集合卡尔曼滤波估计水文地质参数的局域化修正

集合卡尔曼滤波(Ensemble Kalman Filter,EnKF)作为一种有效的数据同化方法,在众多数值实验中体现优势的同时,也暴露了它使用小集合估计协方差情况下精度较低的缺陷。为了降低取样噪声对协方差估计的干扰并提高滤波精度,应用局域化函数对小集合估计的协方差进行修正,即在协方差矩阵中以舒尔积的形式增加空间距离权重以限制远距离相关。在一个二维理想孔隙承压含水层模型中的运行结果表明,局域化对集合卡尔曼滤波估计地下水参数的修正十分有效,局域化可以很好地过滤小集合估计中噪声的影响,节省计算量的同时又可以防止滤波发散。相关长度较小的水文地质参数(如对数渗透系数)更容易受到噪声的干扰,更有必要进行局域化修正。     

基于EnKF综合水头和浓度观测数据推估地下水流模型参数

地下水反应运移模型具有参数个数众多,观测数据类型多样的特点。为了探究不同类型观测数据在反应运移模拟数据同化中的数据价值,构建了三氯乙烯降解反应运移模型的理想算例,基于水头和浓度两种类型观测数据,采用集合卡尔曼滤波方法推估渗透系数和贮水系数的非均质空间分布,讨论了影响同化结果的因素。结果表明:与仅同化水头数据的结果相比,联合同化水头和浓度观测数据推估渗透系数场和贮水系数场时具有更高的精度,在观测数据拟合和模型预测方面也有更好的表现。与目前溶质运移模型、非饱和流模型等地下水模型中的研究结果相似,数据同化结果受样本数量,观测井的数量和位置的影响,合理优化布置监测井和选择样本数量可有效改善数据同化效果并提高计算效率。     

考虑预报偏差的迭代式集合卡尔曼滤波在地下水水流数据同化中的应用

集合卡尔曼滤波（Ensemble Kalman Filter,EnKF）方法已广泛应用于地下水水流和污染物运移模拟相关问题的求解。但前人研究多建立在同化系统预报模型是准确的基础上,忽视了模型概化的不确定性。当模型概化不准确时,将导致预报偏差,可能带来错误的系统估计。因此,文章提出考虑模型预报偏差的迭代式集合卡尔曼滤波（Bias aware Ensemble Kalman Filter with Confirming Option,Bias-CEnKF）方法。以地下水水流数据同化为例,研究模型概化存在不确定条件下,边界条件、初始条件、源汇项概化不准确时新方法的有效性。结果表明,当预报模型概化不准确时,使用标准EnKF方法进行数据同化,可能会导致滤波发散,造成同化失败。Bias-CEnKF方法不仅保留了较好的同化性能,同时减小了参数、变量、偏差项非线性关系带来的不一致性。针对文章中4种情景,Bias-CEnKF同化获得的含水层渗透系数场以及水头场均接近真实场,且预报结果可靠。本研究进一步提升了模型概化不确定时EnKF方法的适用性,为实际野外复杂条件下地下水水流数据同化问题提供了可靠的方法。     

地下水污染强度及渗透系数场的反演识别研究

在制定地下水污染修复方案时,污染源参数和渗透系数场是最重要的地下水数值模型参数,但前人研究多集中于单一类型参数的识别。文章中采用地下水污染物运移模型（MT3DMS）和数据同化方法（迭代局部更新集合平滑器,ILUES）构成地下水污染源识别的求解框架,并利用Karhunen-Loève展开技术实现渗透系数场的参数降维,最后通过同化水头与浓度数据实现地下水污染源强和渗透系数场的联合反演。结果表明：（1）ILUES算法能精确识别污染源参数和渗透系数场,并且具有很高的普适性;（2）精确表征渗透系数在空间上呈现出的非均质性,是预测污染物迁移路径、反演污染强度的关键;（3）ILUES算法参数影响着反演效果,综合考虑计算效率和计算精度等,可以得到算例的最佳样本集合大小（Ne=4 000）和ILUES算法最佳参数组合（局部临近样本集合占比α=0.4,相对权重b=4）。但在实际工程案例中,如果对精度的要求不是过高,经验组合（α=0.1,b=1）更值得推荐。研究结果对于区域地下水资源调查、评价和管理等工作具有较强的实践意义,并可为后期地下水污染预测及地下水监测井网优化提供技术支撑。     

