Grain yield and N use

A fundamental aspect of intercropping is to avoid unfavorable intra- or inter-specific competition possibly including interspecific facilitationReference Callaway³⁰, where plants increase the growth and survival of their neighborsReference Willey^2, Reference Ofori and Stern^31, Reference Vandermeer³². LER ratios varied between 0.98 and 1.51 and indicate complementarities within the present grain legume–barley IC combinationsReference Willey^2, Reference Ofori and Stern³¹, but with considerable variation over location, years and crops (Fig. 1). Combined IC grain yields were comparable to grain yields of sole cropped pea, but significantly greater than sole cropped lupin, faba bean and barley yields (Fig. 1). In descending order, the greatest grain yields were obtained for IC containing pea, faba bean and lupin. Pea was the dominant IC component on both soil types with no significant difference between cultivars. Faba bean dominated in the IC on the sandy loam soil, but not on the sandy soil. Lupin was suppressed by barley at both locations.

Figure 1.

Average grain yields, yield stability [indicated below the x-axis as % CV (coefficient of variation) on average yields] of SC and IC of two pea cultivars (Agadir; peaA and Bohatyr; peaB), two barley cultivars (Otira=O and Lysiba=L), faba bean (Columbo) and lupin (Prima), grown in a sandy loam soil and a sandy soil during 2001–2003. Measures of intercropping advantage estimated using the LER are given on the top of IC bars. LSD_0.05 between cropping strategies is given by floating bars.

The SC performance supported the interspecific interactions within the IC stand with grain yields of the tested semi-leafless pea cv. Agadir of 520 g m⁻² in average over the 3 years on the sandy loam, 13% more than the normal leafed pea cv. Bohatyr. On the sandy soil there was no effect on choice of pea cultivar when sole cropping, averaging grain yields of 430–440 g m⁻², but with higher general pea–barley LER values due to lower barley yields (Fig. 1). When yields of IC exceed the yield sum of the component species grown aloneReference Willey², it is often as a result of better use of available growth resources Reference Hauggaard-Nielsen, Ambus and Jensen^8, Reference Vandermeer^32, Reference Natarajan and Willey³³ typically controlled by the level of interspecific interactions and the corresponding SC yields. It has been inferred that the trend of LER in legume–cereal IC was associated with the yields of the legumesReference Ofori and Stern³¹, possibly due to the reputation of especially grain legumes having high yield variabilityReference Jensen and Hauggaard-Nielsen^15, Reference Giller and Cadisch^34, Reference van Kessel and Hartley³⁵. However, in the present study, the weaker interspecific competitor was sometimes the grain legume and sometimes barley depending on crop combination and soil type, which means that it is not a species phenomenon. It may just be the growth of the weaker component that is determining the yield efficiency of the IC combination.

Faba bean and lupin had both a yield of 340 g m⁻² in the sandy loam soil, with decreasing yields of 320 and 270 g m⁻² on the sandy soil, respectively. However, especially faba bean showed a high degree of complementarity when intercropped, with LER values between 1.37 and 1.51, indicating up to 50% better utilization of the environmental sources for plant growth by the IC than by the corresponding SC. Faba bean might be a better choice than pea due to better spatial or temporal complementarity towards the barley companion crop, leaving space for both crops to develop and thereby utilize available environmental growth resources. The yield of lupin on the sandy loam soil was the results of yields of 420, 430 and 170 g m⁻² in years 2001, 2002 and 2003; the latter yield due to the inoculation with an ineffective commercial B. lupini strain. Lupin have the greatest protein content of the present grain legumesReference Jensen, Joernsgaard, Andersen, Christiansen, Mogensen, Friis and Petersen³⁶ and extraordinary ability to mobilize phosphorusReference Braum and Helmke³⁷, but when growing a species under environmental conditions without efficient populations of native symbiotic bacteria difficulties with appropriate inoculation are clearly a limitation for farmersReference Peoples, Ladha and Herridge³⁸.

In contrast to results from JensenReference Jensen⁶ yield stability of IC was not greater than that of grain legume SC, with the exception of the IC including faba bean. The grain legume–barley IC showed less or similar variability compared to at least one of the respective SCs (Fig. 1). When working in organic cropping systems evaluation of specific species and cropping strategies should be conducted as an integrated part of the organic farming practices. It is not always appropriate to continue with the general knowledge very often gathered under conventional growing conditions, such as robust and stable yielding cereals as compared to unpredictable and variable grain legumes. On the sandy loam soil the highest yield stability was actually observed for the peas, whereas the barley yield was very variable, especially at the low N level.

Faba bean and lupin had lower yield stability than pea and fertilized barley. In contrast, on the sandy soil the highest yield variability was observed in the pea cultivars and faba bean and the lowest in lupin and barley, where the application of urea-N increased the yield variation significantly (Fig. 1). Grain legume–barley intercropping might not be the highest yielding as compared to the yield of one of the corresponding SCs in a single year, but it can be regarded as insurance against the complex abiotic and biotic stresses influencing crop performance, especially in organic systems. Self-regulation within the IC stand caused by interspecific interactionsReference Willey² can have a compensation effect against temporal or spatial nutritional limitations and/or attack from pest and disease organisms reducing annual yield variability.

Despite accounting for approximately more than half of the total biomass production (Fig. 1) the grain legumes accumulated less soil N when intercropped than could have been expected from SC uptake (Fig. 2). However, the LER values for soil N uptake were all considerably higher than 1, indicating a better utilization of soil N sources by the IC than by SC (data not shown). In accordance with other reported workReference Hauggaard-Nielsen, Ambus and Jensen^5, Reference Jensen^6, Reference Knudsen, Hauggaard-Nielsen, Jornsgard and Jensen⁹ barley obtained proportionately more of the soil N when intercropped (Fig. 2), indicating that barley has a greater competitive ability for inorganic N sourcesReference Hauggaard-Nielsen, Ambus and Jensen^5, Reference Jensen⁶. Likewise, after application of 5 g N m⁻² the grain yield of barley was raised 40–50% at both locations independent of cultivar (Fig. 1). When an intercropped cereal is more competitive for soil inorganic N the legume is forced to rely on N₂-fixationReference Knudsen, Hauggaard-Nielsen, Jornsgard and Jensen^9, Reference Carr, Martin, Caton and Poland^39, Reference Corre-Hellou, Fustec and Crozat⁴⁰. The intercropped grain legumes increased their proportion of plant N derived from N₂-fixation by on average 10–15% compared to the corresponding SC (Fig. 2). Nitrogen fixation was very constant in grain legume SC over species and location, varying from 13.2 to 15.8 g N m⁻², being lowest in peas and highest in faba bean and lupin. In the IC the nitrogen fixation per area decreased with increasing barley suppression of the grain legume, and fixation was more reduced with barley cv. Otira than when intercropped with barley cv. Lysiba, and more for faba bean and lupin than for peas. Especially lupin was suppressed when intercropping, with a reduced N₂-fixation from 15 to 5–6 g N m⁻². The interspecific interactions among the crops are delicate and specific to locations, and the growing of a certain IC combination may derive from several considerations such as: (i) more stable yields, (ii) N inputs to the cropping system and (iii) competitive ability towards traditional heavy weed infestation of a specific field, among several others.

Figure 2.

Total above-ground nitrogen (N) accumulation in SC and IC of grain legumes and barley partitioned in crop soil N, and leguminous symbiotic N₂-fixation at two separate locations during 2001–2003 (for further information see Fig. 1). Measures of percentage of N accumulated in above-ground grain legume originated from fixation are given on the top of bars. LSD_0.05 between cropping strategies is given by floating bars.

Crop N balances were determined to evaluate the effect of cropping on soil N fertility. After subtracting N exported in the harvested grain crop, N balances were +2 g m⁻² for both pea cultivars on both soil types and about +3.5 g m⁻² for the faba bean and lupin. In barley, the crop N balances were all negative, ranging from, on average, −3.5 g m⁻² on both soil types without urea-N application to near 0 after application of 5 g urea-N m⁻² (Fig. 3). As the quantity of N in the harvested grain and the amount of fixed nitrogen is slightly lower for peas than for faba bean and lupin, the N balance is lower for the peas although still positive, meaning that grain legumes make a net contribution of N to the soil. When intercropping, a slightly negative balance was found except when faba bean was intercropped with barley cv. Lysiba on the sandy soil. The same trend was seen for the pea cv. Bohatyr. However, grain legume–cereal IC are not likely to increase soil inorganic N in the long term, but rather deplete it, although at a slower rate than in barley sole cropping.

Figure 3.

