精准农业中土壤养分分析的适宜取样数量的确定  

Proper soil sample numbers for soil nutrient analysis in precision agriculture

作　　者：赵伟[1] 谢德体[1] 刘洪斌[1] 王晓东[2] 丁声源[2] 

机构地区：[1]西南大学资源环境学院,重庆400716 [2]西南大学经济管理学院,重庆400716

出　　处：《中国生态农业学报》2008年第2期318-322,共5页Chinese Journal of Eco-Agriculture

基　　金：重庆市科委重点攻关项目(2006AB1015)资助

摘　　要：为研究适合丘陵地区的土壤养分取样数量,减少采样分析成本,本文采用Cochran(1977)公式对丘陵地区土壤特性的取样点数量进行了探讨。结果发现:在1km2研究区域内,利用Cochran公式对土壤养分进行预估,其取样数目比网格取样法明显减少,土壤中的N以15%的相对误差为宜,在置信度分别为90%和95%条件下,分别只采集2个和3个样本就能反映其含量水平;P以15%的均值误差为宜,在置信度分别为90%和95%条件下,分别采集17个和25个样本就足够反映其基本情况;Mg以5%的均值误差为宜,在置信度分别为90%和95%下,取样数量分别为11个和15个。The Cochran formula was used to determine suitable sampling plot numbers of soil characteristics in order to determine proper soil sampling numbers in hill slope areas and to reduce costs of sampling and analysis. Research result shows that pre-estimation of soil nutrients using Coehran formation can greatly reduce sampling plot number. 2 to 3 samples can reliably reflect soil nitrogen content in 1 km2 within the accuracy range of 90%- 95% , and only 15% proper relative error. 17 to 25 samples can sufficiently reflect soil phosphorus content within the accuracy range of 90% -95 % , and 15% proper relative error. Accordingly, 11 to 15 samples can sufficiently reflect soil magnesium content within the accuracy range of 90% -95% , with 5% proper relative error.

关 键 词：精准农业 置信度 取样数量 土壤养分 

分 类 号：S131.3[农业科学—农业基础科学]