Abstract
Cover crops can be used to reduce leaching and erosion, introduce variability into crop rotation and fix nitrogen (N) for use by the main crops, less is however known about effects on weeds. The effects on weed seed bank, weed growth and grain yield of 4 years of annual undersown clover and ryegrass alone and in combination, and one of the 4 years with clover or clover + grass as green manure, were studied in oat and spring wheat at two experimental sites in south-eastern Norway. These treatments were compared with no undersown crop (control) and with weed harrowing. In contrast to many results in the literature, the undersown clover in this study did not suppress annual weeds, but fertilized the weeds as well as the cereals. Undersown clover resulted in a statistically significant increase of grain yield at the two sites to 116% and 121% of control. During the 4-year period relative seed bank and density of emerged weed (dominated by Spergula arvensis) increased significantly about 4.5 and 10 times respectively in the undersown clover plots at Apelsvoll. At Kise both ryegrass alone and ryegrass mixed with clover significantly suppressed the weed biomass to 70% and 74% of control respectively. It is concluded that fertilization effects of undersown clover may have dominated and overriden the competitive effects. One whole-season clover green manure did not increase the mean yield, but resulted in a significant drop in seed bank size the following year, because of limited weed establishment in an established ley. Only a slight increase in average weed biomass was observed at one of the two experimental sites. The weed seed bank and the weed biomass were essentially kept at steady state during the experimental period in harrowed plots, but harrowing decreased grain yield significantly at both sites.
Introduction
In organic crop rotations dominated by cereals, nutrient supply and weed control seem to be key factors for improving and stabilizing yield. Most of the cereal in Norway is grown in areas with no or only limited livestock farming. For organic cereal production this implies that nutrient supply, especially in a long-term perspective, is a serious challenge. One possible source of nutrients, especially of nitrogen (N), in such farming systems is legume cover crops, either (i) undersown in cereals or (ii) as an annual green manure ley in rotation with the cereals. The former has the advantage that a cash crop is grown together with the cover crop. The latter is without any cash crop, although farmers have some income from government subsidies. As summarized in Teasdale et al. (Citation2007), the most common cover crop practice in organic cereal production in Scandinavia is undersowing of clover or clover-grass. When management is optimized for (i) cereal and cover crop species and cultivars, (ii) sowing time and seeding rates of the cover crop, and (iii) soil fertility, there are often small or insignificant negative cover crop effects on crop yield in these systems (Breland Citation1996, Olesen et al. Citation2002, Molteberg et al. Citation2004).
Cover crops may also have other beneficial effects, such as reducing soil erosion, improving soil quality or minimizing nutrient or pesticide losses through leaching and surface runoff (Sustainable Agriculture Network Citation1998, Blackshaw et al. Citation2005, Teasdale et al. Citation2007). Cover crops may improve soil health by increasing organic matter content, improving soil structure, and facilitating more diverse and biologically active microbial communities. Legume cover crops fix atmospheric N, which becomes available to succeeding crops and inevitably also to weeds.
Most studies on the use of undersowing in cereals have focused mainly on nutrient supply and effects on cereal yield (e.g. Breland Citation1996). Although Breland (Citation1996) and Hartl (Citation1989) partly included assessments on weeds, very few studies (e.g. Bergkvist et al. Citation2010) have focused mainly on weed control. The use of undersown clover in cereals may jeopardize weed control by harrow after sowing the cover crop. Furthermore, a growing cover crop in the autumn also obstructs stubble cultivation, for example for couch grass control (Rasmussen et al. Citation2006), which is otherwise a standard non-chemical method for controlling this weed. The question is to what extent undersown crops, by possibly preventing weed establishment and growth, may reduce the need for weed harrowing. Although a cover crop may suppress weeds growing at the same time (Teasdale et al. Citation2007, Maiksteniene et al. Citation2009) the weed suppression effects differ (Ross et al. Citation2001, Pridham and Entz Citation2008) and intercrops sometimes reduce the cereal yield compared to cereal monoculture due to competition with a cash crop (Pridham and Entz Citation2008). Mostly white clover, but also red clover, have proven to be suitable N-fixing crops for undersowing in cereals and they are commonly used on organic farms. White clover has been suggested as the most suitable cover crop in leek as it balances the trade-off between stress on the main crop and weed control compared with the more aggressive red clover (den Hollander et al. Citation2007). The ability to suppress weeds is dependent on clover species, cultivar, the natural weed flora, regional climate and soil fertility levels present (Ross et al. Citation2001, den Hollander et al. Citation2007, Hiltbrunner et al. Citation2007).