基于集合卡尔曼平滑算法的土壤水分同化

为研究观测资料稀少情况下土壤质地及有机质对土壤水分同化的影响,发展了集合卡尔曼平滑(Ensemble Kalman Smooth, EnKS)的土壤水分同化方案。利用黑河上游阿柔冻融观测站2008年6月1日至10月29日的观测数据,使用EnKS算法将表层土壤水分观测数据同化到简单生物圈模型(Simple Biosphere Model 2, SiB2)中,分析不同方案对土壤水分估计的影响,并与集合卡尔曼滤波算法(EnKF)的结果进行比较。研究结果表明,土壤质地和有机质对表层土壤水分模拟结果影响最大而对深层的影响相对较小;利用EnKF和EnKS算法同化表层土壤水分观测数据,均能够显著提高表层和根区土壤水分估计的精度,EnKS算法的精度略高于EnKF且所受土壤质地和有机质的影响小于EnKF;当观测数据稀少时,EnKS算法仍然可以得到较高精度的土壤水分估计。     

基于集合卡尔曼滤波的多相流模型参数估计——以室内二维砂箱中重质非水相污染物入渗为例

重质非水相有机污染物(DNAPL)泄漏到地下后,其运移与分布特征受渗透率非均质性影响显著。为刻画DNAPL污染源区结构特征,需进行参数估计以描述水文地质参数的非均质性。本研究构建了基于集合卡尔曼滤波方法(EnKF)与多相流运移模型的同化方案,通过融合DNAPL饱和度观测数据推估非均质介质渗透率空间分布。通过二维砂箱实际与理想算例,验证了同化方法的推估效果,并探讨了不同因素对同化的影响。研究结果表明:基于EnKF方法同化饱和度观测资料可有效地推估非均质渗透率场;参数推估精度随观测时空密度的增大而提高;观测点位置分布对同化效果有所影响,布置在污染集中区域的观测数据对于参数估计具有较高的数据价值。     

基于EnKF融合地球物理数据刻画含水层非均质性

含水层非均质性的刻画是模拟地下水中污染物运移的关键。以渗透系数为研究对象,构建了综合集合卡尔曼滤波方法、有效电阻率模型与地下水运移模型的同化框架,通过融合地球物理观测数据与污染物浓度观测数据来推估渗透系数的空间分布。基于理想算例,验证了该同化框架刻画含水层非均质渗透系数场的有效性,并针对不同初始参数信息与观测类型对比了耦合与非耦合水文地球物理方法的适用性。研究结果表明:基于集合卡尔曼滤波方法同化多种类型的观测数据,可有效地推估非均质参数空间分布。当初始信息较准确时,耦合方法的参数推估精度更高;初始信息存在偏差时,非耦合方法有更好的同化效果。由于非耦合方法计算成本较低且对初始信息缺失时适用性更强,在实际应用中可先基于非耦合方法初步估计参数,再利用耦合方法进一步提高参数推估精度。融合多种类型观测数据可有效提高参数推估效果。     

