| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a USDA-ARS Southwest Watershed Research Center, 2000 East Allen Road, Tucson, AZ 85719
b Department of Soil Water and Climate, University of Minnesota, 1991 Upper Buford Circle, St. Paul, MN 55108
c Southwest Research and Outreach Center, 23669 130th Street, Lamberton, MN 56152
* Corresponding author (dthoma{at}tucson.ars.ag.gov)
Received for publication May 20, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: DMRP, dissolved molybdate reactive phosphorus • TP, total phosphorus • TS, total solids
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
It has long been known that tile line effluent is a significant source of NO3–N loading to surface waters. Since the 1960s studies have shown nitrate concentrations commonly exceed the 10 mg L–1 NO3–N federal drinking water standard and are often related to fertilizer N addition (Gast et al., 1978; Baker and Johnson, 1981; Jaynes et al., 2001). Differences in climate, soil properties, and management affect the variability in pollutant concentrations and loadings from drainage effluent (Baker et al., 1975; Sims et al., 1998; Hanway and Laflen, 1974; Kladivko et al., 1991; Randall et al., 2000). It has also long been known that surface soil cover can reduce soil erosion and particulate-associated nonpoint-source pollution. The beneficial effects of 15 to 30% crop residue cover have been shown to reduce soil erosion by as much as 50 to 90% depending on precipitation, slope, and soil properties (Wischmeier, 1973; Ketcheson and Stonehouse, 1983).
Tillage studies in northern latitudes (>40°) have shown chisel plow–based tillage practices (approximately 30% residue cover) generally reduce yield in a continuous corn cropping system due to residue buildup that keeps soils from warming and drying in the spring. Yield reductions of 502 to 565 kg ha–1 were observed in multiyear studies on poorly drained soils under chisel plow compared with moldboard plow tillage systems. However, yield was greater for a chisel plow–based system on a well-drained soil (Randall et al., 1996).
In one of the few studies on flat landscapes, Ginting et al. (2000) showed that storm events large enough to cause ponding at the surface inlet allowed sufficient time for entrained particles to settle thus limiting sediment and sorbed P transport to surface waters. However, prolonged ponding caused P desorption from soil colloids resulting in higher concentrations of soluble P leaving fields. In a simulated 10-yr return interval rainfall, Ginting et al. (2003) reported relatively minor losses of sediment and both forms of P from flat landscapes regardless of surface residue cover.
The chemical forms of applied nutrients, the cropping and tillage management systems, and the climatic factors all interact to affect off-site delivery of pollutants. For instance, there is a concern that using manure as an N source adds P in excess of crop needs thus becoming more of a problem in surface inlet losses (Sims et al., 1998). Soluble nutrient availability to plants and in turn off-site transport also depends on the extent of manure or fertilizer mixing with the soil. Therefore, tillage and nutrient interactions are important in understanding nutrient losses from both surface inlet flow and tile drainage.
These factors illustrate the complexity of interactions that control pollutant losses via surface inlet flow and tile drainage on both steep and relatively flat landscapes. Since much of the Midwest is relatively flat, research is needed in such landscapes to determine if the conclusions drawn from studies on steeper slopes hold in flat landscapes as well. An area that is relatively flat in the upper Midwest is the Minnesota River basin. In this 3.86-million-ha watershed, 33 and 71% of the area is less than 2 and 6% slope, respectively (University of Minnesota, 2001). However, the Minnesota River is a major carrier of nonpoint-source pollution (sediment, N and P) from southwestern and south-central Minnesota to the Mississippi River. United States Geological Survey monitoring studies have shown that the annual suspended sediment load for the Minnesota River at Mankato, MN, has varied from 0.18 to 3.27 million Mg per year from 1968 to 1992 (Payne, 1994). Therefore, there is an increased interest in finding management practices that can reduce nonpoint source sediment pollution to the Minnesota River without significantly affecting crop yield.