A vigorously growing cover crop will suppress weed growth (Teasdale et al. Citation2007). The present study aims to verify whether this also includes effects on annual weeds in cropping systems of undersown cover crops in organic cereal production. Our hypothesis was that both the repeated undersown crops and a system with 25% of the crop rotation as full-season clover-ley or clover-grass mixture would increase the cereal yields (cf. Løes et al. Citation2011), but also suppress the annual weed growth.
The present paper raises the following three main questions: (i) Does undersown clover suppress annual weeds? (ii) Since undersown ryegrass does not add nitrogen to the system, do grass species compete better with the weeds, and what effects can be expected if clover and ryegrass are mixed? (iii) Annual green manure leys can potentially improve nitrogen balance, soil structure and perennial weed control, but how do they affect annual weeds, and does it matter if the ley is sown by clover alone or a clover-grass mixture is used? In the present study, answers to these questions were sought by comparisons with an untreated and a weed harrowed control. We also investigated effects of undersown clover, annual clover green manure, and weed harrowing on the weed seed bank. In another part of the investigation (Løes et al. Citation2011), the same field experiments were used to quantify the effects of repeated clover undersowing on (1) field N-balance and (2) mobilization of soil mineral phosphorus (P) and potassium (K).
Materials and methods
The field experiment was performed at two sites, located at Bioforsk, Arable Crops Division, Apelsvoll (60°42'N, 10°51'E) and Kise (60°77'N, 10°81'E), both in south-eastern Norway. There is somewhat more clay and soil organic matter at Kise (USDA texture class: loam) than at Apelsvoll (USDA texture class: imperfectly drained loam developed on morainic till), but significantly more AL-soluble phosphorus at Apelsvoll than at Kise. The preceding crops at Apelsvoll were grass-clover ley in 2000 and wheat without catch crop in 2001. At both sites, the experiments were placed on organically managed land. At Kise, barley with white clover catch crop was grown in 2001, and various spring cereals with no catch crops in 2001. Neither farmyard manure nor any other fertilizer had been used at the sites in recent years. Therefore, the nutrient level in the soil was low. More details concerning soil types and nutrient contents are published by Løes et al. (Citation2011).
summarizes the treatments and shows the rotation plan for the years 2002–2005. At both sites each treatment was replicated four times (four blocks). The experimental plots were 36 m2 and a central plot of 8.4 m2 was used for grain harvest. In total, seven treatments with spring cereals were compared (). Four treatments (explained in ) compared repeated undersowing of various catch crops. Two treatments compared red clover with and without timothy (Phleum pratense L.) used as a full-season green manure. The full-season green manure was established in the first year of the study, and the residual effect on cereal yields was measured in two seasons without catch crops. On all plots, except in the year with full-season green manure, a 4-year sequence of spring cereals were grown; oats, wheat, oats and wheat (). Red clover was used as the green-manure catch crop in oats and white clover in wheat, to increase the biodiversity of the study and avoid too heavy competition between clover and wheat. In 2006, residual effects were measured in all plots, with spring barley as the test crop. The full-season green manure was cut in June and September. The varieties used in the experiment were oats Lena, wheat Avle, ryegrass Tove, red clover Nordi, white clover Milkanova, timothy Grindstad and barley Tiril. The amounts of seed were 200, 10, 15, 5 and 22 kg ha-1 of grains, ryegrass, red clover, white clover and timothy. For red clover, the amount was reduced to 7 kg ha-1 when mixed with ryegrass, and to 3 kg ha-1 when mixed with timothy. The spring cereals were sown with 12.5 cm row distance, and at≈ 4 cm depth, and the field was rolled with a Cambridge roller afterwards. In addition to a control treatment, without either any catch crop or any mechanical weed-control measures, a weed-harrowing treatment was included (). In these, the harrowing was performed twice – the first time before the cereal had germinated (blind harrowing), and the second time at the 3–4 leaf stage of the cereal plants. The weed harrowing was performed with a 4.5-m broad spring-tine harrow (Einböck, Dorf an der Pram, Austria) with angled spring-tines, diameter 7 mm, and tine distance c. 25 mm passing at 8 km h−1 during both harrowing stages. At late harrowing, the speed was adjusted to prevent more than 20% covering of cereal. The tine angle and the depth of the wheel were adjusted to achieve the desirable harrowing depth (25 mm) at low speed.
Table I. A summary of the treatments in the field experiments at Apelsvoll and Kise (RC = red clover, WC = white clover, T = timothy, RG = ryegrass, WH = weed harrowing).
