| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Agricultural and Biological Engineering, Purdue University, West Lafayette, IN 47907
b USDA-ARS National Soil Tilth Laboratory, Ames, IA 50011
c USDA-ARS Soil Drainage Research Unit, Columbus, OH 43210
* Corresponding author (kladivko{at}purdue.edu).
Received for publication January 2, 2004.
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Nitrate concentrations and mass losses in subsurface tile drains vary with soil organic matter level, yearly weather variations, fertilizer N rates and timing, drain spacing, cover crop growth, cash crop yield, and water table control practices (Baker and Johnson, 1981; Bergstrom, 1987; Drury et al., 1996; Kladivko et al., 1999; Gentry et al., 2000; Jaynes et al., 2001; Randall and Mulla, 2001; Dinnes et al., 2002). Noncontrollable factors such as precipitation and the mineralization of soil organic matter have a great impact on drainage volumes and nitrate loads (Randall and Goss, 2001), which therefore put limits on the concentrations and loads that can be achieved while using the land for row-crop agriculture. Research on the controllable factors of fertilizer, crop, and drainage system management practices must be done within the larger context of climate and soil organic matter content and ideally over a long enough time period to encompass a typical range of weather variation. There are relatively few long-term (>10 yr) studies of nitrate leaching on a field scale. Our 15-yr study provides an important data set for assessments of nitrate leaching in the Mississippi River basin. Our site is on a low organic matter (approximately 1.3%), loess-derived silt loam soil in southeastern Indiana, which contrasts with the high organic matter soils of most of the drainage studies in Iowa, Minnesota, and Illinois. By comparing results from different soils and climatic zones within the Midwest, scientists and policymakers will hopefully gain greater understanding of the challenges to reducing nitrate loads to subsurface drains.
The objectives of our study were to (i) evaluate the effect of three different drain spacings on water flow and nitrate leaching into subsurface drains over a 15-yr period and (ii) measure changes in nitrate leaching that would result from first converting from conventional monoculture corn with high N fertilizer rates to the same cropping system with lower N rates, and then to a no-till corn–soybean rotation with lower fertilizer N rates and a winter "trap crop."
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Corn was planted each year from 1984 (one year before nitrate N measurements began) through 1993, using conventional tillage (chisel plow to a 20- to 25-cm depth in spring, followed by two passes with a disc or field cultivator). In 1994 a no-till, soybean–corn rotation was begun, with the addition of a winter wheat (Triticum aestivum L.) cover crop after corn as a "trap crop" for N in the soil profile. Fertilizer N rates were gradually reduced during the course of the 16-yr experiment, as new knowledge became available and fertilizer rate "philosophy" changed within the Purdue extension recommendations. Preplant fertilizer N rates were 285 kg N ha–1 for the first 5 yr of monoculture corn, 228 kg N ha–1 for the last 5 yr of monoculture corn, 200 kg N ha–1 in 1995, and 177 kg N ha–1 in 1997 and 1999, all preplant-applied as anhydrous ammonia. The nitrification inhibitor nitrapyrin [2-chloro-6-(trichloromethyl) pyridine] was used at a rate of 0.56 kg a.i. ha–1 with the anhydrous ammonia applications from 1984 through 1995 (excluding 1994 as soybean was grown and no N was applied in that year). A small amount (8–28 kg N ha–1) of "starter" fertilizer N (as 18–15–0 of N–P–K in 1984–1989, and 19–7–0 liquid blend starting in 1990) was also applied during the planting operation for corn. Table 1 lists field management practices and crop yields for the 15-yr period of nitrate N measurements.
|
Flow
To test for differences in flow among drain spacings, values for annual flow were calculated for each drain. In each case, the annual drainflow record was reasonably stationary (without a long-term trend). Because the differences in flow between the two blocks were large, a simple paired observation comparison (t test) was used to test for significant differences in flow within each pair of spacings within each block (e.g., Steel and Torrie, 1980).