The Minnesota Pollution Control Agency has estimated that a 40% reduction in sediment load is necessary to achieve federally mandated water quality goals in the lower Minnesota River. The agency has recommended the use of conservation tillage practices such as chisel plowing that leave about 30% residue cover at the soil surface (Minnesota River Assessment Project, 1994). This study was designed to quantify the effect of a conservation tillage practice (chisel plowing) vs. moldboard plowing with the addition of manure or commercial fertilizer (urea and triple superphosphate). The study focused on water, sediment, mineral N, and P losses in surface inlet flow and in tile drainage, and on corn yield from a flat landscape (<3% slope) in the Minnesota River basin. Soil and crop management in this study was intended to simulate practices of local producers.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
Instrumentation and Data Collection
The instrumentation was similar to that of Zhao et al. (2001a) with the exception of sump pumps installed to remove excess drainage water, and installation of Isco (Lincoln, NE) 3700 water samplers housed in weather-resistant plywood shelters. Surface inlet flow was measured with 3.6-L-capacity tipping buckets while tile flow was measured with 0.36-L tipping buckets. Inlet flow rates greater than 3.6 L min–1 initiated a signal pulse to the water sampler that drew a 100-mL sample volume from the surface inlet drain pipe. The surface inlet flow samplers were programmed to collect composite samples at 10- and 20-min periods for rain and snowmelt events, respectively. The samplers each contained 24, 1-L bottles, and were programmed to composite flow samples for 2 h per bottle during snowmelt runoff events and for 1 h per bottle for rain events. Tile line flow samples were collected by hand daily, Monday through Friday. Runoff and drainage data continuously recorded on a data logger were used to calculate flow rate, water volume, and pollutant load.
Surface inlet water samples were analyzed for sediment (TS) by drying and weighing, total phosphorus (TP) via perchloric acid digestion (USEPA, 1981), soluble P via dissolved molybdate reactive method (DMRP) (Wendt and Corey, 1980), and nitrate and ammonium concentrations conductimetrically (Carlson, 1978, 1986). Tile line water samples were only analyzed for mineral N (NH4–N and NO3–N). Manure samples were analyzed for total N using a modification of the Kjeldahl method (Bremner, 1986), where a heating block was used in lieu of a distillation apparatus. Total P in manure was measured via perchloric acid digestion (USEPA, 1981), whereas mineral N was determined conductimetrically (Carlson, 1978, 1986) after 2 M KCl extraction (Keeney and Nelson, 1982).
Soil cores (0- to 15-, 15- to 30-, 30- to 60-, and 60- to 90-cm depth) and manure samples were collected each fall before tillage and analyzed for nutrient management planning purposes. Soil mineral N was determined conductimetrically (Carlson, 1978, 1986) similar to analysis for manure samples. Soil P was determined using the Olsen P method (Olsen and Sommers, 1982). Soil pH was measured using a 1:1 soil to water mixture (McClean, 1986). In fall 2001, one surface (0- to 15-cm depth) bulk density measurement was made per plot at random locations using the excavation method (Blake and Hartge, 1986). Crop residue cover was measured after fall tillage and after spring planting using the line transect method described by Morrison et al. (1993). The plot slopes were measured in 2003 with a clinometer.
Site Characteristics
The study was conducted at the Southwest Research and Outreach Center near Lamberton, MN (44.2° N, 95.3° W). The soils at the experiment site were Webster–Canisteo clay loams (fine-loamy, mixed, superactive, mesic Typic Endoaquolls and fine-loamy, mixed, superactive, calcareous, mesic Typic Endoaquolls, respectively). Both are highly productive but poorly drained soils, developed in depressions from calcareous glacial till (Table 2). Surface slopes ranged from 1.5 to 5% with an average of 3% for the 16 plots. The surface slope of moldboard-plow plots (3.75%) was significantly (p = 0.003) greater than chisel-plow plots (2.25%). Average surface bulk density for the moldboard plow treatments was 1.32 Mg m–3 compared with 1.28 Mg m–3 for the chisel plow treatments and was not statistically different by treatment or interactions. Assuming a particle density of 2.65 Mg m–3, these bulk densities were equivalent to porosity of 50 and 52% for the moldboard plow and the chisel plow treatments, respectively.
|
|
Most of the runoff events in 1999, 2000, and 2002 were relatively small (a few tips of the tipping buckets per event). However, over the duration of the experiment, 10 rain events greater than 18 mm induced surface inlet flow large enough to trigger sample collection by the Isco samplers. The average depth of precipitation for these 10 flow-inducing events was 57 mm.
