Introduction
Dry beans (Phaseolus vulgaris), also known as common beans, edible beans and field beans, have well-documented human health benefits (Mitchell et al., Reference Mitchell, Lawrence, Hartman and Curran2009; Hayat et al., Reference Hayat, Ahmad, Masud, Ahmed and Bashir2014; Messina, Reference Messina2014), which has resulted in an increase in production to meet demands from national programs (USDA-FNS, 2012). Consumers are also demanding local, organic foods; thus US sales of organic dry beans increased from US$8.4 million in 2008 to US$29.9 million in 2014 (USDA-ERS, 2016). Organic dry beans were harvested from about 300 ha in Minnesota in 2014, which accounted for 3% of total US organic dry bean cropland (USDA-NASS, 2014). By contrast, organic soybeans were harvested from 4830 ha in Minnesota in the same year (USDA-NASS, 2014). Growers in the region are more accustomed to producing soybeans than dry beans given historical commodity crop trends in the region, and a large market exists for organic soybeans (McBride and Greene, Reference McBride and Greene2009).
Despite the strong interest and relatively large production of soybean in Minnesota, there has been a fourfold increase in organic dry bean production in the state since 2008, with the majority of growth occurring in the dark red kidney, navy and pinto bean types. Given rising demand, prices for organic food grade dry beans are generally two to three times greater than for conventionally produced dry beans (USDA-AMS, 2016). Organic growers in the Upper Midwest could improve revenues from the pulse phase of their rotation by substituting dry bean for soybean (Glycine max), while producing food-grade grains for local food systems. Furthermore, the incorporation of alfalfa and dry bean into organic rotations could contribute to greater landscape diversity in the Midwestern USA, where 2-year corn-soybean rotations predominate. The dominance of corn-soybean rotations has been criticized for contributing to losses of landscape-scale biodiversity (Karlen et al., Reference Karlen, Hurley, Andrews, Cambardella, Meek, Duffy and Mallarino2006; Meehan et al., Reference Meehan, Werling, Landis and Gratton2011), whereas introducing dry bean and forage crops into rotations can increase biodiversity. With greater diversity comes the potential for enhanced biocontrol of crop pests (Gardiner et al., Reference Gardiner, Landis, Gratton, DiFonzo, O'Neal, Chacon, Wayo, Schmidt, Mueller and Heimpel2009), increased soil quality (Karlen et al., Reference Karlen, Hurley, Andrews, Cambardella, Meek, Duffy and Mallarino2006), lower nitrogen (N) leaching (Coiner et al., Reference Coiner, Wu and Polasky2001) and decreased weed soil seedbanks (Teasdale et al., Reference Teasdale, Mangum, Radhakrishnan and Cavigelli2004).
Dry beans fit well into existing crop rotations in Minnesota, but they are relatively inefficient at biologically fixing atmospheric nitrogen (N2) compared with other pulses (Isoi and Yoshida, Reference Isoi and Yoshida1991; Peoples et al., Reference Peoples, Brockwell, Herridge, Rochester, Alves, Urquiaga, Boddey, Dakora, Bhattarai, Maskey, Sampet, Rerkasem, Khan, Hauggaard-Nielsen and Jensen2009). Unlike soybean, dry beans benefit from supplemental N, especially if soils are low in organic matter, or if nodulation is poor due to insufficient inoculation by Rhizobium (Kaiser et al., Reference Kaiser, Lamb and Eliason2011; Farid and Navabi, Reference Farid and Navabi2015). Only 50% of pulse crops’ N requirement is generally derived from biological N2 fixation (compared with >80% in forage legumes) and much of this fixed N is removed through grain harvest (Watson et al., Reference Watson, Atkinson, Gosling, Jackson and Rayns2002). In organic systems, legume crops are widely used in crop rotations to build soil fertility and increase subsequent crop yields due to their N fixing ability (Blackshaw et al., Reference Blackshaw, Molnar, Clayton, Harker and Entz2007; Heilig and Kelly, Reference Heilig and Kelly2012). However, typically little long-term soil N benefit is derived from adding pulse crops to a rotation, as large proportions of fixed N are removed via grain harvest (Kumar and Goh, Reference Kumar and Goh2000). When grown in rotation with wheat and other N-responsive crops, dry beans have been found to contribute only 5–12 kg N ha−1 to soils for subsequent crop production (Miller et al., Reference Miller, Waddington, McDonald and Derksen2002; Kaiser et al., Reference Kaiser, Lamb and Eliason2011). Alternatively, organic growers could incorporate perennial forage legumes like alfalfa (Medicago sativa) into rotations to reduce the need for additional N inputs. In Minnesota, alfalfa N credits for bean production range from 25 to 80 kg N ha−1, depending on the stand density of the terminated alfalfa (Kaiser et al., Reference Kaiser, Lamb and Eliason2011). Christenson et al. (Reference Christenson, Gallagher, Harrigan and Black1995) observed 10% greater yields for dry bean when grown following alfalfa in a rotation, compared with following corn.
