Final Report for LS91-035
[Note to online version: The report for this project includes tables, figures, and special characters (superscripts, subscripts, etc.) that could not be included here; full-sized numerals have been used instead. The regional SARE office will mail a hard copy of the original report at your request. Just contact Southern SARE at (770) 412-4787 or email@example.com.]
The inherent inefficiency of fertilizer N utilization by corn can lead to a relatively large pool of residual soil N subject to leaching and possible contamination of groundwater supplies, particularly on sandy soils in the southeastern Coastal Plain. The objectives of this research were to: (1) evaluate the potential of several cover crops to capture residual fertilizer N from a corn production system, and (2) study the field and laboratory decomposition of cover crops for the purpose of developing a simulation model to describe N release from cover crops over a wide range of soil and climatic environments.
Field experiments on a Coastal Plain soil (Norfolk loamy sand) served as a basis for characterizing NO3 leaching potential and the subsequent potential of winter annual cover crops (crimson clover, rye, spring oat, wheat, and native weeds) to recover and recycle residual fertilizer N. Two approaches were used to evaluate these N dynamics. The first method established two levels of residual soil N via fertilizer N applied to the previous corn crop, while the second method employed 15N- enriched potassium nitrate applied to microplots immediately prior to planting rye and crimson clover cover crops in early fall. Field decomposition of cover crop residues was monitored by placing air-dried plant material on the soil surface in nylon mesh bags for 1,2,4,8, and 16 weeks. Following retrieval, mesh bag contents were analyzed for dry weight and N concentration. As a corollary to the field decomposition study, laboratory experiments were conducted to determine decomposition rate constants under non-limiting conditions for carbohydrates and cellulose pools of the CERES-N model for cover crops decomposing on the soil surface.
Prior N fertilizer rate for corn profoundly influenced subsequent cover crop DM accumulation both years, indicating the ability of cover crops to increase N accumulation under conditions favoring carryover of fertilizer N. At both residual soil N levels, rye demonstrated the ability to reduce profile soil inorganic N compared to legume cover crops. Soil inorganic N to a 90-cm depth was greater under hairy vetch compared to crimson clover in April (prior to corn planting) and September of both years, reflecting differences in total N content and decomposition dynamics between the two legumes.
Estimates of the subsequent cover crop N pool potentially available to corn by 16 weeks was in the order of crimson clover>rye>spring oat>wheat. Based on a recommended fertilizer N rate of 150 kg/ha for corn grown in the North Carolina Coastal Plain, the percentage of this N requirement met by cover crop N release ranged from 12% with wheat to 51% with crimson clover.
In general, results from the 15N experiment confirmed findings from the unlabeled N experiment. Rye recovery of fall-applied 15N-enriched fertilizer was 35% by the following April compared to 10% by native weeds and only 5% by crimson clover. As a percentage of the total residue 15N, nearly 23% of the crimson clover N was released by corn maturity compared to 14% of the native weed N and 7% of the rye N. Low total N accumulation by rye and crimson clover, however, severely limited actual N contributions to corn growth and yield.
Under controlled environment conditions, the dynamics on C and N mineralization of a mix of leaves and stems was different from the patterns predicted from isolated leaves and stems. Laboratory results indicated a strong interaction between stems and leaves during early stages of decomposition, which may be relevant for predicting N mineralization from cover crop residues. The best predictors for N mineralization were residue C/N ratio and the reciprocal of residue N concentration.
In summary, there are some clear patterns that emerged in these experiments. Results of studies using two approaches to assess NO3 leaching potential and the subsequent cover crop recovery of residual fertilizer N in soil indicated the ability of rye over that of spring oat, wheat, crimson, clover, and native weeds to fill this niche. In situations (environmental stress or pest-related pressures) where low yields of a high-N requirement crop such as corn result in relatively high levels of residual soil inorganic N, rye should be the cover crop of choice for remedial action. With respect to cover crop release of N, crimson clover would serve better in this capacity compared to the grasses evaluated. Therefore, in order to optimize the inherent capabilities of grasses and legumes, a grass-legume biculture may be more appropriate in cover-crop based production systems. Further investigation is warranted on the role of bicultures in soil water and nutrient dynamics.
Finally, if cover crops are to be widely used as a soil management tool, the potential scope of application should be expanded. This expansion might include the evaluation of rye cultivars with respect to nutrient recovery and the use of cover crops to recycle nutrients contained in animal wastes.
With the aformentioned factors in mind, the objectives of this research were to:
1. Evaluate the potential of several cover crops to capture residual fertilizer N from a corn production system.
2. Study the field and laboratory decomposition of cover crops for the purpose of developing a simulation model to describe N release from cover crops over a wide range of soil and climatic environments.
Environmental concern regarding nitrate (NO3) pollution of groundwater is a major problem facing agriculture in the 1990's. Winter annual cover crops, as a component of conservation production systems, can provide a means of utilizing residual or mineralized NO3 in soils during non-crop periods and thereby reduce the amount of NO3 leaching. The subsequent availability of N recovered by cover crops is also of interest when one considers the importance of resource-use efficiency in sustainable production systems. The overall objective of this proposal was to characterize differences among cover crops regarding residual fertilizer NO3 use-efficiency in sustainable production systems. A secondary objective focused on the development of simulation technology to describe N release from decomposing cover crops. Considerably higher concentrations of residual soil inorganic N occurred following corn fertilized with 300 vs 250 kg/ha. Because the greatest concentrations at the high fertilizer N rate were found in the 45- to 75-cm depth interval, it is likely that further downward movement during winter months would remove this inorganic N from the effective crop rooting zone. In this regard, a winter annual cover crop of rye was quite effective in accumulating residual fertilizer N and thereby minimizing further N losses from the plant-soil system. Spring oat showed some moderate potential for accumulating soil N and reducing NO3 leaching while wheat, crimson clover, and native weeds were relatively ineffective in this role. Field decomposition of cover crop residues indicated that crimson clover released N faster to the subsequent corn crop than the grass cover crops. This faster N release by crimson clover was more evident during the early part of the growing season (4 to 6 weeks). As a corollary to the field decomposition studies, laboratory studies were conducted to assess the dynamics of C and N mineralization from leaves and stems of crimson clover, rye, oat, and wheat. In general, CO2 emission rates were highest for leaves and lowest for stems. The total CO2-C evolved and remaining C after 160 days for leaves and stems of all residues was similar to the amounts predicted from isolated leaves and stems. In contrast, net N mineralized from leaves and stems of oat and wheat was higher than the amount predicted from isolated plant parts. Results from this research illustrate the role winter annual cover crops can play in conserving N within the soil-plant-system while maintaining soil productivity. Additionally, a relatively simple model describing N release from cover crop residues will enable further streamlining of summer crop N requirements.