For economic reasons, the analysis of the weed seed bank was conducted in only four of the seven treatments (). The most fundamentally different treatments were chosen: control, clover undersown each year in cereal, clover-ley in one of the 4 years (2003) and weed harrowing. In each plot (8 × 4.5 m), representative soil samples were collected each year before the growing season started (), as described by Sjursen (Citation2001). The soil samples were analysed by the seedling emergence method (Sjursen Citation2001), but only two germination periods were used. The number of emerged seedlings for each species from the soil samples, germinated in pots in the greenhouse, was converted to the number of seeds 1 m-2 to a depth of 20 cm (i.e. seeds 200 L−1). This figure is referred to as seed bank size.
Table II. Dates of management and sampling operations.
Weed emergence in the fields was assessed at Zadoks 49 for cereals (Zadoks et al. Citation1974), and just before harvesting the cereal plants (). Weeds were not assessed in plots with green manure, grown as leys without cereals as a cash crop in 2003. Only soil samples for seed bank analysis were collected (see above). Weeds were counted by species on 0.25 m2 at four random places in each plot. The number of plants m-2 of each species and total number of weeds m-2 (weed density) were calculated. All operations and corresponding dates at the two experimental sites are also shown in .
Total biomass of plant material in the same 0.25 m2 plot was sorted into weeds, undersown crop and cereal plants and harvested after having counted weeds. The material was dried in an airflow dryer at 80 °C and weighed. Harvesting of the grain is described by Løes et al. (Citation2011).
shows monthly precipitation and temperature during the growing seasons 2002–2005 for the two experimental sites. The normal values for the 30-years period 1961–1990 are included. Detailed description of the weather conditions is given by Løes et al. (Citation2011).
Table III. Monthly mean precipitation and temperature at the experiment sites.
Analysis of variance (ANOVA) was carried out for seed bank, plant density and biomass data. Site, year and treatments were main effects in the model (SAS 2004). To explore relationships between the seed bank, weed emergence and biomass, Pearson correlations (r) were calculated. Correlation within data for the most frequent species (Spergula arvensis, Stellaria media and Erodium cicutarium) and correlations between data for weeds, undersown crop and grain yields were calculated separately (SAS Institute Citation2004).
Results
Effects on weed seed bank
Significant differences between locations in the seed bank size were observed (p<0.001) as well as between years (p<0.05). There were no significant differences for treatments. The only significant interaction was between year and location (p<0.05).
In the control plots, the weed seed bank increased to 244 and 139% during the period 2002–2005 respectively at Apelsvoll and Kise (A). The use of undersown clover at Apelsvoll resulted in the largest and significant increase (to 452% of control) in seed bank size (see B). At Kise, on the other hand, the seed bank decreased to 82% of control from 2002 to 2005, with a major drop in 2004 (to 46%). Clover green manure ley in 2003 resulted in a significant drop in the seed bank size at Kise in 2004 (C), as well as when analysing both locations, Kise and Apelsvoll, together (results not shown). A tendency of decline of the seed bank size was observed in the weed harrowing plots during the 4 years at both locations (D).
Figure 1. Effect on weed seed bank to 0.20 m soil depth (left y-axis) and emergence of weeds in the field at Zadoks 49 (right y-axis) in control plots without clover undersowing (A), with clover undersowing (B), with clover ley (C) and with weed harrowing (D) at Apelsvoll and Kise.
Effects on field weed emergence
The emergence of weeds, measured as number of plants m−2, was significantly influenced by location (p<0.001), year (p<0.001) and treatment (p<0.01). The interactions between location and year as well as treatment and year were highly significant (p<0.001).
The number of emerged weed in control plots at Apelsvoll was about 4 times larger in 2005 than in 2002 and significantly higher than all other years (A). The number of weeds increased to 240% at Kise in 2005, compared with 2002 (A). Undersown clover also resulted in the largest increase in number of emerged weed plants (to 1031%) at Apelsvoll from 2002 to 2005 (B). At Kise the increase was only to 157% during the same period. Because of repeated mowing in the year of clover-ley (2003), no weed plants were counted this year. The same plots, however, were counted in 2002, 2004 and 2005 – years in which only cereals were grown. Weed density at Apelsvoll was about 600% higher in 2005 than in 2002 (see C). At Kise the weed density increased to about 270% from 2002 to 2005. Weed harrowing resulted in the lowest increase, to 205% and 187% in plant density during the 4 years at Apelsvoll and Kise, respectively (D). Averaged over both locations, plant density increased significantly from 2002 (322 plants m-2) to 2003 (726 plants m−2). There was a minor decline to 616 and 629 plants m−2 in 2004 and 2005, respectively.
Averaged over the 4 years, repeated undersown ryegrass alone, clover alone or clover + ryegrass only resulted in minor deviations in number of weeds compared with control at both locations. There was only a tendency of increased numbers of weeds (+5% and +6% at Apelsvoll and Kise, respectively) when clover was undersown (A). When ryegrass was undersown, alone or in combination with clover, there was a small reduction in number of weeds to 98% and 93% of control at Apelsvoll respectively, and to 98% and 86% at Kise. In the treatment with clover-ley, the average number of plants was reduced to 80% at Apelsvoll and significantly to 62% at Kise. When clover was grown together with timothy as green manure ley in 2003, the reduction in weed density was significant at both sites, to 70% at Apelsvoll and 65% at Kise. Weed harrowing every year also resulted in a significant reduction of plant density to 66% at Apelsvoll and 77% at Kise.
Figure 2. Average across all years of emerged number of weeds at Zadoks 49 (A), weed biomass at Zadoks 49 (B), weed biomass at harvest (C) and grain yield (D) with control treatment, undersown ryegrass (RG), undersown clover (C), undersown clover and ryegrass (C + RG), clover and timothy (C + T ley), clover ley (C ley) and weed harrowing (WH) at Apelsvoll and Kise compared with control (100%) at the two experimental sites. Columns with same letters (lower case: Apelsvoll; upper case: Kise) are not significantly different (p<0.05). Bars indicate±SE (n = 16).
Effects on weed biomass
Similar to weed emergence, weed biomass at Zadoks 49, measured as kg dry matter ha−1, was significantly influenced by location (p<0.001), year (p<0.001) and treatment (p<0.01). The interactions location * year as well as treatment * year were also significant (p<0.05 and p<0.001, respectively).
At Apelsvoll the weed biomass increased (to 139% of control) averaged over the 4 years when clover alone was undersown compared with sown together with ryegrass (to 116% of control; B). When ryegrass was undersown, the weed biomass decreased slightly to 95% of control. At Kise, on the other hand, the weed biomass decreased to 92% of control when clover was undersown. When ryegrass was undersown alone or together with clover, the weed biomass decreased significantly to 70% and 74% of control, respectively. Clover + timothy or clover alone as green manure ley in 2003, resulted respectively in an insignificant decrease (to 94%) or increase (to 116%) in weed biomass at Apelsvoll. At Kise, however, the weed biomass decreased significantly to 58% and 57% of control, for the two treatments, respectively. Plots with weed harrowing resulted in 86% of control at Apelsvoll, which was significantly lower than plots with undersown clover, but not significantly different compared with control plots. At Kise, however, the weed biomass decreased significantly (to 60%) compared with the control treatment, when the plots were weed harrowed.
The weed biomass at harvest was significantly influenced by year as the only main factor (p<0.001). There was a decline in weed biomass at harvest in 2003 at both experimental sites (data not shown). In average over the 4 years at Apelsvoll, the weed biomass varied between 91% of control in plots with undersown clover + ryegrass and 126% in plots with clover as green manure (C). At Kise the weed biomass varied significantly from 57% in the similar plots as at Apelsvoll (undersown clover + ryegrass) to 105% in plots with clover + timothy as green manure (C).
Effects on grain yield
Treatment and year (p<0.001), but not location, significantly affected grain yields, measured as kg ha−1 with 15% water. Furthermore, there were significant interactions between location and year and between treatment and year (p<0.001).
Undersown clover resulted in a significant increase in the grain yield, to 3510 kg ha−1 (116% of control) at Apelsvoll and 3494 kg ha−1 (121%) at Kise (D), averaged over the 4 years. Undersown clover + ryegrass also resulted in increased grain yields at both locations, 3473 kg ha−1 (114%) and 3511 kg ha−1 (121%) at Apelsvoll and Kise, respectively. The relative increase was only significant for Kise. The use of green manure leys in 2003, both clover with and without timothy, gave lower grain yields, both at Apelsvoll and Kise, than when cereals were grown each year with undersown clover. The reductions were significant at Apelsvoll, but not at Kise. A grain yield of 2530 kg ha−1 (83% of control) was recorded in weed-harrowing plots at Apelsvoll, the greatest reduction at that location. On the other hand, 2833 kg ha−1 was recorded at Kise, which was only slightly lower than the control treatment.