Concentration
To model the trend in each drain's concentration record over the 15-yr period, annual flow weighted mean concentrations (C in mg N L–1) were calculated for each drain. Based on Emerson and Stoto's test (e.g., Chapter 4 in Hoagland et al., 1983), the concentration data were reasonably characterized by a lognormal distribution; hence the data were log-transformed. Time response (repeated measures) models were developed for the concentration vs. time behavior of each individual drain. At least two model forms including polynomial regressions were considered, but spline forms with segmented polynomials and/or rational polynomials were selected (e.g., Jaynes et al., 2001, 2004). A segmented polynomial is a function defined by a simple polynomial on each of two or more separate intervals of the independent variable (here, the time in years), and a rational polynomial is the quotient of two simple polynomials (Rivlin, 1969; Kimball, 1976). Robust regression methods using Ramsay's Ea robust weight, with a = 0.3 were employed in an iteratively reweighted nonlinear least squares regression procedure. Model selection and comparison were based on multiple performance statistics (including mean square error, prediction error sum of squares [PRESS] and their corresponding coefficients of determination, variance ratio, and Akaike Information Criterion [AIC]) and residual diagnostics (including autocorrelation tests). The final selected form of the regression was:
 |
 |
Load
Two different types of analyses were used to assess the nitrate N loads over the 15-yr period. The first analysis was used to model the trend with time in nitrate N load for each drain, in a similar procedure as described for the concentration vs. time regressions. Annual total nitrate N load values (L in kg N ha–1) were calculated for each drain. The 1985 data were excluded because the record was incomplete (9 mo only). Load values were rescaled with the square root transformation to improve normality. The final selected model form was:
 |
 |
The second type of analysis tested for significant differences in loads between drain spacings within each block. This was done by fitting a regression model to the measured differences in annual loads within each pair of drains. The 95% confidence interval of the fitted difference for each year was then calculated, and for periods when the confidence interval did not include zero, the differences between the two drains were considered statistically significant.
Seasonal Patterns
To test whether concentrations varied consistently on a seasonal basis within years, several types of analyses were conducted. For each drain record, monthly flow-weighted mean concentrations were log-transformed for the same reasons discussed for the annual concentrations. The data for each drain were then examined for seasonal differences with several methods including graphical exploratory data analysis and ANOVA procedures that considered both long-term trend and autocorrelation structure. Unfortunately, diagnostics revealed violations of the usual error structure assumptions even after simple corrective procedures were applied (e.g., Littell et al., 1996). Because the purpose of the seasonal analysis was simply and briefly to compare with general trends found by other researchers, more complicated tests were not performed and exact probabilities are not reported. Instead, the tendency for the occurrence of seasonal patterns in concentration will be summarized.
Concentration versus Flow
To test whether there was a relationship between concentrations and flow rate, several different types of analyses were performed. Data for these analyses were from Tuesday through Friday only, since Saturday through Monday data represented a flow-weighted concentration from three days, rather than one day. Simple correlations of daily nitrate N concentration vs. daily flow were performed for each drain, for the overall record and by year, month, and quarter. An overall correlation for all drains and years was also calculated. In addition, a frequency analysis was performed to test for changes in concentration vs. changes in flow rate. Both flow and concentration data were grouped by whether they had increased or decreased since the previous measurement point. Both frequency analysis (e.g., Steel and Torrie, 1980) and simple correlation tests were performed with the resulting data set.
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Hydrology
Annual rainfall over the 15-yr period ranged from a low of 800 mm for 1987 to a high of 1366 mm in 1995 (Table 2), with the 15-yr average (1118 mm) being nearly equal to the 30-yr "normal." Drainflow volumes varied among years as a result of the differences in annual rainfall and the timing of the rainfall within each year. Drainflow per unit area decreased as drain spacing became wider, as expected (Table 2). Averaged across both blocks, drainflow varied from a low of 6.7 cm (8% of annual rainfall) for the 20-m spacing in 1987, which was the driest year in the study (800 mm rainfall), to a high of 32.5 cm (26% of annual rainfall) for the 5-m spacing in 1985. A similarly wide range of drainflows between 0 and 40% of annual rainfall has been found across a range of studies in the humid regions of North America (Kladivko et al., 2001).
