 1998. The contour strip treatment had a 0.15 kg ha-1 reduction in nitrate loss while the agroforestry treatment lost 1.92 kg ha-1 more than the predicted losses. The reduction in nitrate loss on the contour strip represents 4% of the predicted loss.
|
The largest single nitrate loss occurred on 28 June 1998 (2.57, 3.69, and 2.37 kg ha-1 loss on the control, agroforestry, and contour strip watersheds, respectively). This accounted for 46, 70, and 52% of the total loss during the treatment period for the control, agroforestry, and contour strip treatments. Again, timing of fertilizer application, land preparation, and heavy precipitation are believed to be related to these losses. This largest single loss was associated with 2.2 and 0.41 kg ha-1 more nitrate N loss on the agroforestry and contour strip plots than was predicted. With the exclusion of the nitrate N loss during this one event, agroforestry and contour strip treatments show a 0.13 kg ha-1 nitrate reduction over the treatment period. The second largest loss on treatment watersheds occurred soon after treatments were established (22 June 1997). On the control treatment, it occurred on 8 Apr. 1999. During this period heavy precipitation occurred when the ground was fully recharged and free of vegetation. Nitrate loss that occurred on our watersheds suggests the importance of nutrient management, timing of fertilizer application, ground conditions, and precipitation. Heavy precipitation events are especially important as they accounted for the greater percentage of the total loss of nitrate in our study.
Our results indicate that during the first year of our treatment period, soon after treatments were established, grass strips and agroforestry treatments were ineffective in controlling nitrate loss. However, during the second year, the treatments began showing an effect. If the largest single loss on both the agroforestry and contour strip plots was excluded, they would have had 28 and 41% reductions in nitrate N losses, respectively, during the second year. The benefits from having grass strips and trees became even more apparent in the third year of sampling (Fig. 6). Furthermore, during the early part of 1999, when heavy rains occurred and the ground was fully charged, a positive effect was found from having the grass strips and trees. This occurred in spite of the existence of the worst combination of conditions on the site. Even though our buffers are only three years old, they are already showing good potential as a management practice to reduce sediment and nutrient losses from row-cropped fields. With the exclusion of the largest single loss, agroforestry and contour strip treatments reduced nitrate N loss by 26 and 39%, respectively, on average during the second and third years of treatments.
Ammonium Loss
During the treatment period, the control, agroforestry, and contour strip treatments lost 0.5, 0.3, and 0.4 kg ha-1 ammonium, respectively. Based on calibrations, the contour strip treatment reduced ammonium loss by 32% during the treatment period. Only one runoff event produced samples in 1997 after treatments were established (Fig. 7)
. During this event, the treatments lost more ammonium than the control. The contour strip treatment had a 40% reduction in ammonium loss in 1998 while the agroforestry lost more than that predicted. In 1999, the contour strip watershed recorded a 33% reduction in ammonium loss. In contrast, the agroforestry treatment consistently lost more than the predicted loss in 1997 and 1998. Out of those 17 sampling events, only three events in 1998 and the first event of 1999 produced less than the predicted loss on the agroforestry treatment. However, it also began to show declining losses in ammonium.
|
Treatments have reduced ammonium loss on row-cropped watersheds, especially during large catastrophic events. To date, our results have not shown that trees in a grass strip are more effective in reducing ammonium loss compared with grass strips alone. However, Fig. 7 indicates that reductions in ammonium loss occur on both watersheds with time. During the first three years of treatment, the grass strip treatment appeared to be more effective in controlling ammonium loss from row-cropped watersheds. However, we anticipate greater reductions from the agroforestry watershed as trees occupy greater soil volumes and grow larger.
|
SUMMARY AND CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
In this study, we examined the agroforestry and contour strip effects on runoff, sediment, and nutrient loss reductions on corn–soybean rotations. The agroforestry treatment, after only three years, reduced runoff and total phosphorus losses by 1 and 17% based on calibration relationships. The contour grass strip treatment reduced runoff and total phosphorus losses by 10 and 8%, respectively. Most reductions occurred in second and third years after treatment establishment, as the vegetation cover increased and roots of the vegetation occupied more soil volume.
Extreme precipitation events were found to contribute significantly to the export of nutrients and runoff. The largest three runoff events accounted for 36% of the total runoff during the treatment period. The contour strip and agroforestry treatments reduced the runoff of the largest three events by 10 and 11%, respectively. The largest single TP and TN losses accounted for 19 to 22% and 25 to 41% of the total on the three watersheds. Our results clearly indicate that agroforestry and contour strip practices, when incorporated directly into corn–soybean watersheds in the Midwest, can be used to effectively reduce runoff volume and sediment and nutrient loss. These decreases in runoff, sediment, and nutrients following treatment application are especially significant given the relatively small number of runoff events and the short time that treatments have been in place. In particular, the contribution made in reducing N and P loss should increase with tree growth on the agroforestry watershed.
Buffer strips can be used to control degradation of stream water quality from agricultural nonpoint source pollution. In particular, site-specific slope and precipitation factors should be factored in when determining the appropriate width and tree density of buffer strips. Until such design factors become widely studied, however, landowners and policymakers can effectively use buffers similar to those of our study to reduce NPS pollution significantly.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. P. Udawatta, C. J. Gantzer, S. H. Anderson, and H. E. Garrett Agroforestry and Grass Buffer Effects on Pore Characteristics Measured by High-Resolution X-ray Computed Tomography Soil Sci. Soc. Am. J., January 25, 2008; 72(2): 295 - 304. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Yihe Lu, Bojie Fu, Liding Chen, Guohua Liu, and Wei Wei Nutrient transport associated with water erosion: progress and prospect Progress in Physical Geography, December 1, 2007; 31(6): 607 - 620. [Abstract] [PDF]
|
|
|
|
|
|
R. P. Udawatta, S. H. Anderson, C. J. Gantzer, and H. E. Garrett Agroforestry and Grass Buffer Influence on Macropore Characteristics: A Computed Tomography Analysis Soil Sci. Soc. Am. J., August 22, 2006; 70(5): 1763 - 1773. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. L. Bishop, W. D. Hively, J. R. Stedinger, M. R. Rafferty, J. L. Lojpersberger, and J. A. Bloomfield Multivariate Analysis of Paired Watershed Data to Evaluate Agricultural Best Management Practice Effects on Stream Water Phosphorus J. Environ. Qual., May 11, 2005; 34(3): 1087 - 1101. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Seobi, S. H. Anderson, R. P. Udawatta, and C. J. Gantzer Influence of Grass and Agroforestry Buffer Strips on Soil Hydraulic Properties for an Albaqualf Soil Sci. Soc. Am. J., May 6, 2005; 69(3): 893 - 901. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. P. Udawatta, P. P. Motavalli, and H. E. Garrett Phosphorus Loss and Runoff Characteristics in Three Adjacent Agricultural Watersheds with Claypan Soils J. Environ. Qual., September 1, 2004; 33(5): 1709 - 1719. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||
