基于轮廓相似度的土壤肥力时空变异性判别

采用几何轮廓相似度模型，分别对彰武县境内柳河流域所辖乡村1985年与2009年土壤肥力等级进行判别．结果表明：研究区所辖43个乡村土壤肥力普遍处于中下等水平；与1985年相比，2009年有15个乡村土壤肥力级别降低即土壤肥力变差，9个乡村土壤肥力级别升高即土壤肥力变好，19个乡村土壤肥力级别未发生变化．该模型可在缺乏多标度数据分布信息或小样本条件下难以估计统计特征的应用问题中，通过分析多标度数据的几何特征对数据进行判别；且模型采用了凹函数求平均值，从而减小了一些突变型数据对判别的干扰．因此，采用几何轮廓相似度模型判别土壤肥力时空变异性准确且简便可行．

Soil Fertility Temporal and Spatial Variability Discrimination Based on Contour Similarity

Soil fertility orders of villages of the Liuhe basin in the Zhangwu county in 1985 and 2009 were discriminated via geometric contour similarity model. The results showed that soil fertilities of 43 villages of study area were in the middle and lower levels. With compared 1985 soil fertility orders of 15 villages dropped（deteriorated）, those of 9 villages rose（got better）and those of 19 villages remained unchanged. The model could discriminate data through analyzing geometric characteristics of multi scale data in the absence of multi scale data distribution information or under the condition of small sample, which is difficult to estimate statistical characteristics. And the model reduced distinguishing interference of some mutant data by employing concave func- tion for the average value. So it is exact, simple and feasible for soil fertility temporal and spatial variability discrimination by geometric contour similarity model.