Along with soil fertility considerations, organic growers must design crop rotations with weed management in mind. Weed prevention and suppression is important in organic systems, particularly in dry bean, as yield reductions of up to 99% have been reported if weeds are not controlled throughout the growing season (Amador-Ramirez et al., Reference Amador-Ramirez, Wilson and Martin2001). Integrating alfalfa into a crop rotation can influence weed dynamics, as alfalfa has a different growth habit, harvest regime and life cycle than annual row crops, potentially reducing the need for mechanical weeding operations that disturb the soil and stimulate weed seed germination (Teasdale et al., Reference Teasdale, Mangum, Radhakrishnan and Cavigelli2004; Tautges et al., Reference Tautges, Burke, Borrelli and Fuerst2017). Seed production of both grass and broadleaf weeds has been observed to be lowest in alfalfa, compared with corn (Zea mays) and soybean where weeds were controlled with tillage (Clay and Aguilar, Reference Clay and Aguilar1998; Kegode et al., Reference Kegode, Forcella and Clay1999). Weed seedling emergence has been observed to decrease by 90% over the life of a 3-year alfalfa stand (Anderson, Reference Anderson2010), and residual suppressive effects of alfalfa on weeds have been observed in subsequent crops in a rotation (Tautges et al., Reference Tautges, Burke, Borrelli and Fuerst2017).
Although dry bean production has been studied in conventional cropping systems in Minnesota (Burnside et al., Reference Burnside, Ahrens, Holder, Wiens, Johnson and Ristau1994; de Jensen et al., Reference de Jensen, Kurle and Percich2004), there is a lack of information on production of dry beans in organic crop rotations. As both environment and dry bean genotype variation influence N2 fixation, grain yields and subsequent N contribution (Farid and Navabi, Reference Farid and Navabi2015), there is a need to evaluate dry bean productivity in common Midwestern crop rotations, productivity among bean types and subsequent rotational effects of dry bean in diverse environments and soils. Therefore, our objectives were to (1) determine the effect of preceding corn and alfalfa crops on yields of five dry bean types compared with food grade soybean; (2) evaluate the effect of preceding alfalfa and corn crops on weed biomass and residual soil nitrate in dry bean and soybean, and (3) evaluate the effects of dry bean compared with soybean, on following wheat yields.
Methods
Location description
The field trials were conducted at four locations (referred to throughout as ‘Lamberton’, ‘Madison’, ‘Becker’ and ‘Rosemount’; Table 1) in Minnesota from 2010 to 2015. The trials at the University of Minnesota's Southwest Research and Outreach Center near Lamberton, Minnesota and an on-farm site near Madison, Minnesota were established on certified organic land on a Webster clay loam (fine-loamy, mixed, superactive, mesic Typic Endoaquolls). The two field trials at the University of Minnesota UMore Park near Rosemount, Minnesota and the Sand Plains Research Farm near Becker, Minnesota were not established on certified organic land, but were managed organically with no synthetic chemical fertilizer or pesticide inputs for 3 years prior to the trial's establishment. The trials at Rosemount and Becker were on a Tallula silt loam (coarse-silty, mixed, superactive, mesic Typic Hapludolls) and a Hubbard loamy sand (sandy, mixed, frigid Entic Hapludolls), respectively.
