以海涂围垦区为例,利用普通克立格插值法和序贯高斯条件模拟方法对土壤盐分的空间分布进行估值和模拟,并利用序贯指示条件模拟进行不确定性评价。结果表明,样区东南区域土壤盐分含量较高而北部区域盐分含量低。由普通克立格法得到的土壤盐分的空间分布整体比较连续,具有明显的平滑效应,估值结果数据的分布频率趋于平缓。序贯高斯条件模拟结果整体分布相对离散,突出了原始数据分布的波动性,其模拟结果数据的分布频率相对集中。预测精度上,序贯高斯条件模拟的预测结果精度相对较高。以评价标准204mSm^-1作为土壤盐分含量的阈值进行的序贯指示条件模拟结果显示,在土壤盐分含量较高的东南部地区,超过阈值的概率超过75%,而北部很多盐分相对含量低的地区,超过阈值的概率值都低于25%。以超阈值概率为0.9、0.85和0.8三个值来选取盐分的高值风险区进行空间不确定评价,结果表明,联合概率比单点统计的概率更为严格,在划分较大范围高盐风险区域时,最好同时采用联合概率来进行信度评价。
Ordinary Kriging spatial and sequential Gaussian conditional simulation (SGCS) were used to estimate and simulate distribution of soil salinity in a coastal saline land in Zhejiang province, and sequential indicator conditional simulation (SICS) was used to assess the probability of salt content exceeding the predefined threshold. The results indicated that the southeast region in the study area had high salinity degree, while low salinity degree w the northern and southwest section. The spatial distribution of salinity obtained by ordinary Kriging was con as found in tinuous and smooth, whereas that obtained by the SGCS was relatively discrete and fluctuant. In terms of prediction precision, the SGCS method had higher precision. The results of the SICS indicated that the conditional probability of salt content exceeding the threshold of 204mS m^-1 was greater than 75% in the southeast area, whereas in the north the probability of salt content exceeding 204mS m^-1 was less than 25%. Based on the joint probability statistics, the spatial uncertainty (multi-points uncertainty) obtained from the SICS was used to assess the reliability of delineation for regions suspected of high risk. The result showed that the assessment of spatial uncertainty was more strict than that of single-point uncertainty. Therefore, it is important that those two approaches of uncertainty assessment were used simultaneously for risk analysis.