Summary of the composition of the weed flora
Summaries of the species composition of the weed flora are given in (seed bank) and V (density of emerged weeds in the fields) during the four experimental years. Spergula arvensis was the most frequent species in the seed bank and in the fields at both experimental sites, and it increased significantly during the experimental period. Erodium cicutarium was especially found at Kise, both in the seed bank and in the field. The number of annual species in the seed bank was essentially at the same level during the 4-years period (12–14 at Apelsvoll and 14–17 at Kise; ). A maximum of two perennial species were identified in the seed bank at Kise in 2002, 2004 and 2005 (). On the other hand a maximum of nine emerged perennial species were identified at same site in 2003 (). The most frequent perennial species was Cirsium arvense at Kise in 2003–2005, but it was missing at Apelsvoll.
Table IV. Soil seed bank (number of seeds m−2) down to 0.20 cm depth. The first 20 species are annual, the last three species are perennial. Unidentified + clovers are not counted in the number of species, but are included in the sum of all species. Sums with same letter (Apelsvoll: lower case; Kise: upper case) are not significantly different at p ≤ 0.05 (n = 16).
Table V. Weed emergence (number of plants m−2) in fields with seed bank analysis. The first 24 species are annual, the last 9 species are perennial. Unidentified, clovers and Brassica spp. are not counted in the number of species, but are included in the sum of all species. 0 = mean value between 0 and 0.5,− = not observed. Sums with same letter (Apelsvoll: lower case; Kise: upper case) are not significantly different at p ≤ 0.05 (n = 16). 0 = mean values between 0 and 0.5.− = not detected.
Correlations between weed data
The correlation coefficients for ‘seed bank × emergence’ for total number of all species varied between –0.116n.s. in plots with clover green manure at Kise to 0.752*** for undersown clover at Apelsvoll (). Corresponding correlations for S. arvensis varied between 0.292n.s. and 0.871***, thus indicating a close positive relationship for this species between soil seed bank and plant density in the field. There are also strong correlations for ‘seed bank × biomass production in the field’ and ‘emergence × biomass’, essentially explained by S. arvensis. The only significantly positive correlation for S. media (0.706**) was found for ‘seed bank × emergence’ in control plots at Kise. Weed biomass production in control plots at Kise was closely related to the seed bank of E. cicutarium (r=0.738**). In weed harrowing plots, emergence (plant density) and biomass production were closely related (r=0.700**).
Table VI. Pearson correlation coefficients (r) for weed data 2002–2005 (n = 16). *: p ≤ 0.05; **: p ≤ 0.01; ***: p ≤ 0.001; n.s.: not significant.
Correlations between weed biomass, undersown crops and grain yield data
At Apelsvoll there were positive correlations between weed biomass and grain yield for all three treatments with undersown crops. The correlation coefficient (r) varied between 0.493 n.s. and 0.670** (). This means that both weed growth and grain yield varied in the same manner throughout the experiment (A). These correlations were not observed for the Kise data (B). The correlations of biomass between the undersown crop (clover or clover + ryegrass) and the weed or grain yield were, however, negative (r between -0.419n.s. and -0.924***; see A). This implies that the undersown crop could compete with weeds or the cereal crop. On the other hand, the correlation between clover biomass in 2002 and the grain yield the subsequent year (2003) was significantly positive at both locations (r=970*** and 0.807* for clover at Zadoks 49 or at harvesting vs. grain yield respectively; n = 8). The following effect of undersown clover was thus very clear. The only significantly positive correlation clover biomass vs. weed biomass, was between the clover biomass in 2002 and weed biomass in 2005 (r=888**).
Figure 3. Biomass of weed and clover at Zadoks 49, and grain yield and biomass of clover at harvesting, at Apelsvoll (A) and Kise (B).
Table VII. Pearson correlation coefficients (r) for data on weeds, undersown crops and main crop 2002–2005 (n = 16). W = weed, Us = undersown crop, Z49 = at stage Zadoks 49, Harv = at harvesting. *: p ≤ 0.05; **: p ≤ 0.01; ***: p ≤ 0.001; n.s.: not significant.
Discussion
Undersowing clover did not suppress annual weeds in the present study. This very clear outcome was true both regarding short-term effects, measured as field weed emergence, as well as for the ‘long-term’ perspective, measured as weed seed bank development. This result contradicts with Teasdale et al. (Citation2007), who stated that it is generally agreed that ‘a vigorous living cover crop will suppress weeds growing at the same time as the cover crop’. One example was that a spring-sown subterranean or white clover cover crop in white cabbage suppressed weeds by c. 50% (Brandsæter et al. Citation1998). We suppose, however, that the evaluation of weed suppression in cover crop studies is highly related to the type of control treatment used in the study, because of the dynamics and interactions between two dominating factors, namely (i) the competitive effects of the cover crops on weeds, and (ii) the successive fertilization effect of using legumes as cover crop. In the present study, the latter factor seems to be more important than the competitive effect in any given year. Studies (e.g. Alkämper Citation1976, Bostrøm, unpublished) have shown that many weed species have more effective soil nutrients uptake than crops.