Locations of four organically managed field sites in Minnesota, with stand age of alfalfa and nitrogen (N) fertilization rates of corn crops, preceding planting of six bean types.
1 Composted cow manure.
2 Composted hog manure.
Experimental design and crop management
Five dry bean types and soybean were planted following corn and alfalfa in a randomized complete split plot design with four replicates, with preceding crops alfalfa and corn as the whole-plot treatment and bean type as the split-plot treatment. The trial was established in 2011 at the Rosemount location and in 2013 at the Becker, Lamberton and Madison locations. The locations were selected to reflect a range of fertility scenarios, to enable examination of the effects of preceding crops alfalfa and corn on bean yields in a range of organic fertility management programs. Alfalfa varied in stand age but received no N fertilizer at any of the locations, whereas fertility management of corn varied by location (Table 1). Following the final cutting of alfalfa and corn grain harvest, residues were disked and incorporated in the fall by chisel plowing.
In the spring following alfalfa termination and corn harvest, the seedbed was prepared with a field cultivator and bean crops were planted in early June after being inoculated with the proper Rhizobium strain. A food grade soybean (cv. MN 1505SP), and five dry bean types: black (cv. Eclipse), heirloom (cv. Peregion), dark red kidney (cv. Red Hawk at Rosemount; cv. Montcalm at Becker, Lamberton, and Madison), navy (cv. OAC Rex) and pinto (cv. Lariat) were grown. Seeding rates were 370,658 seeds ha−1 for soybean, 222,395 seeds ha−1 for small-seeded black, heirloom, and navy beans, and 172,974 seeds ha−1 for large-seeded kidney and pinto. Soybean and dry beans were seeded in four rows to split plots of 3 × 6 m2 in early June. Beans were rotary hoed and inter-row cultivated twice to control weeds when soil moisture conditions permitted. Organic crop rotations in the region typically follow a pulse crop with a small grain; therefore, the effects of preceding bean types on subsequent wheat yields were assessed. Following the bean crop treatments, the seedbed was prepared with a field cultivator and spring wheat (cv. RB07 or Prosper) was drilled across all plots at a rate of 135 kg ha−1 in the beginning of May the following year.
Data collection
Bean yields were determined by hand-harvesting a 3.75-m2 area in the center of each split plot in the second or third week of September following physiological maturity. Grain was threshed and dried at 35 °C, and yields were calculated based on a 13% moisture level. Wheat was hand-harvested from a 1-m2 area from the center of each plot, threshed and dried at 35 °C. Yields were adjusted to 12% moisture content. Weed biomass was measured by clipping weeds from a 1-m2 area within each plot at both bean and wheat harvests, though weed sampling in wheat did not occur at Rosemount. Weeds were dried at 35 °C and yields reported on a dry matter basis. At the beginning of the experiment and before planting bean and wheat, soil samples were collected to 15 cm depth and analyzed for pH, P and K, which indicated adequate levels for beans and wheat based on University of Minnesota Guidelines (Kaiser et al., Reference Kaiser, Lamb and Eliason2011). Also, in the spring before planting beans and wheat, soil samples were collected from each plot to a depth of 61 cm and samples were extracted for nitrate-N with potassium chloride and analyzed using the cadmium reduction method (Dorich and Nelson, Reference Dorich and Nelson1984).
Statistical analysis
As the four locations differed greatly in soil type and fertility, locations were analyzed separately. The effects of preceding crops alfalfa and corn and bean type on bean yield and weed biomass were analyzed with analysis of variance (ANOVA), using the ‘nlme’ package (Pinheiro et al., Reference Pinheiro, Bates, DebRoy, Sarkar, Heisterkamp, Van Willigen and Maintainer2017) in R version 3.2.2 (R Core Team, 2015). The main effects of ‘preceding crop’ and ‘bean type,’ and interactions between main effects, were treated as fixed effects and ‘block’ was treated as a random effect. All data were verified visually for normality and constant variance of residuals using QQ and residuals versus fitted values plots. ANOVA was performed to compare soil nitrate-N levels following alfalfa and corn, prior to bean planting, with ‘preceding crop’ as a fixed effect and ‘block’ as a random effect. Analysis of covariance was performed to analyze the effect of weed biomass and dry bean type on bean yield. Pearson's correlation coefficients were calculated for bean yield, bean weed biomass and bean pre-plant soil nitrate-N; and for wheat yield, wheat weed biomass and wheat pre-plant soil nitrate-N, across all four locations. The effect of preceding bean type on following wheat yield and weed biomass was analyzed with ANOVA, where ‘bean type’ was treated as a fixed effect and ‘block’ as a random effect.
Results and Discussion
Effects of alfalfa and corn on bean yields and soil nitrate
Preceding crop (alfalfa versus corn) and bean type main effects on bean yields were significant at all four locations, though there was no interaction between preceding crop and bean type. Bean yields were greater following alfalfa compared with corn at Madison (P < 0.001) and Rosemount (P < 0.001), but were similar following alfalfa and corn at Becker (P = 0.894) and were greater following corn compared with alfalfa at Lamberton (P = 0.009; Fig. 1). Bean yields were not correlated with soil nitrate-N measured at bean planting (r = −0.09; P = 0.220). The lack of correlation between soil nitrate-N and bean yields was not consistent with the findings of several studies, where bean yields were observed to increase with N fertilization (Edje et al., Reference Edje, Mughogho and Ayonoadu1975; Liebman et al., Reference Liebman, Corson, Rowe and Halteman1995). Pre-plant soil nitrate tests have been used as a diagnostic tool for adaptive N management on some soils (Yost et al., Reference Yost, Coulter, Russelle and Davenport2014), but may not be a good predictor of soil N status on all soils or as an indicator of crop performance (van Kessel and Hartley, Reference van Kessel and Hartley2000; Farid and Navabi, Reference Farid and Navabi2015). There is some evidence that dry bean yield response is greater for post-emergence N applications than for soil N status at planting (Da Silva et al., Reference Da Silva, Tsai and Bonetti1993). Differences in bean yield response to previous crop and fertilization regimes may be a function of differences in soil N release.
Bean yields, averaged across six bean types, following alfalfa and corn at four locations in Minnesota. Within a site, means followed by the same letter are not significantly different at α = 0.05, as determined by Tukey's HSD.
Greater bean yields following alfalfa at Rosemount was likely due to the beans benefiting from residual soil N, as alfalfa can contribute residual soil fertility to following crops in a rotation (Yost et al., Reference Yost, Coulter, Russelle and Davenport2014). Soil nitrate-N levels measured prior to bean planting suggest N fertility was high following alfalfa at Rosemount, where soil nitrate-N was greater following alfalfa, compared with following unfertilized corn (Table 2). At Madison, despite alfalfa's benefit to subsequent bean yields, no differences in soil nitrate-N were observed following alfalfa and manure-fertilized corn (Table 2). It is possible that alfalfa was contributing to soil N fertility but that the additional soil N was not captured in the soil nitrate measurements, as soil nitrate is highly spatially and temporally variable (Bruckler et al., Reference Bruckler, de Cockborne, Renault and Claudot1997). Additionally, mineralization rates of alfalfa residues may have been low as a result of low spring temperatures prior to soil sampling, but accelerated later in the season as soil temperatures increased, leading to increased nitrate availability for bean uptake. At Madison, bean yields were greater following alfalfa compared with corn despite the 208 kg N ha−1 in liquid hog manure applied to corn. Residual fertility effects from the hog manure may have been limited in the following bean crop due to the high liability of nutrients in hog manure, which generally has a C:N ratio of <10 (Malley et al., Reference Malley, Yesmin and Eilers2002). In contrast, the C:N ratio of 3-year-old alfalfa root residues ranges from 19 to 27 (Louarn et al., Reference Louarn, Pereira-Lopès, Fustec, Mary, Voisin, Carvalho and Gastal2015), which results in net immobilization of N (Wedin and Tilman, Reference Wedin and Tilman1990).