|
|
0.05) in drainflow in all drain pairs (5 vs. 10, 10 vs. 20, 5 vs. 20 m) within each block, when considering the 15-yr period as a whole. Differences in flow between drain spacings occurred in most of the 15 yr for all drain pairs except the 10- vs. 20-m east plots, which had five years with similar drainflows and the first year of the study with greater flow in the 20-m plot than in the 10-m plot. Drainflow as a percent of annual rainfall ("drainage efficiency") was calculated for each year and each drain and averaged over the various time periods of interest. Averaged over the 15-yr period, drainage efficiencies were 20.6, 14.9, and 12.1% for the 5-, 10-, and 20-m spacings, respectively. A comparison was made between the two different cropping systems used in different time periods of the study, namely continuous corn with chisel tillage in 1985–1993 vs. the 1994–1999 system of soybean–corn rotation with no-till and a winter wheat cover crop after corn. Due to the lack of on-site rainfall data during the 1991–1995 and 1998–1999 time periods, however, the calculated drainage efficiency values for those years are not as certain as the other time periods. The actual measured drainflow (cm) is not affected by the inaccuracies of the rainfall data and is discussed at the same time.
Average annual drainage efficiencies were 20.4, 13.7, and 11.2% (23.3, 15.6, and 12.8 cm flow) for the 5-, 10-, and 20-m spacings, respectively, for the 1985–1993 continuous corn years, and 20.8, 16.5, and 13.2% (22.8, 18.1, and 14.5 cm flow), respectively, during the 1994–1999 soybean–corn years. The data suggest that the drainage efficiency did not change over time for the 5-m spacing, but that the 10- and 20-m spacings showed an increased efficiency during the later time period. This increased efficiency for the wider spacings may reflect a maturation of the drainage system over time, with flow paths to the drains developing from greater distances over the years after drain tiles were installed. Detailed analysis of individual storm hydrographs would be needed to further evaluate this hypothesis and is beyond the scope of this paper. Because the 5-m spacing showed no evidence for a change in drainage efficiency with time, it suggests that the changes for the 10- and 20-m plots were not due to evapotranspiration differences resulting from the new cropping system. We cannot separate the effects of changes in tillage (chisel to no-till), crop (continuous corn to soybean–corn rotation), and winter cover crop (none to winter wheat) in this experiment, but the combination of changes apparently caused no major change in drainage efficiency for this field.
Within the six years of soybean–corn rotation, there were no apparent trends in drainflow or drainage efficiency between the corn years and the soybean years. Each year had a different combination of rainfall timing, intensity, and amount, and so it was not possible to discern any consistent effects between the two crops because they were not both grown in the same year.
Nitrate Nitrogen Concentrations
Nitrate N concentrations in drainflow decreased considerably over the 15-yr period. Concentrations were consistently in the 20 to 30 mg L–1 range in the 1985 to 1988 period and in the 7 to 10 mg L–1 range in the 1996 to 1999 period (Table 3). Concentrations did not vary with block or drain spacing.
|
0.05) between spacings within each block, in any of the three regression parameters. Thus, the conclusion is that drain spacing does not affect the rate of change of concentrations resulting from a field management change, within the range of spacings studied here.
|
|
The 71% decrease in nitrate N concentrations from the 1986 to 1988 period (28 mg L–1) to the 1997 to 1999 period (8 mg L–1) is probably a result of a combination of the management practice effects along with yearly weather and crop yield variations. Reduction of fertilizer N rates (Baker and Johnson, 1981; Randall and Goss, 2001; Jaynes et al., 2004) and the growth of a winter "trap crop" after corn (Owens et al., 1995; Dinnes et al., 2002; Strock et al., 2004) have both been shown to reduce nitrate N concentration in drainage and are probably the major factors in our study. A corn–soybean rotation compared with continuous corn (Randall et al., 1997) or the soybean phase compared with the corn phase of a corn–soybean rotation (Jaynes et al., 2001; Randall et al., 2003) have sometimes shown lower nitrate N concentrations and may also be a factor at our site. Conversion from spring chisel tillage to no-till on this field probably had minor influence on nitrate leaching, since most of the annual drainage during the "tilled" years occurred before the tillage operation was performed. Tillage systems have generally been found to not have much impact on nitrate losses except for fall tillage followed by warm and wet conditions (Randall and Goss, 2001).
