Weed suppression and crop yield performance in sole and intercrops of common vetch and spring wheat depending on seed density ratio in organic farming

Unkrautunterdrückung und Ertragsleistung in Reinsaat und Gemengen von Saatwicken mit Sommerweizen in Abhängigkeit vom Saatstärkenverhältnis im ökologischen Landbau

Journal für Kulturpflanzen, 72 (1). S. 12–24, 2020, ISSN 1867-0911, DOI: 10.5073/JfK.2020.01.02, Verlag Eugen Ulmer KG, Stuttgart

Correspondence

Roman Kemper, Agroecology and Organic Farming Group, Institute of Crop Science and Resource Conservation, University of Bonn, Katzenburgweg 3, 53115 Bonn, Germany, e-mail: rkemper@uni-bonn.de

Accepted

25 November 2019

This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/deed.en).
Dies ist ein Open-Access-Artikel, der unter den Bedingungen der Creative Commons Namensnennung 4.0 International Lizenz (CC BY 4.0) zur Verfügung gestellt wird (https://creativecommons.org/licenses/by/4.0/deed.de).

Materials and methods

The 30-year (1986–2016) mean annual precipitation at the experimental site is 693 mm with a mean temperature of 9.0 °C. In the long term average the mean temperature from April to August is 14.2 °C and the precipitation rate is 306.5 mm. In the growing period from April to August 2018 the mean temperature (16.6 °C) was 2.4 °C warmer than the 30-year (1986–2016) mean and the experimental site received only half the amount (157.9 mm) of the longterm average precipitation. Weather data were recorded at the nearby weather station Lübeck-Blankensee from the German Meteorological Service.

Primary tillage was carried out with a mouldboard plough in autumn 2017. Prior to seeding secondary tillage was performed with a cultivator and followed by a rotary harrow. Crops were sown on 21^st April 2018 with a plot seeder (PZT D3–24 Quatro, Agrar Markt Deppe, Germany), which is especially constructed for row intercrops, at a sowing depth of 3.5 cm and with a row distance of 12.5 cm. The intercrops were sown in alternating rows. No mechanical weed control was performed and no fertilizer was applied. Perennial weeds, in particu­lar Canada thistle (Cirsium arvense (L.) Scop.) and Coltsfoot (Tussilago farfara L.), were removed three times by hand until EC 33 of spring wheat as they occurred in patches and are not spread over all plots unlike annual weeds.

Every 7 to 12 days, beginning at stem elongation stage (20 d.a.s.), measurements of the photosynthetically active radiation (PAR) above crop canopy and PAR transmitted to weed canopy level were taken to determine the fraction of PAR transmitted to weed canopy level. A SunScan Canopy Analysis System type SS1 and a BF5 Sunshine Sensor (Delta-T Devices Ltd, UK) were used. PAR measurements were taken at weed canopy level and four measurements per plot were taken. Weed canopy level was defined as a dynamic height as weeds grow and PAR was measured on ground level at the first and second measurement, at 5 cm at third and fourth measurement and at 10 height from the fifth measurement onwards.

RWB (equation 1) was calculated according to Szumigalski and Van Acker (2005). WBI is the weed biomass in the intercrop and WBS_cv and WBS_sw are the weed biomasses in the sole crops of common vetch and spring wheat, respectively. RWB < 1indicates synergistic weed suppression by the intercrop partners.

[1]

WR (equation 2) shows the weed suppression of the intercrops compared to the sole crop of common vetch. Relative weed biomass was calculated according to Carton (2017).

[2]

LER (equation 3, 4, and 5) is the relative land area under sole crops that is needed to produce the same yields as in intercropping. The land use efficiency will be greater of an intercrop compared to the corresponding sole crops, if the LER values are > 1 (Willey, 1979). I_cv and I_sw are the biomasses in the intercrop of vetch and wheat, respectively, and S_cv and S_cw are the biomasses of the vetch and wheat sole crop, respectively. In case of the final harvest grain yield instead of biomass was used.

[3]

[4] [5]

CR (equation 6) shows the number of times one crop is more competitive than the other. It is the ratio of the LER of both crops, corrected by the proportion in which the crops were sown as an intercrop (Willey and Rao, 1980). As the values of the two crops are reciprocals of each other, only the CR for common vetch is calculated. P_cv and P_sw are the seed density ratios of common vetch and spring wheat, respectively.

