Mineral Potential Assessment of Sedimentary Deposit using Frequency Ratio and Logistic Regression of Gangreung Area, Korea

S Lee,N.-W Park,H.-J Oh

doi:10.1109/igarss.2006.406

Abstract

Mineral resource potential mapping is an important procedure in mineral resource assessment. The aim of this study is to analyze relationships between sedimentary deposits and related factors and integrated the relationships using probabilistic and statistical models in GIS environment to identify areas that have not been subjected to the same degree of exploration. For this, a variety of spatial geological data were compiled, evaluated and integrated to produce a potential map for deposits in the Gangreung area, Korea. This empirical approach assumes that all deposits share a common genesis and comprises three main steps such as identification of spatial relationships, quantification of identified spatial relationships and integration of multiple quantified spatial relationships. For this, a spatial database including sedimentary mineral deposit, topographic, geologic, geophysical and geochemical data were constructed for the study area using Geographic Information System (GIS). The used 55 sedimentary mineral deposits and the related to factors, geological data such as lithology and fault, geochemical data such as Al, As, Ba, Ca, Cd, Co, Cr, Cu, Fe, K, Li, Mg, Mn, Mo, Na, Ni, Pb, Si, Sr, V, W, Zn, Cl, F - , PO 4 2- , NO 2 - , NO 3 - and SO 4 2- , HCO 3 - , pH, Eh, Conductivity, geophysical data such as Bouguer and magnetic anomaly were used. Using the constructed spatial database, the relationships between minerals deposit areas and related 36 factors were identified and quantified by frequency ratio and logistic regression models which are probabilistic and statistical model. All factors were used for mapping of regional mineral potential using overlay method in GIS environment. Then, the mineral potential map was verified using existing mineral deposit area. The verification results showed 89.53% and 92.83% in frequency ratio and logistic regression models each.

Full Text