The conclusion that fertilization effects of legumes dominate and override competitive effects is supported by the fact that undersown ryegrass suppressed weeds to some degree, especially at Kise. Because undersown ryegrass does not add nitrogen to the system, only the competitive factor is demonstrated in this treatment. At Kise, both ryegrass alone and ryegrass mixed with clover significantly suppressed weeds measured at Zadoks 49 (B). The significant reduction of weed biomass in the Kise plots with undersown ryegrass, either alone or in mixture with clover (B), indicates a strong suppressing effect of these cover crops. This result agrees with Breland (Citation1996), who found a lower percentage of soil covered by weeds on catch-crop plots (including ryegrass) than on plots with cereal crops only. In our study the grain yield was reduced slightly as well (96% of control; D). Lack of water in drought periods may also be an explanation (cf. ). These effects were not seen at Apelsvoll, where the experiment area was irrigated. When clover or clover + ryegrass were undersown in the main crop, increases in the weed biomass of 39% and 16%, respectively, were found at Apelsvoll, but decreases respectively of 8% and 16%, at Kise (B). The increased N supply may have benefited weed growth at Apelsvoll and thereby obscured suppressive effects from the undersown clover (cf. significant negative correlation between biomass of undersown crop vs. weed biomass at Zadoks 49 for Apelsvoll-data; ). The reason for increased weed biomass at Apelsvoll may be the same as for increased seed bank, i.e. better water and nitrogen availability (cf. Albrecht Citation2005).
When main-crop green manure was included in the rotation, a decline in seed bank size the following year was found at both sites (C). Decrease of seed bank during periods with grass-clover mixtures is well documented in an earlier study by one of the authors (Sjursen Citation2001), and others (cf. Paatela and Erviö Citation1971, Lawson et al. Citation1994, Davies et al. Citation1997, Albrecht Citation2005). The reason for the decrease is limited weed establishment in an established ley, consequently also a low production of annual weed seeds during the period. The use of an annual green manure ley is probably in most cases incorporated in cropping systems because of the potential for improving the N balance and soil structure. Recently, however, more focus has been directed at the potential for control of perennial weeds in 1-year green manure, and studies have shown that, e.g. creeping thistle (Cirsium arvense) can be significantly suppressed by frequent cutting during one season (Dock-Gustavsson Citation1994, Graglia et al. Citation2006).
The essentially steady state of the seed bank in harrowed plots (D) at both sites during the experimental period agrees with the results of Stenerud (Citation2009), who found a lower increase (27%) of seed bank size in harrowed plots without undersown crop than in control plots (58%). A Danish study indicated that undersown cover crops gave weed control equivalent to low-intensity weed harrowing in plots without undersown cover crops. High-intensity weed harrowing, however, gave better weed control than the use of cover crops (Rasmussen et al. Citation2006). In the present study, we sowed the cereals and the cover crops the same day. To improve the suppression of weeds, as well as to include a legume in the cropping system, it seems favourable to sow the cover crop the same day as the last harrowing operation. Modern weed harrows are equipped with pneumatic seeders and thus allow harrowing and seeding in the same operation.
Although weed harrowing gave decreased cereal yields (D) in the present study, this aspect should not be stressed too much, because we had to harrow across the seeding direction due to the experimental design. This will not harm the cereal plant at the blind harrowing stage (pre-emerge weed harrowing), but the selectivity of harrowing this way is probably decreased when harrowing after cereal emergence at the 3–4 leaf stage. Most studies on weed harrowing show that proper harrowing does not decrease cereal yields (e.g. Mangerud et al. Citation2007).
Spergula arvensis was the most frequent weed species and the main constituent of the weed biomass, which at Apelsvoll was positively correlated with grain yield for the undersown treatments (). At increasing levels of nitrogen, Andreasen et al. (Citation2006) found that S. arvensis was the only species of six species tested in greenhouse experiments that increased its dry weight at the same rate as barley, when grown alone without competition from barley. They also found that S. arvensis accumulated most N and P per cent of dry matter at increasing N- and P-levels. In competition with barley, however, they found poor growth of S. arvensis. In contrast to this finding, the S. arvensis-dominated weed biomass at Apelsvoll increased in spite of a simultaneous increase in the grain yield when N-fixing clover was undersown. In Sweden, S. arvensis is one of the most frequent weed species in organic farming (Rydberg and Milberg Citation2000). It was also the most frequent weed species in organically grown vegetables in another study by one of the authors (not yet published).
