云南植烟土壤肥力状况的组合评价法研究

  目的  为建立更加科学、综合的植烟土壤评价方法，  方法  以江苏中烟云南12个基地单元的土壤样本为研究对象，采用单一与组合评价模型对其肥力状况进行评价，并结合组合评价的事前事后检验，最终筛选出最优的组合评价方法。  结果  结果表明:（1）不同烟区土壤肥力状况指标的差异规律不同，尤其是隶属度法、主成分分析法以及灰色关联度法的评价结果规律各异；（2）组合前不同方法得分排序的标准差最大值为1.53，组合后排序的标准差最大值为1.4。3种方法组合后得分最高和最低的烟区均为烟区8和烟区12，较组合前更加一致；（3）7种评价方法的Spearman系数分别为平均值组合法（0.92）、Copeland组合法（0.91）、Borda组合法（0.88）、模糊Borda法（0.86）、灰色关联度法（0.85）、隶属度函数法（0.77）、主成分分析法（0.68），组合后的评价结果普遍好于单一评价结果；（4）平均值组合法的t_k值为7.42，评价结果表现最优。  结论  组合评价法集合单一评价得分排序结果，代表性、可靠性更高，适合于云南不同烟区植烟土壤肥力状况的综合评价。 

Study on method combination for evaluation of fertility status of tobacco growing soil in Yunnan

To realize more scientific and comprehensive evaluation of the fertility status of tobacco growing soil, samples were collected as research objects from 12 base units of China Tobacco Jiangsu Industrial Co., Ltd. These soil samples were analyzed by single and combination evaluation models. Combined with pre- and post-test of combinatorial evaluation, the optimal combinatorial evaluation method was selected. Results showed that: (1) Soil fertility indices vary between different tobacco-growing areas, especially for results obtained by membership degree, principal component analysis and grey correlation degree method. (2) The maximum standard deviation of ranking result of different methods before combination was 1.53, and the value after combination was 1.4. The tobacco growing areas with the highest and lowest value were consistently area 8 and 12 after combination of 3 different methods. (3) The Spearman coefficients of 7 evaluation methods were: mean value combined method (0.92), Copeland combined method (0.91), Borda combined method (0.88), fuzzy Borda combined method (0.86), grey correlation method (0.85), fuzzy mathematics membership function model method (0.77), and principal component analysis method (0.68). The evaluation results of combination evaluation methods were generally better than the results of single ones. (4) The tk value of mean value combination method was 7.42, can be considered the optimal combination method. The combination evaluation methods integrated with score and ranking results of single evaluation method is more representative and reliable, which was suitable for comprehensive evaluation of fertility status of tobacco growing soil from different areas in Yunnan.