In addition to management practice changes, the weather and resulting crop yields had an impact on year-to-year variations in nitrate N concentrations (and load, as discussed later). During the first 5 yr of the drainage study (beginning in 1984, one year before the nitrate measurements began), preplant fertilizer N rates were 285 kg ha–1, which was the recommended application rate at that time for a yield goal of 12.5 Mg ha–1. Several years of poor crop yields (1986, 1988; see Table 1) probably resulted in high residual soil N and contributed to the increasing trend in concentrations over the 1985 to 1989 period. Preplant fertilizer N rates were reduced from 285 to 228 kg N ha–1 in the 5-yr period from 1989 to 1993, and concentrations started to show a decrease in 1990, in the first "flow season" after the reduction in fertilizer application. A rise in concentrations in 1992 probably reflects the poor crop yield in 1991, but concentrations decreased again in 1993 following a high crop yield in 1992. These relatively rapid responses to low crop yield, during the next flow season, have also been observed by others (Randall and Mulla, 2001). The 1994 change to a soybean–corn rotation and lower fertilizer N rates for the corn did not result in an immediate decrease in concentration, but by 1996 the concentrations had declined again. The lower concentrations are probably a result of both the winter wheat "trap crop" after the corn and the lower fertilizer N rates.
The precise fertilizer N rate and crop management system needed for optimal crop growth and environmental quality is region specific and varies from year to year, and this remains a major challenge for agriculturalists worldwide. The results from this site in southeastern Indiana suggest that it is possible to decrease average annual nitrate N concentrations to below 10 mg L–1 (drinking water standard), while still growing corn and soybean, on soils similar to the Clermont silt loam. The Clermont has a much lower organic matter content (approximately 1.3%) than many of the more productive, high organic matter (approximately 4–5%) soils of the U.S. Midwest. Research on higher organic matter soils suggests that it may not be possible to grow corn and soybean on those soils while consistently maintaining concentrations below 10 mg L–1, even with the best management practices currently available (Jaynes et al., 2001; Randall and Mulla, 2001). The results of all these types of studies underscore the necessity for long-term field experiments in different regions and on different soils, to understand the impacts of yearly weather variations, long-term climate, soils, and management on nitrate N leaching.
Nitrate Nitrogen Load
Annual nitrate N loads to drainage water decreased significantly over the 15-yr experimental period (Table 5), due to the large decrease in nitrate N concentrations over the same time period. Annual nitrate N loads averaged 38 kg N ha–1 in the 1986 to 1988 period and 15 kg N ha–1 in the 1997 to 1999 period. This 60% reduction in load occurred in spite of the fact that drainflow was 29% greater in the 1997 to 1999 period (18.4 cm) than in the 1986 to 1988 period (14.3 cm). The 71% decrease in concentrations, from 28 mg N L–1 in the 1986 to 1988 period, to 8 mg N L–1 in the 1997 to 1999 period, resulted in a large decrease in loads even with a moderate increase in flow in those years. Regression analysis was used to test the significance of the decline in loads with time, in a similar manner as was done for concentrations. The 95% confidence interval of the regression slopes did not include zero, again suggesting a significant decrease in loads over the last 10 to 11 yr.
In addition to the long-term trends in nitrate N loads, year-to-year variations in loads occurred as a result of variation in weather and crop yields. Loads were particularly high in 1989 after the low corn yields in the 1988 drought year. The higher residual nitrate that was probably remaining in the soil profile in autumn 1988, coupled with the high drainflow volumes in 1989, led to the highest nitrate N loads of the 15-yr study. Logan et al. (1994) and Randall et al. (1997) have also found greater N losses to drains in years following a drought due to greater residual N in the soil profile.