[6]

AYL (equations 7, 8 and 9) is the proportionate yield loss or gain of intercrops compared to the corresponding sole crop respecting sowing proportions. A positive AYL shows a yield gain whereas as a negative AYL shows yield loss. The value of the AYL shows the relative quantity of yield loss or gain (Banik, 1996; Banik et al., 2000).

[7]

[8] [9]

REIc (equations 10, 11 and 12) is the relative efficiency of common vetch and spring wheat assessed over the time period. It compares the change in dry matter (K) of both intercrop partners within the time interval (t1 to t2). If both intercrop partners have the same proportional growth in one interval, the ratio of their change (REIc) is equal to 1 (Connolly, 1987). The time interval is given between t1 and t2.

[10]

[11]

[12]

REIc was calculated for three intervals: from sowing to first biomass harvest, from first to second biomass harvest, and from second to third biomass harvest. For the first time interval, the seed weight for the corresponding area was taken as biomass.

[13]

y_ijk is the dependent variable. The subscript i indicates the ratio of vetch, j the vetch cultivar, k the block. μ is the general mean, b is the effect of the block, α of the vetch cultivar and x is seed density ratio of vetch in the crop stand. In the regression model, the wheat sole crop is represented by a crop stand with a ratio of 0% vetch.

PAR transmission to weed canopy level at each date was analysed by analysis of variance in Proc GLM. A nested two factorial model was used. By using a control factor, the treatments were categorized into wheat sole crop and treatments with vetches. Within the treatments with vetches there was a two-factorial structure. The first factor was the ratio of vetch in the intercrop and the second factor was the vetch cultivar which is not examined in this publication. If necessary, data were transformed with logarithm to achieve homogeneity of variances and normal distribution. The model of the analysis of variance is shown in equation 14.

[14]

y_ijkm is the dependent variable. The subscript i indicates the ratio of vetch, j the vetch cultivar, k the block, and m the control factor level. μ is the general mean, d is the effect of the block, g of the control factor, α of the vetch cultivar and β of the seed density ratio of vetch in the crop stand. Means were compared with a Tukey test at a significance level of α = 0.05.

The indices RWB, WR, LER, CR, AYL, and REIc were calculated for each intercrop plot. The calculations were done according to Oyejola and Mead (1982). The means of each intercrop treatment of RWB, LER, CR and REIc was compared to a value of 1 with a t-test in Proc Ttest. Likewise, the means of each intercrop treatment of AYL and WR were compared to a value of 0. Additionally, an analysis of variance was performed for the different indices in Proc GLM to see the influence of seed density ratio. Subsequently a comparison of means was performed with a Tukey test at a significance level of α = 0.05.

Results

Fig. 1. Weed biomass [g dry matter m^–2] as a functi­on of ratio of common vetch in the seeding mixture in a replace­ment series of spring wheat and common vetch at different bio­mass harvests. Symbols show the means of the measured values.

Two out of three RWB values were significantly lower than 1 at the first, second, and third harvest date, whereas at the fourth harvest date no RWB value differed significantly from 1 (Table 1). At all harvest dates, RWB was significantly lower in intercrops with a vetch ratio of 25 and 50% than in those with 75% vetch. Except for intercrops with a vetch ratio of 75% at the last harvest date all WR values are significantly different from 0%. At the first three harvest dates, WR was significantly higher in intercrops with 25% and 50% vetch than in those with 75% vetch (Table 1). Intercrops with 25% vetch achieved a significantly higher weed reduction than those with 75% vetch at fourth harvest, while WR in intercrops with 50% vetches was intermediate.

Fig. 2. Photosynthetically active radiation (PAR) transmitted to weed canopy level [%] in a re­placement series of spring wheat and com­mon vetch during vege­tation. Line type refers to ratio of vetch in the seeding mixture. Means followed by the same letter are not significant­ly different within one date (α = 0.05).

Regression analysis for the PAR data of the second measurement date 32 d.a.s. against the seed density ratio showed that PAR available to weeds increased with increasing ratio of vetch in the intercrop during juvenile development (Fig. 3).

