Agronomy Journal Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published in Agron J 98:1610-1619 (2006)
DOI: 10.2134/agronj2005.0302
© 2006 American Society of Agronomy
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Miller, P. R.
Right arrow Articles by Holmes, J. A.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Miller, P. R.
Right arrow Articles by Holmes, J. A.
Agricola
Right arrow Articles by Miller, P. R.
Right arrow Articles by Holmes, J. A.
Related Collections
Right arrow Other Legumes
Right arrow Nitrogen
Right arrow Other Forage Crops
Right arrow Crop Systems
Right arrow Water Conservation

Production Papers

Cropping Sequence Effect of Pea and Pea Management on Spring Wheat in the Northern Great Plains

P. R. Miller*, R. E. Engel and J. A. Holmes

Department of Land Resources and Environmental Sciences, P.O. Box 173120, Montana State Univ., Bozeman, MT 59717-3120

* Corresponding author (pmiller{at}montana.edu)

Received for publication November 3, 2005.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Annual legumes permit intensified cropping in no-till systems in the drought-prone northern Great Plains. Our objectives were to compare cropping sequence effects of pea (Pisum sativum L.) with fallow, mustard (Sinapis alba L.), and wheat (Triticum aestivum L.), and to measure the effects of pea harvest timing and shoot biomass presence on soil water use and N contribution, and yield and grain quality of subsequent wheat. Pea, mustard, wheat, and fallow preceded spring wheat at three sites in Montana. In the first year, two harvest timings (anthesis and maturity) were included and managed for presence or absence of crop shoot biomass. In the second year, a wheat test crop was grown at four N fertilizer rates. Regardless of management, pea used equal or less soil water, contributed equal or greater soil N, and had equal or greater positive impact on subsequent wheat growth than mustard or wheat. Compared with maturity, midseason harvest timing of pea increased soil N (30–39 kg NO3–N ha–1) and soil water (19–39 mm) available in the spring to the subsequent wheat test crop at two of three sites. Under severe drought, midseason harvest of pea increased wheat yield 50% and critically increased grain density compared with the mature pea harvest. At the N-limited site, midseason harvest of pea increased wheat yield 14% and grain protein 9% compared with mature pea harvest. Pea shoot biomass presence did not affect soil water or N, or growth of a subsequent wheat crop.

Abbreviations: DAS, days after seeding • LSD, least significant difference • LTA, long-term average • NIR, near infrared • NUE, nitrogen-use-efficiency • PASW, plant-available soil water


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
NO-TILL SYSTEMS promise to increase crop productivity in the northern Great Plains by harvesting a greater proportion of incident precipitation (Peterson et al., 1996; Nielsen et al., 2005). Producers may achieve greater precipitation use-efficiency by reducing fallow frequency and progressing to more intensive annual cropping systems (Farahani et al., 1998). However even within no-till systems, annual cropping carries considerable risk for much of the semiarid northern Plains (Zentner et al., 2002; Lyon and Peterson, 2005). For example, a recent drought cycle in northern Montana resulted in annual crop yields that were less than 40% of comparative fallow crop yields at two of five test locations (Miller and Holmes, 2005). An approach that has been promoted for reducing agronomic risk associated with annual cropping is to grow an annual legume crop early in the summer fallow period and to terminate the crop as green fallow before maturity to conserve soil water (Pikul et al., 1997). However recent research results have been mixed about the value of this practice. Nielsen and Vigil (2005) showed that an annual legume green manure cropping system was not viable in northern Colorado, while Zentner et al. (2004) provided a positive economic assessment of annual legume green manure in a long-term crop rotation study in southern Saskatchewan.

In addition to the termination timing of an annual legume crop, producers have questions about the effect of shoot biomass removal on the subsequent N contribution from the annual legume. Harvesting pea for annual forage is a profitable practice increasingly used in Montana and baling of pea straw after seed harvest has long been a practice associated with dry pea production. It would be useful to understand the impact of forage removal or straw harvesting on rotational benefits from annual legumes. Rotational N benefits from annual legumes have been reported in a review article of the Northern Great Plains region (Miller et al., 2002a), but no information could be found on the consequence of annual legume shoot biomass removal within a no-till cropping system.

Previously published reports regarding annual legume green manures in the northern Great Plains have been conducted in tillage-based systems. Several annual legume species have potential for use as green manure crops. A study conducted at Swift Current, SK, compared four annual legume species for green manure effects and concluded that pea was the most effective N2–fixer and contributed the greatest N response to spring wheat (Biederbeck et al., 1993; 1996). Similarly, Tanaka et al. (1997) concluded that pea was a superior green manure crop.

