考虑边界条件不确定性的地下水污染风险分析

为分析边界条件不确定性对地下水污染质运移数值模拟模型输出结果的影响，运用Monte Carlo方法对一算例进行阐明，并从污染风险预报方面对模拟结果进行分析.为减少重复调用模拟模型产生的大量计算负荷，将边界条件（第一类边界条件-水头值）作为随机变量，建立地下水污染质运移数值模拟模型的Kriging替代模型，在保证较高精度的同时，实现了Monte Carlo模拟.结果表明：边界条件的不确定性，对地下水污染质运移数值模拟模型预报的结果有很大影响，考虑与未考虑边界条件不确定性得到的研究区污染羽分布差别较大.对地下水污染质运移数值模拟模型的Monte Carlo模拟结果进行统计与分析，可以评估研究区观测井1，2，3污染物浓度预报结果的可靠程度，并且可以预报出研究区观测井1，2，3遭受不同程度污染的风险.

Groundwater pollution risk analysis considering the uncertainty of boundary conditions

In order to analyze the influence of boundary conditions' uncertainty on the output results of groundwater contaminant transport numerical simulation model, the Monte Carlo method was used to illustrate an example, and the simulation results were analyzed in term of pollution risk prediction. To reduce the large amount of computational load generated by repeated calls of the simulation model while ensure the high accuracy,a boundary condition (the first type boundary condition-water head value) was used as a random variable to establish a Kriging surrogate model of the groundwater contaminant transport simulation model. The results showed that the uncertainty of boundary conditions had a great influence on the prediction results of groundwater contaminant transport numerical simulation model. The distribution of contaminant plume in the study area was significantly different from the ones that without considering the uncertainty of boundary conditions. Taking statistics and analysing on the Monte Carlo simulation results of groundwater contaminant transport numerical simulation can assess the reliability degree of the predicted pollutant concentration of observation wells 1, 2, 3, and also predicted the pollution risk of observation wells 1, 2, 3 in the in the study area.