Groundwater contamination by agricultural chemicals is one of the major problems facing agriculture in the 1990's. Nitrate is of particular concern because relatively low concentrations (10 ppm N03-N) impair water for human consumption and because many shallow groundwater supplies now exceed recommended NO3-N drinking water standards (Hallberg, 1986; Keeney, 1986). Increases in groundwater NO3 levels have been associated with major increases in N fertilization, primarily with respect to corn (Zea mays L.) (Hallberg, 1987).
Long-term experiments in the Corn Belt have shown that fertilizer N removal by corn grain rarely exceeds 40% at economically optimum corn yields (Blackmer, 1986; Oberle and Keeney, 1990 a,b). In cases where residual N from previous crop fertilization remains in the soil, the potential for N03 leaching also exists. Analysis of long-term climatic data for most of the humid Southeast shows that rainfall exceeds evapotranspiration during the winter and early spring months. With an annual crop such as corn, the land is often bare during this period when the greatest potential for significant leaching exists.
Winter annual cover crops have been recognized as an integral component of southern agricultural systems for many years because of their role in soil erosion control and enhancement of soil productivity. With the refinement of conservation tillage technology, new strategies have evolved with regard to cover crop management. The role of nonleguminous cover crops in efficient use of water and N was reviewed by Wagger and Mengel (1988). Indirect evidence for cover crops utilizing residual N can be found in studies reported by Langdale et al. (1979), Pelchat (1986), and Utomo (1986). In these experiments, N content of winter wheat (Triticum aestivum L.) or rye (Secale cereale L.) cover crops ranged from 12 to 66 kg N/ha. Pelchat (1986) found that increasing N applications to the previous corn crop from 0 to 180 kg N/ha resulted in an increase of 16 kg/ha in N uptake by the subsequent cover crop. Information is limited, however, on the role of cover crops as sinks for residual N. Moreover, a better understanding of the management of cover crops in this role is needed to ensure that the residual N trapped is effectively recycled to subsequent summer crops.
Winter cover crop effects on N03 leaching were evaluated more directly in drainage lysimeters by Karraker et al. (1950). They found only small leaching losses under bluegrass (Poa pratensis L.) sod and rye cover crops. Large losses of NO3 occurred under a lespedeza (Lespedeza. ssp.) sod, but not when a winter rye cover crop was grown. They concluded that most of the NO3 came from decomposing, senescent lespedeza residues. In Sweden, Bertilsson (1988) concluded that a rape (Brassica napus L.) cover crop could greatly reduce NO3 losses, even when farm yard manure was applied in the autumn. These studies point to a potential benefit of cover crops with respect to reduction in NO3 leaching. However, definitive data on the differences among cover crops for their residual N03 use-efficiency are not available.
The decomposition of crop residues and the corresponding release of N is a complex process affected by residue composition, soil type, and climatic conditions. Due to the many factors involved, several researchers have developed computer simulation models to study and describe the decomposition process (Seligman and Van Kuelen, 1981; Parton et al., 1987; Smith, 1979; Hunt, 1977; Van Veen and Paul, 1981; Stroo et al., 1989). Most of these simulation models use the C/N ratio or the N concentration of the residue to control the N mineralization rate. However, good general relationships between C/N ratio or N concentration and N mineralization are still lacking. As a consequence, a relatively simple model for predicting the release of N from different crop residues has yet to be developed. Further work is needed to study the decomposition of crop residues and to determine the main chemical characteristics that control the decomposition process. From this effort should come the development of a general model of the decomposition of cover crop residues under field conditions.
Two methods were employed to accomplish each objective, each providing unique information. The field research was conducted over a two-year period from 1991 to 1993 on a Norfolk loamy sand (fine-loamy, siliceous, thermic Typic Paleudult) at the Lower Coastal Plain Research Station in Kinston. This and similar Coastal Plain soils are quite susceptible to leaching losses of fertilizer N in crop production systems. Selected soil physical and chemical characteristics of the surface 0.15 m prior to the initiation of the experiment were as follows: 86% sand, 8% silt, 6% clay, 2.3 cmolc/kg CEC, and pH 5.8.
Objective 1a: Leaching of Fertilizer N
Unlabeled N Experiment
With the first method, a field experiment was established in a new area each year that had a previous crop of corn fertilized with 150 or 300 kg N/ha. These rates represent 100 and 200%, respectively, of the recommended amount for corn grown in the Coastal Plain. All plots received a broadcast application of 50 kg N/ha as ammonium nitrate at corn planting and either 100 or 250 kg N/ha surface banded 10 cm to the side of each row approximately 6 weeks after planting. Following corn harvest in September 1991 and 1992, 5-cm diameter soil cores (3 per plot) were taken to a depth of 90 cm in 15-cm increments. Soil samples were air dried, extracted with 2M KCI (10 g soil in 100 mL) for 1 hr, and analyzed for NH4 and NO3 on a Lachat autoanalyzer.
15N- Labeled Experiment
A second approach employed 15N methodology under field conditions. Prior to planting rye and crimson clover cover crops in early October, the experimental area was chisel plowed and disked. The experiment was established in an area that had a previous crop of corn fertilized with 150 kg N/ha. A fallow (no cover crop) treatment and the two monocultures comprised the 3 treatments in a randomized complete block design experiment with three replications.