Often, reports on short- to medium-term (2–6 yr) response of, including an annual legume green manure, have been negative, due to excessive soil water use (Zentner et al., 1996; Nielsen and Vigil, 2005) or failure to cycle sufficient N from the crop residues (Pikul et al., 1997; Tanaka et al., 1997). Zentner et al. (2004) reported the first positive economic assessment of an annual legume green manure in a 12-yr crop rotation experiment at Swift Current, SK. They showed that growing lentil (Lens culinaris Medik) as green manure during the early summer fallow period improved net profitability after 6 yr of a 3-yr fallow–wheat–wheat rotation. This delayed response coincided with a change in the timing of lentil termination from midflower to first flower, resulting in less soil water extraction. Similarly, Cochran and Kolberg (2002) showed that spring wheat yields following lentil green manure equaled or exceeded the tilled fallow control after 4 yr of cropping history in a fallow–wheat rotation at Culbertson, MT. However, Cochran and Kolberg (2002) attributed the delayed response to cumulative change in soil N supply.

If positive agronomic effects from green manure accumulate slowly over time, farmers may be reluctant to adopt this practice unless short-term economic benefits are realized. Given that agricultural systems in the northern Great Plains derive significant income from cattle (Bos taurus), growing an annual legume forage crop during the summer fallow period could have value on-farm or in local cash markets. Welty (1984) reported that pea forage yielded greater protein per hectare than barley (Hordeum vulgare L.) or oat (Avena sativa L.) forage in a subhumid environment at Kalispell, MT. Sims et al. (1991) conducted a comprehensive study of forage yield that included 10 large-seeded annual legumes over 12 dryland site-years in Montana and concluded that pea forage yields were consistently greater than most other annual legume forages. Annual legume forage could provide an economic incentive for growing a partial crop during the summer fallow period but it is critical to know how forage removal affects N cycling to a subsequent cereal crop. The objectives in this study were: to compare cropping sequence effects of pea with fallow, mustard, and wheat; and more specifically, to measure the effects of pea harvest timing and shoot biomass presence for soil water use, soil N contribution, yield, and grain quality of a subsequent wheat crop.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Field Site Description and Experimental Design
A 2-yr cropping sequence study was conducted during 1999 to 2001 at three locations; Amsterdam, Denton, and Havre, MT. Site characteristics for this study are summarized in Table 1. All trials were established on commercial farms in wheat stubble fields with at least 3 yr of previous no-till management. All crops were directly seeded into winter wheat stubble. Monthly precipitation amounts (Table 2) were measured differently at each site. Growing season precipitation was collected with on-site manual rain gauges with a 1-cm layer of vegetable oil added to prevent evaporative loss from the gauge. Growing season precipitation data was also measured with an electronic "tipping bucket" rain gauge at Denton. Over-winter precipitation was obtained from the nearest meteorological station located, 2, 4, and 19 km from Havre, Denton, and Amsterdam, respectively. The climatic context ranged from drier than normal to near normal annual precipitation at each location–year.


View this table:
[in this window]
[in a new window]
  		Table 1. General soil characteristics for research sites at Denton and Havre, MT, 1999–2000 and Amsterdam, MT, 2000–2001.

 

View this table:
[in this window]
[in a new window]
  		Table 2. Over-winter and monthly growing season precipitation during consecutive growing seasons at three research sites in Montana.