Fig. 3. Photosynthetically active radiation (PAR) transmitted to weed canopy level [%] 32 days after sowing (d.a.s.) as a function of ratio of vetch in the see­ding mixture in a re­placement series of spring wheat and com­mon vetch. Symbols show the means of the measured values.

Fig. 4. Nitrogen content [%] of weeds as a function of ratio of common vetch in the seeding mixture in a replacement series of spring wheat and common vetch at diffe­rent biomass harvests. Symbols show the me­ans of the measured values.

Weed nitrogen accumulation increased with increasing ratio of vetch in the intercrop according to poly­nomial regression (Fig. 5). Weed nitrogen accumulation increased from first to second harvest. However, weed nitrogen accumulation from second and third harvest scarcely differed. The coefficient of determination (R²) for regression of weed nitrogen accumulation against seed density ratio was lower at third harvest than at first and second harvest (Fig. 5).

Fig. 5. Nitrogen accumulation [g m^–2] in weed bio­mass as a function of ra­tio of common vetch in the seeding mixture in a replacement series of spring wheat and com­mon vetch at different biomass harvests. Sym­bols show the means of the measured values.

For two intercrops in the first interval REIc values were significantly different from 1 (Table 3). In the first interval (T0-T1) vetches in intercrops with a vetch ratio of 25% grew at a higher rate, while wheat grew at higher rate in intercrops with a ratio of 75% of vetch. In the second interval (T1-T2) vetches in intercrops with a vetch ratio of 50% and 75% grew at a higher rate, whereas wheat grew at higher rate in intercrops with a ratio of 25% of vetch. At the third interval (T2-T3) vetches in intercrops with a vetch ratio of 75% grew at a higher rate, while wheat grew at higher rate in intercrops with a ratio of 25% of vetch.

Fig. 6. Grain yield [g dry mat­ter m^–2] as a function of ratio of vetch in the see­ding mixture in a re­placement series of spring wheat and com­mon vetch. Line type re­fers to expected and actual yield (solid lines actual grain yield, das­hed lines expected yield). Symbols show the means of the measu­red values.

Total grain yield varied between 227 to 354 g m^–2. Maximum total grain yield was achieved in intercrops. The curve progression for total grain yield of both vetch and wheat was an optimum curve. 43% of vetch in the seeding mixture achieved the maximum total grain yield. However, the total grain yield between the three different intercrops did not differ. The proportion of vetch grain of the total yield increased with increasing ratio of vetch in the seeding mixture. Moreover, Fig. 6 shows the comparison of actual grain yield of intercrop components and total intercrop with the expected grain yield. Expected grain yield was calculated across the replacement series by a linear curve progression. For total grain yield and wheat grain yield the curves for actual grain yield ran clearly above the curve for expected grain yield. For vetch grain yield the curve of actual grain yield ran slightly below the curve for expected grain yield.

Indices in Table 4 were calculated on the basis of grain yield. All different intercrop ratios achieved a LER which was significantly above 1 (Table 4). The seed density ratio of the intercrops had no effect on LER. For the competitive ratio of common vetch, all intercrops achieved a CR which was significantly below 1 (Table 4). The seed density ratio of the intercrops did not influence the CR. Actual yield gain for wheat was significantly higher than 0 for all intercrops (Table 4). The AYL of wheat increased with increasing ratio of vetch in the intercrop. In intercrops with a vetch ratio of 25% and 50% significant actual yield loss of vetches was found (Table 4). The actual yield loss of vetch was stronger with low ratios of vetch. The AYL was significantly different from 0 in all intercrops (Table 4). Actual yield gain increased with increasing ratio of vetch.

Discussion

The enhanced weed suppression in intercrops can also be seen from the indices RWB and WR (Table 1). Both synergistic weed suppression (RWB < 1) and WR decreased during the growing period. This shows that weed suppression in intercrops is especially effective in early growth stages where crops are very sensitive to weed pressure. The higher synergistic weed suppression in intercrops with a vetch ratio of 25% and 50% reflects the results of the weed biomass, which was much higher in intercrops with 75% vetches. Except third harvest date, WR followed the same trend that weed reduction was stronger in intercrops with fewer vetches. Although RWB and WR are calculated differently, both indices basically show the same trend.

Photosynthetically active radiation (PAR). Differences in PAR transmission to weed canopy level between seed density ratios were enormous (Fig. 2). In the period from 20 to 40 d.a.s. more light was transmitted to weed canopy level in crop stands with higher ratios of vetch, whereas in the period after 40 d.a.s. more light was transmitted to the weed canopy level in crop stands with higher ratios of wheat. The reason might be the slower juvenile development of common vetch in comparison to spring wheat in the beginning and by high biomass production and lodging of the vetch in later growth period. Thus, it can be concluded that PAR transmission to weed canopy level in the period from 20 to 40 d.a.s. is of decisive importance for the establishment of weeds in the different crop stands. The regression analysis for the PAR data of the second measurement date against the seed density ratio (Fig. 3) revealed a similar curve progression like that of weed biomass against seed density ratio (Fig. 1). It needs to be considered that the PAR data are one-year results.

In sole and intercrops of pea and oat, Gronle et al. (2015) determined that the factor light competition between crops and weeds is not responsible for weed suppression. This result was primarily revealed in a pot experiment where shoot and root interference could be controlled. However, Gronle et al. (2015) showed that in one experimental year highest PAR transmission to weeds was found in pea sole crop and lowest in oat sole crop at the beginning of crop growth. In another field trial the same researcher team showed that in intercrops with triticale and winter peas PAR transmission to weed canopy level is a factor for weed suppression (Gronle et al., 2014). This was especially the case for normal-leafed winter peas which are comparable to vetch due to their high leaf and biomass production (Gronle et al., 2014).

Moreover, the availability of PAR in the later stage might be important for late-season weed infestation. However, this effect could not be clearly seen as no late-season weed infestation was found in the year of the field experiment, probably due to summer drought. Corre-Hellou et al. (2011) discussed that peas are sensitive to weed infestation during juvenile development, but that they have a high competitive ability for light in later growth stages. Similar to the results for vetch sole and intercrops in the present study, Gronle et al. (2014) detected that PAR transmission to weed canopy level was very low in sole and intercrops of normal-leafed pea in the later growth period.

Nitrogen. Poggio (2005) determined less PAR at weed canopy level in barley sole and pea-barley intercrops compared to pea sole crops and simultaneously less accumulated nitrogen in weed biomass in barley sole and pea-barley intercrops compared to pea sole crops. Thus, higher PAR interception resulted from a greater leaf area caused by more available nitrogen (Poggio, 2005). Similar results, which showed an interrelation of nitrogen and light use in intercrops, were found in experiments by Corre-Hellou et al. (2006) and Corre-Hellou et al. (2011), as leaf area relates with the capability for absorbing soil nitrogen. The lower availability of nitrogen to weeds in intercrops and wheat sole crops can be assumed as a further reason for weed suppression. The weed nitrogen content and accumulation decreased with increasing ratio of wheat in the intercrop (Fig. 4). As cereals are more efficient in nitrogen uptake (Jensen, 1996) and vetches as legumes are self-sufficient in terms of nitrogen, more nitrogen is available for weeds when the ratio of wheat in the crop mixture is reduced. This result is confirmed by many other studies reporting higher weed suppression in cereal sole and cereal-legume intercrops compared to legume sole crops (Hauggaard-Nielsen et al., 2001; Poggio, 2005; Corre-Hellou et al., 2011; Gronle et al., 2015).

Comparison of different indices evaluating productivity of intercrops. The LER increased from first biomass harvest to grain harvest (Tables 2 and 4). Thus, the extra yield in intercrops is built up over time. LER was not affected by the seed density ratio, neither for the biomass nor for the grain yield data. In a replacement series of pea and barley of Hauggaard-Nielsen et al. (2006), seed density ratio also did not affect LER in the first half of the growing season. However, at their last harvest, the LER tended to decline with increasing ratio of peas in the intercrop. The same trend as shown by LER is presented by AYL (Table 2 and 4). AYL increased over time to a value of 0.73 averaged across grain yield of all intercrops, which indicates a yield gain of 73% in intercrops compared to sole crops. Unlike LER, actual yield gain of intercrops increased with an increasing vetch ratio. Thus, AYL is more sensitive to different sowing proportions as they were considered when computing the index. Also Banik (1996) and Banik et al. (2000) emphasised that AYL gives more precise information than LER. When intercrops are investigated in a replacement series, it is therefore more useful to quantify the extra yield of intercrops by AYL instead of LER. All values of AYL for wheat were positive (Tables 2 and 4), which showed that there was a yield gain of wheat in intercrops compared to sole crops. Actual yield gain of wheat increased with increasing ratio of vetch in the intercrop. Thus, wheat is more productive with less wheat in the mixture due to reduced intraspecific competition of wheat plants. AYL for vetch showed that common vetches in intercrops with spring wheat tended to have actual yield loss instead of actual yield gain in intercropping (Tables 2 and 4). However, the actual yield loss of vetch biomass was not significant in all intercrops. At first harvest, actual yield loss of vetch biomass was lower in intercrops with a lower ratio of vetch, whereas at third harvest date and for the grain yield data a higher ratio of vetch resulted in a lower actual yield loss of vetch biomass. Thus, intraspecific competition of vetches might be lower than interspecific competition with wheat. The CR values for both biomass and grain yield showed that vetch was always less competitive compared to wheat (Tables 2 and 4). In an 50:50 intercrop of common vetch and spring wheat, Šarunaite et al. (2010) also determined spring wheat as the dominant component in the intercrop, whereas other studies did not clearly show that vetch is less competitive than the cereal (Dhima et al., 2007; Erol et al., 2009).

Results of REIc showed that the intercrop partner with a lower proportion grew at a higher rate at the beginning. This can be explained due to reduced intraspecific com­petition. At the later stage, the intercrop partner with the lower proportion grew at a lower rate, because this partner might be already suppressed. Hauggaard-Nielsen et al. (2006) calculated REIc for two intercrops of pea and barley and reported that barley grew at higher rate in the first growth phase, whereas peas grew at higher rate in the last growth phase in both intercrops with high and low proportions of pea.

Comparison of actual and expected yield. The regression curve of grain yield against seed density ratio revealed that intercrops achieved higher total yield and the indices quantify this yield gain. The comparison of actual grain yield with the expected grain yield (Fig. 6) visualised these results and pointed out spring wheat as the major source for extra yield. It should be noted, that this is the result of a one-year study and could have been caused by the dry weather conditions. Yield benefits of intercrops are caused by more efficient acquisition and use of resources (Hauggaard-Nielsen et al., 2008; Jensen et al., 2015), for instance due to nitrogen fixation grain legumes and cereals can use complementary nitrogen pools in intercropping (Jensen, 1996; Bedoussac et al., 2015). Thus, the legume will rely on fixation of atmospheric nitrogen and the cereal acquires more mineral nitrogen than the legume (Jensen, 1996). Due to improved nitrogen nutrition of intercropped cereals, their leaf area per cereal plant is greater. Thus, intercropped cereals have a higher competitive ability for light than of sole cropped cereals, resulting in a higher relative yield of cereals in intercrops (Corre-Hellou et al., 2006).

Maximum total yield was achieved in intercrops and not in sole crops. The optimal ratio differs for maximum yield, high proportions of vetch in grain yield and highest weed suppression. Weed infestation was low in intercrops with a ratio of less than 60% vetch. High total grain yield was achieved with all three intercrops. Higher vetch proportions in grain yield were found in intercrops with higher ratio of vetch in the seeding mixture. Thus, intercropping common vetch with cereals in ratios between 30 and 60% vetch in the mixture seem to be possible. This emphasises the advantages of replacement series compared to field trials with just one intercropping treatment, as replacement series give the possibility to calculate optima for different criteria.

Weed nitrogen content and accumulation increased with higher vetch ratios in the intercrop. In the first 40 days less light reached the weed canopy level in intercrops with higher ratios of wheat compared to intercrops with high vetch ratios. Thus, crop-weed competition for nitrogen and for light in the early growth stages was determined as the major factors contributing to weed suppression in vetch-wheat intercrops.

The mean land equivalent ratio (LER) for grain yield was 1.32. The index actual yield loss or gain (AYL) was predominantly positive. The mean AYL for grain yield indicated a yield gain of 73% compared to the sole crops. AYL was more precise than LER, as it took the different seed density ratios into account. Thus, it is better suited for field experiments with replacement series.