Data from: Maize-nutrient response information applied across Sub-Saharan Africa

The profit potential for a given investment in fertilizer use can be estimated using representative crop nutrient response functions. Where response data is scarce, determination of representative response functions can be strengthened by using results from homologous crop growing conditions. Maize (Zea mays L.) nutrient response functions were selected from the Optimization of Fertilizer Recommendations in Africa (OFRA) database of 5500 georeferenced response functions determined from field research conducted in Sub-Saharan Africa. Three methods for defining inference domains for selection of response functions were compared. Use of the OFRA Inference Tool (OFRA-IT; http://agronomy.unl.edu/OFRA) resulted in greater specificity of maize N, P, and K response functions with higher R2 values indicating superiority compared with using the Harvest Choice Agroecological Zones (HC-AEZ) and the recommendation domains of the Global Yield Gap Atlas project (GYGA-RD). The OFRA-IT queries three soil properties in addition to climate-related properties while the latter two options use climate properties only. The OFRA-IT was generally insensitive to changes in criteria ranges of 20–25% used in queries suggesting value in using wider criteria ranges compared with the default for information scarce crop nutrient response functions.

可通过具有代表性的作物养分响应函数，估算特定化肥施用投资的盈利潜力。当响应数据稀缺时，可借助相似作物种植环境下的研究结果，提升代表性养分响应函数的确定精度。本研究从非洲化肥推荐优化（Optimization of Fertilizer Recommendations in Africa, OFRA）数据库中选取玉米（Zea mays L.）养分响应函数；该数据库包含5500个经地理标记的响应函数，均源自撒哈拉以南非洲的田间试验数据。研究对比了三种用于划定响应函数筛选推理域的方法。与采用收获选择农业生态区（Harvest Choice Agroecological Zones, HC-AEZ）及全球产量缺口图谱项目（Global Yield Gap Atlas, GYGA-RD）的推荐域相比，使用OFRA推理工具（OFRA-IT; http://agronomy.unl.edu/OFRA）可获得特异性更强的玉米氮（N）、磷（P）、钾（K）养分响应函数，其更高的决定系数（R²）体现了该方法的优越性。OFRA-IT除查询气候相关属性外，还会查询三项土壤属性；而后两种方法仅使用气候属性。OFRA-IT对查询中采用的20%~25%标准范围变化总体不敏感，这表明针对养分响应数据稀缺的作物，使用宽于默认值的标准范围具备应用价值。