 
The study was a randomized complete block design with a split-split-split plot arrangement, with Year 1 crops as main plots (29.2 by 14.6 m), harvest timing as subplots (14.6 by 14.6 m), shoot biomass removal as the subsubplot (7.3 by 14.6 m), and four N rates in the Year 2 test crop as the subsubsubplot (7.3 by 3.6 m). However, at Havre the subsubsubplot was 1.8 by 14.6 m. All treatments were replicated four times. Year 1 crops were pea, yellow mustard, and spring wheat and included a chemical fallow control within each block (7.3 by 14.6 m). Harvest timing occurred midseason (i.e., forage) or at crop maturity (i.e., grain) for all crops. The midseason harvest timing coincided with early bloom of pea. This timing for pea coincided with midflower for yellow mustard and the late boot stage for spring wheat (approximately Zadoks 50). For the maturity stage at Denton and Havre, all crops were harvested after the longest-maturing crop, wheat, was mature due to the remoteness of the sites and restricted local access to plot harvesters. Preharvest shattering of mustard did not occur at any of the sites, while preharvest shattering of pea did occur at Havre but not at the other two sites. This shattering at Havre likely decreased yields slightly (visually estimated to be 10–20%), but the shattering was similar across all treatments. At the midseason harvest timing, shoot biomass was sprayed with glyphosate [460 g N-(phosphonomethyl)glycine ha–1] and left in place ("present") or cut with a sickle-bar mower at 0.1 m and removed from the subsubplot by hand ("absent"), 0 (Havre), 6 (Denton), or 14 (Amsterdam) d after glyphosate application. Similarly, at maturity, each crop was harvested for grain at a 0.1-m stubble height and residual shoot biomass (i.e., straw) deposited on the plot ("present"), or the mature shoot biomass was cut with a sickle-bar mower at 0.1 m and removed unthreshed from the subsubplot by hand ("absent"). In Year 2, a spring wheat test crop was sown at four N rates (0, 22, 45, and 90 kg ha–1), oriented transverse to the Year 1 plots at Denton and Amsterdam. At Havre, fertilizer N rates were applied within subsubplots in the same direction as the previous year's sowing. Details of agronomic management are presented in Table 3. During Year 1, all crops were directly seeded into wheat stubble. Mustard seed was treated with 0.3 g captan {3a,4,7,7a-tetahydro-2-[(trichloromethyl)thio]-1H-isoindole-1,3(2H)-dione} and 20 g thiram (tetramethylthiuram disulfide) kg–1, pea seed was treated with 150 mg metalaxyl [methyl-N-(methoxyacetyl)-N-(2,6-xylyl)-DL-alaninate] kg–1, and wheat seed was treated with 15 mg tebucanozole [(RS)-1-p-chlorophenyl-4,4-dimethyl-3-(1H-1,2,4-triazol-1-ylmethyl)pentan-3-ol] and 20 g metalaxyl kg–1. Crop densities at all sites met or exceeded reported threshold densities.


View this table:
[in this window]
[in a new window]
  		Table 3. Agronomic factors for cropping sequence study at Denton and Havre, 1999–2000, and Amsterdam, MT, 2000–2001.

 
Cultural Practices, Field, and Laboratory Measurements
At each location, soil samples were collected near seeding in the spring and after harvest in the fall during Year 1, and near seeding in the spring of Year 2 of the 2-yr crop sequence (Table 3). Soil depth above petrocalcic or gravelly layers varied among locations preventing us from sampling a typical 1.2-m soil profile at all sites. Thus, the soil sampling depth was limited to 0.6 m at Denton and 0.9 m at Havre. Soil cores (30-mm diam.) were divided into 0.3-m depth segments and placed in plastic-lined paper sample bags, or in plastic freezer bags, and placed in storage coolers for transport from the field to the laboratory. At the laboratory, the samples were either weighed "wet", opened, and placed upright in a drying oven, or placed in a freezer (–20°C) until recording wet weights and placing in the drying oven. Soil samples were oven-dried (50°C), "hot-weighed" immediately from the oven, ground to pass a 2-mm sieve, and analyzed for NO3–N according to the procedure of Hamm et al. (1970). Bulk density values for the sampling depths were determined at each site by dividing the oven-dried soil mass values by volumes of soil cores. Bulk density values were averaged across all cores (by depth) and applied to the soil NO3–N concentration and water content values to determine soil NO3–N content and volumetric water content. In Year 1, spring soil N and water values were averaged by replicate, and considered uniform within a replicate. Eight soil cores per replicate were taken at Havre, and 14 cores per replicate at Amsterdam and Denton. After that, postharvest and before seeding, two soil cores were taken and combined to form the sample for each subsubplot.

Crops were seeded at each site with plot-scale no-till seeders according to machinery availability. All provided satisfactory crop emergence. All fertilizer was applied at seeding as a side-band, or as a combination of side-band and seed-placed fertilizer, except at Havre in 2000, where N was applied on the soil surface above the seed row in a dribble band (Table 3). During Year 1, N fertilizer was applied to mustard and wheat plots to achieve 90, 78, and 67 kg ha–1 of available N (soil NO3–N to 60 cm + fertilizer N) at Havre, Denton, and Amsterdam, respectively. Although these represent modest levels of available N, wheat grain protein was consistently greater than 147 g kg–1, indicating that N did not limit wheat yield at any site (Engel et al., 1999; Selles and Zentner, 2001), nor likely mustard yield. Pea seed was treated with an appropriate strain of rhizobia to ensure nodulation and provide for plant N nutrition (Table 3). Monitoring of root nodulation confirmed successful inoculation of pea at all sites, although the nodulation response was slow at Amsterdam and pea exhibited symptoms of N deficiency early in the season. To correct this problem, 22 kg N ha–1 (urea formulation) was broadcast applied 50 d after seeding and before a low intensity rain event (8 mm). Symptoms of plant N deficiency disappeared within 4 d, and plants remained green the remainder of the season at this site. Fertilizer P, K, and S, were applied to the field sites to ensure these nutrients were not limiting (Table 3).