Field microplots consisting of galvanized steel flashing were used to prevent the lateral movement of water and fertilizer 15N. The microplots measured 2 by 3m, with flashing installed approximately 10 mm into the soil and extending 10 mm above the soil surface. Microplots were divided into 4 quadrants to facilitate fertilizer application. Potassium nitrate labeled with 10 atom % 15N was uniformly added as a solution to the soil surface at 50 kg N/ha (ca. 30 g 15N per microplot) on 11 October 1991 and 13 October 1992. In order to simulate a profile distribution of residual fertilizer N that might occur after a corn growing season, the fertilizer solution was moved into the soil over a 10-day period by approximately 8.0 cm of natural rainfall and applied water. Four soil cores (1.59-cm diameter) were obtained per microplot (1 core per quadrant) and composited in 15-cm increments to a depth of 90 cm following the last application of water in mid-October. Holes were filled with soil from the adjacent untreated area to minimize any alteration in soil water dynamics.
Soil samples were air-dried, ground with a mortar and pestle, and analyzed for total inorganic N as previously described. Once the inorganic N concentration was determined, an extract volume containing 40 to 100 µg N was incubated in 120-mL specimen cups with 0.4 g MgO and 0.2 g Devarda's alloy for 6 days (Sorensen and Jensen, 1991). Glass fiber disks (ca. 7 mm diameter) were acidified with 20 µL KHS04 and sealed in a Teflon packet to trap gaseous ammonia. Following the 6-day incubation period, glass fiber disks were removed from the Teflon packet and desiccated over concentrated H2SO4 for 48 h. All 15N analyses were conducted on a Europa Scientific Tracer Mass Stable Detector.
Objective 1b: Utilization of Residual N
For each approach outlined under the N leaching component of Objective 1a, winter annual cover crops were evaluated for their ability to recover residual fertilizer N that might otherwise be lost to leaching during the winter months.
Unlabeled N Experiment
The experimental area was chisel plowed and disked prior to planting cover crops in early October. Crimson clover (Trifolium incarnation L.), rye, wheat, and spring oat (Avena sativa L.) were drilled in plots measuring 5.8 by 15.2 m. Seeding rates were 28 kg/ha for crimson clover and 56 kg/ha for the grasses. A fallow (no cover crop) and 4 cover crops comprised the 5 main plot treatments and two levels of residual soil N represented the split-plot factor in a randomized complete block design with 4 replications. Winter annual weeds in the fallow treatment were allowed to grow during the cover crop phase of the study and consisted primarily of henbit (Lamium amptexicaule. L.) and chickweed (Stellaria L.).
Aboveground cover crop DM was determined in December, early March, and mid April by harvesting a 0.5-m² quadrat from each plot on all sampling dates. After the first sampling date, care was taken to provide adequate distance between newly and previously harvested areas. Plant samples were dried at 65°C, weighed, ground, and analyzed for total N and C on a Perkin Elmer 2400 CHN Elemental Analyzer. Soil sampling was conducted to a depth of 90 cm in 15-cm increments at each plant sampling date. Soil samples were analyzed for total inorganic N in the same manner as previously described.
15N- Labeled Experiment
Cover crop utilization of residual fertilizer N in the 15N experiment followed an approach similar to that previously described; however, a more detailed N balance should be obtained with this method. After profile distribution of 15N-enriched fertilizer occurred, the cover crops and native weeds in the fallow microplots were allowed to grow until corn planting the following April. Plant and soil samples were taken during the cover crop season in December, early March, and mid April. In order to estimate cover crop dry matter and N accumulation for the December and March sampling dates, a 534-cm² area was harvested in each microplot quadrant and then composited into one sample. Care was taken to provide adequate distance between the December and March harvested areas. All plant material was dried at 65°C, weighed, ground, and analyzed for total C, N, and 15N. For the April sampling date, aboveground biomass was completely harvested in each microplot and allowed to air dry on greenhouse benches before it was weighed and subsampled for moisture, total N, and 15N determinations. Soil sampling was conducted to a depth of 90 cm in 15-cm increments at each plant sampling date, with samples processed and analyzed as previously described.
Objective 2a: Field Residue Decomposition
Unlabeled N Experiment
In order to better understand the recycling of N trapped by cover crops, and thereby potentially reduce the N required by the subsequent summer crop, cover crop decomposition was monitored with nylon mesh bags containing plant residue from the respective cover crops following corn fertilized with 150 kg N/ha only. Aboveground whole plant material was collected immediately prior to chemical desiccation in mid April, air-dried on greenhouse benches, and 18.0 g placed in 1-mm mesh nylon bags (15.2 by 30.5 cm). This addition corresponded to a residue loading rate of 3.7 Mg dry matter/ha. Growth stage at harvest was heading for rye, boot for oat and wheat, and mid-bloom for crimson clover. The samples were carefully handled to minimize detachment or breakage of various plant parts. Prior to mesh bag placement in the field, corn was planted via no-tillage in chemically desiccated cover crops. Bags were placed on the soil surface in the corn interrow in early May and retrieved at 1,2,4,6,8, and 16 weeks after field placement. Bag contents were dried at 65°C, weighed, ground, and analyzed for total C and N. Soil contamination was accounted for by ashing a 1-g subsample at 550°C and all plant constituents are reported on an ash-free basis.
Nonlinear regression equations for percent original N remaining at each mesh bag retrieval date were determined by fitting the data to one and two pool models, previously described by Ranells and Wagger (1992), using the Marquardt option of the NLIN procedure developed by SAS (SAS, 1985). A single pool model assumes that all of the plant N will mineralize at the same rate. The two pool model attempts to segregate plant N into two groups of mineralizable N; one that quickly mineralizes and a second, more passive pool that releases N more slowly. The general form of the equations were as follows:
Eq.  PNR = P + (100-P)e-kt
Eq.  PNR = 10OPe -klt +100(1-P)e-k2t
where: PNR = Percent nitrogen remaining at time 't'
P = N pool(s)
k, kl, k2 = rate constant of N release
An appropriate model was chosen for each cover crop treatment based on successful regression and root mean square error values.