基于多点地质统计与集合平滑数据同化方法识别非高斯渗透系数场

在冲积含水层中,由于岩相的非均质分布,渗透系数一般呈现出明显的非高斯特性（例如砂和黏土两种岩相）,非高斯特性给地下水模型参数的推估带来了困难与挑战。目前广泛使用的集合平滑数据同化方法（ESMDA）虽然有效且计算成本较低,但仅适用于高斯场。多点地质统计方法虽已广泛用于模拟非高斯场,但其无法融入动态观测数据推估参数。基于多点地质统计方法中的直接采样法（DS）与集合平滑数据同化方法,构建一种新的数据同化框架（ESMDA-DS）,既可保持参数场的非高斯特性,又可融合多源数据精确推估非高斯场。构建三个理想算例验证ESMDA-DS对非高斯参数场的推估效果,并探讨了不同类型观测数据对推估效果、水位与浓度预测精度的影响。三个理想算例包括仅融合水位数据（Case 1）,同时融合水位与浓度数据（Case 2）,同时融合水位、浓度与对数渗透系数数据（Case 3）。结果表明:ESMDA-DS方法结合了ESMDA与DS的各自优势,能有效融合观测数据推估渗透系数场,并保持参数场的非高斯特性。通过对比三个算例推估结果,Case 3的参数场推估效果最好,水位与浓度预测精度最高,Case 2次之,Case 1最差,表明融合多源数据可改善推估效果,提高预测精度。     

土壤水分同化系统的敏感性试验研究

利用1998年7月6日至8月9日青藏高原GAME-Tibet试验区MS3608站点的4cm、20cm和100cm的土壤水分观测数据同化SiB2模型输出的表层、根区和深层土壤水分,探讨了一个基于集合卡尔曼滤波和简单生物圈模型的单点土壤水分同化方案。分析和评价了集合大小、同化周期、模型误差、背景场误差以及观测误差对同化系统性能的影响。结果表明:①增加集合数目可以减小土壤水分同化系统的误差,但同时又降低了运行效率;②对于集合卡尔曼滤波,初始场的估计是否准确对同化系统性能影响不大;③模型误差和观测误差的准确估计可以提高土壤水分的估计精度;④利用数据同化的方法对土壤水分的估计有显著提高。     

不同滤波算法在土壤湿度同化中的应用

为研究不同滤波算法在土壤湿度同化中的有效性,以及土壤湿度模拟结果对模型参数的敏感性,结合简单生物圈模型SiB2,设置敏感性实验,探求土壤饱和水力传导度对土壤湿度模拟结果的影响;并在此基础上,采用集合卡尔曼滤波（EnKF）、无迹卡尔曼滤波（UKF）和无迹粒子滤波（UPF）开展土壤湿度实时同化实验。结果表明：土壤饱和水力传导度能显著影响土壤湿度模拟精度;利用EnKF、UKF、UPF同化站点观测数据,均能改善土壤湿度模拟结果;3种同化方法在不同土壤层的同化效果不同,在土壤表层,EnKF的有效性优于UKF和UPF,在根域层和土壤深层,3种滤波方法有效性在降雨前后相差较大。因此,针对性地选择同化方法,是提高土壤湿度模拟精度的有效手段。     

Pattern Recognition in a Bimodal Aquifer Using the Normal-Score Ensemble Kalman Filter

The ensemble Kalman filter (EnKF) is now widely used in diverse disciplines to estimate model parameters and update model states by integrating observed data. The EnKF is known to perform optimally only for multi-Gaussian distributed states and parameters. A new approach, the normal-score EnKF (NS-EnKF), has been recently proposed to handle complex aquifers with non-Gaussian distributed parameters. In this work, we aim at investigating the capacity of the NS-EnKF to identify patterns in the spatial distribution of the model parameters (hydraulic conductivities) by assimilating dynamic observations in the absence of direct measurements of the parameters themselves. In some situations, hydraulic conductivity measurements (hard data) may not be available, which requires the estimation of conductivities from indirect observations, such as piezometric heads. We show how the NS-EnKF is capable of retrieving the bimodal nature of a synthetic aquifer solely from piezometric head data. By comparison with a more standard implementation of the EnKF, the NS-EnKF gives better results with regard to histogram preservation, uncertainty assessment, and transport predictions.     