Nitrate N mass loads per unit area varied similarly to drainflow volumes, with a tendency for greater losses from the narrower spacings (Table 5). Annual nitrate N losses were 50, 37, and 27 kg ha–1 for the 5-, 10-, and 20-m spacings, respectively, in the 1986–1988 period and 16, 16, and 13 kg ha–1 in the 1997–1999 period. The differences in loads between spacings were significant in some but not all years. As described in the Materials and Methods section, years with significant load differences were identified by fitting a regression to the difference in loads within each pair of drains within each block. When the 95% (or 90%) confidence interval of the modeled difference for a given year did not include zero, the difference in loads between the two drains was considered to be statistically significant. The pair of drains with the greatest number of years of significant differences in modeled N loads was the 5- and 20-m east plots, and the regression analysis is illustrated in Fig. 3. The figure shows significant (P 0.05) differences between the two spacings in the 1986 through 1993 period. Significant differences between drains were more common in the east block than in the west block. Significant periods of difference for all pairs of drains at both the P 0.05 and the P 0.10 levels are shown in Table 5.
|
Seasonal Effects on Drainflow, Loads, and Concentrations
Loads to the drains exhibit a clear seasonal cycle related to the timing of drainflow in this system. The majority of the drainflow and nitrate N loads occur during the fallow season of November through March, as discussed in previous reports about this site (Kladivko et al., 1991, 1999). Rainfall is relatively uniformly distributed throughout the year, with the driest month (October) receiving 6.0% of the annual rainfall and the wettest two months (July and April) receiving 10.5 and 10.4%, respectively (Fig. 4). However, drainflow varies much more over the year, due to higher evapotranspiration during the growing season. Drains typically begin to flow in November or December at this site, after the soil profile has rewetted following the growing season, although in some years there is no flow until January. Drainflow continues throughout the winter during most years, although in some years the drainflow ceased during parts of January or February. This site thus differs from sites further north in the Midwest, where flow ceases most of the winter and occurs primarily in April through June (Randall and Goss, 2001). Flow at this site usually ceases in May or early June, but occasionally there will be some small flow events in July through October. The 14-yr (1985 was omitted because it was a partial year for nitrate analyses) distribution of the fraction of flow and nitrate N load occurring in each month is shown in Fig. 4 for the 20-m west drain. In climates where the soil does not freeze all winter, most of the drainage and nitrogen loss occurs in the late fall, winter, and early spring (Goss et al., 1993; Drury et al., 1996; Gilliam et al., 1999). The differences in the seasonal distribution of flow must be considered when comparing studies from different regions and when designing and evaluating management strategies for decreasing nitrate loads to drainage water.
|
At our site, the tendency for slightly higher nitrate concentrations in May–June–July has minimal impact when considering total nitrate loads for the year. On average only 8.2% of the annual load and 7.1% of the annual flow occurred in June. Both flow and load were minimal in July through October. On average 64% of the annual flow and 63% of the annual nitrate load occurred in the fallow season of November through March. Another 17% of the flow and 15% of the nitrate load occurred in April, much of which was before any field operations for the next crop. Thus, nearly 80% of the annual drainflow and annual nitrate load occurs in the winter and early spring before fertilization for the next crop. These results underscore the potential importance of cropping systems that would use some of the nitrogen and water during late fall or early spring, such as winter cover crops or perennials that grow later in the fall and earlier in the spring than the typical corn–soybean rotation. Perennial crops generally produce much lower nitrate N concentrations and loads in drainflow than annual row crop systems, but lack of a suitable market limits widespread adoption of such systems (Randall and Mulla, 2001). Winter cover crops may be able to provide some of the benefits of perennials in the midst of an annual cropping system, if sufficient growth of the cover crop can be obtained under the cool conditions of the Midwest.