15N- Labeled Experiment
Air-dried cover crop material from the April harvest was placed in newly established microplots where the same unlabeled cover crops were harvested and removed. Corn ('Dekalb-Pfizer 689') was hand planted on 28 April 1992 and 27 April 1993. All microplots received a broadcast application of 33 kg N/ha as ammonium nitrate at planting and 67 kg N/ha as ammonium nitrate surface banded 10 cm to the side of each row approximately 6 weeks after planting. This was done so that corn recovery of cover crop 15N would not be constrained by low available soil N. Previous research in North Carolina has shown that fertilizer N at 90 kg/ha was sufficient to optimize corn yields when preceded by crimson clover or hairy vetch cover crops (Wagger, 1989b). Lime, P, and K were applied broadcast according to soil test recommendations for corn. Residual weed control was provided with 2.24 kg a.i./ha alachlor [2-chloro-N-(2', 6'-diethylphenyl-N-(methoxymethyl)-acetamide] and 2.24 kg a.i./ha atrazine [6-chloro-N-ethyl-N'-methylethyl)-1,3,5-triazine-2, 4-diamine]. Post-emergent weed control was provided by a broadcast, directed application of 1.08 kg a.i./ha linuron [N'-(3,4-dichlorophenyl)-N-methoxy-Nmethylurea] 6 weeks after planting. At maturity, corn was hand harvested from each microplot, weighed, and subsampled for moisture, total N, and 15N- determinations. Soil sampling was conducted after corn harvest and analyzed for inorganic N and 15N as previously described. Data was analyzed by treatment using SAS GLM procedures (SAS, 1985).
Objective 2b: Laboratory Residue Decomposition
Samples of cover crop residues from our field studies (oat, rye, wheat, crimson clover) were divided into leaves and stem components and each component analyzed for soluble carbohydrates, cellulose, hemicellulose, lignin, acid detergent fiber, and neutral detergent fiber contents with the methods of Goering and Van Soest (1970). In addition, C and N contents of the residues were determined.
In order to characterize the decomposition of the different residues, subsamples of leaves or stems were placed on the surface of a mixture of soil and sand packed in 5-cm PVC tubes. Each tube had a perforated plexiglass plate with cheesecloth at the bottom to retain the soil. The soil-sand mixture was kept at 55% water-filled porosity. Three tubes of each residue were placed inside high humidity chambers in incubators set at 10, 20, 30, and 40°C. The tubes were leached with 0.01 M CaC12 at 14, 28, 56, 84, 112, and 140 days, and the inorganic N in the leachates analyzed by colorimetric methods. The cumulative N mineralized from each residue, at each temperature, was used to adjust our model.
To develop our model, the submodel of N mineralization from PAPRAN (Seligman and Van Kuelen, 1981) was modified to describe the separate decomposition of the two components described, leaves and stems. The rate constants in the model were adjusted so that the laboratory N mineralization from each component is adequately described at each of the temperatures used. This allows us to determine rate constants for each temperature, which can then be used to develop a function to adjust the rate constants in the model according to temperatures measured under field conditions.
Unlabeled N Experiment - 1991 to 1992
Residual Soil N and Cover Crop Performance
The initial soil sampling following corn harvest in September 1991 revealed significant differences in the distribution of inorganic N between N rates applied to the previous corn crop (Fig. la). There was a modest increase with depth in soil inorganic N levels under corn fertilized with 150 kg N/ha, ranging from a low of 5 to 8 mg/kg in the upper 30 cm to 18 mg/kg between 60 to 90 cm. In contrast, residual soil inorganic levels following corn fertilized with 300 kg N/ha ranged from 11 to 33 mg/kg, being higher by approximately two-to threefold at most depths compared to the low N rate treatment. Moreover, the largest differences between N rates occurred in the 45- to 75-cm depth range. Without some form of intervention, these elevated soil inorganic N concentrations (ca. 33 mg/kg) this deep in the soil profile would likely be lost from the plant-soil system before planting of next year's row crop. In this study, precipitation from cover crop seeding in October to termination of cover crop growth and corn planting the following April was 50 cm.
Cover crops were seeded in early October 1991 and by December only rye had shown a dry matter (DM) response (Fig. 2) due to residual soil inorganic N levels. In general, DM accumulation values were in the order of rye > oat > wheat > fallow = crimson clover. The corresponding cover crop N accumulation values reflected a similar pattern, with the N content in rye following the high N rate nearly double (39 vs. 21 kg N/ha) that of rye after corn fertilized with 150 kg N/ha (Fig. 3). There was no difference in N accumulation among the other cover crop treatments due to prior N rate, with mean values across N rates ranging from 4 kg N/ha in crimson clover to 17 kg N/ha in spring oat.
Even though cover crop N accumulation differed between species by December, there was little difference in soil inorganic levels due to cover crop. Consequently, soil inorganic N results for this date are presented as mean values across cover crops for the 150 and 300 kg/ha N rates (Fig. lb). From September to December there was a marked decline in soil inorganic N concentrations under both N rates; however, differences were still evident between N rates. Soil inorganic levels in the upper 45 cm were low (1 to 3 mg/kg) and similar between N rates. These same relatively low concentrations persisted through the remainder of the profile under the low N rate but increased below 45 cm under the high N rate. It is also of note that the center of highest inorganic N concentration (ca. 17 mg/kg) had shifted from 60 to 75 cm in September to 75 to 90 cm in December.
Cover crop DM accumulation increased appreciably from December to March as spring regrowth proceeded (Fig. 2a and b). Averaged across N rates, rye accumulated the highest DM (2.64 Mg/ha) and wheat accumulated the least (0.87 Mg/ha). Also noteworthy is the result that mean DM in the fallow treatments, which was composed of winter annual weeds, exceeded DM values for both wheat and crimson clover. With regard to prior N rate, and similar to December results, only rye showed a marked DM increase (3.27 vs 2.01 Mg/ha) when following the 300 vs 150 kg/ha N rate. Cover crop N accumulation paralleled the DM results, with mean rye N content (40 kg/ha) approximately two to three fold greater than the other cover crop treatments and rye N accumulation increasing 55% under high residual soil N (Fig. 3). Spring oat at the high residual soil N level accumulated the same amount of N as rye at the low residual N level. Nitrogen accumulation by wheat, crimson clover, and fallow treatments were similar and unaffected by prior N rate.
