成矿预测研究三维技术方法资料.pptVIP

中国地质大学（北京） 河南栾川钼多金属矿区三维成矿预测——主要成果深化 多金属——矿种复杂 多类型——成因类型复杂 多元性——成矿信息复杂：多尺度、多类型 多空间——地层、构造、岩体 多平台——2D、2.5D、3D交互应用 基于成矿系列理论、地质异常理论建立三维成矿模式； 基于三维可视化技术与非线性方法高效集成地学多元信息，定位定量评价多金属矿产的资源量（resource）。 河南栾川200km2三维地质图 河南栾川200km2三维地质图（地表500m以下） 河南栾川200km2三维地质图（地表1000m以下） 三维Mo化探异常信息叠加图 三维重力异常（延拓）信息叠加图 三维磁异常（延拓）信息叠加图 基于概率神经网络法靶区圈定 三维Mo成矿潜力靶区叠加图 Eric GrunskyEditor-in-ChiefComputers GeosciencesReviewers comments:Reviewer #1: NOTE: UPLOAD WAS NOT FUNCTIONAL. REVIEWER COMMENTS FOR ALL EYES ARE BELOW:Manuscript CAGEO-D-10-00251: Mineral resources assessment based on 3D modeling and probabilistic neural networks in Luanchuan region, ChinaReview by K. Sprague1.? SummaryNonlinear classification methods coupled with advanced 3D geological model building and query capabilities were employed to predict likely locations of significant Mo and Pb-Zn-Ag mineral deposits in the Lunchuan region of China and furthermore assess their mineral resource potential. The first of two nonlinear methods used was probabilistic neural networks (PNN) in order to generate a continuous mineral potential field map that required classification. The PNN relied on a myriad of geological inputs from the study area including strata, structure, granite porphyry, gravity, magnetism and principal component analysis (PCA) scores for Mo and Pb-Zn-Ag. The second non-linear method employed was fractal concentration-area (C-A) analysis to determine classification thresholds that reduced the number of classes generated using the PNN results from generally hundreds via standard techniques (case quoted is 187) to a more tractable set of 6 classes. The fractal C-A technique exploits a power-law relationship between area and the potential field value to determine breakpoints in the class boundaries.The paper comments on insights and pitfalls encountered while using the various techniques described above to further the objectives of the study and concludes with estimates of locations for Mo and Pb-Zn-Ag deposits and