Plant-available soil water (PASW) before seeding was estimated to depths of 0.6 m at Denton, 0.9 m at Havre, and 1.2 m at Amsterdam, by subtracting lower limits (Ritchie, 1981) of 86 mm at Denton, 123 mm at Havre, and 113 mm at Amsterdam (Table 3). At Denton and Havre it is likely that additional water was accessed by plants below the depths of our soil measurements but the water holding capacity in those gravelly layers would be expected to be minimal.

Weeds were managed through use of recommended herbicides for all crops and fallow was managed chemically with glyphosate applied at least three times during the growing season at 690 to 920 g a.i. ha–1. Herbicides were applied in 94 L ha–1 of water with flat fan nozzles at a pressure of 207 to 276 kPa using a custom-made 7.3-m wide shrouded sprayer (Brad Gregoire, Havre, MT) at all sites. Minimal supplemental hand weeding was conducted to ensure weeds did not affect crop yield. In 1999, pea was injured by MCPA-amine (2-methyl-4-chlorophenoxyacetic acid) applied at the high end of the range for recommended application in Montana (830 g a.i. ha–1) that delayed crop growth for 2 to 3 wk and reduced midseason biomass and likely reduced grain yields.

Tiller density was determined at maturity by counting two 0.5-m row segments at opposite ends of each subsubsubplot. Shoot biomass was measured at midseason and maturity by hand-harvesting (0.1 m above the ground) two randomly chosen locations equivalent to approximately 1 m2 each. Biomass samples were oven dried at 50°C and weighed to estimate yield. Shoot N concentrations (i.e., green forage at midseason harvest and straw only at maturity) were determined using a LECO CNS combustion analyzer (LECO Corporation, St. Joseph, MI) to estimate the amount of N returned to plots where shoot residues remained. All crops were harvested directly with a plot combine for grain yield measurement using a swath width of 1.5 m, except during Year 2 at Denton and Havre where the swath width was 1.2 m. Plot "edge effects" were minimized by leaving edge rows unharvested and the number of crop rows harvested were recorded to ensure accurate yield measurement. Thus, the actual harvested area was 22 m2 in Year 1, and varied among sites from 9 to 22 m2 in Year 2. Grain samples were oven dried at 50°C for 72 h, cleaned, weighed, and reported on a dry matter basis. Seed density (i.e., test weight) was determined on grain subsamples using a 0.95-L container. Grain N concentrations for Year 1 crops were determined using a LECO CNS combustion analyzer and for the spring wheat test crop using whole-grain near infrared (NIR) reflectance spectroscopy with an Infratec 1225 instrument (FOSS Analytical A/S, Denmark). Grain N concentrations were converted to protein concentration using a factor of 5.7 for wheat and 6.25 for mustard and pea (Jones, 1941). Nitrogen-use efficiency (NUE) was estimated by dividing the grain N yield by available N or soil NO3–N at seeding plus applied fertilizer N. This estimation of NUE was used for comparative purposes only since it ignores N contained in the wheat straw, mineralization of soil N during the growing season, and potential soil extraction of N by crop rooting below 0.6 m.