Statistical Analysis
Analysis of variance (ANOVA) of tillage, nutrient source, and their interactions on water quality parameters, crop yield, residue cover, and surface slope was performed using the MINITAB13 (Ryan et al., 2000) statistical package. Statistical significance was checked at the 0.1 probability level. Soil depth was used as a split assuming randomization in the analysis of soil N by depth.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Soil Nitrogen
There were no significant differences in soil NO3 or soil NH4 at any depth due to tillage treatments in any year of the study; hence data by tillage are not presented. However, there was a significant decrease (p < 0.10) in soil NO3 and soil NH4 with depth in most years (Table 4). This is expected because manure and urea fertilizer were applied in the top part of the soil profile. The only exception to this trend occurred in 2002 where there was no significant difference with depth for soil NH4.
|
Although under application of manure NO3–N in some crop years, 2000 and 2002, corresponds to statistically lower soil test NO3–N levels in manure plots it does not always appear to be the main factor causing lower soil NO3–N levels in manure plots. This is evident by noting that in the 2001 crop year NO3–N application rates for manure (167 kg ha–1) and urea (161 kg ha–1) were similar, but soil test NO3–N levels in manure plots were still much less than urea plots (0.9 and 3.0 mg kg–1 for manure and urea, respectively).
Soil Phosphorus
There was no statistical difference in soil P level after fall harvest for 1999 through 2002 by tillage treatment but there was a significantly (p < 0.005) higher soil P level in the plots treated with manure than in plots treated with triple superphosphate (Table 5). The 4-yr average soil P level was 24 and 12 mg kg–1 for manure- and urea-treated plots, respectively. Soil P levels did not appear to track P addition from either manure or triple superphosphate. The difficulty of applying the recommended amount of manure as a fertilizer source was echoed by Randall et al. (2000). It is likely that soil P levels in manure plots would have been even higher if manure application in 2000 and 2002 had been sufficient to meet crop N needs.
|
|
The 20% of average annual precipitation lost from the plots as a combination of surface inlet flow and tile flow (Table 6) falls between 22% annual combined surface inlet and tile losses observed by Zhao (1998) for 1996 on the same plots and 14% runoff for the Cottonwood watershed, in which the plots are located, based on river gauge data (Baker et al., 1979). In most years, tile flow exceeded surface inlet flow in spite of high soil clay content. This difference was due to low intensity storms, gentle slopes, strong soil aggregation, and macropore flow, as evidenced by turbid tile line flows during some events. The exception to this trend was 2002, when surface inlet flow exceeded tile flow. This may have been due to the high intensity storm events that occurred June through August that year. For the duration of the study water loss was split about 25 and 75% between surface inlet flow and tile drainage. This suggests that most precipitation events were low intensity storms and likely did not exceed the infiltration capacity of the soil.
Water Quality—Storm Event Basis
Tracking sediment and nutrient pollutant losses on a storm event basis for a single year revealed important physical, phenological, and climate processes that interacted to affect annual pollutant loading of surface waters. Four major rain events occurred in 2000 on 26 February (31 mm), 18 May (56 mm), 1 July (32 mm), and 9 August (36 mm) (Table 7). These events roughly corresponded to slightly thawed soil conditions, bare soil right after planting, medium canopy cover, and full canopy cover, respectively. Maximum rainfall intensity on 1 July and 9 August was less than 25 mm h–1. Rain intensities were not recorded on 26 February and 18 May. However, average rain intensity on 18 May at a station about 16 km (10 mi) from the experimental site was 3.8 mm h–1 for a period of 23 h.
|
After planting (18 May), there was a continuous decrease in both surface inlet flow and sediment loss for subsequent rain events (Table 7). This was because of the increase in canopy cover that protected soil from detachment and transport as well as due to increase in soil water storage capacity as the crop depleted soil water and provided storage capacity for subsequent rains. For the 18 May storm event, there was a significant tillage by nutrient source interaction effect on tile flow. Chisel plow–urea (2.3 cm) and moldboard plow–manure (1.5 cm) plots had greater drainage than chisel plow–manure (1.1 cm) or moldboard plow–urea (1.0 cm) plots (Table 7). The reason for these interactions on flow is not clear. However, they did not produce any significant difference in pollutant losses.
