Jan Bocianowski, Piotr Szulc, Kamila Nowosad

1 Department of Mathematical and Statistical Methods, Poznań University of Life Sciences, Wojska Polskiego 28, Poznań 60-637,Poland

2 Poznań University of Life Sciences, Department of Agronomy, Dojazd 11, Poznań 60-632, Poland

3 Wroc?aw University of Environmental and Life Sciences, Department of Genetics, Plant Breeding and Seed Production,Grunwaldzki 24A, Wroc?aw 53-363, Poland

Abstract The objective of this study was to assess soil tillage methods by years interaction for dry matter of plant yield of maize (Zea mays L.) grown in West Poland by the additive main effects and multiplicative interaction model. The study comprised four soil tillage methods, analysed in 12 years through fleld trials arranged in a randomized complete block design, with four replicates. Dry matter of plant yield of the tested soil tillage methods varied from 86.7 dt ha–1 (for no-plough tillage in 2005) to 246.4 dt ha–1 (for complete conventional tillage in 2012), with an average of 146.6 dt ha–1. In the variance analysis,49.07% of the total dry matter of plant yield variation was explained by years, 12.69% by differences between soil tillage methods, and 10.53% by soil tillage methods by years interaction. Dry matter of plant yield is highly influenced by soil tillage methods by years factors.

Keywords: AMMI model, biplot, Zea mays L., dry matter of plant yield, stability

1. Introduction

So far, the dominant way of farming in all agricultural production systems is full plowing, regardless of soil and economic conditions (Verhulstet al. 2011). It is the most durable and at the same time the most widespread way of preparing the soil for sowing in every climate zone,regardless of the intensity of plant cultivation (Kitonyoet al.2018). The method has very well-developed theoretical foundations and is characterized by many advantages. Full plow cultivation is expensive and energy-consuming, as it absorbs 25–40% of the total expenditure on plant production and requires a rich park of machines and tools for farming.The autumn-covered soil left behind by winter until early spring, without cover, is subject to watery and windy erosion,excessive thickening of deeper layers soil (Czy? 2004;Bojarszczuket al. 2017). Due to the signiflcant increase in the intensiflcation of plant production and changes in the method of harvesting plants, the time required for pre-sowing cultivation has been shortened. Therefore, it was necessary to simplify agrotechnical procedures, consisting mainly in reducing the crop or limiting the number of treatments, until their total elimination, to the so-called zero crop. Direct sowing is used in many countries around the world (Czy? 2004; Bojarszczuket al. 2017). In the system of cultivation after harvesting grains (ears), the remaining maize straw in the fleld is not collected for fodder but after it has been minced, it is plowed. Almost half of the produced biomass returns to the soil, increasing the content of organic matter and nutrients. However, when growing silage, the whole plant biomass is removed from the fleld, helping to reduce the amount of organic matter in the soil. Many breeders have understood the importance of soil tillage methods by years interactions in plant breeding programme for a long time(Fischer and Edmeades 2010). Dry matter of plant yield is greatly influenced by soil tillage methods, years, and complex soil tillage methods by years interactions.

In fleld crop trials, the soil tillage methods by years (SY)interaction maybe analyzed by the additive main effects and multiplicative interaction (AMMI) model (Zobelet al. 1988).This model was originally developed for analysis in social sciences and physics (Mandel 1961; Gollob 1968), and later adjusted for research in plant sciences (Gauch 1988;Cornelius 1993). The AMMI model is a combination of the analysis of variance for the SY main effects and the principal component analysis (PCA) with multiplicative parameters in a single analysis. The AMMI biplot graphic display simultaneously both main and interaction effects, and enables a single analysis of the SY interaction.

The objective of this study was to assess soil tillage methods by years interaction for dry matter of plant yield in maize (Zea maysL.) by the additive main effects and multiplicative interaction model.

2. Materials and methods

The experiment was located at the Experimental and Educational Center in Swadzim in the experimental fleld of the Department of Agronomy at the Poznań University of Life Sciences. The static fleld experiment on various variants of the preparation of the role for maize cultivated in grain technology was established in 1996. The experiment was founded as one-factor in four fleld repetitions. The size of a single plot was 34.72 m2. The study presents and discusses the yield of the dry mass of whole plants (grain+straw) as a result of the long-term use of the following tillage methods in monoculture maize cultivation. The following cultivar variants were examined: A, full traditional cultivation, deep autumn plowing (30 cm), spring cultivator with a string roller;B, simplifled traditional cultivation, shallow autumn plowing(15 cm), spring cultivator with a string roller; C, only spring cultivation, spring plowing sowing (15 cm), cultivator with a string roller; D, no tillage cultivation, zero tillage, and direct sowing in stubble. In 1996, after harvesting winter wheat and collecting straw on designated plots, the cultivation procedures provided in the experimental diagram were performed. Autumn plowing (methdos A and B) in all years of research was carried out in mid-November, while spring plowing, sowing (method C) at the end of March. In 1997,maize was sown for the flrst time, while in 1998, maize was flrst grown in monoculture. The yield of dry matter of whole maize plants presented and analyzed in the work came from the years 2004–2015 (i.e., from 7 to 18 years of maize cultivation in monoculture). Yields of dry matter were made on plants from 2 m2. Separately, the straw yield, the crop of flasks without shirts were determined. The dry matter yield of whole plants was the sum of two previous crops.

A two-way flxed effect model was fltted to determine the magnitude of the main effects of variation and their interaction on dry matter of plant yield. Least-squares means were simultaneously produced for the AMMI model.The model flrst flts additive effects for the main effects of soil tillage methods (S) and years (Y) followed by multiplicative effects for SY interaction by principal component analysis.The AMMI model (Gauch and Zobel 1990; Nowosadet al.2016, 2017) is given by:

Where,ygeis the dry matter of plant yield mean of soil tillage methodgin yeare, μ is the grand mean, αgis the soil tillage method mean deviations, βeis the year mean deviations,Nis the number of PCA axis retained in the adjusted model, λnis the eigen value of the PCA axisn, γgnis the soil tillage method score for PCA axisn,is the score eigenvector for PCA axisn,Qgeis the residual, including AMMI noise and pooled experimental error.

3. Results and discussion

In the analysis of variance, the sum of squares for years effect represented 49.07% of the total, and this factor had the highest effect on dry matter of plant yield (Table 1). The differences between soil tillage methods explained 12.69%of the total dry matter of plant yield variation, while the effects of SY interaction explained 10.53%. Values for the two principal components were also signiflcant. The flrst principal component (IPCA 1) accounted for 57.64% of the variation caused by interaction, while IPCA 2 accounted for 32.72% (Table 1; Fig. 1).

Dry matter of plant yield of the tested soil tillage methods varied from 86.7 dt ha–1(for method D in 2005) to 246.4 dt ha–1(for method A in 2012) throughout the 12 years, with an average of 146.6 dt ha–1(Table 2). The method A of soil tillage had the highest average dry matter of plant yield(158.0 dt ha–1), and the method D had the lowest (122.7 dt ha–1). The average dry matter of plant yield per year also varied from 110.1 dt ha–1in 2008 to 215.1 dt ha–1in 2012.

The stability of tested soil tillage methods can be evaluated according to biplot for dry matter of plant yield(Fig. 2). The soil tillage method A interacted positively with the 2007, method B with 2012 and 2013, method C with 2009, and method D with 2004, 2006, and 2008 (Figs. 1 and 2). Soil tillage methods on the highest point in certain sections of the graph have the best results in years located in the same section (Fig. 2). Soil tillage method B, with average dry matter of plant yield close to the general mean of 146.6 dt ha–1, is distinguished on the biplot. This method of soil tillage had the highest stability.