Groundwater contaminant source identification based on QS-ILUES

When groundwater pollution occurs,to come up with an efficient remediation plan,it is particularly important to collect information of contaminant source(location and source strength)and hydraulic conductivity field of the site accurately and quickly.However,the information can not be obtained by direct observation,and can only be derived from limited measurement data.Data assimilation of observations such as head and concentration is often used to estimate parameters of contaminant source.As for hydraulic conductivity field,especially for complex non-Gaussian field,it can be directly estimated by geostatistics method based on limited hard data,while the accuracy is often not high.Better estimation of hydraulic conductivity can be achieved by solving inverse groundwater problem.Therefore,in this study,the multi-point geostatistics method Quick Sampling(QS)is proposed and introduced for the first time and combined with the iterative local updating ensemble smoother(ILUES)to develop a new data assimilation framework QS-ILUES.It helps to solve the contaminant source parameters and non-Gaussian hydraulic conductivity field simultaneously by assimilating hydraulic head and pollutant concentration data.While the pilot points are utilized to reduce the dimension of hydraulic conductivity field,the influence of pilot points’layout and the ensemble size of ILUES algorithm on the inverse simulation results are further explored.     

Integration of Tracer Test Data to Refine Geostatistical Hydraulic Conductivity Fields Using Sequential Self-Calibration Method

On the basis of local measurements of hydraulic conductivity,geostatistical methods have been found to be useful in heterogeneity characterization of a hydraulic conductivity field on a regional scale. However,the methods are not suited to directly integrate dynamic production data,such as,hydraulic head and solute concentration,into the study of conductivity distribution. These data,which record the flow and transport processes in the medium,are closely related to the spatial distribution of hydraulic conductivity. In this study,a three-dimensional gradient-based inverse method-the sequential self-calibration (SSC) method-is developed to calibrate a hydraulic conductivity field,initially generated by a geostatistical simulation method,conditioned on tracer test results. The SSC method can honor both local hydraulic conductivity measurements and tracer test data. The mismatch between the simulated hydraulic conductivity field and the reference true one,measured by its mean square error (MSE),is reduced through the SSC conditional study. In comparison with the unconditional results,the SSC conditional study creates the mean breakthrough curve much closer to the reference true curve,and significantly reduces the prediction uncertainty of the solute transport in the observed locations. Further,the reduction of uncertainty is spatially dependent,which indicates that good locations,geological structure,and boundary conditions will affect the efficiency of the SSC study results.     

Hydraulic conductivity prediction based on grain-size distribution using M5 model tree

The hydraulic conductivity, K_s, is one of the most important hydraulic properties which controls the water and solute movement into the soil. It is measured on soil specimens in the laboratory. On the other hand, sometimes it is obtained by tests carried out in the field by a number of researchers. Therefore, several experimental formulas have developed to predict it. Recently, soft computing tools have been used to evaluate the hydraulic conductivity. However, these tools are not as transparent as empirical formulas. In this study, another soft computing approach, i.e. model trees, have been used for predicting the hydraulic conductivity. The main advantage of model trees is that, unlike the other data learning tools, they are easier to use and represent understandable mathematical rules more clearly. In this paper, a new formula that includes some parameters is derived to estimate the hydraulic conductivity. To develop the new formulas, experimental data sets of hydraulic conductivity were used. A comparison is made between the estimated hydraulic conductivity by this new formula and formulas given by other’s researches.     

A stochastic collocation based Kalman filter for data assimilation

In this paper, a stochastic collocation-based Kalman filter (SCKF) is developed to estimate the hydraulic conductivity from direct and indirect measurements. It combines the advantages of the ensemble Kalman filter (EnKF) for dynamic data assimilation and the polynomial chaos expansion (PCE) for efficient uncertainty quantification. In this approach, the random log hydraulic conductivity field is first parameterized by the Karhunen–Loeve (KL) expansion and the hydraulic pressure is expressed by the PCE. The coefficients of PCE are solved with a collocation technique. Realizations are constructed by choosing collocation point sets in the random space. The stochastic collocation method is non-intrusive in that such realizations are solved forward in time via an existing deterministic solver independently as in the Monte Carlo method. The needed entries of the state covariance matrix are approximated with the coefficients of PCE, which can be recovered from the collocation results. The system states are updated by updating the PCE coefficients. A 2D heterogeneous flow example is used to demonstrate the applicability of the SCKF with respect to different factors, such as initial guess, variance, correlation length, and the number of observations. The results are compared with those from the EnKF method. It is shown that the SCKF is computationally more efficient than the EnKF under certain conditions. Each approach has its own advantages and limitations. The performance of the SCKF decreases with larger variance, smaller correlation ratio, and fewer observations. Hence, the choice between the two methods is problem dependent. As a non-intrusive method, the SCKF can be easily extended to multiphase flow problems.     

