观测数据权重对地下水有机污染物生物降解模型参数估计的影响

地下水多组分反应运移模型是刻画解释微生物降解地下水有机污染物的水文地质、生物地球化学耦合过程并对各组分分布特征进行预测的重要工具.反应运移模型往往包含多个参数,并含有多种不同生物化学组分的观测数据,高维参数和观测数据类型的多样性给模型参数的反演带来了极大的困难.因此,本文以反应运移软件PHT3D模拟某污染场地地下水中有机污染物生物降解的理想正问题为例,对比讨论了不同观测数据权重对反应运移模型参数估计的影响.结果表明:未考虑观测数据权重时参数估计结果与真实值相差较大,而采用反映观测误差的权重时参数估计值与真实值更接近.敏感性分析表明,参数对于不同观测组数据的敏感度有明显差异.对于本例,生物降解模型中的降解反应参数最敏感,其次是溶解速率迁移常数,多相溶解度摩尔分数敏感度较小.

Effect of the observation weights on model parameter estimation for biodegradation of organic containments in groundwater

Multi-component reactive transport model is an important tool to characterize and explain the coupling processes between hydrologic tranpsort and biogeochemical reactions of biodegradation of organic pollution for groundwater remediation. A reactive transport model often contains lots of parameters and different observation types for different biochemical components. The large number of parameters and the diversity of observation types have brought great difficulty for highly parameterized inversion. Taking a case of organic pollutant biodegradation in a contaminated site simulated with the reactive transport model PHT3D as an example, the effects of different observation weights on parameter estimation are discussed and compared. Results show that the estimated parameter values are close to the true values when considering the observation weights defined as the inverse of the variance of the measurement error. Sensitivity analysis result indicates that the sensitivities of different parameters are varied depending on different observation groups. For this biodegradation model, the maximum microbial uptake rates are most sensitive parameters and play an important role, followed by the transfer rates between the immobile (NAPL) and the aqueous concentration of the hydrocarbon compound, which are relatively less sensitive, and the molar fractions of the NAPL composition are insensitive.