Erodium cicutarium was one of the most frequent weed species at Kise. This weed species prefers warm, dry, sandy and N-rich soils (Korsmo et al. Citation2001). Dryer and warmer weather conditions at Kise than at Apelsvoll () could explain the higher frequency at Kise. In a botanical field study near Kise, Often et al. (Citation2008) found an increased frequency of E. cicutarium from 1961 to 2004. Riesinger and Hyvönen (Citation2006) suggest that the ruderal species E. cicutarium is a result of cropping history, rather than an outcome of the organic farming system, since they found this species only in a few fields of organically cropped spring cereals in Finland.
The results suggest that annually repeated undersowing with nitrogen-fixing clovers increases grain yields, but also weed biomass and seed bank size. The subsequent fertilization effect of using legumes as cover crop seems to enhance weed growth more than competition from the cover crop suppressed it in any given year. However, when ryegrass was sown alone or mixed with clover, weed growth in one of the experimental sites was suppressed, which demonstrates the strong suppressing effect of that cover crop. Grain yields only increased, as expected, when ryegrass was mixed with clover. Main-crop clover green manure resulted in a significant drop in seed bank size the following year, because of limited weed establishment in an established ley. But the average weed biomass increased slightly at only one of the two experimental sites, probably because of better water and nitrogen availability of that site. The weed seed bank and the weed biomass were essentially kept at steady state during the experimental period in harrowed plots. Although harrowing resulted in decreased grain yield in our study, several studies have shown that the treatment is selective when carried out in a proper way and under good conditions.
Acknowledgements
The Research Council of Norway funded this study. The authors wish to thank the field technicians Oddvar Bjerke, Marit Helgheim and the staff at Kimen Seed Testing Laboratory AS for the technical assistance.
References
- Albrecht , H. 2005 . Development of arable weed seedbanks during the 6 years after the change from conventional to organic farming . Weed Research , 45 : 339 – 350 .
- Alkämper , J. 1976 . Einfluz der Verunkrautung auf die Wirkung der Düungung . Pflanzenschutz-Nachrichten Bayer , 29 : 191 – 235 .
- Andreasen , C. , Litz , A-S. and Streibig , J. C. 2006 . Growth response of six weed species and spring barley (Hordeum vulgare) to increasing levels of nitrogen and phosphorus . Weed Research , 46 : 503 – 512 .
- Bergkvist , G. , Adler , A. , Hansson , M. and Weih , M. 2010 . Red fescue undersown in winter wheat suppresses Elytrigia repens . Weed Research , 50 : 447 – 455 .
- Blackshaw , R. E. , Moyer , J. R. and Huang , H. C. 2005 . Beneficial effects of cover crops on soil health and crop management . Recent Research Developments in Soil Science , 1 : 15 – 35 .
- Brandsæter , L. O. , Netland , J. and Meadow , R. 1998 . Yields, weeds, pests and soil nitrogen in a white cabbage-living mulch system . Biological Agriculture and Horticulture , 16 : 291 – 309 .
- Breland , T. A. 1996 . Manuring with clover and ryegrass catch crops undersown in small grains: crop development and yields. Acta Agriculturae Scandinavica . Section B – Soil & Plant Science , 46 : 30 – 40 .
- Davies , D.H. K. A. , Christal , M. , Talbot , H. M. , Lawson , & Wright , G. McN . 1997 . Changes in weed populations in the conversion of two arable farms to organic farming . In Brighton crop protection conference 1997: Proceedings of an international conference held in Brighton, UK in November 1997: Weeds , pp. 973 – 978 . Alton : British Crop Protection Council
- Den Hollander , N. G. , Bastiaans , L. and Kropff , M. J. 2007 . Clover as a cover crop for weed suppression in an intercropping design – II. Competitive ability of several clover species . European Journal of Agronomy , 26 : 104 – 112 .
- Dock-Gustavsson , A-M. 1994 . Åkertistelns reaktion på avslagning, omgrävning och konkurrens . Fakta – Mark/växter 13 . Uppsala : SLU . (In Swedish)
- Graglia , E. , Melander , B. and Jensen , R. K. 2006 . Mechanical and cultural strategies to control Cirsium arvense in organic arable cropping systems . Weed Research , 46 : 304 – 312 .
- Hartl , W. 1989 . Influence of undersown clovers on weeds and on the yield of winter wheat in organic farming . Agriculture Ecosystems and Environment , 27 : 389 – 396 .
- Hiltbrunner , J. , Jeanneret , P. , Liedgens , M. , Stamp , P. and Streit , B. 2007 . Response of weed communities to legume living mulches in winter wheat . Journal of Agronomy and Crop Science , 193 : 93 – 102 .
- Korsmo , E. , Vidme , T. and Fykse , H. 2001 . Korsmos ugrasplansjer , 3rd edition , Oslo : Landbruksforlaget AS. (Including text in English) .
- Lawson , H. M. , Wright , G. McN , Bacon , E. T. G. , & Yeoman , D. 1994 . Impact of set-aside management strategies on the soil seedbank and weed flora at Woburn . Aspects of Applied Biology , 40 , 453 – 459 .
- Løes , A-K. , Henriksen , T. M. , Eltun , R. and Sjursen , H. 2011 . Repeated use of green-manure catch crops in organic cereal production – grain yields and nitrogen supply . Acta Agriculturae Scandinavica Section B – Soil & Plant Science , 61 : 164 – 175 .
- Maiksteniene , S. , Arlauskiene , A. , Velykis , A. and Satkus , A. 2009 . Enhancement of competitive ability of cereals towards weeds by means of crop rotation . Zemdirbyste-Agriculture , 96 : 23 – 34 .
- Mangerud , K. , Brandsæter , L. O. and Wærnhus , K. June 2007 2007 . “ Criteria for improved weed harrowing in cereals ” . In 14th EWRS Symposium , June 2007 , 82 Norway : Hamar .
- Molteberg , B. , Henriksen , T. M. , & Tangsveen , J. 2004 . Use of catch crops in cereal production in Norway . Planteforsk Grønn kunnskap , 8 12 , 1 – 57 . (In Norwegian with an English summary)
- Often , A. , Bruserud , A. , & Stabbetorp , O. 2008 . Floraen på Nes og Helgøya. Ugras . Nes og Helgøya Lokalhistorisk skrift, 2008 , 38 – 68 . (In Norwegian)
- Olesen , J. E. , Rasmussen , I. A. , Askegaard , M. and Kristensen , K. 2002 . Whole-rotation dry matter and nitrogen grain yields from the first course of an organic farming crop rotation experiment . Journal of Agricultural Sciences , 139 : 361 – 370 .
- Paatela , J. and Erviö , L. R. 1971 . Weed seeds in cultivated soils in Finland . Annales Agriculturae Fenniae , 10 : 144 – 152 .
- Pridham , J. C. and Entz , M. H. 2008 . Intercropping spring wheat with cereal grains, legumes, and oilseeds fails to improve productivity under organic management . Agronomy Journal , 100 : 1436 – 1442 .
- Rasmussen , I. A. , Askegaard , M. , Olesen , J. E. and Kristiansen , K. 2006 . Effects on weeds of management in newly converted organic crop rotations in Denmark . Agriculture Ecosystems and Environment , 113 : 184 – 195 .
- Riesinger , P. and Hyvönen , T. 2006 . Impact of management on weed species composition in organically cropped spring cereals . Biological Agriculture and Horticulture , 24 : 257 – 274 .
- Ross , S. M. , King , J. R. , Izaurralde , R. C. and O'Donovan , J. T. 2001 . Weed suppression by seven clover species . Agronomy Journal , 93 : 820 – 827 .
- Rydberg , N. T. and Milberg , P. 2000 . A survey of weeds in organic farming in Sweden . Biological Agriculture and Horticulture , 18 : 175 – 185 .
- SAS Institute 2004 . Base SAS, 9.1 Procedures Guide . Cary NC: SAS Institute Inc .
- Sjursen , H. 2001 . Change of the weed seed bank during the first complete six-course crop rotation after conversion from conventional to organic farming . Biological Agriculture and Horticulture , 19 : 71 – 90 .
- Stenerud , S. G. 2009 . Combined effects of weed harrowing and undersown clover on weed growth and cereal yield . Master Thesis, Norwegian University of Life Sciences, Department of Plant and Environmental Sciences, Ås, Norway .
- Sustainable Agriculture Network 1998 . Managing cover crops profitably, . Handbook Series Book 3 , 2nd edition . Beltsville , MD : United States Department of Agriculture .
- Teasdale , J. R. , Brandsæter , L. O. , Calegari , A. and Skora Neto , F. 2007 . “ Cover crops and weed management ” . In Non-chemical weed management , Edited by: Upadhyaya , M. K. and Blackshaw , R. E. 49 – 64 . Wallingford : CAB International .
- Zadoks , J. C. , Chang , T. T. and Konzak , G. F. 1974 . A decimal code for the growth stages of cereals . Weed Research , 14 : 415 – 421 .