Crop growth models such as DSSAT-CERES-Maize have proven to be useful for analysing plant growth and yield within homogenous land units. The paper presents results of newly developed model-based site-specific Soil Profile Optimisation (SPO) tools in combination with an updated version of an already published Nitrogen Prescription Model (NPM). Site-specific soil profiles were generated through an inverse modelling approach based on measured site-specific yield (point-based) and tops weight (above-ground biomass time-series) and evaluated. Site-specific soil profiles generated based only on measured yield variability were able to explain 72% (R² 0.72) of yield variability (dependent variable) based on selected soil profile input parameters (independent variable). Site-specific soil profiles generated based on measured yield and tops variability simultaneously (multiple target variable) explained 68% of yield variability (R² 0.68). The NPM uses the SPO generated site-specific soil profiles for economic evaluation of site-specific N application rates. NPM simulated N application rates, aiming at the maximisation of marginal net return (MNR) were 25% lower compared to the uniform N application rates with an assumed grain and N price of 0.17 and 0.3 Euro kg⁻¹ respectively, under rainfed conditions over three years based on soil profiles generated via an inverse modelling approach only from measured yield variability (one target variable). N application rates were 28% lower when based on soil profiles generated from simultaneously included grain and tops variability in the inverse modelling approach. The results highlight the importance of site-specific fertilizer management when maximising MNR.

DSSAT-CERES-Maize 等作物生长模型已被证明可用于分析同质土地单位内的植物生长和产量。该论文介绍了新开发的基于模型的特定地点土壤剖面优化（SPO）工具与已发布的氮处方模型（NPM）的更新版本相结合的结果。基于测量的特定地点产量（基于点）和顶部重量（地上生物量时间序列），通过逆向建模方法生成特定地点的土壤剖面并进行评估。仅根据测量的产量变异性生成的特定地点土壤剖面能够解释基于选定的土壤剖面输入参数（自变量）的 72% (R ² 0.72) 产量变异性（因变量）。根据同时测量的产量和顶部变异性（多个目标变量）生成的特定地点土壤剖面解释了 68% 的产量变异性 (R ² 0.68)。 NPM 使用 SPO 生成的特定地点土壤剖面来对特定地点施氮量进行经济评估。在超过 3 年的雨养条件下，假设谷物和氮肥价格分别为 0.17 和 0.3 欧元 kg ^-1，NPM 模拟的氮肥施用量以边际净收益 (MNR) 最大化为目标，比统一的氮肥施用量低 25%。年基于仅通过测量的产量变异性（一个目标变量）通过逆向建模方法生成的土壤剖面。根据反演建模方法中同时包含谷物和顶部变异性生成的土壤剖面，施氮量降低了 28%。结果凸显了特定地点肥料管理在最大化 MNR 时的重要性。