There were no differences due to cover crop in the distribution of soil inorganic N with the prior fertilizer N rate of 150 kg/ha by March 1991, consequently, data are presented for the prior 300 kg N/ha rate only (Fig. 4). A distinct pattern was evident in soil inorganic levels, with treatment differences confined to the lower two soil depths and concentrations in the order of rye < spring oat < wheat, crimson clover, and fallow. Soil inorganic N concentrations under rye were generally < 1 mg/kg throughout the profile compared to a mean of 13 mg/kg under wheat, crimson clover, and fallow plots at the 75- to 90-cm depth. These results, and the concomitant N accumulation values, reflect the ability of rye to effectively utilize residual soil N. Just prior to cover crop desiccation and corn planting in April 1992, DM accumulation by rye, spring oat, and crimson clover was similar and greater than wheat and fallow treatments (Fig. 2). These DM values were in the range commonly reported in other studies for the region. Only rye and spring oat responded to residual soil N level, increasing an average of 45% (1.37 Mg/ha) from low to high residual soil N plots. The corresponding N accumulation values, averaged over residual soil N level, were in the order of crimson clover = rye > oat > wheat> fallow (Fig. 3). As with earlier sampling dates, rye N accumulation increased sharply (108%) under high compared to low residual soil N levels. Spring oat also responded to residual soil N level by April, accumulating 27 kg N/ha at the low residual soil N level compared to 50 kg N/ha at the high residual N level . Only the fallow treatment, comprised of winter annual weeds whose growth had terminated by the April sampling date, reflected a net decrease in N accumulation. Associated with these relatively low N accumulation values in fallow plots were the greatest soil inorganic N concentrations, most notably in the 60- to 75- (7 mg/kg) and 75- to 90-cm (9 mg/kg) depth intervals (Fig. 4). Wheat, crimson clover, and spring oat treatments had inorganic N concentrations intermediate to fallow and rye (1 mg/kg). While this range in soil inorganic N concentrations was relatively small just prior to planting corn, NO3 movement below the 90-cm sampling depth appears likely given the differences in cover crop N accumulation.
Cover Crop N Release Rates
Nitrogen release patterns, expressed as a percentage of the initial residue N remaining, are illustrated in Fig. 5. An exponential equation reflected the decline in residue N over time for all curves, with only marginal differences between cover crop residues. These results are somewhat surprising, as previous work has shown distinct differences between cover crop N release rates, particularly between grasses and legumes (Wilson and Hargrove, 1986; Wagger, 1989a). Residue decomposition and N release proceeded at a relatively rapid rate, such that by 4 weeks in the field the N remaining in the respective plant residues ranged from 42 to 50%. By 8 weeks, which corresponded with corn tasseling/silking and a period of high corn N demand, the percentage N remaining ranged from a low of 18% for wheat to 27% for rye. The general order of cover crop N release by 16 weeks was crimson clover > wheat = rye > oat, with a mean residue N remaining value of approximately 8%.
Inspection of the initial residue C:N ratios provides some explanation for the relatively small differences in N release rates between the residues. The N concentration or C:N ratio of plant residues has frequently been used as a tool for predicting the rate of decomposition. Oat, rye, and wheat had C:N ratios in a very narrow range of 35 to 37:1 while the C:N ratio of crimson clover was 19:1. Although the C:N ratio of crimson clover was below the theoretical value (25:1) above which net N immobilization occurs (Allison, 1966), and the grasses were above this threshold value, these differences are not overly large. Another contributing factor may have been the cover crop developmental stage at the cessation of growth. Ranells and Wagger (1992) reported that a greater proportion of crimson clover N was released from clover collected at the late vegetative stage compared to harvest dates ranging from early bloom to late seed set. Averaged over 2 years, C:N ratios increased from 14:1 to 18:1 between late vegetative and late bloom while cellulose and lignin concentrations increased an average of 48%. Of the structural carbohydrates, lignin is the most resistant to decomposition by microorganisms and its presence in cell walls can retard the degradation of cellulose (Szegi, 1988; Waksman and Hutchings, 1936).
It is important to obtain some index of N availability based on the initial residue N content. Estimates of N released from each cover crop were calculated from values generated by the decomposition curves and then multiplied by the cover crop N content at the time of desiccation for cover crops following the previous low fertilizer N rate (150 kg/ha). Using this approach, the potentially available N by 16 weeks was 59 kg N/ha for crimson clover, 32 kg N/ha for rye, 25 kg N/ha for oat, and 17 kg N/ha for wheat.
15N-Labeled Experiment - 1991-1992
Cover Crop Recovery of 15N
Following the application of 15N labeled KNO3 to microplots planted to the respective cover crops, soil sampling was conducted to determine the residual profile N03 distribution. Averaged across microplots, approximately 55% of the labeled fertilizer was in the surface 15 cm of soil and 34% in the 15- to 30-cm depth (data not presented). Below 45 cm the distribution of 15N was relatively uniform, including detectable enrichment in the 75- to 90-cm soil layer. The intent was to facilitate a somewhat deeper and more uniform distribution of 15N in the soil profile but apparently 7.5 cm of irrigation water was not sufficient for this task. In this regard, it is also important to characterize the precipitation environment from cover crop planting to termination of growth the following spring. From October 1991 to April 1992 generally below normal precipitation prevailed (Table 1), which would moderate further downward movement of NO3.
Estimates of fertilizer 15N recovered by cover crops in December and March were based on small area samples in each microplot. By mid December 1991, < 1% of the applied 15N had been recovered in each cover crop treatment (Table 2). The results were unexpected since there were substantial differences in cover crop DM estimates, ranging from approximately 0.75 Mg/ha for fallow and crimson clover treatments to 1.85 Mg/ha rye. However, profile soil inorganic N was considerably lower under rye (16 kg N/ha) compared to crimson clover (53 kg N/ha) and fallow (71 kg N/ha). By early March 1992, estimates of cover crop 15N recovery had increased considerably from December values. Winter annual weeds in fallow microplots had accumulated 10%, crimson clover 5%, and rye 20% of the fertilizer 15N applied the previous fall (Table 2). The associated soil inorganic 15N values in March reflected the continued ability of rye to better utilize residual soil N than crimson clover or winter annual weeds. Profile soil inorganic N in March decreased sharply from December levels, yet there was nearly a fourfold difference between rye (6 kg N/ha) and crimson clover/fallow (22 kg N/ha) treatments.
Just prior to termination of cover crop growth in April, recovery of residual 15N in crimson clover and fallow treatments was unchanged from March estimates (Table 2). In contrast, rye 15N recovery increased from 20 to 35%. These results are similar to those from a study conducted on the Maryland Coastal Plain, where average percent recoveries of fall N were 8% for native weed cover, 8% for crimson clover, and 45% for rye (Shipley et al., 1992). Based on the initial fertilizer 15N application of 50 kg/ha, N recovery was 2.5, 5, and 17.5 kg/ha for crimson clover, fallow, and rye, respectively. As might be expected, profile soil inorganic N declined to relatively low amounts from March to April but levels were approximately threefold greater under fallow and crimson clover treatments compared to rye.
Corn Recovery of Residue 15N
Total cover crop N at desiccation in April was 19, 21, and 61 kg/ha for fallow, crimson clover, and rye treatments, respectively (data not presented). These values represent the cover crop N pool potentially available to the subsequent corn crop. Poor stands in crimson clover microplots severely limited its potential N contribution to corn.
Approximately 5 wk after returning 15N labeled cover crop residues to new microplots, estimates of corn recovery of residue N were < 1% for all treatments (Table 3). Corn was at the V3 to V4 developmental stage, a period when plant demand for N is relatively small compared to later stages of development. The inorganic soil 15N pool, which also represents N mineralized during cover crop decomposition, indicates very distinct differences in available residue N. As a percentage of the total residue 15N, about 6% of the rye N, 15% of the native weed N, and 34% of crimson clover N resided in the soil inorganic N pool. The results illustrate the relative differences in grass vs. legume C:N ratios with respect to governing residue decomposition. Three weeks later (25 June), corn had recovered nearly 11% of the N in crimson clover compared to 6% in fallow and 2% in rye treatments. The amounts of residue 15N in soil inorganic N pools at this date were low, such that treatment differences were generally of no agronomic consequence.
Corn recovery of residue 15N by physiological maturity in early September followed a pattern similar to the mid-season sampling date (Table 2). In summing up the total residue 15N available during the corn growing season, including an estimate of 15N in corn roots, approximately 23% of the crimson clover N was released compared to 14% of the native weed N and 7% of the rye N. Based on the initial N content of the respective cover crops, these percentages equate to only 3, 5, and 4 kg/ha of available N from fallow, crimson clover, and rye, respectively. These N contributions would have no impact on corn yield potential.
Finally, these corn recovery values of residue 15N were less than half of the estimated 80 to 90% of initial N that was released from grass and legume residues during the corn growing season in the mesh bag experiment at the same location. It should be noted that N-availability estimates using a mesh bag approach only provide information with respect to a potentially available N pool, and therefore do not afford a direct measurement of residue-N contributions to a cropping system. Nevertheless, results obtained with 15N-labeled residues in this study suggest considerable overestimation of N contributions from cover crop residues based on a mesh bag technique.
Modeling Residue Decomposition
Determining Best Decay Rate Constants for the Unmodified CERES N
When fitting the model to either all the data or only to data from stems and leaves and stems, the maximum carbohydrate decay rate constant estimated for the unmodified CERES model was 0.8/d, which is the same as that in the original model. The decay rate of the cellulose(CELL)pool was lower than in the original model (0.01/d when using all the data or 0.005/d when using only data from stems and leaves and stems, versus 0.05/d). These results agree with other CERES validation studies in which one or more of the decay rates needed to be reduced because the model was over predicting residue decomposition (Vigil et al., 1991; Bowen et al., 1993). It is important to note that even though we used a wide range of decay rates in our procedure for fitting constants, we could not get a good fit of the data (Fig. 1, Table 3).
Assigning Actual values to the CARB, CELL and LIGN pools
The estimated, pool sizes of the cover crop residues varied widely (Table 1). In all cases carbohydrate (CARB) pool was larger than the 20% used in the CERES model, and it
varied from 26.7 to 66.2%. The CELL pool varied from 28.2% for clover leaves to 67% for
rye stems. The lignin (LIGN) pool was as high as 14% for clover stems and as low as 2.6% for rye leaves. Data from the literature show a broad range of cover crop residue composition, which makes it difficult to reach generalizations such as those used in CERES-N (Janzen and Kucey, 1989; Müller et al; 1988). Furthermore, chemical characteristics of the residues regarding pool sizes are greatly affected by the maturity stage of the crop (Müller et al., 1988; Wagger, 1989a; Honeycutt et al; 1993). Given the important effect of chemical composition on N release patterns (Wagger, 1989a), it would appear to achieve a close estimation of pool sizes when modeling N mineralization from crop residues.
When the actual sizes of the CARB, CELL and LIGN pools (Table 1) were used in the model, the new decay rate for carbohydrates was 0. 1 4/d (Table 2). The decay rate for CELL was 0.0034/d when the model was fitted to all data, and 0.0023/d if only data from stems and leaves and stems were used for the validation (Table 2). The decay rate of carbohydrates is similar to the 0.16-0.19/d range proposed for glucose by Paul and Van Veen (1978) and to the 0.2/d used for carbohydrates by Verberne (1990) and by Bowen et al. (1993). The decay rate constant for CELL is close to the lower limit of the range (0.03 to 0.004/d) reported by Paul and Van Veen (1978). The fact that decay rates were below the reported ranges might be explained by the slower decomposition associated with surface application of residues (Parker, 1962; Smith
and Sharpley, 1990).
Using the actual values for each pool greatly improved the prediction of net N mineralized and N remaining (Tables 3 and 4). The root mean square error (RMSE) for net N mineralized decreased from 1.00 to 0.28 g/m², for the whole data set. At the same time, the SE of the mean deviation was lower, and the correlation coefficient improved from 0.85 to 0.96. A similar trend in the evaluation parameters was observed for N remaining.
It should be noted that the model tended to underpredict net N mineralization from isolated leaves (Fig. 2). In cover crops, however, this is a purely theoretical situation, because most of the time the composition of the residues will be between the leaves and stems treatment and the isolated stems. When we fitted the model to leaves and stems and stems data, the RMSE for net N mineralized was as low as 0. 1 8 g/M² versus 0.44 g/M² for the original model.
Model with Separate Decomposition of Leaves and Stems
Simulating the separate decomposition of stems and leaves with a common point of interaction through the inorganic N pool slightly improved the prediction of net N mineralization from cover crops (Table 3). When considering net N mineralized from the individual cover crops, this improvement only took place in the case of cereals, in which the RMSE for the original CERES model was 0.51 g/m² versus the 0.46 g/m² for the modified model (data not shown). These results support the existence of interaction between leaves and stems in cereals during the decomposition process (Collins, 1991; Quemada and Cabrera, 1995). Prediction of N remaining was also slightly improved when decomposition was simulated separately (Table 4). In agreement with Collins et al. (1991), we conclude that simulating interactions between residue components might be useful in developing mechanistic models, but might not improve significantly the prediction of the decomposition. Nevertheless, some studies have reported that after one year of decomposition most of the leaves had decomposed while the straw was still remaining in the field (Stott et al., 1990), and recent experiments emphasized the important role of cereals stems as a sink of N (Schomberg et al., 1994). Therefore, more research might be needed to develop and test models that split the decomposition of plant components.
Model with Separate Decomposition of Leaves and Stems Using Actual Values for the
CARB, CELL and UGN Pools
When using the actual values for crop residue pools (Table 1) in the model with separate decomposition of leaves and stems, there was a slight improvement on prediction. The RMSE decreased from 0.55 to 0.50 g/m² for net N mineralized and from 0.88 to 0.83 g/m² for N remaining when the whole data set was used. At the same time there was a slight decrease of the standard error of the mean deviation and the correlation coefficient was improved.
These results lead to the former conclusion, i.e. they support the experimental hypothesis of interactions between residue components in the decomposition process, but the improvement on the prediction of N released was very small.
Following Addiscott and Withmore (1991), CERES-N in which actual values of the CARB, CELL and LIGN pools were used, presented a statistically satisfactory simulation of net N mineralized. They had a high correlation coefficient and the mean difference was not significantly different from zero (Table 3).
The F-test comparing the mean squares due to pure error and lack of fit was significant in all cases at a 0.01 level (Tables 3 and 4), which shows that the model could still be improved (Withmore, 1991). However, because the data used were obtained under controlled conditions, the experimental error was very small, making the lack of fit test too strict for a model that will be mainly applied under field conditions.
Educational & Outreach Activities
1. Quemada, M., and M.L. Cabrera. 1995. Carbon and nitrogen mineralized from leaves and stems of four cover crops. Soil Sci. Soc. Am. J. 59:471-477.
2. Quemada, M., and M.L. Cabrera. 1995. CERES-N model predictions of nitrogen mineralized from cover crop residues. Soil Sci. Soc. Am. J. 59:1059-1065.
3. Ranells, Noah N., and Michael G. Wagger. 1996. Nitrogen release from grass and legume cover crop monocultures and bicultures. Agron. J. 88:777-782.
There were two field days at the research site in the summers of 1992 and 1993. A combined total of approximately 175 people attempted these field days, including extension agents, farmers, and conservation agencies personnel.
Winter annual cover crops can and should play a vital role in the conservation of natural resources and the maintenance of a sustainable agricultural enterprise. These benefits come in the form of reduced soil erosion, improved soil water availability for crop needs, and less reliance on synthetic fertilizer inputs. The magnitude of these potential benefits will also be greater in Piedmont and Mountain regions because of the erosion hazard and water infiltration limitation.
To date, cover crop research has focused on monoculture based systems with respect to cover crop performance. Information is scarce regarding the role grass/legume bicultures can play as winter cover crops in maintaining viable and resource conservative agricultural systems. Research on grass/legume mixes has been primarily conducted in pasture systems. In order to maximize the potential of cover crops as a multipurpose management tool, it is essential to develop basic information on the management of grass/legume mixes for cover crop based production systems. This approach can ultimately be used to manage plant-available nutrients more efficiently, thereby minimizing the potential for environmental concerns and enhancing farm profitability.
Areas needing additional study
Addiscott, T. M., and A. P. Withmore. 1987. Computer simulation of changes in soil mineral nitrogen and crop nitrogen during autumn, winter and spring, J. Agric. Sci. Camb. 109:141-157.
Allison, F. E. 1966. The fate of nitrogen applied to soils. Adv. Agron. 18:219-258.
Bavel, C. H. M., and F. T. Verlinden. 1956. Agricultural drought in North Carolina. North Carolina Agric. Exp. Stn. Tech. Bull. No. 122.
Bertilsson, G. 1988. Lysimeter studies of nitrogen leaching and nitrogen balances as affected by agricultural practices. Acta Agric. Scand. 38:3-11.
Blackmer, A. M. 1986. Potential yield response of corn to treatments that conserve fertilizer nitrogen in soils. Agron. J. 78:571-575.
Bowen, W. T., J. W. Jones, R. J. Carsky, and J. O. Quintana. 1993. Evaluation of the nitrogen submodel of CERES-Maize following legume green manure incorporation. Agron. J. 85:153-159.
Clark, A. J., A. M. Decker, and J. J. Meisinger. 1994. Seeding rate and kill date effects on hairy vetch-cereal rye cover crop mixtures for corn production. Agron. J. 86:1065-1070.
Collins, H. P., L. F. Elliot, R. W. Rickman, D. F. Bezdicek, and R. I. Papendick. 1990. Decomposition and interactions among wheat residue components. Soil Sci. Soc. Am. J. 54:780-785.
