基于最小数据集的植烟土壤质量评价及障碍因子诊断

Quality assessment and obstacle factor diagnosis of tobacco planting soil based on minimum data set

为准确评价四川烟区植烟土壤质量,采集四川省烟田土壤样品,通过构建最小数据集计算土壤质量指数,并使用障碍因子诊断模型明确主要障碍因子。结果表明:(1)根据相关分析、主成分分析和Norm值计算结果,筛选碱解氮、pH值、氯离子、酸性磷酸酶、速效钾、容重、有效铁和脲酶8项评价指标进入最小数据集。(2)根据总数据集计算的土壤质量指数(总数据集土壤质量指数,SQI_(TDS))与根据最小数据集计算的土壤质量指数呈极显著(P<0.01)正相关;随机森林分类算法计算结果表明,最小数据集评价指标可对根据SQI_(TDS)划分的土壤质量档次进行分类,分类准确率为97.3%,说明该最小数据集可代替总数据集对植烟土壤质量进行评价。(3)四川省各县(市、区)的植烟土壤质量指数平均值分布在45.76~65.36之间,除喜德县和普格县外,其他植烟县(区、市)的土壤质量指数平均值均大于50.00;四川省植烟土壤质量提升的主要障碍因子为总孔隙度、容重、水溶态钙、水溶态镁和有效铁。

In order to accurately evaluate the quality of tobacco planting soil in Sichuan Province,tobacco planting soil samples were collected.The soil quality index was calculated by developing a minimum data set.The diagnostic model of obstacle factors was used to identify the main obstacle factors.The results showed that:1)According to the results of correlation analysis,principal component analysis and Norm value calculation,eight evaluation indicators including alkali-hydrolytic nitrogen,pH value,chloride ion,acid phosphatase,available potassium,bulk density,available iron and urease were selected in the minimum data set.2)The soil quality index calculated by the total data set(Total data set soil quality index,SQI_(TDS))was significantly(P<0.01)positively correlated with the soil quality index calculated by the minimum data set.The calculated results of the random forest classification algorithm showed that the evaluation indicators in the minimum data set could classify the soil quality grades classified by SQI_(TDS),and the classification accuracy was 97.3%,indicating that the minimum data set could replace the total data set to evaluate the quality of tobacco planting soil.3)The average values of soil quality indexes of Sichuan tobacco planting soil samples were 45.76-65.36.Except for Xide County and Puge County,the average soil quality index of other tobacco planting counties(districts or cities)was higher than 50.00.The main obstacle factors of tobacco planting soil in Sichuan were total porosity,bulk density,water-soluble calcium,water-soluble magnesium and available iron.