摘 要 ：调查了云南省水稻主产区沾益、寻甸两县（两试验点）106块稻田水稻种植管理、有害生物为害（病虫草害）和产量等信息，并利用非参数多变量统计方法对数据进行了聚类分析和对应分析。主要结果：（1）确定了6种种植管理模式（PR），其中PR1和PR4跨区域存在，其它模式是各试验点特有的；PR1、PR2和PR4对应高产水平；从化肥、农药、水分供应、劳动力投入方面看，PR1农业投入最高，PR2农业投入较高，PR4农业投入水平中等。（2）确定了5种有害生物为害模式（IN），其中IN2在两县同时存在，其它模式是各自特有的；IN3整体为害水平最高，对应低产，IN1整体为害水平最低，对应高产。（3）在区域水平上，种植管理模式和有害生物为害模式存在较强的关联性( χ2, P < 0.0001)，二者能很好地解释水稻产量的变化诱因，多维列联表对应分析图提供了合成的观点，不仅反映了三者在区域之间的差异，同时说明了不同区域种植管理和有害生物为害模式的多样性，从产量增加路线可以看出寻甸县具有进一步提升水稻产量的潜力，为制定有害生物防治策略、提高区域稻作水平提供了理论依据。
Abstract:The rice cropping practices, pest injuries (due to pathogens, insects, and weeds) and yields were surveyed in 106 individual farmers′ fields, in the main rice-producing area of Yunnan Province: Zhanyi and Xundian counties (two experimental sites). The characterization of patterns of rice cropping practices and multiple pest injury profiles and their relation with the rice yield were analyzed by using the nonparametric multivariate techniques: cluster analyses and correspondence analyses. The main results are: (1) six patterns of cropping practices (PR) are identified; PR1 and PR4 are common across sites, whiles others are site-specific; PR1, PR2 and PR4 are related with high yield levels; PR1 corresponds to the highest agricultural input, PR2 to the higher input and PR4 to the medium input from the view of mineral fertilizers, pesticides, water supply and labor input; (2) five pest injury profiles (IN) are determined; IN2 is common across sites, whiles others are site-specific; IN3 is the highest injury level of pest combinations in five profiles, related with the lowest yields in these profiles, whiles IN1 is the lowest injury level, associated with the highest yields; (3) patterns of cropping practices and injury profiles are strongly associated at regional scale ( χ2, P < 0.0001); both of them are available for a good description of the variation in actual yield; the plot of correspondence analyses on multi-dimension contingency table provides a framework that accurately reflects difference in cropping practices, pest injuries and yields between two sites, illustrates the diversity of patterns of cropping practices and injury profiles, and indicates the potential for increased productivity in Xundian from the path of increasing yield levels (Y1-Y5). Results will provide some foundations for developing pest management strategies and improving rice production level at the regional scale.