Soybean response to nitrogen application across the United States: A synthesis-analysis

Abstract

A B S T R A C TThe effects of supplemental nitrogen (N) on soybean [Glycine max (L.) Merr.] seed yield have been the focus of much research over the past four decades. However, most experiments were region-specific and focused on the effect of a single N-related management choice, thus resulting in a limited inference space. Here, we composited data from individual experiments conducted across the US that examined the effect of N fertilization on soybean yield. The combined database included 207 environments (experiment × year combinations) for a total of 5991 N-treated soybean yields. We used hierarchical modeling and conditional inference tree analysis on the combined dataset to establish the relationship and contribution of several N management choices on soybean yield. The N treatment variables were: N-application (single or split), N-method (soil incorporated, foliar, etc.), Ntiming (pre-plant, at a reproductive stage, etc.), and N-rate (from a 0 N control to as much as 560 kg ha). Of the total yield variability, 68% was associated with the effect of environment, whereas only a small fraction of that variability (< 1%) was attributable to each N variable. Averaged over all experiments, a single N application and the split N application were 60 and 110 kg ha −1 greater yielding than the zero N control treatment, respectively. A split N application with more than one method (e.g., soil incorporated and foliar) resulted in 120 kg ha −1 greater yield than zero N plots. Split N application between planting and reproductive stages (Rn) resulted in greater yield than zero N and single application during a Rn; however, the effect was not significantly different than N application at other growth stages. Increasing the N rate increased the environment average soybean yield; however, 93% of the environment-specific N-rate responses were not significant which suggested a minimal effect of N across the examined region. A large yield variability was observed among environments E-mail address: mourtzinis@wisc.edu (S. Mourtzinis).Abbreviations: BNF, biological nitrogen fixation; C, check (no nitrogen was applied); MM, major management practices; N, nitrogen; N-rate, nitrogen rate; N-application, number of nitrogen applications; N-method, method of nitrogen application; N-timing, timing of nitrogen application (growth stage/s); P, all nitrogen was applied at planting only; PR, split nitrogen application at planting and reproductive growth stages; pP, all nitrogen was applied at pre-planting only; Rn, reproductive growth stage; R, all nitrogen was applied at a reproductive growth stage only; RR, split nitrogen application at two reproductive growth stages; V, all nitrogen was applied at a vegetative growth stage only; Vn, vegetative growth stage MARKwithin the same N rates, which was attributed to growing environment differences (e.g., in-season weather conditions, soil type etc.) and non-N related management (e.g., irrigation). Conditional inference tree analysis identified N-timing and N-rate to be conditional to irriga...

Agronomic Practices for Reducing Wheat Yield Gaps: A Quantitative Appraisal of Progressive Producers

“…First, we fitted a threelevel unconditional (no predictors) hierarchical nested linear model to assess the random effects associated with field (Level 1) nested within region (Level 2) nested within year (Level 3) on the intercept (i.e., wheat yield). Similar approaches have been previously used in agricultural sciences (Long et al, 2017;Mourtzinis et al, 2018a). Similar approaches have been previously used in agricultural sciences (Long et al, 2017;Mourtzinis et al, 2018a).…”

“…This method is used to evaluate unbalanced data, and its use has recently increased in agricultural sciences (e.g., Tittonell et al, 2008a;Mourtzinis et al, 2018aMourtzinis et al, , 2018b due to properties such as the ability to explore interactions, lack of statistical distribution assumptions, the ability to handle both continuous and categorical variables, and the ability to handle outliers, multicollinearity, and heteroscedasticity (Tittonell et al, 2008a). However, one weakness of this approach is that the analysis of interactions between management effects was limited to hypothesis testing, which can be subjective in nature and depend on the researcher's understanding of the crop's response to input.…”

“…At this step, we used the covariance parameter estimates to compute the intraclass correlation coefficient (ICC) and quantify how much of the total variability in grain yield was accounted for by year and region nested within year, using the ICC of the residuals as an indication of the field-level effect (Bell et al, 2013). Similar approaches have been previously used in agricultural sciences (Long et al, 2017;Mourtzinis et al, 2018a). Next, we gradually estimated more complex models checking for model improvement after each step using the Bayesian information criterion (BIC).…”

“…Thus, we used conditional inference trees to better explore Level 1 factor interactions. This method is used to evaluate unbalanced data, and its use has recently increased in agricultural sciences (e.g., Tittonell et al, 2008a;Mourtzinis et al, 2018aMourtzinis et al, , 2018b due to properties such as the ability to explore interactions, lack of statistical distribution assumptions, the ability to handle both continuous and categorical variables, and the ability to handle outliers, multicollinearity, and heteroscedasticity (Tittonell et al, 2008a). Conditional inference trees consider data distribution to avoid overfitting issues and selection bias towards covariates with many splits or missing values (Hothorn et al, 2006b).…”

Does gypsum increase crop grain yield on no‐tilled acid soils? A meta‐analysis

“…The main drivers of crop response to gypsum were identified by using the conditional inference regression tree procedure in the software Jmp 13 (SAS Institute, 2016). This method is widely recommended for meta-analyses when some level of the variables considered was not used in a random manner in the same trial (Mourtzinis et al, 2018). In addition, the data to be processed need not exhibit a normal distribution, and the results are scarcely influenced by the presence of outliers or multicollinearity (Tittonel Shepherd, Vanlauwe, & Giller, 2008).…”