There were few significant (p < 0.1) differences between tillage or nutrient source treatments for both surface inlet flow and tile drainage water quality parameters for other storm events. The absence of significant differences between tillage treatments suggests that (i) residue cover differences between moldboard plow and chisel plow treatments after secondary tillage were too small to have a major influence on soil detachment or on sediment transport, and/or (ii) soil properties (including slope) were such that there was minimal difference in the amount of runoff and drainage between the treatments.
The importance of an individual storm event at an inopportune time on annual pollutant loads is apparent in the contribution from the 18 May 2000 event that occurred shortly after planting when nutrients were most susceptible to leaching and runoff because soil nutrient concentrations were high and surface cover was low. This event resulted in 47 and 38% of the annual surface inlet flow and tile flow, respectively, and accounted for 62, 41, and 44% of the annual surface inlet sediment load, tile NO3–N, and NH4–N loads, respectively (Tables 7 and 8).
|
Loads—Annual Basis
There were several significant differences in pollutant loads leaving the plots during the course of the study (Table 8), but the differences were not consistent for each year due to climate variability and soil heterogeneity. In 1999 there was a significant difference (p = 0.075) by nutrient source treatment in tile flow with urea-treated plots losing more water (3.4 cm) than manure-treated plots (2.9 cm). Consequently, there were also significant differences in tile drainage losses of NO3–N (p = 0.07) and combined NO3 + NH4–N (p = 0.094) losses by nutrient source treatments. In both cases more N was lost from the urea-treated plots. The NO3–N and NO3 + NH4–N losses from urea-treated plots were 0.97 and 1.63 kg ha–1 respectively, compared with 0.24 and 1.02 kg ha–1, respectively, from manure-treated plots. The significantly greater losses of mineral N from urea-treated plots in 1999 were partly attributable to flow differences and partly to the susceptibility of inorganic forms of N to leach and the potential of organic forms (manure) of N to decrease N leaching due to slow availability of mineralized N throughout the growing season. High ammonium losses in 1999 and 2001 were most likely due to preferential flow paths providing a direct conduit to the tile lines. Indirect evidence of preferential flow was observed as turbid ("lightly colored") water flowing through the tile lines immediately after heavy rain or rapid snow melt events.
In 2000 and 2001 there were no significant differences by tillage or nutrient source treatment or their interactions on losses or concentrations of pollutants leaving the plots through either surface inlets or tile drainage (Table 8). In 2002 there was a significant difference (p = 0.054) by nutrient source treatment for surface losses of NO3–N and combined NO3 + NH4–N. In both cases more N was lost from the urea-treated plots. Losses of surface NO3–N and NO3 + NH4–N from urea-treated plots were 0.63 and 0.77 kg ha–1, respectively, while losses of surface NO3–N and NO3 + NH4–N from manure-treated plots were 0.26 and 0.37 kg ha–1, respectively. The lower losses from manure compared with urea plots suggested slow but continuous release of manure organic N that was taken up by the crop more efficiently. Additionally, the inorganic fertilizer was not incorporated as deeply as the injected liquid hog manure (approximately 15 cm). This may have left it more susceptible to surface transport, especially in a year like 2002, which had more intense storms than previous years as indicated by the greater surface inlet flow losses (Table 8).
During the review of this paper, a concern was raised about the difficulty of comparing N and P losses when N and P application rates are not the same between manure and urea treatments in some years (2000 and 2002). Our results show that even in years when N rates were similar (1999 and 2001), there was no difference in N losses. These two years included the first year (1999) when the carryover effect was minimal and the third year (2001) when there was some carryover effect. Nearly similar N losses between the manure and the urea plots (both when manure N application rate was low and matching) suggest that these high organic matter soils are contributing a substantial amount of N through mineralization. This is further supported by the fact that corn grain yields (discussed later) between manure and urea plots were nearly similar over all years. Higher soil NO3–N and NH4–N levels in both nutrient source treatments in 2000 and 2002 (low N manure application years) and lower soil NO3–N and NH4–N levels in 2001 (high manure N application year) further suggest that soil mineralization may have been a more important factor in controlling available N and thus N leaching than other factors, including the N application rates. This observation is consistent with the results from another study in Minnesota (Dr. Gary Malzer, personal communication, 2004) where manure application rates varied from 0 to 74670 L ha–1. Even in the strip that received no manure application, corn grain yield was as high as 11.5 Mg ha–1, which was most likely because a large quantity of N was available from soil mineralization. This strip happened to be on a Webster clay loam soil, the same soil type used this study.
