多元统计分析在四川退耕还林区土壤肥力评价中的运用

Multivariate Statistical Analysis Application in Evaluation Soil Fertility of Coverting Land into Forests in Sichuan

以四川洪雅退耕还林区4种退耕还林模式和农耕地17个土壤肥力指标为例,采用多元统计的主成分分析法对土壤肥力等级进行综合评价.借助SPSS统计软件得出土壤肥力指标前5个主成分的累计贡献率为84.17%,已足够反映原指标信息.这5个主成分分别概括为:①土壤有效磷,铵态氮,pH水浸,水解氮,全氮,硝态氮和全磷指标;②土壤全钾、速效钾和缓效钾指标;③土壤总孔隙度、毛管孔隙度和最大持水量指标;④土壤容重,毛管持水量,pH盐浸指标;⑤土壤有机质指标.借助SPSS统计软件计算主成分因子的综合得分.结果显示:就各模式土壤上层比较,模式Ⅲ(牛鞭草)上层在综合得分中排名第一.就各模式土壤下层比较,模式Ⅱ(苦竹+牛鞭草)下层在综合得分中排名第一.

Multivariate statistical principal component analysis was adopted to evaluate soil gertility rank, taking example for 17 soil fertility factors of four patterns of the coverting lands into forests and farmland in Hongya, Sichuan Province. The cumulative contribution of the first 5 principal components of soil fertility factors is 84.17%, and therefore they are sufficient to reflect the information of original factors. The 5 principal components include： O soil available phosphate, soil NH4 ^＋ -N, pH（ H2 O）, available N, total N, total P and soil NO3^- -N; 2） soil total K, soil rapidly available K and soil slowly available K; （3） total soil porosity, capillary porosity and max soil moisture capacity; （4） soil bulk density, capillary capacity and pH（KCI） ; and （5） soil organic matter. The Integrated score was calculated in virtue of SPSS statistical software. As far as the improved effects on forest soil after converting lands into forests was concerned, pattern m （Hemarthria） was the best one among all patterns about the top level soil; while pattern Ⅱ （Gramineae ＋ Hemarthria） was the best one about the lower soil.