Concentrations versus Flow Rates
Exploratory graphical analysis of concentration and flow data revealed some periods where concentration declined rapidly as the flow increased rapidly. An example from the 20-m west plot in 1998 (Fig. 5) shows that on Days 99, 106, 120, and 127, the flow peaked while the concentration dropped. This behavior is opposite of the typical preferential flow behavior of pesticides (Kladivko et al., 1991, 1999) or newly applied tracers (Kung et al., 2000), which tend to move with the water in the preferential flow paths and have higher concentrations as the hydrograph rises. Nitrate and other chemicals that are well distributed in the soil matrix tend to move more by matrix flow, however, and therefore drain concentrations are diluted by the relatively clean water flowing in the preferential pathways (Smettem et al., 1983; Hallberg et al., 1986). This behavior does not occur consistently in our drains, however, and we do not know why it occurs sometimes but not all the time.
|
0.01) but weak negative correlation of –0.0614 between daily concentration and daily flow rate (Saturday through Monday excluded; see the Materials and Methods section). Correlations run on subsets of the data found significant negative correlations for some years, months, or quarters in each drain with an occasional positive correlation. The frequency analysis of change in concentration vs. change in flow highlights changes within events and also gave a similar general result, with a significance level of P = 0.012 across all drains. These weak relationships were similar to the results of Jaynes et al. (1999), who found a weak negative correlation between change in nitrate N concentration and change in flow in a stream site in Iowa but inconsistent trends for the tile drains. |
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Current concerns about hypoxia in the Gulf of Mexico have focused attention on nitrate N loads from tile-drained soils of the U.S. Midwest. As researchers and policymakers evaluate the current nitrate N loads from these systems and explore options for reducing the loads, past and current studies from different Midwest locations should be carefully compared and contrasted. Some key points from our current study that should be kept in context when comparing results across the region are highlighted here:
Additional research comparing the low organic matter soils represented here and the high organic matter soils of much of the U.S. Midwest is essential for understanding the system and designing appropriate management practices for different regions.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Related articles in JEQ:
This article has been cited by other articles:
|
|
 |
|
  D. B. Jaynes, T. C. Kaspar, T. B. Moorman, and T. B. Parkin In Situ Bioreactors and Deep Drain-Pipe Installation to Reduce Nitrate Losses in Artificially Drained Fields J. Environ. Qual., March 1, 2008; 37(2): 429 - 436. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. B. McDonald, J. W. Singer, and M. H. Wiedenhoeft Establishment and Growth of Self-Seeded Winter Cereal Cover Crops in a Soybean-Corn Rotation Agron. J., February 29, 2008; 100(2): 432 - 439. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. B. McDonald, J. W. Singer, and M. H. Wiedenhoeft Self-Seeded Cereal Cover Crop Effects on Interspecific Competition with Corn Agron. J., February 29, 2008; 100(2): 440 - 446. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T.C. Kaspar, D.B. Jaynes, T.B. Parkin, and T.B. Moorman Rye Cover Crop and Gamagrass Strip Effects on NO3 Concentration and Load in Tile Drainage J. Environ. Qual., August 31, 2007; 36(5): 1503 - 1511. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. D. Nance, L. R. Gibson, and D. L. Karlen Soil Profile Nitrate Response to Nitrogen Fertilization of Winter Triticale Soil Sci. Soc. Am. J., June 29, 2007; 71(4): 1343 - 1351. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. S. Algoazany, P. K. Kalita, G. F. Czapar, and J. K. Mitchell Phosphorus Transport through Subsurface Drainage and Surface Runoff from a Flat Watershed in East Central Illinois, USA J. Environ. Qual., April 5, 2007; 36(3): 681 - 693. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. W. Stone and J. T. Wilson Preferential Flow Estimates to an Agricultural Tile Drain with Implications for Glyphosate Transport J. Environ. Qual., August 9, 2006; 35(5): 1825 - 1835. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. R. Ball Coelho, R. C. Roy, and A. J. Bruin Nitrogen Recovery and Partitioning with Different Rates and Methods of Sidedressed Manure Soil Sci. Soc. Am. J., February 2, 2006; 70(2): 464 - 473. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. J. Kladivko, G. L. Willoughby, and J. B. Santini Corn Growth and Yield Response to Subsurface Drain Spacing on Clermont Silt Loam Soil Agron. J., September 19, 2005; 97(5): 1419 - 1428. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||
for flow-weighted annual mean nitrate N concentrations vs. time for all six drains.
