No differences in soil structure under winter wheat grown in different crop rotational positions

Abstract Yield decline in wheat grown after wheat is frequently attributed to fungal disease occurrence, but it is also found without visible disease infection. Thus it is hypothesized that other factors such as N supply or soil structural degradation may lead to wheat yield decline when grown after wheat. The aims of this study were to analyze if (i) the crop rotational position of winter wheat causes differences in soil structure at the beginning of the growing season and (ii) the soil structure is related to differences in wheat biomass formation by this date. Different soil structural properties under winter wheat as well as total aboveground biomass of wheat grown in different crop rotational positions (monoculture, first, second and third wheat after oilseed rape) were investigated in two long-term field experiments with contrasting soil texture. At both field sites, no significant effect of the crop rotational position in any of the analyzed soil structural parameters was found. Wheat biomass in spring was on average 54% higher for wheat grown after oilseed rape compared to second and third wheat after oilseed rape or monoculture. In conclusion, growth reduction of wheat cultivated after wheat was not linked to soil structure as measured in spring.


Introduction
Wheat (Triticum aestivum L.) is one of the most important staple food crops worldwide and the most cultivated crop in Germany for many years.After decades of increasing wheat yields in Europe, yields stagnate since around 2000 (Brisson et al. 2010).Possible causes include an increasing share of wheat production at less favourable locations (Moore and Lobell 2015) due to the relatively high economic advantage of wheat cultivation over other crops, resulting in repeated cultivation of wheat after wheat.This was confirmed by Weiser et al. (2018) who emphasized that the increasing area of winter wheat (WW) cultivation during the past two decades in Germany was only possible with growing wheat after wheat in the same field.
Wheat grown after wheat often causes substantial yield decline (Weiser et al. 2018) and increased susceptibility towards drought stress compared to rotational cropping (Sieling et al. 2005) which is most commonly caused by an infection of the wheat plants with the fungus Gaeumannomyces graminis var.tritici (Ggt) and following root senescence.However, wheat yield decline is also found without severe Ggt infection (Sieling et al. 2007), raising the question if other drawbacks, such as a soil structural degradation in wheat grown after wheat, contribute to yield decline (Sieling et al. 2007).In a previous study, Sieling et al. (2005) reported that biomass losses for wheat grown after wheat occurred at early growth stages and were persistent until later growth stages.Therefore, it can be assumed that the yield decline is based on biomass losses in early growth stages.Regarding this, in our study we focused on differences in soil structure and wheat aboveground biomass at the beginning of the growing season.
Soil structure can be influenced by the crop rotation, for example by improving soil macroporosity and hydraulic conductivity by crops with a high rooting intensity like legumes (McCallum et al. 2004).Further, penetration resistance (PR) and dry bulk density (BD) can be decreased by the cultivation of oilseed rape (Brassica napus L.) and lupin (Lupinus angustifolius L.) (Chan and Heenan 1996).Several soil structural properties may be influenced by roots and residues of break crops in the rotation in comparison to monoculture wheat, due to exudation or release of stabilizing or destabilizing substances in the rhizosphere and production of stable biopores (Kirkegaard et al. 2008).Studies in southern Germany showed positive effects of spring rapeseed on aggregate stability (AS), soil porosity and PR, leading to an improved root density of subsequent WW which could explain the higher wheat yield (Schönhammer and Fischbeck 1987).However, crop properties and soil management practices may modify the effects of the crop rotation on soil structure (Ball et al. 2005).
As soil structure affects the soil water and air balance as well as rootability, it may directly influence the water and nutrient acquisition of plants and therefore the plant biomass yield (Kautz et al. 2013).Whalley et al. (2008) found a decrease in wheat yield in case of soil compaction and increased PR, due to the plants not being able to develop a deep root system.Atkinson et al. (2009) reported that increasing soil porosity led to a decreased plant population based on a poor soil-seed contact.In addition, Bölenius et al. (2017) pointed out that plant-available water capacity was the most important soil physical property affecting the yield of WW.
Concerning WW crop rotations, Feng et al. (2020) found in a 17-year-long study that crop diversity of wheat rotations improves AS, while lowering dry BD.Pranagal and Wo źniak (2021) did not detect differences between wheat monoculture (WM) and wheat grown as rotational crop regarding soil physical properties after 30 years.Agomoh et al. (2020) reported that WM may decrease yields, yet not soil health parameters, especially soil biochemical properties.As soil microbial activity is a crucial predecessor of soil structural improvements, this seems contradictive to other results.Thus, the effect of wheat crop rotations on soil structural properties is not clear and is likely dependent on soil texture, which varied among studies from sand to clay soil (Agomoh et al. 2020;Feng et al. 2020;Pranagal and Wo źniak 2021).Furthermore, the above studies focused on long-term winter WM compared to the first wheat after a break crop.In practice, however, wheat is also grown as second or even third wheat after break crops which may have different effects on soil structure.
In this study we analyzed the soil structure under WW grown as monoculture, first, second and third wheat after oilseed rape as break crop in two long-term field experiments with silty and sandy soil texture to investigate if (i) the crop rotational position causes differences in soil structure at the beginning of the growing season and (ii) soil structural properties are related to differences in wheat biomass formation at this timepoint.
The crop rotational trial at Hohenschulen was established in 1989 (Sieling et al. 2005) with one crop rotation: WOR-WW-WW-WW-faba bean-oats.The first (W1) and third (W3) wheat after oilseed rape as break crop were considered.Each crop rotation element was cultivated every year.This trial contains no field replicates, thus the study years serve as replications.Plot size was 54 m 2 .
The sown WOR cultivars were "Arazzo" (2019) and "Puzzle" (2020 and 2021) at Harste and "Penn" (2018-2020) and "Smaragd" (2021) at Hohenschulen, while as WW cultivar "Nordkap" was chosen for all study years.Regarding the crop residues left by the pre-crops, the C input varied between 2.6 and 5.2 Mg C ha −1 , while the C:N ratio ranged between 63 and 70 in 2020 and between 84 and 122 in 2021 (Table 1).After WOR and WW harvest in Harste, reduced soil tillage was performed with a disk harrow to 4 cm soil depth.Before WW sowing in 2020 and 2021, reduced soil tillage was conducted with a cultivator to 12 cm soil depth in Harste and to 15 cm after WOR and 25 cm after WW in Hohenschulen.Sowing density was 400 seeds m −2 on 13 October 2020 and 18 October 2021 in Harste and 320 seeds m −2 on 1 October 2020 and 8 October 2021 in Hohenschulen.In both trials, WW N supply was 265 kg N ha −1 , including soil mineral N content in spring and N fertilization.Further crop management followed the recommendations of the regional extension services, partially adapted according to the personal expertise of the responsible technicians.