Statistical Analysis
An ANOVA was performed by location-year with JMP IN (SAS Institute, 2005). Sites were analyzed independently due to heterogeneity of error variances and different treatment responses at each site that caused large year x treatment interactions. Blocks (replicates) were considered a random effect while crop, harvest timing, shoot biomass presence, and N fertilizer rate were considered "fixed" effects. Data were analyzed in two stages. In the first stage, all data were included and analyzed simply for crop effect, of which chemical fallow was one of four categories for the crop classification variable. Year 1 analyses used residual error to test significance among simple crop effects while Year 2 analyses, which included the N fertilizer rate variable, used the replicate x crop interaction to test crop main effects. At Havre, since the midseason harvest timing was absent for wheat, only the mature harvest timing was analyzed to determine crop effect. In the second stage, only pea management treatments were analyzed. For Year 1 soil water and N analyses, chemical fallow treatments were omitted and orthogonal contrasts were used to test pea management effects. Separate analyses of pea treatments only were performed to test specifically the interaction of pea harvest timing and shoot biomass presence or absence. For Year 2 wheat test crop analyses, only pea management plots were included in a split-split-plot analysis with harvest timing as the main plot, shoot presence as the subplot, and N fertilizer rate as the subsubplot. Effects were declared significant at P < 0.10 unless otherwise specified.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Year 1 Crop Productivity
In Year 1, Amsterdam and Denton experienced more severe drought than Havre (Table 2) causing different shoot biomass accumulation patterns among crops within each location (Table 4). For example mustard shoot biomass accumulation reached its maximum by the midseason sampling date, 66 and 68 d after seeding (DAS) at Denton and Havre, while wheat shoot biomass increased marginally at Denton and more than doubled at Havre between midseason and crop maturity sampling times (88 and 105 DAS). At Amsterdam, all crops increased biomass markedly between the midseason sampling date (67 DAS) and maturity (98 DAS). However, the remarkable ninefold increase in shoot biomass accumulation for pea was due to a temporary N deficiency that severely limited early plant growth, and that was corrected with a rescue broadcast N application described in the methodology above. No clear pattern emerged with respect to the impact of crop type on biomass N yields. Ranking of biomass N yields for pea, mustard, and wheat varied with each location and harvest timing. Plant C/N ratios were narrowest for pea, except at the Amsterdam mature sampling, and became wider as crops matured. At Amsterdam, a significant rainfall event (50 mm) between physiologic maturity and the actual date of pea harvest probably resulted in significant soluble N losses from the standing dead plant tissue.


View this table:
[in this window]
[in a new window]
  		Table 4. Biomass and grain attributes for Year 1 for dry pea, yellow mustard, and spring wheat at Denton and Havre, MT, 1999, and Amsterdam, MT, 2000.

 
Seed yields were low for all crops at Denton due to a midseason hail event (22 June). Mustard seed yields were low at all locations due to poor synchrony of crop growth and water use. Mustard grew vigorously early in the season, rapidly depleting soil water in the rooting zone, and resulting in very low harvest indices at all locations (0.07–0.15). At Havre, pea seed yield was likely constrained because of herbicide injury due to a recommended rate of MCPA-amine that delayed growth by 2 to 3 wk (see methodology section). At Amsterdam, relative seed yields under drought stress were consistent with previous reports where pea equaled spring wheat yield (Miller et al., 2001, 2002a). Seed N yield was equal for pea and wheat, both markedly greater than mustard, at all sites.

Year 1 Crop Effects
Soil Nitrogen
The effects of Year 1 crop and pea management effects on soil N were inconsistent among sites (Table 5). At Denton and Havre, postharvest soil NO3–N was greater (22–39 kg N ha–1) under chemical fallow than under cropped plots. By the following spring soil NO3–N levels had increased proportionately more under the cropped areas than under fallow. Hence, no differences were evident among any treatments at Denton in the spring. At Havre, spring soil NO3–N levels were greater under pea than mustard and wheat stubbles. Soil NO3–N under pea stubble was also even greater than under chemical fallow. Soil under the chemical fallow treatment was saturated in early spring and substantial N losses probably occurred as a result of denitrification. This is evidenced by the fact that spring soil NO3–N levels were significantly (P < 0.05) lower than those observed the preceding fall. An identical response was observed in a chemical fallow treatment at another study located at the same time and location (Miller and Holmes, 2005). It is important to note that soil N was measured in the 0.6- to 0.9-m soil layer to see if measurable N leaching occurred from upper to lower soil layers and no evidence of this was found. In fact, decreases of NO3–N of similar proportion also occurred in the 0.6- to 0.9-m soil layer.


View this table:
[in this window]
[in a new window]
  		Table 5. Means for crop effect and pea management effects for fall and spring soil NO3–N to 0.6-m soil depth at Denton and Havre, MT, 1999–2000, and Amsterdam, MT, 2000–2001.

 
We hypothesized that soil nitrate N would be greater for the midseason compared with the mature harvesting timing of pea, based on a narrower C/N ratio (midseason pea biomass averaged a C/N ratio of 13 compared with 46 at maturity), a longer period for biomass decomposition, and less extensive utilization of indigenous soil N. Pea harvested midseason resulted in greater soil nitrate than pea harvested at maturity, for both sampling dates at Havre and Amsterdam, but this did not occur at Denton (Table 5). However, these differences in soil N between harvest timings did not relate to pea biomass N yield, which did not differ between harvest timings at Havre or Amsterdam. In contrast, pea biomass N yield at Denton was 22 kg ha–1 greater at the midseason compared with the mature harvest timing but no soil N differences were detected.

Contrary to our hypothesis, pea shoot/straw presence generally did not affect soil NO3–N. At one of three sites, an increase of 13 kg NO3–N ha–1 was associated with pea shoot/straw presence in the spring at Havre. The results of this study indicate that in the short-term, soil NO3–N tests were not very sensitive to differences in pea residue N content or presence. Although somewhat greater management (shoot presence/absence) effects on spring soil nitrate N levels were anticipated, the results are perhaps not too surprising given that large size and dynamic nature of the soil organic N pool relative to the magnitude of differences in N either added, or removed, by pea. The lack of relationship between shoot biomass and soil N dynamics has been reported previously (Bremer and van Kessel, 1992; Stevenson and van Kessel, 1996; Beckie and Brandt, 1997; Beckie et al., 1997).

Soil Water
Postharvest plant-available soil water was greatest under chemical fallow at all sites, and PASW under pea stubble was greater than mustard stubble at Havre and Amsterdam, locations with moderate to deep soil profiles (Table 6). By spring, no differences in PASW were evident at Denton due to the shallow soil profile at that location, small differences remained at Amsterdam, and large differences remained at Havre. At Havre, crop stubbles held 57 to 74 mm less PASW than chemical fallow, with PASW under pea stubble 19 mm greater than mustard stubble. At Amsterdam, PASW under pea stubble equaled chemical fallow and was 15 mm greater than mustard stubble. Consistent with these results, comparatively low soil water use by field pea has been reported previously in this region (Miller et al., 2002b, 2003; Miller and Holmes, 2005).


View this table:
[in this window]
[in a new window]
  		Table 6. Means for crop effect and pea management effects for fall and spring plant available soil water (PASW) at Denton and Havre, MT, 1999–2000, and at Amsterdam, MT, 2000–2001.

 
The midseason harvest timing for pea conserved 41 and 47 mm greater PASW than the mature harvest timing, at Havre and Amsterdam, respectively (Table 6). By spring, harvest timing effects remained large at Havre (39 mm) but diminished at Amsterdam (19 mm). Shoot biomass presence/absence generally had no effect on PASW, except at Havre in 2000 when, contrary to our hypothesis, the absence of shoot biomass caused 9 mm greater PASW. This result, however, was an experimental artifact. Glyphosate was applied to terminate growth at the midseason harvest timing and, at Havre, the shoot biomass "absence" plots were removed later the same day, ceasing soil water use immediately. Glyphosate is a translocated herbicide that generally requires 10 to 14 d for complete tissue necrosis and during this period the crop continues to use water. Thus, at Havre, the shoot biomass "presence" plots had a longer period of water use than the "absence" plots. To prevent this artifact at Denton and Amsterdam, midseason crop biomass was not removed until 6 and 14 d after glyphosate application, respectively. Interaction of harvest timing with shoot biomass/presence generally was not detected and the lone interaction at Havre in 2000 was due to the experimental artifact described above. There, shoot biomass absence caused 19 mm greater PASW than shoot biomass presence at the midseason harvest timing, but did not differ at the mature harvest timing. Thus, at Havre, the net effect of midseason harvest timing with biomass removal resulted in 98 mm of spring PASW for pea, equal with chemical fallow (107 mm) and more than double that under spring wheat harvested at maturity (43 mm). This highlights the potential for midseason shoot biomass (i.e., forage) removal of pea for conserving soil water.

Year 2 Wheat Test Crop
Representative of conventional farming practice, spring wheat yield in the chemical fallow control ranged from 1.7 to 2.6 Mg ha–1 among sites. These sites differed markedly in the soil effects caused by different crop sequence strategies. On the shallow soil at Denton, there were no treatment differences in soil nitrate N or PASW at the time of seeding for the test crop (Tables 5 and 6). At Havre, the previous crop and harvest timings affected both soil nitrate N and PASW but a 34-d drought from 2 June to 5 July made water the chief limiting factor. Conversely, more plentiful PASW and timely growing season rainfall at Amsterdam, made soil N an important limiting factor. Thus, spring wheat test crop response was considered independently for each site.

Denton: Soil Nitrogen and Water Not Yield Limiting
At Denton, the low early season biomass, grain yield, grain protein, and grain N yield for the Year 2 wheat test crop grown on wheat stubble was due to a critical harvest timing x shoot biomass presence interaction in the Year 1 wheat crop (Table 7). For example, wheat grown on wheat stubble harvested at maturity, with shoot biomass present (i.e., straw remaining on the plots), yielded only 1.08 Mg ha–1, 38% less than the chemical fallow control. This yield reduction was not related to soil N or water. These wheat plots exhibited symptoms of chlorosis localized neatly within the previous year's plot boundaries, easily distinguished from the deep green foliage characteristic of other plots at this site. Samples of these chlorotic plants were analyzed at Montana State University Plant Diagnostics Lab and Fusarium oxysporum was isolated from the tissue (J. Riesselman, MSU Plant Diagnostics Lab Director, personal communication). Apparently volunteer wheat seedlings served as a critical bridge to this disease because where the wheat plots were terminated midseason or when shoot biomass was removed at maturity (i.e., wheat stems and heads removed unthreshed), no yield loss occurred. This commercial farm field at Denton had a history of aggressive N fertility management and soil nitrate N levels were high, preventing N from being yield limiting. However, grain protein concentration and grain N yield responded linearly to increased fertilizer N indicating successful plant uptake (Table 7). As expected in a case where yield was not N-limited, the NUE declined with each successive rate increment.


View this table:
[in this window]
[in a new window]
  		Table 7. Means for crop effect and pea management effects on subsequent wheat productivity and grain quality parameters at Denton, MT, 2000.

 
There were no important effects of pea management at Denton. Head density in the wheat test crop was 5% greater following the midseason harvest timing of pea, and 3% greater when shoot biomass remained on the plots, but in neither case did these small differences translate to grain yield differences. At Denton, the chief economic constraint for the wheat test crop would have been the grain density (i.e., test weight) values below 747 kg m–3 (58 pounds bushel–1), the minimum test weight for U.S. No. 1 dark red northern spring wheat. All grain density values at Denton fell considerably below this standard.

Havre: Soil Water Was the Chief Limiting Factor
At Havre, a 34-d drought from 2 June to 5 July presented a key water constraint to the Year 2 wheat test crop response. Spring wheat biomass harvested 68 DAS on 16 June showed no effect due to the previous crop. However, head density, grain yield, grain density, grain N yield, and NUE measured at maturity were all much greater for chemical fallow (Table 8). Few differences were observed among the crop stubbles, but where they occurred, wheat test crop productivity/quality was lowest on mustard stubble, while pea and wheat stubbles did not differ (note: at Havre, crop comparisons were made based on the mature harvest timing only due to the absence of a midseason harvest treatment for Year 1 wheat). At Havre, important economic parameters for the wheat test crop were grain yield and grain density. Wheat grown on Year 1 crop stubbles yielded 36% (mustard) to 56% (pea) of the chemical fallow control. Previous economic research in the neighboring province of Saskatchewan reported the break-even yield for recropped spring wheat to be within a range of 67 to 84% of the yield attained on fallow at the same location (Zentner et al., 1986). Using this guideline, it was clearly unprofitable to grow wheat on any crop stubble at Havre in 2000. Further, grain density values were below 747 kg m–3 on all crop stubbles potentially compounding yield loss with additional grade loss. This suggests that water stress by wheat occurred during the grain fill period.


View this table:
[in this window]
[in a new window]
  		Table 8. Means for crop effect and pea management effects on subsequent wheat productivity and grain quality parameters at Havre, MT, 2000.

 
However, at Havre, pea harvest timing had important effects on the productivity and quality of the wheat test crop. Midseason pea harvest timing increased grain yield by 50% compared with the mature harvest timing, resulting in a grain yield that was 84% of the chemical fallow control. This yield response was correlated with a 34% increase in final head density (r = 0.75, P < 0.01) and a 40% increase in grain N yield (r = 0.83, P < 0.01). Importantly, the grain density following the midseason pea harvest timing (737 kg m–3) was near the minimum standard of 747 kg m–3. Pea shoot biomass presence had little effect on wheat test crop parameters. Midseason biomass of the wheat test crop was 7% higher where shoot biomass was absent, due mainly to greater soil water under the midseason pea harvest treatment with shoot biomass removed. However, final head density, grain yield, grain density, and grain N yield were unaffected by pea shoot biomass presence. Grain protein was slightly higher in the presence of shoot biomass (P = 0.08) but this did not translate to greater grain N yield (P = 0.31). Two notable interactions affected grain protein. The pea harvest timing x shoot biomass presence interaction resulted in lower protein concentration in the absence of shoot biomass at the midseason pea harvest timing (148 vs. 163 g kg–1), while shoot biomass presence had no effect at the mature harvest timing (170 vs. 167 g kg–1). This may be due to differential soil water content from shoot biomass presence/absence at the midseason pea harvest, discussed above. The pea harvest timing x N fertilizer rate interaction was due to a moderate increase in grain protein following the midseason pea harvest timing (140–168 g kg–1) compared with a steep increase following the mature pea harvest timing (135–201 g kg–1). This differential response to N fertilizer rate likely relates to greater PASW following the midseason pea harvest timings, resulting in 50% greater grain yield, and therefore, protein dilution. As at Denton, NUE varied inversely with N fertilizer rate because N was not the chief yield-limiting factor at Havre either.

Amsterdam: Soil Nitrogen Was the Chief Limiting Factor
Due to timely growing season rainfall at Amsterdam, the wheat test crop responded to soil N differences. The crop effects of pea were in the highest statistical categories for all wheat test crop parameters (Table 9). Midseason biomass following pea was 15% greater than chemical fallow, translating to a final grain yield that was 5% greater than chemical fallow. This positive crop sequence effect occurred only for pea, as mustard and wheat stubbles resulted in grain yields that were 7 and 11% less than chemical fallow. Nitrogen-use-efficiency was 30 to 42% lower on chemical fallow than all Year 1 crop stubbles, indicating that a greater proportion of soil NO3–N measured before seeding, or N fertilizer applied at seeding, was not taken up by the wheat test crop in chemical fallow.


View this table:
[in this window]
[in a new window]
  		Table 9. Means for crop and pea management effects on subsequent wheat productivity and grain quality parameters at Amsterdam, MT, 2001.

 
Pea harvest timing caused important differences in all wheat test crop parameters, due to the greater availability of soil N and water following the midseason harvest timing. Midseason pea harvest timing resulted in 14% greater grain yield and 9% greater grain protein, integrated as 24% greater grain N yield, compared with the mature pea harvest timing. Clearly, the pea crop effect, combined with midseason harvest timing, resulted in superior availability of N to the wheat test crop. However, contrary to our hypothesis, pea shoot biomass presence did not affect any wheat test crop parameter at this highly N responsive site.

All wheat test crop parameters, with the notable exception of NUE, were highly responsive to N fertilizer rates. Grain yield responded linearly with N fertilizer rate, reaching the yield "sufficiency" level only at the highest (90 kg ha–1) N rate in this study (Engel et al., 1999; Selles and Zentner, 2001). Constant NUE across widely varying N rates reflected efficient crop uptake of this important yield-limiting factor. The additive effects of previous crop type and harvest timing caused wheat following pea harvested at midseason to yield 12% greater than the chemical fallow control, highlighting the potential cropping systems benefits of growing pea as an annual forage in wheat-based systems.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Pea conferred stronger rotational benefits to wheat than mustard by conserving greater soil water and contributing greater soil N, especially when growth was terminated midseason. The occurrence of a severe drought at one location in this study highlighted the potential for midseason growth termination strategies in pea to mitigate drought risk. Contrary to conventional wisdom, pea shoot biomass presence/absence had no important effect on soil water, soil N, or growth of a subsequent wheat crop. The failure of shoot biomass presence to confer positive soil and crop growth benefits was observed within the context of a 2-yr cropping sequence, and a different response may be observed in a long-term cropping systems study.


   ACKNOWLEDGMENTS
 
This research was supported by the Montana Fertilizer Tax Advisory and the Montana Agricultural Experiment Station. Gregg Carlson, Dave Wichman, Jody McConnell, and Brad Gregoire supplied valuable technical assistance. We appreciate the farm field access provided by Rich Barber (Denton), Alec McIntosh (Havre), and Matt Flikkema (Amsterdam).


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 J. Environ. Qual.Home page
M. P. Dusenbury, R. E. Engel, P. R. Miller, R. L. Lemke, and R. Wallander
Nitrous oxide emissions from a Northern Great Plains soil as influenced by nitrogen management and cropping systems.
J. Environ. Qual., March 1, 2008; 37(2): 542 - 550.
[Abstract] [Full Text] [PDF]

Home page
 Agron. J.Home page
F. L. Walley, G. W. Clayton, P. R. Miller, P. M. Carr, and G. P. Lafond
Nitrogen Economy of Pulse Crop Production in the Northern Great Plains
Agron. J., November 6, 2007; 99(6): 1710 - 1718.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Miller, P. R.
Right arrow Articles by Holmes, J. A.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Miller, P. R.
Right arrow Articles by Holmes, J. A.
Agricola
Right arrow Articles by Miller, P. R.
Right arrow Articles by Holmes, J. A.
Related Collections
Right arrow Other Legumes
Right arrow Nitrogen
Right arrow Other Forage Crops
Right arrow Crop Systems
Right arrow Water Conservation

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Crop Science Vadose Zone Journal
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome