| 采样数量和分布模式对土壤有机质空间插值精度的影响 |
| |
| 引用本文： | 蒋威,郜允兵,刘玉,潘瑜春,邢世和.采样数量和分布模式对土壤有机质空间插值精度的影响[J].福建农业学报,2016(3):273-279. |
| |
| 作者姓名： | 蒋威 郜允兵 刘玉 潘瑜春 邢世和 |
| |
| 作者单位： | 1. 福建农林大学资源与环境学院，福建 福州 350002; 国家农业信息化工程技术研究中心，北京 100097;2. 国家农业信息化工程技术研究中心，北京 100097; 农业部农业信息技术重点实验室，北京 100097; 北京市农业物联网工程技术研究中心，北京 100097;3. 福建农林大学资源与环境学院,福建 福州,350002 |
| |
| 基金项目： | 国家科技支撑计划课题(2013BAJ05B01)；国家自然科学基金项目(41201173) |
| |
| 摘 要： | 基于北京市大兴区南部郊区平原农用地的高密度土壤样点,结合GIS和空间插值方法,对研究区的样点数据进行随机抽样生成不同数量的数据集,探讨采样点数量和分布模式对土壤有机质空间预测插值精度的影响。研究结果表明:(1)利用不同方式确定的合理采样点数量由高到低依次是插值方法的精度要求、变异函数拟合效果、正态分布法和均值及变异函数的满足小于5%相对误差;(2)变异函数拟合效果总体随着采样点数量的增加呈梯度式增强,但495点集优于825点集;相同样点分布模式下变异函数拟合效果相似;(3)相同样点集的不同插值方法预测精度存在差异,最佳插值方法不同;2272、825、115点集BEK都是最优,1485、495和83点集分别是SWT、SK和LPI最优;(4)样点间最近邻距离分布模式中P值小于0.05,插值模型对土壤有机质插值精度低;P0.05,R比率接近1,样点分布强于样点数量对插值模型对预测精度的影响;其P0.05且R比率大于1.20,各常规插值方法能得到较稳定的插值精度。
|
| 关 键 词： | 土壤有机质 采样数量 插值模型 分布模式 精度 |
Effect of Sampling Quantity and Distribution on Precision of Spatial Interpolation on Soil Organic Matters |
| |
| Abstract: | Based on the randomly sampled data collected from the high density soil sampling spots at the agricultural lands in Beijing southern suburban plain of Daxing ,the geographic information system and the spatial interpolation were applied to determine the effect of sampling quantity and distribution on the precision of the interpolation on soil organic matters .The results showed that (1) the appropriate sampling quantity from high to low was included in the tests for precision ,variation function fitting ,normal distribution ,as well as average and variation functions that met the required relative error of less than 5% ;(2) although the imitative effect of variation function gradually increased with increasing sampling number ,495 sampling sets was found to be better than 825 ,while the variation function fitting effect was similar under same sampling distribution ;(3) the precision differed according to the interpolation methods at same sampling set ,BEK was the best for 2 272 ,825 and 115 sampling sets ,while SWT ,SK ,and LPI for 1 485 ,495 and 83 sets ,respectively ;and ,(4) as far as sampling distribution was concerned ,the precision of interpolation model was low when the P value of the nearest neighbor distance was less than 0.05 ,it improved when the P values was greater than 0.05 and the R ratio close to 1 ,and stabilized when the P value was greater than 0.05 and the R ratio greater than 1.20 . |
| |
| Keywords: | soil organic matter sampling quantity interpolation model distribution pattern precision |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