Soil and biomass sampling
At both sites, undisturbed soil core samples (250 cm 3 ) were taken in April 2021 and 2022 from 5 to 10 cm, 20 to 25 cm and 40 to 45 cm soil depth with 8 replicates per plot and depth, four cones each from two distinct subplots.The depths were chosen as representative for relevant layers of the soil, that is, the upper and lower topsoil as well as the subsoil.The sampling positions were located be-Table 2. Effect of the crop rotational position of winter wheat (W1 = first wheat after oilseed rape, W2 = second wheat after oilseed rape, WM = wheat monoculture) on bulk density (BD), total pore volume (PV), air capacity (AC), field capacity (FC), pneumatic conductivity (PC) and aggregate stability (AS) in three soil depths in Harste.

Soil depth (cm)
Crop rotational position BD (g cm −1 ) Total PV (Vol%) AC (Vol%) FC (Vol%) PC (cm s −1 ) A S( % ) tween the WW rows which were spaced 12.5 cm apart.Soil cores were dried at 105 • C for 24 h for determination of soil BD, from which total pore volume (PV) was calculated with a particle density of 2.62 and 2.65 Mg m −3 at Harste and Hohenschulen, respectively.Before drying, all samples were saturated to a matrix potential of pF = 1.8 in a sandbox (Eijkelkamp, Giesbeek, the Netherlands) to determine field capacity (FC) and air capacity (AC), the latter by deducting FC from the total PV.Afterwards, pneumatic conductivity (PC) was determined with an air permeability meter PL-300 (Umwelt-Geräte-Technik GmbH, Müncheberg, Germany).
For soil AS analysis, one composite soil sample per plot was taken at the same time and depths and from the same subplots as the soil cores on both field sites in 2022.The soil was air dried and sieved subsequently with a 2 mm and a 1 mm sieve to obtain the fraction of 1-2 mm dry sieved soil aggregates.For this fraction, the share of water stable aggregates was determined by a wet sieving device (Eijkelkamp, Giesbeek, the Netherlands).To do so, 4 g of the sieved soil was placed on a 250 μm sieve in a small can filled with deionized water.The sieves were moved up and down 100 times in 3 min, leading to the slaking of unstable aggregates.The soil material released this way was collected in the can.This procedure was repeated with the remaining soil material and a dispersing solution (2 g sodium hexametaphosphate L −1 ) in the cans, leading to a dispersion of all aggregate structures.The material from the stable aggregates was collected in the can, leaving only sand particles >1 mm on the sieve.All cans were dried at 105 • C for 24 h.After determination of dry matter weight, the portion of water stable aggregates was calculated.
Soil PR was measured in situ in March in 2021 (only Harste) and 2022 (both sites) down to a depth of 60 cm with a penetrologger (Eijkelkamp, Giesbeek, the Netherlands), which was equipped with a cone having a cross-sectional area of 1 cm 2 and an angle of 60 • .Ten measurements per plot were performed at a speed of 1 m s −1 .Aggregated values were calculated for 0-15, 15-30, 30-45 and 45-60 cm soil depth.
For determination of soil mineral N concentration (N min , nitrate and ammonium) in three soil depths (0-30, 30-60 and 60-90 cm), five soil cores per plot with a diameter of 2.8 cm were randomly taken in December and March at both sites and mixed to a composite sample.The soil N min was extracted with 0.0125 mol L −1 CaCl 2 in the ratio of 50 g soil to 200 mL solution and analyzed according to VDLUFA A 6.1.4.1 (Association of German Agricultural Analytic and Research Institutes) (VDLUFA 2020).
Wheat biomass was harvested at growth stage BBCH 29/30 (end of tillering/beginning of stem elongation; mid-to end of April).Plants were cut off right above the ground with an electric cutter in 2 × 0.5 m 2 per plot.The whole samples were weighed, and a mixed subsample was dried at 105 • C for 24 h to determine dry matter weight content used to calculate dry matter biomass.Afterwards, the samples were analyzed for total N by flash combustion with the FlashSmart Elemental Analyzer (ThermoFisher Scientific, USA).Plant N uptake was calculated by multiplying total N and dry matter biomass.