Nitrogen balance for each cropping strategy of sole cropping (a) and grain legume–barley intercropping (b) on a sandy loam soil and a sandy soil during 2001–2003 (for further information see Fig. 1). Balances were calculated according to equation (5). LSD_0.05 between cropping strategies is given by floating bars.

Grain N concentration

The N content of grain legumes was highest on the sandy soil as compared to the sandy loam (Table 2) whereas the barley grain N content for the two soils went in both directions. The peas had an average N level of 3.6%, the faba bean 5% and the lupin 5.3%. A significant effect of the barley cultivar was noted where barley cv. Lysiba had a N concentration in the grain which was 0.2% higher than barley cv. Otira, and it responded to increased N level by a larger increase in grain N concentration compared to barley cv. Otira, which responded by a higher increase in total grain yield (Fig. 1). Competition from barley had little effect on grain legume grain N concentration despite a general reduced grain legume total N accumulation and increased proportion of N derived from fixation (Fig. 2). When barley was intercropped with pea significantly higher grain N concentration was found compared to the barley SC (Table 1), also when applying fertilizer N. On the contrary, faba bean only increased barley cv. Otira grain N concentration on the sandy loam whereas intercropped lupin did not influence measured barley grain qualities.

Table 2.

Grain nitrogen concentration (%) of sole cropped and intercropped grain legumes with barley at two locations during 2001–2003.

¹ For further information on species cultivars see Figure 1.

² The symbol+indicates application of 5 g urea-N m⁻² to the barley sole crops.

Grain legume–cereal interspecific competition can modify cereal grain N concentrationReference Hauggaard-Nielsen, Ambus and Jensen^5, Reference Jensen⁶ because grain legumes in general compete less efficiently for soil N sources. Relatively more soil N becomes available to the intercropped cereal as compared to the respective SC (Fig. 2). However, since grain legumes compete for other growth factors, such as light, water and non-N nutrients, cereals may not increase their yield in direct proportion to the amount of N available and an increased concentration of grain N can be observed. This is presumably the case for the present productive intercropped pea, whereas the less productive faba bean and lupin are not able to create enough interspecific competition towards growth factors other than N to raise barley N concentration (Table 1).

Other studies show how the N content of the wheat (Triticum spp.) grain increases when intercropped with faba beanReference Bulson, Snaydon and Stopes^21, Reference Jensen²². From a recent study including intercropping of wheat with faba bean in Denmark, Germany, Italy and UK in both additive and replacement designs it was concluded that the increase in protein concentration of wheat grain in IC could be of economic benefit when selling wheat for breadmaking, but only if the bean crop was also marketed effectivelyReference Godding, Kasyanova, Ruske, Hauggaard-Nielsen, Jensen, Dahlmann, Von Fragstein, Dibet, Corre-Hellou, Crozat, Pristeri, Romeo and Monti⁴¹. Other cereal species such as rye (Secale spp.) or oat (Avena spp.) are also possible cereal IC components, depending on the specific target for the IC (grain quality, soil fertility, weed infestation level, etc.). Knowledge of the effect of intercropping lupin and cereals for maturity is limitedReference Palmason, Danso and Hardarson²³, but the limited branching ability of the present lupin may reduce the ability to obtain sunlight which can translate into major competitive limitationsReference Midmore⁴² that strongly influence the interspecific competitive ability.

Competitive ability towards weeds

IC that are particularly effective at suppressing weeds capture a greater share of available resources than SCReference Liebman and Dyck¹⁹, which was clearly the case in the present study (Fig. 1). Any part of the soil surface that is not occupied by crop species is potentially subject to invasion by weedy species. Therefore, a typically more vigorous barley canopy structure as compared especially to faba bean and lupin in the present study provides a quicker, greater and more extensive soil coverage. A lower grain legume seeding density and initial growth rate, as compared to cerealsReference Hauggaard-Nielsen, Ambus and Jensen⁸, can fuel a rapid and intensive early weed resource uptake and thereby dominance throughout the rest of the growing seasonReference Liebman and Davis⁴³.

Weed infestation in the different crops were comparable; however, it tended to be highest in sole cropped faba bean, lupin and unfertilized barley, where the application of urea to barley reduced the weed infestation by around 50% (Fig. 4). Especially on the sandy loam soil considerable weed infestation levels were found within faba bean and lupin as well as barley, except for the fertilized barley cv. Otira which was able to suppress the weeds. The weed biomass on the sandy soil was one-quarter that on the sandy loam soil due to efficient mechanical weeding management and less infestation by volunteer red clover (Trifolium pratense), which was a major problem on the sandy loam soil, especially in 2002. In general, the intercrops average the weed infestation levels in such a way that they are always lower than the levels of weeds in one of the IC component crops. This shows a more resilient crop stand able to respond to actual growing conditions as compared to grain legume sole cropping. Such a trait might be important to include as a management tool when the quantity and diversity of the weeds are high, as in the case of organic farming systemsReference Rydberg and Milberg⁴⁴.

Figure 4.

Weed above-ground DM production below grain legumes and barley sole cropping (SC) as compared to below grain legume–barley intercrops at two separate locations during 2001–2003 (for further information see Fig. 1). LSD_0.05 between cropping strategies is given by floating bars.

Effect on diseases

Components of IC are often less damaged by pest and disease organisms than when grown as SC, but this often varies unpredictablyReference Trenbath¹⁶. As an example, barley net blotch infestation levels are highlighted in Figure 5, as it was the most serious disease on barley during all 3 years. Interestingly, increasing the number of grain legume components reduced the amount of disease. As a general picture, reduction in disease was observed in all IC systems compared to the corresponding SC (Table 3). For all diseases (with the exception of brown spot on lupin) disease reduction was in the range of 20–40% (Table 3). Pathogens varied in dispersal mechanism and type (biotrophic or necrotrophic) and crops varied in height and anatomy. This suggests that there are mechanisms operating, possibly in all IC systems, whereby disease levels are reduced. However, for one disease in particular (brown spot on lupin) disease reduction was almost 80% in the IC. Furthermore, the well documented complementarity between intercropped barley and pea with respect to NReference Jensen^6, Reference Hauggaard-Nielsen, Ambus and Jensen^8, Reference Corre-Hellou, Fustec and Crozat⁴⁰ may in turn influence plant healthReference Kinane and Lyngkjær⁴⁵.

Figure 5.

Effect of two-, three- and four-component IC of barley, lupin, faba bean and pea on incidence of barley net blotch (Pyrenophora teres) during 2002 when grown in the sandy loam soil. * Disease incidence measured as area under disease progress curve (AUDPC). Different letters indicate significant (P<0.05) differences using the Kruskal–Wallis test. For further information see Figure 1.

Table 3.

Diseases observed during the growth seasons 2001–2003 on the sandy loam soil and their percentage severity on the SC and median disease percentage reductions in disease in the dual intercrop systems (IC) as compared to the corresponding SC.

¹ For further information on species cultivars see Figure 1.

² The amount of disease observed when sole cropping (SC) in percentage leaf area covered.

³ Median percentage disease reduction in the present dual IC.

⁴ *, **, *** indicate significant differences (P<0.05, 0.01, 0.001) from SC using the Kruskal–Wallis test. ns, indicates no significant difference.

To be able to predict which crop–disease combination would give the largest disease reduction, the mechanisms behind the disease reduction in IC must be understood, which was not possible during the present study. However, it is without doubt that under some conditions, intercropping can usefully contribute to the control of disease populations, but do the reductions in disease have any real effect on yield? This is a very difficult question to answer. We believe it is fair to say that if disease levels are high, then reduction in disease levels will have a greater effect on yield than if the initial disease levels are low. To test that, it is necessary to carry out specific yield experiments, where there is a disease-free control, which is impossible when working in organic farming systems.

Intercropping on the market

Most grain legume–cereal mixtures with similar ripening times are easy to combine-harvest using traditional on-farm equipment, but few buyers purchase mixed grains. Farmers are often left with the options of harvesting the mixture for animal feed. However, during the past 3–5 years a few Danish buyers working in the organic market have purchased mixed grains, often used for seed because of less damage from pest and diseases. The buyers charge 15–20€ per ton mixed grain for separation and cleaning, but when it is contracted for seed, a premium is given, making it profitable for the farmers to grow.

Grain legume–cereal intercropping is regarded as a cropping strategy based on the manipulation of plant interactions in time and space to maximize growth, with the possibility of increasing input of leguminous N₂-fixation into cropping systems and reducing the need for fertilizer N applications, and reducing pesticides due to improved competition towards weeds and less general damage by pest and disease organisms. Intercropping strategies offer a number of agroecological functions and services to the market, which are of increasing importance taking into account present environmental and energy issues on the global political agenda. Improved harvest technologies may in future make intercropping more attractive in the intensive agricultural areas of developed countries.