Goering, H. K., and P. J. Van Soest. 1970. Forage fiber analyses: Apparatus, reagents, procedures, and some applications. USDA Agricultural Handb. No. 379. U. S. Government Printing Office, Washington, DC.
Hallberg, G. R. 1986. From hoes to herbicides: Agriculture and goundwater quality. J. Soil Water Conserv. 41:357-364.
Hallberg, G. R. 1987. Agricultural chemical in groundwater: Extent and implications. Am. J. Altern. Agric. 2:3-15.
Honeycutt, C. W., L. J. Potaro, K. L. Avila, and W. A. Halteman. 1993. Residue quality, loading rate and soil temperature relations with hairy vetch (Vicia villosa Roth.) residue carbon, nitrogen and phosphorus mineralization. Biological Agriculture and Horticulture 9:181-199.
Hunt, W. H. 1977. A simulation model for decomposition in grasslands. 1977. Ecology 58:469-484.
Janzen, H. H., and R. M. N. Kucey. 1988. C, N, and S mineralization of crop residues as influenced by crop species and nutrient regime. Plant and Soil 106:35-41.
Kamprath, E. J. 1986. Nitrogen studies with corn on Coastal Plain soils. North Carolina Agric.
Karraker, P. E., C. E. Bortner, and E. N. Fergus. 1950. Nitrogen balance in lysimeters as affected by growing Kentucky bluegrass and certain legumes separately together. Bull. No. 557. Kentucky Agric. Expt. Sta., Univ. of Ky., Lexington, KY.
Keeney, D. R. 1986. Sources of nitrate to groundwater. CRC Crit. Rev. Environ. Control. 16:257-304.
Langdale, G. W., A. P. Barnett, R. A. Leonard, and W. G. Fleming. 1979. Reduction of soil erosion by the no-till system in the Southern Piedmont. Trans. Am. Soc. Agric. Eng. 22:82-86 and 92.
Müller, N. M., V. Sundman, O. Soininvaara, and A. Merilainen. 1988. Effect of chemical composition on the release of nitrogen from agricultural plant materials decomposing in soil under field conditions. Biology and Fertility of Soils 6:78-83.
Oberle, S. L. and D. R. Keeney. 1990a. Factors influencing corn fertilizer N requirements in the Northern U.S. Corn Belt. J. Prod. Agric. 3:527-534.
Oberle, S. L. and D. R. Keeney. 1990b. Soil type, precipitation, and fertilizer N effects on corn yields. J. Prod. Agric. 3:522-527.
Parker, D. T. 1962. Decomposition in the field of buried and surface applied cornstalk residue. Soil Sci. Soc. Am. Pro. 28:368-370.
Parton, W. J., D. S. Schimel, C. V. Cole, and D. S. Ojima. 1987. Analysis of factors controlling soil organic matter levels in Great Plains grasslands. Soil Sci. Soc. Am. J. 51:1173-1179.
Paul, E. A., and J. A. Van Veen. 1978. The use of tracers to determine the dynamic nature of organic matter. Transactions of the 11th International Congress of Soil Science 1978, vol. III:61-102.
Pelchat, J. A. 1986. Effects of tillage and winter cover crops on nitrogen requirements of corn in Indiana. Ph.D. Diss. Purdue Univ., West Lafayette, Ind. Diss. Abst. DA8/00940. 186 pp.
Quemada, M., and M. L. Cabrera. 1995. Carbon and nitrogen mineralized from leaves and stems of four cover crops. Soil Sci. Soc. Am. J. 59:471-477.
Ranells, N. N., and M. G. Wagger. 1992. Nitrogen release from crimson clover in relation to plant growth stage and composition. Agron. J. 84:424-430.
Ranells, N. N., and M. G. Wagger. 1996. Nitrogen release from grass-legume cover crop monocultures and bicultures. Agron. J. 88:777-782.
SAS Institute, Inc. 1985. SAS user=s guide: Statistics. 5 ed. SAS Institute, Inc.
Schomberg, H. H., J. L. Steiner, and P. W. Unger. 1994. Decompostion and nitrogen dynamics of crop residues: residue quality and water effects. Soil Sci. Soc. Am. J. 58:372-387.
Seligman, N. G., and H. Van Kuelen. 1981. PAPRAN: A simulation model of annual pasture production limited by rainfall and nitrogen. In Simulation of nitrogen behavior of soil-plant systems, E. M. J. Frissel, J. A. Van Veen. Pudoc, Wageningen.
Shipley, P. R., J. J. Meisinger, and A. M. Decker. 1992. Conserving residual corn fertilizer nitrogen with winter cover crops. Agron. J. 84:869-876.
Smith, S. J., and A. N. Sharpley. 1990. Soil nitrogen mineralization in the presence of surface and incorporated crop residues. Agron. J. 82:112-116.
Smith, O. L. 1979. An analytical model of the decomposition of soil organic matter. Soil Biol. Biochem. 11:585-606.
Sorensen, P., and E. S. Jensen. 1991. Sequential diffusion of ammonium and nitrate from soil extracts to a polytetrafluroethylene trap for 15N determination. Anal. Chim. Acta. 252-201-203.
Stroo, H. F., K. L. Bristow, L. F. Elliott, R. I. Papendick, and G. S. Campbell. 1989. Predicting rates of wheat residue decomposition. Soil Sci. Soc. Am. J. 53:91-99.
Szegi, J. 1988. Cellulose decomposition and soil fertility. Akademiai Kiado, Budapest, Hungary.
Utomo, M. 1986. Role of legume cover crops in no-tillage and conventional tillage corn van
Van Veen, J. A., and E. A. Paul. 1981. Organic carbon dynamics in grassland soils. 1. Background information and computer simulation. Can. J. Soil Sci. 61:185-201.
Verberne, E. L. J., J. Massink, P. de Willigen, J. J. R. Groot, and J. A. Van Veen. 1990. Modelling organic matter dynamics in diffferent soils. Netherlands Journal of Agricultural Science 38:221-238.