Delta yield–based optimal nitrogen rate estimates for corn are often economically sound

Abstract

Corn producers often overapply nitrogen (N) fertilizer to minimize risk of yield loss because of uncertainty regarding actual corn N requirements. On-farm N trials, which are simple to deploy and interpret, may enable producers to better understand their corn N fertilizer requirements. We analyzed a database of corn yield response to N trials conducted in Ontario, Canada to determine if delta yield (dY) N trials can reliably estimate economic optimum N rates (EONRs) and to assess the financial liability of dY-EONR estimates. Delta yield is calculated as the yield difference between nonlimiting and very low (starter only) N rates. The dY-EONR estimation relationship is derived from a rectangular hyperbolic relationship between agronomic efficiency and dY. The derived dY-EONR relationship has a rapid initial estimated EONR rate of increase that diminishes with increasing dY and that approaches a constant increase rate of 16.1 kg N ha -1 . At a N-corn price ratio of 7, the dY-EONR estimation model has RMSE = 29.1 kg N ha -1 (R 2 = .64). Within range of recent corn and N prices, 64-80% of dY-EONR estimates in combined calibration and validation data (n = 746) had return losses less than $25 ha -1 relative to the actual EONR.Return losses exceeding $50 ha -1 occurred for 8-18% of the dY-EONR estimates, most of which (83-98%) were due to overestimation. Delta yield trials provide an easily implemented method for corn producers to conduct on-farm yield response trials repeatedly over years in order to obtain a better idea of their N requirements. INTRODUCTIONNitrogen (N) fertilizer contributes significantly to corn (Zea mays L.) production costs, but the actual requirement varies substantially among years and fields and spatially within fields (Scharf et al., 2005;Shanahan, Kitchen et al., 2008). Ransom et al. (2020) reported that corn producer-selected Abbreviations: AE, agronomic efficiency; dEY, grain yield increase associated with applying EONR; dY, delta yield or grain yield increase associated with applying nonlimiting N rates; EONR, economic optimum nitrogen rate; OMAFRA, Ontario Ministry of Agriculture, Food and Rural Affairs; PR, ratio of N cost and corn price.

Drip Fertigation with Relatively Low Water and N Input Achieved Higher Grain Yield of Maize by Improving Pre- and Post-Silking Dry Matter Accumulation

“…Finding ways of simultaneously increasing grain yield and nitrogen use efficiency as well as reduce environmental impact has been a major challenge in maize production in the NCP. Irrigation practice and N management, and their interaction, are important parameters influencing maize productivity [8,[43][44][45][46][47]. Previous studies have shown that drip fertigation, which integrates drip irrigation and N management, can improve maize grain yields and nitrogen use efficiency concurrently [48].…”

Comparing Random Forest to Bayesian Networks as nitrogen management decision support systems

“…Using the delta-yield approach (Janovicek et al, 2021), the observed EONR values varied from 105 to 199 kg N ha −1 (Table 1). Delta yield, meaning the difference in grain yield between a non-limiting and very low (<34 kg N ha −1 ), can be used to predict EONR accurately by taking advantage of the rectangular hyperbolic relationship between delta yield and agronomic efficiency (Janovicek et al, 2021;Lory & Scharf, 2003). Briefly, we used the following equation to predict EONR using delta yield:…”

“…We developed a RF model to estimate delta yield using variables which were used for developing the BN model. The RF model first predicted delta yield, which was then used to predict sidedress N rate using delta-yield equations (Janovicek et al, 2021). RF uses numerous non-parametric and classification decision trees following a process called "bootstrap aggregation" or "bagging."…”

“…The ability of the recommendation models (BN, RF, provincial) to provide reasonably accurate N recommendations was based on the frequency of obtaining recommended N rates that produced net returns within CAN$25 ha −1 of maximum profit or net return at EONR. Quadratic-plateau response equations and their associated EONR were derived from the size of delta yield as outlined by Janovicek et al (2021). The delta-yield derived EONR and yield response equations were used to estimate maximum net returns and the N rates that produce net returns within CAN$25 ha −1 of maximum net returns.…”

“…Therefore, we classified soils drainage classes into four categories namely: very‐poor, poor, imperfect (imperfect and moderately well), and well drained (very‐rapid, rapid, and well) (Table 2). Using the delta‐yield approach (Janovicek et al., 2021), the observed EONR values varied from 105 to 199 kg N ha −1 (Table 1). Delta yield, meaning the difference in grain yield between a non‐limiting and very low (<34 kg N ha −1 ), can be used to predict EONR accurately by taking advantage of the rectangular hyperbolic relationship between delta yield and agronomic efficiency (Janovicek et al., 2021; Lory & Scharf, 2003).…”