Concentrations—Annual Basis
As with pollutant loads, there was considerable variability in pollutant concentrations leaving the plots in surface inlet flow and tile drainage due to climate and soil heterogeneity (Table 9). However, there were no significant differences due to tillage or nutrient source effects or their interactions for any year in this study. The relatively high N concentration in 1999 through surface inlets was due to spring application of manure and urea at the start of the experiment. In all other years, nutrient sources were fall-applied.
|
|
Dissolved P losses were only 5.8% of the TP losses during the 2001 snowmelt period (Table 10). This is contrary to the findings of Hansen et al. (2000) who found soluble P losses were the dominant form of P loss in snowmelt. Ginting et al. (2000) reported a three-year average soluble P loss of 0.47 kg ha–1 compared with 0.1 kg ha–1 found in one snowmelt event in this study. This discrepancy is most likely due to rapid melting of an unusually deep snow pack in 2001 thus reducing the interaction time of snowmelt with plant residue and soil particles. These results suggest that soluble P losses during snowmelt depend not only on water losses but also on the rapidity with which the snowmelt occurs. Another reason could be the difference in soil test P levels between various studies.
In 2001, the surface inlet flow from snowmelt represented 73% of the annual surface inlet flow losses (Tables 8 and 10), yet it accounted for only 11% of the annual sediment load. The low sediment loss in spite of high water loss was likely due to the slower velocities of melt water and lack of rain drop impact compared with rain events. Sharratt et al. (2000) measured much higher sediment loss from two recently thawed soils (16650 kg ha–1 average) under simulated rain intensities of 96 mm h–1, thus indicating that high intensity spring rains when soils are slightly thawed could induce large sediment loss.
Corn Yield
There was large variation in corn yield primarily due to climate variation over the four years of the study (Fig. 2)
. In general, 1999 and 2000 were better growing seasons due to more uniform distribution of moisture throughout the year, while 2001 and 2002 each had wind events that physically damaged the crop late in the growing season.
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Generally, a lack of major differences in N losses between urea- and manure-treated plots appears to be related to soil mineralization. It seems that soil mineralization overwhelmed the N addition to soil either from manure or from urea applications. This is because even in years when the manure N rates were under-applied, there was no difference in N losses between the manure and the urea plots. Furthermore, corn grain yields were about the same even in years when manure N application rates were lower than urea N rates. For the year (1999) when manure and urea N application rates were similar, the beneficial effect of manure as opposed to urea application was a decrease in combined NO3–N and NH4–N losses via tile drainage. This may be partially due to slow but continuous release of manure organic N that was taken up by the crop more efficiently. Another beneficial effect of manure application was increased corn grain yield for 1 yr. This increase in corn yield was possibly due to additional nutrients that may have been present and/or from the nutrients that may have accumulated after many years of manure application. There was some NH4–N in tile drainage from manure-applied plots and that appeared to be associated with preferential flow paths.
The results show that a chisel plow–based system with approximately 30% residue cover will not be sufficient to dramatically reduce sediment loads from poorly drained flat lands. Taken in the context of the Minnesota River, adoption of conservation tillage such as chisel plowing in the basin will have minimal positive effect on Minnesota River water quality especially to the extent (40% reduction in sediment load) desired by the regulatory agencies (Minnesota River Assessment Project, 1994).
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
, Q. Dortch, and W.J. Wiseman, Jr. 1999. Characterization of hypoxia: Topic 1 report for the Integrated Assessment of Hypoxia in the Gulf of Mexico. NOAA Coastal Ocean Program Decision Analysis Ser. no. 15. NOAA Coastal Ocean Program, Silver Spring, MD.This article has been cited by other articles:
|
|
 |
|
  K. A. Oquist, J. S. Strock, and D. J. Mulla Influence of Alternative and Conventional Farming Practices on Subsurface Drainage and Water Quality J. Environ. Qual., June 27, 2007; 36(4): 1194 - 1204. [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]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Vadose Zone Journal | Journal of Plant Registrations | ||||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal |
