Soybean (Glycine max L. Merril) seed costs have steadily increased over the past decades and triggered a growing interest in reducing seeding rates, while improving profitability via seed savings and better yields. However, little is known on the implications of early-season spatial plant-to-plant uniformity on soybean yield under the current reduced seeding rates and modern varieties. Therefore, the objectives of this study were to i) explore traditionally used and new metrics to characterize early-season spatial plant-to-plant uniformity, ii) select and identify the best metrics relating the spatial plant-to-plant uniformity and soybean yield, and iii) evaluate the effect of stand uniformity under varying yielding environments and seeding rates. Soybean trials evaluating different planters (Max Emerge® and Exact Emerge®) and seeding rates (160, 215, 270 and 321 thousand seeds ha-1) were conducted in rainfed and irrigated conditions during 2019 and 2020 in the US and Argentina. Early-season plant-to-plant spacing was recorded in all trials and multiple stand uniformity metrics were calculated. Spacing standard deviation (sd) and the sd of the achieved versus target evenness index (ATEI = observed spacing/theoretical plant spacing) showed the greatest correlation with soybean yield (R2 = 0.26 and 0.32, respectively). However, only ATEI sd exhibited a more consistent relationship with soybean yield across environments. Non-uniform stands, described by ATEI sd, negatively impacted soybean yields under low and medium yield environments (mean of 2.7 and 3.0 Mg ha-1, respectively), while high yielding environments (mean of 4.3 Mg ha-1) were not sensitive to stand uniformity. The effect of unevenness spacing was not affected by the tested seeding rates. Our study provides new insights on the ability of soybeans to compensate yield at low yield environments and suggests uniformity of spacing between plants as a target to maximize yields regardless of the seeding rate.
Article
Early-season spatial plant-to-plant uniformity can affect soybean yields
https://doi.org/10.21203/rs.3.rs-1503199/v1
This work is licensed under a CC BY 4.0 License
You are reading this latest preprint version
spatial stand uniformity
spacing uniformity metrics
yield environments
seeding rates
Soybean (Glycine max L. Merril) seed costs have doubled since 1997 1 mainly due to the increase in genetic yield potential 2, the introduction of herbicide traits 3,4, and more recently due to the use of seed treatments 5. Concurrently, farmers have adopted lower seeding rates 6 strongly supported by the improvement of planting equipment 7 and the increase in yield contribution from branches in modern soybean varieties 8. The property of reducing seeding rates while maintaining yield is largely associated with the growth plasticity of soybean plants 9, i.e., the ability of individual plants to produce branches and compensate yield at the community-scale. However, there is a plant density threshold from below which soybean plants cannot compensate and thus yield at community-scale is penalized, defined as the agronomical optimum seeding rate and plant density (AOSR, and AOPD, respectively) 10. Although AOSR and AOPD are commonly assessed at the community-scale, a non-uniform plant spacing distribution may impose excessive intraspecific competition or lack of growth compensation at the plant-level penalizing yield at the community-scale. Therefore, the distribution and spatial variability of plants within the field may be as crucial under current soybean production systems, characterized by high yield potential and reduced seeding rates 10.
Spatial stand uniformity (plant-to-plant spacing distribution) seems to affect yield in soybeans mainly when the genetic (G) by environment (E) interaction (GxE) hinder the crop ability to compensate growth 11–13. For instance, soybean yield in rainfed conditions decreased between 6-9% when non equidistant patterns were imposed 11–14. However, longer maturity groups (MG) showed capacity to fill in skips and yield more relative to the shorter MG 13 or similar to the control treatment (equidistantly spaced) 11. While, yield of irrigated soybeans at optimal planting dates did not differ under equidistant or non-equidistant spacings, even though plant size variability did increase 15. Similarly, within-row alternated planting dates increased plant size variability, affecting the soybean yield of a rainfed system 11, but no effects were observed in irrigated systems 15,16. These results suggest that contrasting resource availability due to rainfall, planting dates, or nutrients, affects soybean phenotypic plasticity to compensate yield. The initiatives on investigating the association between soybean spatial stand uniformity and yield are still scarce, with a common feature of lacking a comprehensive variable for accessing plant-to-plant spacing distribution. This field is becoming increasingly vogue due to the recent development and improvement in high-precision planter systems. Enhancing our understanding of these systems would unveil the potential of precision planting techniques through their benefits on seeding rate recommendations, and efficiency of soybean production 17.
The plant-level variable commonly used to assess spatial stand uniformity is the spacing between plants (average distance to neighboring plants) and is commonly summarized at the community-scale by calculating their standard deviation (sd) 15 or coefficient of variation (cv) 11. The use of these methods to characterize stand spatial uniformity can be problematic for three reasons: i) the spacing between plants does not account for the different distances (or unevenness) to each neighboring plant, or the closeness to the theoretical plant spacing (of a given seeding rate and row width); ii) opposing plant distance outcomes (i.e., skip and double) can generate similar statistical values while their effect on community-scale yield being contrasting; and iii) values are strongly influenced by a few large spacings and by seeding rates (i.e. lower seeding rates obtain greater spacing sd) 18. Thus, there exists a need to develop and test community-scale variables that i) are more robust to detect different spacing outcomes; ii) provide insights into both spacing and overall distribution of plants, and iii) are based on data collected from on-farm fields that are representative of actual producing conditions.
Therefore, the objectives of this study were to i) explore traditionally used and new metrics to characterize early-season spatial plant-to-plant uniformity, ii) select and identify the best metrics relating the spatial plant-to-plant uniformity and soybean yield, and iii) evaluate the effect of those metrics under varying yielding environments and seeding rate levels.
2.1. Sites description and field operations
A total of six field studies were conducted in two different regions over two years. Four studies (two dryland and two irrigated) were located in the State of Kansas, United States (dryland: 39° 4' 30’’ N, -96° 44' 43" W, irrigated: 39° 4' 25" N, -96° 43' 12" W) during the 2019 and 2020 growing seasons (hereafter referred as USDry19, USIrr19, USDry20, and USIrr20 studies). The remaining two studies (dryland) were located in the province of Entre Rios, Argentina (31° 50’ 49’’ S; 60° 32’ 16’’ W) during the 2018/2019 and 2019/2020 growing seasons (hereafter referred as the Arg19 and Arg20 studies). The mean cumulative seasonal precipitation was 560 and 630 mm for the USDry19 and USDry20, respectively; and 770 and 450 mm for the Arg19 and Arg20, respectively. Total water availability was 712 mm for USIrr19 (152 mm from irrigation) and 757 mm for USIrr20 (127 mm from irrigation). The soils were fluventic hapludolls [silt loam, 40% sand, 13% clay, 47% silt, OM 1.7%, 7.7 pH, 31.1 ppm P(Bray-1)] at the US dryland studies, and pachic Argiudolls [silty clay loam, 10.1% sand, 30.6% clay and 59.3% silt, OM 3.2%, 6.8 pH, 34.7 ppm P (Bray-1)] at the US irrigated studies. At the Argentinian studies soil was a Vertic Argiudoll in 2019 [silty clay loam to clay loam, 3.9% sand, 27.6% clay, 67.9% silt, OM 2.65%, 7.2 pH, 12.5 ppm P (Bray-1)] and an Acuic Argiudoll in 2020 [silt loam to silty-clay-loam, 5.6% sand, 28.6% clay, 65.8% silt, OM 3.33%].
The US dryland and irrigated studies were sown on June 4 in 2019 and May 20 in 2020 (genotypes P40A47X- in 2019 and P39A58X- in 2020, Pioneer, U.S.). In 2019, the dryland study was replanted on June 29 due to poor emergence after first sowing. The studies in Argentina were sown on December 5 in 2018 and November 20 in 2019 (genotype AW5815IPRO in both years, Monsanto). At all six studies, plots were kept free of weeds, pests, and diseases through recommended chemical control.
2.2. Study design
The studies carried out in the US were arranged as a split plot design with three replicates in 2019 and 2020. In 2019, the main plot treatment factor was planter type with two levels [John Deere Max Emerge® planter (ME, 12 rows), and John Deere Exact Emerge Planter (EE, 16 rows)], and the split-plot treatment factor was seeding rate with two levels (160 and 321 k seeds ha-1). In 2020 the main plot treatment factor was also planter type with two levels (ME and EE), and the split-plot treatment factor was seeding rate with four levels (160, 215, 270 and 321 thousand seeds ha-1). Plots were 24 and 32 rows wide when planted with ME and EE, respectively, with 0.76 m row spacing. Plot length was 80 m in the dryland studies and 160 m in the irrigated studies.
The studies from Argentina were arranged as a single replicate of each seeding rate treatment level (100, 230, 360 and 550 thousand seeds ha-1) in both years. Plots were 10 rows wide with 0.52 m row spacing and 350 m in length. Argentinian studies were only used for selection of stand uniformity variables due to the simple experimental design. Stand uniformity variables were first fitted using the data from US studies (details below), and then the best explanatory variables were selected to re-fit the relationships combining both data sets from US and Argentina.
2.3. Data collection and spacing uniformity variables
Two segments of 2-m in length were delimited early in the season inside each plot. At the V5 (US studies) and R1 (Arg studies) soybean growth stage 19, the cumulative distance of the plants within each segment was measured and then used to calculate multiple derived variables. Plant spacing (cm) was calculated as the average between the preceding and succeeding neighboring plants distance. In addition, the distance from a plant to each neighboring plant was classified as shorter or longer than the plant spacing (named nearest and farthest neighbor distance, respectively). Achieved versus Target Evenness Index (ATEI, dimensionless) was calculated as the ratio between the observed plant spacing and the theoretical plant spacing (TPS, cm), where TPS is the expected plant spacing derived from a specific seeding rate and row width (Eq. 1).
Equation 1
The ATEI index was designed to account for the closeness of the observed plant spacing to the TPS. Obtained plant densities that are below the seeding rates will have a mean spacing greater than the TPS, hence, to further sense the meaning of ATEI, the relative density (rd) was calculated as the ratio between plant density (based on the number of plants in the 2 m segment) and seeding rate. To account for the evenness of distance to both neighboring plants, we used the Evenness Index (EI), calculated as the ratio between the distance to the nearest neighbor (cm) and the plant spacing of a given plant. Evenness Index values range from zero to one, where a value of one indicates that the plant is equidistant from both of its adjacent neighboring plants (Eq. 2).
Equation 2
In addition, the distance from a plant to its preceding neighboring plant, and the TPS were used to classify the position of each plant into one of eight classes (Fig. 1). Plants were classified in classes ranging from “double” (preceding plant distance < 17.5% of TPS), through “perfect” (preceding plant distance within 83.5 and 116.5% of TPS), to “greater than double skip” (preceding plant distance greater than 215.5% of TPS).
Plant variables (Eq. 1 and 2, Fig. 1) were further summarized at the segment level. Plant spacing (cm), ATEI, and EI were summarized by calculating their mean (except for plant spacing) and sd. Plant spacing sd and cv (directly proportional to sd) have been used in previous studies 11,15 to estimate spatial variability. Hence, plant spacing sd (cm) was included as the base line variable to compare the performance of the new metrics proposed in this study. Lastly, plant position classes (Fig. 1) were summarized by calculating the mean percentage of plants in a segment classified within each of the eight classes.
At physiological maturity, 10 (US studies) or 20 (Arg studies) consecutive plants within each segment were hand cut, individually identified, bagged and oven-dried at 65ºC until constant weight. Total weight of shoot biomass, and grain weight per plant were recorded. Grain weight was reported on a 13 g kg-1 water content. Yield (Mg ha-1) was calculated at the segment level by summing the grain biomass (g) of all harvested plants and correcting it to the segment area.
2.4 Statistical analysis and software
Segment level data (hereafter called community-scale) from the four US studies (site-years) were combined to model yield and the 13 stand uniformity variables (ATEI mean, ATEI sd, EI mean, EI sd, Spacing sd (cm), Double, Mis 33, Mis 66, Perfect, Short skip, Long skip, Double skip, >Double skip) as a function of seeding rate, planter type and their interaction (fixed effects), and block nested in site-year (random effect) (Table 1). Independent models for each of the 4 US studies were built assessing the effects of planter type, seeding rate and their interaction (fixed effects), and seeding rate nested in planter type, and nested in block (random effects) on the same variables previously mentioned (Supplementary Table 2). The models were run using lmer function from lme4 package. In addition, the US and Arg studies were combined to evaluate the effect of site-year on yield, plant density, and all stand uniformity variables (Supplementary Figure 1) using lm function from package stats (R Core Team, 2021). Means of variables were compared by Fisher’s LSD test (alpha= 0.05).
Community-scale data from the four US studies were combined and used to fit bivariate linear regression models with yield as the response variable and each of the spatial stand uniformity variables as the explanatory variable. Significant models (alpha=0.05) were further assessed by calculating the coefficient of determination (R2) and root mean squared error (RMSE). Models with lower RMSE and higher R2 were selected as those that best captured the effect of non-uniform stands on soybean yield (Fig. 2). Both US and Arg data sets were combined and the linear regressions between the selected variables and yield were re-fitted to assess the relationship on an independent data set. Yield was modelled as a function of stand uniformity variables, country (US and Arg), and their interaction (fixed effects) (Fig. 3). The variable showing a similar relationship across the two data sets (i.e., non-significant interaction between stand uniformity metric and country), was selected to continue the analysis. The bivariate linear regression models were run with function lm from package stats (R Core Team, 2021).
Different yielding conditions and seeding rate levels may modify the effect of spatial stand uniformity on yield. To explore this, Arg and US studies were separated in Low (USDry19 and ArgDry20, mean of 2.7 Mg ha-1), Medium (USIrr19, USDry20 and ArgDry19, mean of 3.0 Mg ha-1), and High (USIrr20, mean of 4.3 Mg ha-1) yield environments based on the effect of site-year on yield (Supplementary Figure 1). While, seeding rates were separated in Low (<200 thousand seeds ha-1), Medium (between 200 and 300 thousand seeds ha-1) and High (>300 thousand seeds ha-1). This classification was used to model yield as a function of the stand uniformity metric, yield environment or seeding rate levels, and their interaction (i.e., uniformity metric*yield environment, and uniformity metric*seeding rate level interaction) (fixed effects). If the interaction was significant, Akaike information criteria (AIC) was used to judge if the higher complexity added to the model was improving its quality relative to the simpler model without yield or seeding rate levels.
3.1. Treatment effects on yield and spatial stand uniformity metrics
The effect of the treatments analyzed in this study (planter type and seeding rate) were significant on yield and in 8 out of 13 stand uniformity metrics (alpha = 0.05) (Table 1). From the four seeding rates evaluated, attention will be given only to the lowest (161 thousand seeds ha-1) and highest (321 thousand seeds ha-1) seeding rates since only those levels were present at all four US studies. Yield was not affected by planter type, while a significant effect of seeding rate was observed, yield was greater at the lower seeding rate (3.3 Mg ha-1) compared to the highest seeding rate (3.0 Mg ha-1). In terms of the stand uniformity variables EE planter improved, in general, the uniformity between plants compared to ME planter. EE planter increased the number of perfect spaced plants (24%), decreased the number of double plants (1.3%), and plants misplaced by 33% (4.3%) compared with ME planter (12, 7.8, and 13% respectively). In addition, equidistance to both neighboring plants was improved with EE as it obtained higher evenness index (EI) mean and lower evenness index sd (0.7 and 0.2, respectively) compared to ME (0.5 and 0.3, respectively). In terms of the variables affected by seeding rate, the lower seeding rate (161 thousand seeds ha-1) obtained, in general, a superior stand spacing uniformity compared to the higher seeding rate (321 thousand seeds ha-1). For instance, the low seeding rate showed more perfect spaced plants (23%), less double plants (4%), decreased the ATEI sd (0.6) and evenness index sd (0.2), and increased the evenness index mean (0.6) relative to the high seeding rate (14, and 7.0% of perfect spaced and double plants, respectively; 1.1, and 0.3 for ATEI sd and evenness index sd, respectively; and 0.5 for evenness index mean). However, long skipped plants were more in the low (9%) relative to the high (6%) seeding rate. Thus, the low relative to the high seeding rate depicted more uniform spacing between plants (Table 1).
Treatment effects (planter type and seeding rate) on yield and stand uniformity variables for each of the four US studies are presented in Supplementary Table 1. Yield was only affected by seeding rate (alpha = 0.016) on the US dryland site in 2019, where yield was found to be higher with 161 thousand seeds ha-1 (2.9 Mg ha-1), rather than 321 thousand seeds ha-1 (2.2 Mg ha-1). Lastly, the effect of site-year was significant on yield, plant density, and in 10 out of 13 stand uniformity variables: Spacing sd, ATEI mean and sd, evenness index mean and sd, double plants, plants misplaced by 33%, perfect spaced plants, short skips, and greater than double skip plants (Supplementary Figure 1). Average yield was greatest in the USIrr20 (4.3 Mg ha-1), medium in Arg19, USIrr19 and USDry20 (2.9 to 3.2 Mg ha-1), and lowest in the Arg20 and USDry19 (2.8 and 2.6 Mg ha-1, respectively).
Table 1. Effect of planter type, seeding rate, and their interaction on yield and stand uniformity variables for all site-years. References: Spacing between plants standard deviation (Spacing sd), percentage of perfect spaced plants (Perfect), percentage of plants misplaced by 33% (Mis 33), percentage of plants misplaced by 66% (Mis 66), percentage of double plants (Double), percentage of short skips plants (Short Skip), percentage of long skip plants (Long Skips), percentage of double skips plants (Double Skip), percentage of greater than double skip plants (> Double Skip), achieved versus targeted evenness index mean and standard deviation (ATEI, and ATEI sd), and evenness index mean and standard deviation (EI and EI sd).
|
Yield |
Spacing |
Perfect |
Mis 33 |
Mis 66 |
Double |
Short |
Long |
Double Skip (%) |
>Double |
ATEI |
ATEI sd |
EI |
EI sd |
||
|
Planter1 |
EE |
3.19 |
4.0 |
23.8 a2 |
4.3 b |
15.5 |
1.3 b |
17.0 |
11.4 |
8.1 |
18.9 |
1.6 |
0.7 |
0.7 a |
0.2 b |
|
(P) |
ME |
3.2 |
4.5 |
11.8 b |
13.1 a |
12.2 |
7.8 a |
14.1 |
9.1 |
8.5 |
23.7 |
1.5 |
0.8 |
0.5 b |
0.3 a |
|
p-value |
0.948 |
0.371 |
<0.001 |
<0.001 |
0.070 |
<0.001 |
0.126 |
0.165 |
0.812 |
0.057 |
0.489 |
0.383 |
<0.001 |
<0.001 |
|
|
Seeding |
161 |
3.34 a |
5.2 |
22.7 a |
6.4 |
12.7 |
3.9 ab |
15.7 |
9.0 ab |
6.8 |
22.3 |
1.5 |
0.6 b |
0.6 a |
0.2 b |
|
rate |
215 |
3.36 ab |
4.1 |
21.1 a |
9.6 |
10.6 |
3.9 b |
17.0 |
13.1 a |
7.9 |
17.7 |
1.5 |
0.7 b |
0.6 a |
0.2 b |
|
(k seeds |
269 |
3.06 ab |
3.3 |
13.2 b |
10.0 |
16.5 |
3.1 b |
13.7 |
13.1 a |
10.0 |
21.5 |
1.5 |
0.7 b |
0.6 ab |
0.2 b |
|
ha-1) (SR) |
321 |
3.02 b |
4.4 |
14.1 b |
8.9 |
15.6 |
7.0 a |
16.0 |
5.8 b |
8.5 |
23.8 |
1.7 |
1.1 a |
0.5b |
0.3 a |
|
p-value |
0.04 |
0.058 |
<0.001 |
0.168 |
0.108 |
0.023 |
0.766 |
0.003 |
0.445 |
0.493 |
0.051 |
<0.001 |
<0.001 |
0.01 |
|
|
P x SR |
p-value |
0.831 |
0.978 |
0.011 |
0.085 |
0.414 |
0.249 |
0.032 |
0.122 |
0.787 |
0.885 |
0.760 |
0.977 |
0.229 |
0.020 |
1Planter type: Max Emerge® (ME) and Exact Emerge® (EE)
2Letter and ANOVA threshold. Different letters mean significant differences among treatments.
3.2. Spatial stand uniformity effects on soybean yield
Soybean yield from the four US studies (n = 190) was significantly correlated with seven out of the 13 stand uniformity metrics presented (alpha = 0.05). In descending R2 order, yield was negatively correlated with ATEI sd (R2 = 0.32), Spacing sd (cm, R2 = 0.27), ATEI mean (R2 = 0.16), > Double skipped plants (%, R2 = 0.04), and Evenness Index (EI, R2 = 0.04) sd. While soybean yield was positively correlated with EI mean (R2 = 0.03) and Perfect spaced plants (R2 = 0.02) (Fig. 2). Spacing sd and ATEI sd were the variables selected to continue with the analysis due to their greater R2 and lower RMSE.
Spacing sd and ATEI sd were correlated with yield of US and Arg studies summarized at the community-scale (n = 262). Yield from Arg was also significantly correlated with both spacing sd and ATEI sd. Nevertheless, while the interaction between Spacing sd and Country was significant (alpha= 0.05), the interaction between ATEI sd and Country was not. Hence, ATEI sd showed a stable relationship across two independent data sets (US and Arg) while spacing sd did not. In this sense, for each unit of increase in spacing sd US showed a greater reduction on yield (110 kg ha-1) compared to Arg (30 kg ha-1), even though the observed range of spacing sd was smaller in US (0.5 to 12.8 cm) than Arg (1.2 to 21.7 cm). The latter results shows that a greater variability in the spacing between plants did not translate in a stronger yield reduction. Whereas ATEI sd and soybean yield relationship had a different intercept (due to yield levels US 3.8, and Arg 3.7 Mg ha-1) but shared the same slope between US and Arg (960 kg ha-1 yield decrease for every unit increase on ATEI sd) (Fig. 3). Moreover, ATEI sd ranges were similar between US (0.1 to 2.9). and Arg (0.3 to 2.4). These results showed that the effect of spacing sd on soybean yield was not stable across independent data sets, while ATEI sd had a consistent effect on yield from different environments. Hence, ATEI sd was selected as the variable that best explained spatial stand uniformity effects on soybean yield.
In order to assess the effect of the availability of resources in each site and their impacts on spatial stand uniformity and yield, a yield environment classification was established based on the letter separations of pairwise comparisons of the mean yield by site-year (Supplementary Fig. 1). Site-years were divided into Low (USDry19 and Arg20, 2.7 Mg ha-1), Medium (USIrr19, Arg19, and USDry20, 3.0 Mg ha-1), and High (4.3 Mg ha-1) yield environments. Soybean yield was significantly affected by the interaction between yield environment and ATEI sd (alpha= 0.05), hence, at least one of the three yield environments had a different slope in the relationship between ATEI sd and yield (Fig. 4). Low yield environment had the lowest RMSE (0.64 Mg ha-1) and the greatest R2 (0.32), with a rate of 760 kg ha-1 of yield reduction for every unit increase in ATEI sd. Medium yield environment had the greatest RMSE (0.71 Mg ha-1) and a medium R2 (0.26) with a reduction of 820 kg ha-1 of yield for every unit increase on ATEI sd. Lastly, high yield environment had a medium RMSE (0.68 Mg ha-1) and the lowest R2 (0.01) together with the lowest reduction of yield (140 kg ha-1) for every unit increase in ATEI sd. Moreover, the observed range of ATEI sd was greater in low (0.3 to 2.7) and medium (0.1 to 2.9) relative to high (0.1 to 1.2) yield environments. The relative density (plant density/seeding rate) is visualized in the figure with the point sizes, from which can be noticed that increasing values of ATEI sd denotes a decreasing relative density, attaining minimum values of 0.29, 0.28, and 0.38 in low, medium, and high yield environments, respectively. Even though lower values of relative density were observed in low and medium yield environments, decreases in relative density were accompanied by yield reductions in low and medium, but not in high yield environments. The AIC value of the model including the effect of yield environment was 100 units less than that one accounting for only the effect of ATEI sd, hence, the higher complexity added by a second variable (yield environment) still improved the model. Lastly, the interaction between seeding rate levels and ATEI sd was not significant (alpha = 0.05), therefore, yield decreased under non-uniform stands at the same rate regardless the seeding rate (Supplementary Table 3).
This study provides new insights on the relevance of soybean plant-to-plant spacing uniformity, highlighting the potential for negative impacts on yield mainly under low- and medium- yielding environments. Soybean breeding efforts shifted the overall contribution of branches to yield, reducing the overall dependency to high seeding rates over time 8,9. Still, low yielding environments limited the availability of resources for the soybean plants as previous reports demonstrated a need for increasing seeding rates to compensate for the lack of branching 10,20. In this sense, our results showed that both in high and low yield environments soybean yield did not increase beyond seeding rate of 121 thousand seeds ha-1. Instead, yield variability from low and medium yield environments was explained by within row spacing uniformity, i.e., non-uniform stands decreased yield regardless of the seeding rates. Hence, our study highlights that in low and medium yield environments i) yield optima occurred, either in high and low seeding rates, when plants were more uniformly distributed, and ii) branching plasticity (herein defined as the ability of the plant to produce branches) at low seeding rates can maximize yield provided that a uniform spacing between plants, and an adequate plant establishment (achieved plant density/seeding rates) are achieved 10,20.
Spatial stand uniformity effects on yield can be understood from the onset of competition at the plant level, ultimately shaping yield at the community-scale. Soybean seedlings unravel shade avoidance responses when they sense reductions in the red:far red ratio of the light spectrum (i.e. shading), causing reductions in branch development, overall growth and yield per plant 21–23. Similarly, non-uniform stands will array a continuum of shade avoidance responses with contrasting competitive abilities and yields at the plant level. At a community-level, the yield plasticity associated with branching development 24,25 is sensitive to restrictive environments such as late planting dates 26 or drought stress 25, which reduced yield due to lack of branch growth. In summary, early interference among plants can determine the onset of a differential competitive ability between individuals 27, hence, a plausible effect of plant hierarchies can contribute to and explain the higher yield of the uniform versus non-uniform stands at low and medium yield environments. Uniform spacing between plants allot a similar use of space and resources to each individual plant matching their competitive abilities and decreasing the yield gap across plant hierarchies and, finally, optimizing yield at community-scale. In the other hand, non-uniform stands will array large yield in plants with greater space (i.e., non-shaded) and negligible yield in crowded plants (i.e., shaded 22) without further ability to compensate (branching) at the community-scale.
The most promising metric of spatial stand uniformity, ATEI sd, explained the highest variability of soybean yield among the tested indexes and depicted similar relationships on independent data sets (Figure 3). Furthermore, indexes of stand uniformity presented (Fig. 1, Eq. 1, Eq. 2) were able to detect differences in plant distribution across planter type and seeding rates, whereas the most common uniformity index used in previous literature, spacing sd, did not (Table 1). A previous study has shown that 77% of spacing sd variability is explained by clusters (3 or more plants together), double plants, short and long gaps, with the latter making the greatest contribution to its variability28. Meanwhile, when the spacing sd is related to the TPS (forming the ATEI sd) the confounding effects of spacing sd can be partially overcome, minimizing the influence of seeding rates, long gaps, and distinguishing between doubles and long skips 18,28. High spatial stand uniformity of seedlings is dependent not only on a high uniform sowing pattern, but also on high emergence ratios 29. In this sense, ATEI sd is driven by both spacing uniformity and plant stand achievement (plant density / seeding rate) (Figure 4), meaning that increases in ATEI sd depicts non-uniform spacing as well as plant loss. Therefore, ATEI sd arises as a better spatial stand uniformity predictor of soybean yield and can provide useful information in future research exploring within row stand uniformity.
The findings of this study introduce a first attempt on anticipating that spacing uniformity affects soybean yield when the G x E interaction reproduce phenotypes that are not able to express the plant’s potential growth, penalizing yield at the community-scale. In agreement, a review from Lu et al. (2020), concluded that spatial stand uniformity by precision sowing is an effective way to increase yields of soybean at all seeding rates. Moreover, lowering seeding rates along with improvements in planting quality that delivers seed at a uniform spacing and depth (ultimately optimizing plant emergence and establishment), are becoming core elements of modern soybean farming systems. Farmers should adopt these considerations as an insurance tool against yield losses in poor yielding conditions. Future studies exploring agronomical optimum seeding rate (AOSR) at low yielding environments should consider spacing uniformity and relative density (plant density/seeding rate) 10,20 as factors that may confound the yield response to changes in seeding rates. Lastly, study of plant hierarchies 27 has not been formally explored in soybean crop and could help to clarify the influence of plant types determined early in the season on plant growth and yield formation under real field conditions.
Early-season spatial plant-to-plant uniformity maximizes soybean yield mainly under low- and medium- yield environments. The negative impact of non-uniform stands on yield suggests that plants possess a restriction to compensate yield under more resource-limited yield environments. These results were consistent at all tested seeding rates; hence, soybean yield was maximized with seeding rates as low as 121 thousand seeds ha-1 provided that a uniform spacing between plants was obtained. Spatial stand uniformity under field conditions is a consequence of i) uniformity in the sowing pattern, and ii) emergence and early-season establishment of plants. To capture both effects, relating the observed spacing with the TPS (obtaining ATEI) removed the confounding effects of spacing sd. Hence, ATEI sd arises as a satisfactory variable for exploring early-season spatial plant-to-plant uniformity in soybean crop.
Acknowledgments
The researchers thank all students, technicians, and others who helped with the individual studies summarized in this soybean database. This work was supported by John Deere and USDA-CARE award number 123, the Kansas Soybeans Commission, and INTA EEA Paraná, Proyecto Estructural Desarrollo y aplicación de tecnologías de mecanización, precisión y digitalización de la Agricultura (I177). Contribution no. 22-789 from the Kansas Agricultural Experiment Station.
Ethics declarations
Experimental research and field studies on plants including the collection of plant material, complied with relevant institutional, national, and international guidelines and legislation.
Additional information
The authors declare they have not any competing financial interests or personal relationships that could have influenced the research and this scientific communication.
Data availability statement
The data that supports the findings of this study are available from the authors, upon reasonable request.
Credit authorship contribution statement
V.M.P., L.M.B., A.F.B.R., methodology, Formal analysis, Visualization, Writing - original draft. R.J.M.M., Data curation, Project administration, Funding acquisition, Writing – review & editing. N.E.M., S.C.A., P.V.V. P., Y.W., E.B., Writing - review & editing. A.S., Project administration, Funding acquisition, Writing - review & editing. I.A.C., Conceptualization, Methodology, Writing - original draft, Supervision, Project administration, Funding acquisition.
1. USDA - ERS. Commodity Costs and Returns. https://www.ers.usda.gov/data-products/commodity-costs-and-returns/ (2021).
2. Rincker, K. et al. Genetic Improvement of U.S. Soybean in Maturity Groups II, III, and IV. Crop Science 54, 1419–1432 (2014).
3. Fernandez-Cornejo, J., Wechsler, S., Livingston, M. & Mitchell, L. Genetically Engineered Crops in the United States. SSRN Journal (2014) doi:10.2139/ssrn.2503388.
4. Shi, G. & Chavas, J.-P. Pricing of Herbicide-Tolerant Soybean Seeds: A Market-Structure Approach. 8 (2009).
5. USDA - ERS. Operating costs for soybean production vary by region of the United States. https://www.ers.usda.gov/data-products/chart-gallery/gallery/chart-detail/?chartId=100151 (2021).
6. David A. Hennessy, Yuyuan Che and Hongli Feng. Rising seed costs & a tale of two very different seeding rates. https://www.michiganfarmnews.com/rising-seed-costs-a-tale-of-two-very-different-seeding-rates (2019).
7. Mark Jeschke and Nanticha Lutt. Row Width in Soybean Production. https://www.pioneer.com/us/agronomy/row_width_soybean_prod.html#CurrentRowSpacingTrends_4%20(accessed%20January%2010,%202022). (2016).
8. Suhre, J. J. et al. Soybean Yield Partitioning Changes Revealed by Genetic Gain and Seeding Rate Interactions. Agron.j. 106, 1631–1642 (2014).
9. Carpenter, A. C. & Board, J. E. Growth Dynamic Factors Controlling Soybean Yield Stability across Plant Populations. Crop Sci. 37, 1520–1526 (1997).
10. Carciochi, W. D. et al. Soybean Seed Yield Response to Plant Density by Yield Environment in North America. Agron.j. 111, 1923–1932 (2019).
11. Masino, A., Rugeroni, P., Borrás, L. & Rotundo, J. L. Spatial and temporal plant-to-plant variability effects on soybean yield. European Journal of Agronomy 98, 14–24 (2018).
12. Moore, S. H. Uniformity of Plant Spacing Effect on Soybean Population Parameters. Crop Sci. 31, 1049–1051 (1991).
13. Stivers, R. K. & Swearingin, M. L. Soybean Yield Compensation with Different Populations and Missing Plant Patterns 1. Agron.j. 72, 98–102 (1980).
14. Xu, C. et al. High Density and Uniform Plant Distribution Improve Soybean Yield by Regulating Population Uniformity and Canopy Light Interception. Agronomy 11, 1880 (2021).
15. Andrade, F. H. & Abbate, P. E. Response of Maize and Soybean to Variability in Stand Uniformity. Agron. J. 97, 1263–1269 (2005).
16. Egli, D. B. RELATIONSHIP OF UNIFORMITY OF SOYBEAN SEEDLING EMERGENCE TO YIELD. Journal of Seed Technology 17, 22–28 (1993).
17. Lee, C. D., Egli, D. B. & TeKrony, D. M. Soybean Response to Plant Population at Early and Late Planting Dates in the Mid-South. Agron. J. 100, 971–976 (2008).
18. S. D. Kachman & J. A. Smith. Alternative Measures of Accuracy in Plant Spacing for Planters Using Single Seed Metering. Transactions of the ASAE 38, 379–387 (1995).
19. Fehr, W. R. & Caviness, C. E. Stages of soybean development. 13 (1977).
20. Gaspar, A. P. et al. Defining optimal soybean seeding rates and associated risk across North America. Agron.j. 112, 2103–2114 (2020).
21. Casal, J. J., Sanchez, R. A. & Gibson, D. The significance of changes in the red/far-red ratio, associated with either neighbour plants or twilight, for tillering in Lolium multiflorum Lam. New Phytol 116, 565–572 (1990).
22. Green-Tracewicz, E., Page, E. R. & Swanton, C. J. Shade Avoidance in Soybean Reduces Branching and Increases Plant-to-Plant Variability in Biomass and Yield Per Plant. Weed sci. 59, 43–49 (2011).
23. Yang, F. et al. Growth of soybean seedlings in relay strip intercropping systems in relation to light quantity and red:far-red ratio. Field Crops Research 155, 245–253 (2014).
24. Board, J. E. & Settimi, J. R. Photoperiod Effect before and after Flowering on Branch Development in Determinate Soybean 1. Agron.j. 78, 995–1002 (1986).
25. Frederick, J. R., Camp, C. R. & Bauer, P. J. Drought‐Stress Effects on Branch and Mainstem Seed Yield and Yield Components of Determinate Soybean. Crop Sci. 41, 759–763 (2001).
26. Settimi, J. R. & Board, J. E. Photoperiod and Planting Date Effects on the Spatial Distribution of Branch Development in Soybean. Crop Sci. 28, 259–263 (1988).
27. Maddonni, G. A. & Otegui, M. E. Intra-specific competition in maize: early establishment of hierarchies among plants affects final kernel set. Field Crops Research 85, 1–13 (2004).
28. Liu, W., Tollenaar, M., Stewart, G. & Deen, W. Within‐Row Plant Spacing Variability Does Not Affect Corn Yield. AGRONOMY JOURNAL 96, 6 (2004).
29. Lu, P., Jiang, B. & Weiner, J. Crop spatial uniformity, yield and weed suppression. in Advances in Agronomy vol. 161 117–178 (Elsevier, 2020).
No competing interests reported.