多尺度数据整合方法以贝叶斯(Bayesian)层次模型为框架,融合了尺度提升的思想并能够有效地结合马尔科夫链蒙特卡洛抽样算法,它在将所有相关信息进行整合的同时考虑到先验信息的贡献,为不同来源、不同尺度间动态和(或)静态数据的整合提供了有效的模拟框架。本文基于土壤饱和与非饱和水力特性参数空间定量化研究的重要意义,综述了多尺度数据整合方法的发展及其在该研究中的应用现状,展望了它在进一步整合不同尺度的静态和(或)动态数据从而进行土壤水力参数空间定量化中的应用前景,并分析了延拓应用中所面临的理论和技术突破点,为结合专业知识、运用多尺度数据整合方法探究土壤水文学中的关键科学问题提供了理论分析和技术支撑的参考依据。
Within the framework of Bayesian hierarchical model, the multiscale data integration method combined with upscaling is ideally suited to Markov Chain Monte Carlo (MCMC) simulation, which integrates all relevant information including the prior information, and thus provides an effective framework for integration of dynamic and (or) seismic data from multi-source and multi-scale datasets. Considering the importance of the spatial characterization of the soil property parameters, this paper summarizes the development of the multiscale data integration method and its application in this research field. Its further application in integration of static or dynamic data from different scales, so as to spatially modeling soil hydraulic parameters has been expected. Several sticking points related to such further applications are also discussed, so as to provide theoretic analysis and technical support.