摘要
本研究采用盆栽试验和田间试验方法对国内常用的 2种施肥模型 (二次和平方根 )和 3种试验设计 (D饱和设计、3× 3设计 ,5× 5设计 )进行了比较研究。确定灌溉农田以二次模型最优 ,旱地以平方根模型最佳。 3× 3设计简单、实用、准确、易于在农户中推广应用。研究认为对多点分散试验的结果 ,可采用动态聚类分析的方法 ,获得类肥料效应方程以指导推荐施肥。研究中建立的 3HCCFS - 90曲周施肥系统可实现 90 %的预报吻合度。
Two kinds of fertilization models (quadratic and square root polynomial) and three kinds of experimental designs (saturated design, 3×3 design and 5×5 design) were comparatively studied by both of the pot and field experiments. It was concluded that the quadratic polynomial model was the best one for irrigated land and the square root polynomial model for the dry land. The 3×3 design possessed advantages of simplicity, practice and accuracy, and therefore could be easily used by farmers. The results from the multi-site dispersed fertilizer experiments could be analyzed by using the method of dynamic classification statistics, in order to obtain the general function of fertilizer effects for guiding the fertilizer recommendation. The 3HCCFS-90-Quzhou model of fertilization specifically developed for Quzhou County could estimate wheat and summer corn yields in the region with about 90% of accuracy.
出处
《中国农业大学学报》
CAS
CSCD
北大核心
2003年第z1期53-56,共4页
Journal of China Agricultural University
关键词
动态聚类
施肥模型
施肥系统
dynamic classification statistics
fertilization model
fertilization system