Table 1 Analysis of variance of main effects and interaction for maize soil tillage methods dry matter of plant yield

The obtained results of numerous studies indicate that the cultivation method has a strong impact on the level of maize yield. However, the evaluation of the results on the effect of the intensity of tillage on the yield of maize indicates a large discrepancy in the results-obtained (Gilleret al. 2009).Grifflthet al. (1988) found stabilization of maize yielding in direct sowing in relation to the yield in mechanical cultivation.The authors also emphasize that the longer use of direct sowing results in stabilization of some physical, chemical,and biological features of the soil and creates favorable conditions for yielding maize (Thierfelderet al. 2013).The results of research carried out in the USA by other reseachers indicate that crop simpliflcations, especially zero cultivation applied repeatedly for several years, resulted in changes in the soil environment which initially had a negative impact on the growth, development, and yielding of maize,after which in the following years they stabilized at the level ensuring positive economic effects (Pierceet al. 1994).Earlier research indicates that the yield of maize obtained in a system without mechanical tillage is comparable to that obtained with traditional plow cultivation and in some years even greater (Gilleret al. 2009). The statements cited above were also obtained in the own research.

Fig. 1 Biplot for soil tillage methods by environment interaction of maize (Zea mays L.) in 12 years, showing the effects of primary and secondary components (IPCA 1 (principal component of interaction) and IPCA 2, respectively). A, full traditional cultivation, deep autumn plowing (30 cm), spring cultivator with a string roller; B, simplifled traditional cultivation,shallow autumn plowing (15 cm), spring cultivator with a string roller; C, only spring cultivation, spring plowing sowing (15 cm),cultivator with a string roller; D, no tillage cultivation, zero tillage and direct sowing in stubble.

The AMMI model which incorporates analysis of variance and PCA into a single model and enables simple visual interpretation of the SY interaction is more effective for understanding and evaluating complex data from multi-years dry matter of plant yield trials than traditional and statistical methods (analysis of variation, principal component analysis and linear regression). AMMI model is usually constructed from the flrst two IPCA axes. Measuring SY interaction is very important to determine an optimum breeding strategy for releasing soil tillage methods with an adequate adaptation to target years. SY interactions can result from differential fltness of soil tillage methods across years, but also from year-related differences in fltness of individuals.The AMMI biplot (Fig. 1) shows the stability of soil tillage methods and years, as well as speciflc SY interactions.The zone of stability corresponds to the central region of the biplot, at the interaction of zero scores on the flrst and second principal components axes. Soil tillage methods and years close to each other in any graph area represent speciflc adaptation of a soil tillage method to the year.Among the tested soil tillage methods, the D method had the highest IPCA 1 value of 6.606, while the highest value of IPCA 1 was 5.345 in 2007 (Fig. 1).

4. Conclusion

The AMMI biplot allows the visualization of the main effects of the soil tillage methods and years, in addition to the most important SY interactions. The AMMI model was often used in study of many species. Results obtained from AMMI analyses are very important for developing and recommending best soil tillage methods for production in a speciflc year. The AMMI model provides a useful tool in diagnosing SY interaction patterns and improving the accuracy of response estimates. It enables clustering of soil tillage methods based on similarity of response characteristics and identifying potential trends across years.The suggested strategy could extract more information from the SY interaction, thereby aiding researchers in identifying speciflc soil tillage methods with competitive yields across diverse years.

Table 2 Dry matter of plant yield (dt ha–1) for soil tillage methods(A–D) by years, principal component of interaction (IPCA)values of tested soil tillage methods

Fig. 2 Biplot for the primary component of interaction (IPCA 1)and average dry matter of plant yield (dt ha–1). Vertical line at the centre of biplot is the general grand mean. A, full traditional cultivation, deep autumn plowing (30 cm), spring cultivator with a string roller; B, simplifled traditional cultivation, shallow autumn plowing (15 cm), spring cultivator with a string roller; C, only spring cultivation, spring plowing sowing (15 cm), cultivator with a string roller; D, no tillage cultivation, zero tillage and direct sowing in stubble.

The method of reduced conventional tillage (autumn shallow ploughing (15 cm), cultivator with string roller in spring) was most stable for the dry matter of plant yield among studied soil tillage methods. This result conflrms the opinion on the possibility of shallowing autumn ploughing in the cultivation of maize for grain.