农业产值空间变异性研究——以广西壮族自治区为例

Study on the spatial variability of agricultural output: take Guangxi Zhuang Autonomous Region as an example

关键词:  农业产值  经典统计学  地统计学  空间变异特征  空间集聚性

Key words:  agricultural output  classical statistics  geostatistics  spatial variability charac-teristics  spatial agglomeration

摘要:

【目的】 以广西壮族自治区（以下简称广西）农业产值为研究对象，利用地统计学方法探究广西农业产值的空间变异特征，并揭示其空间分异格局对相关因素的响应规律，为农业决策管理部门的农区分类管理和精准农业提供科学参考。【方法】 文章首先通过四倍法去除异常值并通过样本代表系数校验获取采样样本70个，然后利用经典统计学方法得出样本的统计特征值并进行分析，最后运用地统计学理论分析广西地区农业生产信息的空间变异性特征，并运用Kriging插值法和莫兰指数对广西农业产值的空间格局进行分析。【结果】 广西农业产值的原始数据和对数变换的变异系数分别是61%和21%，属于中等变异性；通过理论模型拟合结果对比选取Gaussian模型进行拟合，其块金效应为0.226%；最后，使用莫兰指数进行空间自相关分析，其全局莫兰指数为0.327 661，表明其空间分布不是随机的，内部存在明显空间集聚性。【结论】 广西的农业产值的空间变异特征主要受其结构性因素影响，随机性因素对该区域的空间变异性影响较小，分析其空间发现，农业产值的空间分布特征与地形地貌的空间分布具有一定相似性，通过空间自相关分析探究其集聚性，总体上广西农业产值高高值集聚区在西北地区，低低值集聚区在中部和南部地区，其他地区集聚性不明显。

Abstract:

[Purpose] To investigate the spatial variation characteristics of agricultural output value in Guangxi by using geostatistical methods and to reveal the response pattern of its spatial variation pattern to related factors，this paper provides scientific reference for agricultural decision - making management department's agricultural area classification management and precision agriculture.[Method] Firstly，70 samples were obtained by removing abnormal values through the quadruple method and checking the sample representation coefficients，then the statistical characteristic values of the samples were obtained and analyzed using classical statistical methods，and finally the spatial variability characteristics of agricultural production information in Guangxi were analyzed using geostatistical theory，and the spatial pattern of agricultural production value in Guangxi was analyzed using Kriging interpolation method and Moran index.[Result] The analysis results show that the coefficients of variation of the raw data and log-transformation of agricultural output value in Guangxi are 61% and 21% respectively，which belong to medium variability； the Gaussian model was selected for fitting by comparing the theoretical model fitting results，and its block gold effect is 0.226%； finally，the Moran index was used for spatial autocorrelation analysis，and its global Moran index is 0.327 661，which indicates that its spatial distribution is not random. indicating that its spatial distribution is not random and there is obvious spatial agglomeration within it.[Conclusion] The spatial variability of agricultural output value in Guangxi is mainly influenced by its structural factors，and the random factors have less influence on the spatial variability of the region. In general，the high-value areas of agricultural output in Guangxi are in northwestern Guangxi，the low-value areas are in central and southern Guangxi，and the clustering in other areas is not obvious.