Spring total soil nitrate-N measured at combined depths of 0–61 cm in April prior to planting six bean types and measured prior to spring wheat planting the following year, at four Minnesota locations.
Location averages are depicted for wheat pre-plant soil nitrate-N levels, as soil nitrate-N was similar following the six bean types.
1 Within rotation years and locations, means followed by the same letter are not significantly different at α = 0.05, as determined by Tukey's HSD.
2 Not significantly different at α = 0.05, within a location.
In contrast, no difference in bean yields was observed following alfalfa versus corn (P = 0.894; Fig. 1) at Becker. Bean yields following alfalfa may have been similar to yields following corn because the alfalfa stand was terminated after only 1 year of growth at Becker. As alfalfa contributes little N to the soil in the year of establishment (Kelner et al., Reference Kelner, Vessey and Entz1997), there was likely little residual soil N remaining to meet bean N requirements. Contrary to Rosemount and Madison, a lack of fertility benefits from alfalfa resulting in low bean yields at Becker is supported by low soil nitrate-N levels observed following alfalfa, compared with corn (Table 2).
Bean yields were greater following corn compared with alfalfa at Lamberton (Fig. 1). The high levels of composted cattle manure (4–5 t ha−1) applied to corn at Lamberton likely led to carryover fertility effects in the bean crop, though those effects were not detected in the soil nitrate-N measurements following corn. When comparing nitrate leaching among several organic fertilizers applied at high rates, Beckwith et al. (Reference Beckwith, Cooper, Smith and Shepherd1998) found the lowest soil nitrate values following application of solid cattle manure, which may explain why low nitrate was observed at Lamberton after a high rate of cattle manure was applied. Solid cattle manure applied at high rates can also increase yields by contributing to crop phosphorus and potassium demands (Evans et al., Reference Evans, Goodrich, Munter and Smith1977), which could explain the bean yield increase despite similarities in soil nitrate values following the corn and alfalfa crops. Phosphorus is an important factor for N2 fixation in dry bean (Ssali and Keya, Reference Ssali and Keya1983) and greater phosphorus availability to dry bean following corn compared with alfalfa at Lamberton may have increased symbiotic N fixation and plant N availability, which resulted in greater yields. Although bean yields were similar following alfalfa and corn, nitrate-N levels were greater following alfalfa than the manured corn (Table 2), indicating a difference in N input between the two preceding crops.
Effects of alfalfa and corn on weed biomass in bean
No differences in weed biomass were observed following alfalfa and corn at Lamberton, Madison and Rosemount (Table 3). The similarity in weed biomass following both alfalfa and corn suggests that the soil weed seed bank was similar following both corn and alfalfa (Clay and Aguilar, Reference Clay and Aguilar1998; Kegode et al., Reference Kegode, Forcella and Clay1999; Anderson Reference Anderson2010), or that neither corn nor alfalfa provided a greater benefit to the bean crop in terms of competitiveness with weeds. Weed biomass was greater following alfalfa compared with corn at Becker (Table 3), where alfalfa was only grown for 1 year.
Total weed biomass in six bean types collected at dry bean and soybean harvest following preceding alfalfa and corn crops, and weed biomass in spring wheat following six bean types, at four locations in Minnesota.
1 Within main effects and locations, means followed by the same letter are not significantly different at α = 0.05, as determined by Tukey's HSD.
There was no correlation between weed biomass in bean and soil nitrate-N at bean planting (P = 0.539), so differences in weed biomass in bean were likely driven by other factors, such as the weed seedbank, bean competitiveness, or other soil factors. Weed biomass is often high in alfalfa at establishment (Hall et al., Reference Hall, Curran, Werner and Marshall1995), but declines appreciably over 3 years of production (Anderson, Reference Anderson2010). At Becker, high weed biomass in establishing alfalfa likely led to weed seedbank buildup, which can contribute to weed infestation in the subsequent crop (Tørreson et al., Reference Tørresen, Skuterud, Tandsaether and Hagemo2003). Three-year alfalfa at Lamberton, Madison and Rosemount did not decrease weed biomass the following year compared with the annual corn crop, which contrasts with reports that alfalfa is effective in reducing weed populations relative to annual crops (Moncada and Sheaffer, Reference Moncada and Sheaffer2010; Tautges et al., Reference Tautges, Burke, Borrelli and Fuerst2017). However, while beans are generally highly susceptible to yield loss when competing with weeds (Blackshaw et al., Reference Blackshaw, Molnar, Muendel, Saindon and Li2000), it was notable that greater weed pressure following alfalfa at Becker did not result in bean yield differences between preceding crops (Table 4). Apart from the effects of preceding crops, weed biomass in bean may also have been affected by other soil nutrient levels that were not measured in this study.
Grain yields of six bean types following preceding crops of alfalfa and corn, at four Minnesota locations.
1 Within a column, means followed by the same letter are not significantly different at α = 0.05, as determined by Tukey's HSD.
Bean types: yield and competitiveness
Soybean was among the highest-yielding bean types at all locations, but black bean yields were consistently high and were similar to soybean at three of the four locations (Table 4). Navy, pinto and heirloom bean yields were similar to soybean at Madison and Rosemount, but lower at Becker and Lamberton (Table 4). Dark red kidney bean was the lowest-yielding type at three of the four locations. These results agree with previous organic research demonstrating that black and pinto are often the highest yielding dry bean types, while dark red kidney is among the lowest in conventional and organic systems (Heilig and Kelly, Reference Heilig and Kelly2012; Michaels, Reference Michaels2016). Given these results, black bean is likely to produce the greatest yields in Minnesota. However, navy, pinto and heirloom bean varieties also demonstrated the ability to produce high yields at some locations, relative to soybean and could be financially viable for an organic grower depending on market demand for those bean types. The similarity among yields of soybean and those of several dry bean types across locations indicates that dry beans are a viable alternative to soybeans in organic systems in Minnesota.
Weed biomass was lower in soybean than the five dry bean types at all locations (Table 3). There was no difference in weed biomass among dry bean types at Lamberton, Madison and Rosemount, and weed biomass differed between only navy and dark red kidney bean types at Becker (Table 3). While few differences were observed in weed biomass among dry bean types, the relationship of weed biomass to bean yield differed among the five dry bean types (P = 0.0454). Soybean yield was not correlated with weed biomass (r = −0.36; P = 0.0934), but dark red kidney bean yield was strongly correlated with weed biomass (r = −0.74; P < 0.001). Weed biomass was consistently high in navy bean across locations, though the correlation between navy bean yield and weed biomass (r = −0.65; P = 0.001) was lower than for dark red kidney bean. Moderate correlations between weed biomass and pinto bean yield (r = −0.59; P = 0.002), black bean yield (r = −0.58; P = 0.003) and heirloom bean yield (r = −0.49; P = 0.016) were observed.
Variability in yield response to weed pressure among bean types may have been related to morphological differences among bean types. The indeterminate semi-climbing or vining (Type III) growth habit has been identified as a bean morphology that confers greater yield stability and competitiveness with weeds, whereas the determinate ‘bush’ type growth habits (Types I and II) generally result in lower competitiveness (Kelly et al., Reference Kelly, Adams and Varner1987; Singh, Reference Singh1989; Wortmann, Reference Wortmann1993). The heirloom dry bean type with a Type III growth habit displayed lower yield response to weed biomass, whereas the navy (Type II) and kidney (Type I) bean types displayed greater yield response to weed biomass. Organic growers should select bean varieties with indeterminate semi-climbing or vining growth habits if they desire maximum competitiveness with weeds.
Bean type effect on spring wheat
Bean type did not affect spring wheat yields the following year (P = 0.115; Fig. 2). Other researchers, such as Przednowek et al. (Reference Przednowek, Entz, Irvine, Flaten and Thiessen-Martens2004) and Miller et al. (Reference Miller, Waddington, McDonald and Derksen2002), reported no consistent effect of dry bean and soybean on subsequent wheat yields. Most of the N within the bean biomass was likely removed in the form of grain at harvest and any N left in the residue decomposed slowly or became immobilized by the decomposing microflora (Lupwayi and Kennedy, Reference Lupwayi and Kennedy2007); consequently, differences in the effect of bean type on soil nitrate-N might be expected to be small. Further, masking of apparent N benefits from legumes by high soil N status has been observed by others (Przednowek et al., Reference Przednowek, Entz, Irvine, Flaten and Thiessen-Martens2004), which could explain why little effect of bean type on wheat yields was detected.
Spring wheat yields following six bean types (‘Black’ = black bean; ‘Heirloom’ = heirloom bean; ‘Kidney’ = dark red kidney bean; ‘Navy’ = navy bean; ‘Pinto’ = pinto bean; ‘Soy’ = soybean) at four locations in Minnesota. Error bars depict the standard error of the mean.
Soil nitrate-N levels prior to wheat planting were similar following all bean types (P = 0.203; Table 2). Wheat yields were positively correlated with soil nitrate-N at planting (r = 0.14; P = 0.048) and negatively correlated with weed biomass (r = −0.50; P < 0.001). Weed biomass in wheat was not affected by bean type grown the previous year (P = 0.869), but weed biomass in wheat was positively correlated with soil nitrate-N at wheat planting (r = 0.18; P = 0.029). Soil nitrate-N levels were relatively stable between pre-plant bean and pre-plant wheat samplings 1 year later at Becker, Lamberton and Madison (Table 2). However, soil nitrate-N decreased from pre-plant bean to pre-plant wheat samplings by 70% at Rosemount (Table 2). Mineralization of N from manures applied at Becker, Lamberton and Madison may have contributed to soil nitrate-N stability from bean to wheat growing years, whereas soil nitrate-N decreased during the bean phase at Rosemount, where no additional N was applied to corn.
As supported by the results of this study, wheat yield is highly responsive to soil N at planting (Justes et al., Reference Justes, Mary, Meynard, Machet and Thelier-Huche1994) and weed growth is favored by high soil N availability (Blackshaw et al., Reference Blackshaw, Brandt, Janzen, Entz, Grant and Derksen2003). These relationships create a paradox for organic growers, because soil N fertility must be built to levels great enough to support high wheat yields; however, maintaining high soil N increases weed pressure, which decreases wheat yields. Maximizing crop competitiveness, which can be increased by soil fertility and steady release of N, is essential for weed management in organic systems (Liebman and Davis, Reference Liebman and Davis2000). The results of this study suggest that management of soil N by incorporating legumes like alfalfa into organic rotations, rather than high rates of manure application, can support high-yielding pulse and wheat crops while limiting weed biomass. Future studies should identify methods for determining alfalfa N credits to pulses, as well as investigating other non-N-related fertility benefits of alfalfa to subsequent pulse crops. Based on our results, we recommend that either alfalfa or corn prior to dry bean production is acceptable in rotations if soil fertility is adequate. However, if soil N is limited, growing alfalfa prior to dry beans will likely increase yields compared with growing beans after corn.
Some dry bean types demonstrated the potential to produce yields similar to soybean. Specifically, black, navy and pinto beans consistently demonstrated the ability to yield similarly to soybean across locations, supporting the idea that dry beans can be a viable alternative to soybeans in organic rotations in Minnesota. However, greater weed biomass in several dry bean types, relative to soybean, suggests that more weed control efforts will likely be required in dry bean on the part of organic growers. More long-term studies on rotations should be conducted to expand understanding of legume rotational effects, particularly N and weed seedbank dynamics, as well as the benefits of diversifying cropping systems with perennial forages and pulses.
Acknowledgements
This research was supported by the National Institute of Food and Agriculture, US Department of Agriculture, under grant number 2011-51300-30743. Partial funding for this research was provided by Ceres Trust, NCR-SARE, and the Minnesota Agricultural Experiment Station.