History matching time-lapse seismic data using the ensemble Kalman filter with multiple data assimilations

The ensemble Kalman filter (EnKF) has become a popular method for history matching production and seismic data in petroleum reservoir models. However, it is known that EnKF may fail to give acceptable data matches especially for highly nonlinear problems. In this paper, we introduce a procedure to improve EnKF data matches based on assimilating the same data multiple times with the covariance matrix of the measurement errors multiplied by the number of data assimilations. We prove the equivalence between single and multiple data assimilations for the linear-Gaussian case and present computational evidence that multiple data assimilations can improve EnKF estimates for the nonlinear case. The proposed procedure was tested by assimilating time-lapse seismic data in two synthetic reservoir problems, and the results show significant improvements compared to the standard EnKF. In addition, we review the inversion schemes used in the EnKF analysis and present a rescaling procedure to avoid loss of information during the truncation of small singular values.     

Oscillatory Pumping Test to Estimate Aquifer Hydraulic Parameters in a Bayesian Geostatistical Framework

Comprehensive information about the spatial distribution of the subsurface hydraulic properties is crucial to model groundwater flow, to predict solute transport in aquifers and to design remediation actions. In this work, a Bayesian Geostatistical approach, as implemented in bgaPEST, was adopted to estimate the hydraulic properties of a well field located at the Campus of Science and Technology of the University of Parma (Northern Italy), in a contest of a highly parameterized inversion. Head data, collected by means of multi frequency oscillatory pumping tests, were used to both estimate the hydraulic parameters and validate the results. The groundwater flow processes were modelled by means of MODFLOW 2005 and an adjoint-state formulation of the same software was used to efficiently calculate the sensitivity matrix, required by the inverse procedure. The Bayesian Geostatistical approach estimated the hydraulic conductivity and specific storage fields, handling a large number of parameters. The results of the inversion are consistent with the alluvial nature of the investigated aquifer and the preliminary traditional pumping tests carried out at the site.     

Hydraulic tomography analysis of municipal-well operation data with geology-based groundwater models

The sustainable management of groundwater resources is essential to municipalities worldwide due to increasing water demand. Planning for the optimized use of groundwater resources requires reliable estimation of hydraulic parameters such as hydraulic conductivity (K) and specific storage (S_s). However, estimation of hydraulic parameters can be difficult with dedicated pumping tests while municipal wells are in operation. In this study, the K and S_s of a highly heterogeneous, multi-aquifer/aquitard system are estimated through the inverse modeling of water-level data from observation wells collected during municipal well operations. In particular, four different geological models are calibrated by coupling HydroGeoSphere (HGS) with the parameter estimation code PEST. The joint analysis of water-level records resulting from fluctuating pumping and injection operations amounts to a hydraulic tomography (HT) analysis. The four geological models are well calibrated and yield reliable estimates that are consistent with previously studies. Overall, this research reveals that: (1) the HT analysis of municipal well records is feasible and yields reliable K and S_s estimates for individual geological units where drawdown records are available; (2) these estimates are obtained at the scale of intended use, unlike small-scale estimates typically obtained through other characterization methods; (3) the HT analysis can be conducted using existing data, which leads to substantial cost savings; and (4) data collected during municipal well operations can be used in the development of new groundwater models or in the calibration of existing groundwater models, thus they are valuable and should be archived.