基于马尔可夫链的多元指示地质统计模型  

Markov Chain-based multi-indicator geostatistical model

作　　者：何芳[1] 吴吉春[1] 

机构地区：[1]南京大学地球科学系,南京210093

出　　处：《水文地质工程地质》2003年第5期28-32,共5页Hydrogeology & Engineering Geology

基　　金：教育部优秀青年教师奖励基金;国家自然科学基金(40272106).

摘　　要：与传统基于交叉变差函数(cross variogram)的多元指示地质统计模型相比,基于马尔可夫链的多元指示地质统计模型采用转移概率(transition probability)来描述区域化变量的空间变化,可以很好地反映复杂空间的连续性,弥补了传统地质统计模型没有考虑地质体分布存在不对称性的缺陷,简化了各向异性的处理过程,且在建立模型的过程中,融入了地质类型分布的比例、平均长度和相互间的迁移转换趋势等地质统计信息,使建模过程更直观,更易于理解。用这种方法建立的模型在理论上有比较成熟的随机理论支持,在实际应用中能更准确地反映地质体的空间分布特征。由于该模型考虑了地质体空间分布的连续性、不对称性和各向异性等特征,因此将其用于模拟地下含水层的空间结构,可以真实反映实际含水层的空间分布,为进一步模拟地下水运动提供有力支持。Comparing with traditional crossvariogrambased multiindicator geostatistical models, Markov chainbased multiindicator geostatistical model regard transition probability as indicator variable for describing spacial change of scope variable. Markov chainbased model can be used to described continuity of complex space, compensating flaw of traditional model models which neglected nonsymmetry of geological body, simplifying manage process of anisotropy. With regarding statistical geological information, such as distribution proportions,average lengths and juxtaposition tendencies of geological types, Markov chainbased geostatistical model was more visual and easy to understand. Model built by this method has theoretical support of relatively mature random theory and it can practically reflected special distribution characters of geological body more accurate and more visual. Considering continuity, nonsymmetry and anisotropy of special distribution, Model can really reappear special distribution of actual aquifers when it is used to simulate special structure of those aquifers and provided powerful support for further simulation of groundwater flow. 

关 键 词：马尔可夫链 地质统计模型 交叉变差函数 转移概率 交叉变差函数 地下水运动 

分 类 号：P641.2[天文地球—地质矿产勘探] O141.4[天文地球—地质学]