Statistical analyses
The statistical data analysis was conducted with R version 4.1.2(R Core Team, Vienna, Austria)).For Harste, the effects of the crop rotational position on PR, BD, PV, AC, FC, PC, N min , wheat biomass and N uptake were analyzed by a linear mixed model ANOVA with year, crop rotational position, and their interaction, as well as replication and the interaction of replication and year as fixed effects and plot as random effect.The effects on AS were analyzed by a linear mixed model ANOVA with crop rotational position as fixed effect and replication as random effect.For Hohenschulen, the effects on BD, PV, AC, FC, PC, N min and wheat biomass were analyzed by a linear mixed model ANOVA with crop rotational position as fixed effect and year as random effect.The effects on AS in Hohenschulen were not statistically analyzed, because at this site AS was analyzed in one year only and, due to the lack of field replicates, the mean values are based on pseudoreplicates.All linear mixed models were calculated with package "lmerTest".The residuals of the models were checked for normal distribution graphically as well as with the Shapiro-Wilk Table 3.Effect of the crop rotational position of winter wheat (W1 = first wheat after oilseed rape, W3 = third wheat after oilseed rape) on bulk density (BD), total pore volume (PV), air capacity (AC), field capacity (FC), pneumatic conductivity (PC) and aggregate stability (AS) in three soil depths in Hohenschulen.test and for homoscedasticity graphically as well as with the Levene's test.If the factor crop rotational position was significant (p < 0.05) for Harste, means were compared by a post-hoc Tukey's test with package "emmeans".

Results
At Harste, there was no significant effect of the factors crop rotational position, year, or their interaction on any soil structure parameter.BD, PV, FC, AC and PC were similar for all crop rotational positions (W1, W2 and WM) in each soil depth (Table 2).AS was not significantly different among the crop rotational positions either but numerically higher in 5-10 and 20-25 cm soil depth for W2 compared to WM and W1, while in 40-45 cm soil depth AS was nearly the same for W1, W2 and WM.Soil water content was similar for W1, W2 and WM and varied over all depths between 0.22 and 0.26 g H 2 O 100 g −1 dry soil, corresponding to between 94% and 98% of FC (data not shown).
Similarly to Harste, at Hohenschulen, no significant effect of the crop rotational position on any soil structure parameter was found.BD and PV were similar for W1 and W3 in 5-10 and 20-25 cm soil depth, with a slightly higher BD for W3 in 40-45 cm soil depth (Table 3).AC tended to be higher for W3 compared to W1 in 5-10 cm soil depth and lower for W3 compared to W1 in 20-25 and 40-45 cm soil depth.FC was similar in all soil depths for W1 and W3.In 5-10 cm soil depth PC was lower for W1 compared to W3, while in 40-45 cm soil depth values were higher for W1.AS decreased for both treatments with depth and was higher for W1 compared to W3 in all soil depths.Soil water content was similar for W1 and W3 and varied over all depths between 0.15 and 0.17 g H 2 O 100 g −1 dry soil, corresponding to between 66% and 102% of FC (data not shown).
No statistically significant effect of crop rotational position or year was found for PR at Harste in all of the analyzed depths.Values increased from 0-15 to 30 -45 cm soil depth and decreased again from 30−45 to 45-60 cm soil depth (Fig. 1).Slight differences between the crop rotational positions were only found in the 30-45 cm layer in the order W1 < W2 < WM.At Hohenschulen, PR increased from 0-15 to 45-60 cm soil depth (Fig. 1).In 0-15 and 15-30 cm soil depth, PR was numerically higher for W1 compared to W3, while in 30-45 and 45-60 cm soil depth PR was slightly higher for W3 compared to W1. Due to the lack of replication at this site, these differences could not be tested for statistical significance.
At both sites, N min was not significantly different between crop rotational positions or years, both in December and March (Table 4).Values were mostly very similar, only in 0-30 cm soil depth in December in Harste N min was slightly higher for WM (71 kg N ha −1 ) and W2 (68 kg N ha −1 ) than for W1 (56 kg N ha −1 ).

Effects of the crop rotational position of wheat on soil structure
In general, the results show that the crop rotational position of WW had no effect on the soil physical properties both in a silty and a sandy loam soil, which was also reported by Pranagal and Wo źniak (2021) when comparing a 30-year-old WW monoculture with a pea-WW-winter barley crop rotation in a Rendzic Phaeozem.In their study, the topsoil total porosity measured in spring was slightly lower under WW monoculture compared to spring wheat monoculture, which might have been caused by soil settlement over winter in the WW monoculture while for spring wheat, the soil was tilled shortly before sowing.Consistently, Sieling et al. (2005) also did not observe significant changes in physical parameters of soil sampled from the first, second and third WW after WOR in a study conducted at the Hohenschulen trial used in our study as well.
AC was generally low in our study, which might have been caused by very high precipitation in April 2021 and 2022 at both field trial locations, leading to short-term variations in BD and consequently AC (Bach and Hofmockel 2016; Linsler ).An AC value of 6 Vol% as found in our study is in the typical range for sandy to silty loam soils (Ad-hoc-Arbeitsgruppe Boden 2005).The overall higher BD at Hohenschulen compared to Harste was likely caused by the higher sand content at Hohenschulen.At Harste, AS was slightly higher for WM than for W1, which is in contrast to Feng et al. (2020), who reported growing wheat in a crop rotation compared to monoculture cropping increased the AS in a loamy soil in 0-5 cm soil depth, while this effect was not detectable in 5-15 cm soil depth.However, since in our study values for W2, also from the crop rotation, were even higher than for W1, the effect of rotational diversity seems to be modified by other factors.Contrastingly, in Hohenschulen, AS values were higher for W1 than for W3, with both rotational positions originating from the same rotation.On a Cambisol with loamy sand texture, Schönhammer and Fischbeck (1987) found a higher AS under WOR than in soil under WW monoculture.In our study, no such positive legacy effect of WOR on subsequent WW was detectable in Harste.Nevertheless, there was a tendency for higher AS in W1 than W3 in Hohenschulen.Possibly, the effect of WOR as pre-crop on AS is larger for sandy than silty soil texture.
In 30-45 cm soil depth, PR was slightly lower for W1 than for W2 and WM (Harste) or W3 (Hohenschulen), which might be due to the strong taproot of preceding WOR causing intensive WW rooting in the upper subsoil layer.Soil after WOR in comparison to after barley or field pea showed a lower PR down to 20 cm soil depth in a previous study (Chan and Heenan 1996).While the first WW after WOR may benefit directly from the taproot of WOR, the second WW after WOR might still benefit indirectly from the positive effects on porosity (Schönhammer and Fischbeck 1987).In our study, although soil total porosity was not significantly different among the crop rotational positions, slightly higher values were found in W1 than in the other treatments in 40-45 cm soil depth.Thus, there might be a small positive effect of preceding WOR compared to WW in the subsoil.Still, none of these differences was significant, thus the overall effect of WW crop rotational position was low.Similarly, Arshad et al. (1998) reported that PR of a silt loam Luvisol under WW was not affected by crop rotation, including fallow, WOR, field pea and WW as pre-crops and a WW monoculture.
Contrary to the missing effect of the crop rotational position on soil physical properties, C input via pre-crop residues varied considerably in the different crop rotational positions regarding total amount and C:N ratio.However, the large contrast between W1/W2 and WM in 2022 might be due to the experimental design as the studied subplots in WM did not receive N fertilizer the year before, resulting in a lower biomass.Otherwise, contrary to the results for 2021, longterm averages for WW and WOR crop residue C input in the trial were found to be similar by Grunwald et al. (2021).Apparently, short-term differences in crop residue input had no effect on soil structure in our study since there was no difference in the pre-crop effect between years.Even though there is little evidence for soil structural differences after winter among different crop rotational positions in our study, soil structure might still have differed before winter.Possibly, such effects were diminished by overwinter rainfall and frost events, and thus sampling before winter might be required in future studies to reveal crop rotational effects on soil structure.
Effects of the crop rotational position on wheat biomass WW biomass at BBCH 29/30 in April was higher for W1 than for W2, W3 or WM on both field sites.Similarly, Sieling et al. (2005;2007) reported a higher early biomass formation for WW grown after WOR compared to WW grown after WW from the crop rotation experiment at Hohenschulen.The higher WW biomass for W1 might have been caused by better root growth conditions after WOR, for example due to a lower PR, leading to an improved root density (Schönhammer and Fischbeck 1987) and the development of a deeper root system (Whalley et al. 2008).However, in our case, the extent of the soil structural effects measured in spring was clearly too low to explain the biomass gains in W1.
Another possible benefit of W1 compared to W2, W3 and WM could be disease suppression in WW resulting from the inclusion of WOR as break crop before W1 (Kirkegaard et al. 2008).As a non-host plant, WOR interrupts pest cycles of different pathogens for WW, for example of fungi (Kirkegaard et al. 2008).In addition, WOR produces allelochemicals suppressing disease organisms; especially root diseases can be controlled via this process (Angus et al. 2015).
Finally, N benefits related to the residuals of WOR as precrop, leading to a higher pre-fertilization N supply of WW grown after WOR than after WW, might play a role in the promotion of WW growth (Kirkegaard et al. 2008).In addition, Sieling et al. (2005) assumed for Hohenschulen that WW following WW might take up lower N because of a smaller root system than WW after WOR.Further they suppose that the early above-and belowground plant growth is of high importance for later biomass formation and thus, investigations to identify the mechanism affecting early growth after different pre-crops are needed.As in March in Harste, WW N uptake was (non-significantly) higher for W1 than for W2 and WM, it may be assumed that N min levels after WOR were higher than after WW, as similarly found by Sieling and Christen (2015).This fits to results from Sieling et al. (2005), who mentioned that the nitrogen supply in autumn must be considered as potential cause for different growth of WW following WW or WOR, as the mineralization potential and soil N min after WOR were higher than after WW.This may explain part of the positive effect of W1 on WW biomass growth by April.
To sum up, we must reject both of our hypotheses that the crop rotational position of WW causes differences in soil structure in April and that the soil structure at this date is related to differences in WW biomass formation.As in our study the WW biomass differences in April cannot be linked to changes in soil structure at this date, further investigations of soil structure and WW biomass formation before April must be considered.Beyond that, future studies are needed to identify the processes related to biomass formation of WW grown after WW.These should include root growth to determine if differences in root development and therefore in water and nutrient uptake are a main cause for WW yield decline.

Conclusion
Crop rotational position of WW did not significantly affect soil structure in April; however, it caused differences in WW biomass by this date with lower values for WW grown after WW than after WOR.Thus, it appears that differences in WW biomass by spring cannot be linked to changes in soil structure at this timepoint.Possibly, stronger differences in soil structure among the crop rotational positions or pre-crops of WW occurred before winter, causing differences in WW growth, and were not detectable anymore after winter.

Fig. 1 .
Fig. 1.Effect of the crop rotational position of winter wheat (W1 = first wheat after oilseed rape, W2 = second wheat after oilseed rape, W3 = third wheat after oilseed rape, WM = wheat monoculture) on soil penetration resistance in four layers in Harste and Hohenschulen.Harste: data from 2021 and 2022 and three replicates each, n = 6, symbols show means with standard deviation.Hohenschulen: data from 2022 and 40 pseudoreplicates, symbols show means.No statistically significant differences between the crop rotational positions were found in all depths (p ≥ 0.05).

Table 1 .
C:N ratio and C input of winter oilseed rape (WOR) and winter wheat (WW) as pre-crop in Harste.
Data from 2021 and 2022, n = 6.AS only from 2022, n = 3. Mean with standard deviation in brackets.No statistically significant differences between the crop rotational positions were found for either parameters in all depths (p ≥ 0.05). Note: Data from 2021 and 2022, n = 2. AS only from 2022.Mean with standard deviation in brackets.No statistically significant differences between the crop rotational positions were found for either parameters in all depths (p ≥ 0.05).
* Mean based on four pseudoreplicates, not statistically analyzed.Note: