Typing mineral deposits using their grades and tonnages in an artificial neural network

A test of the ability of a probabilistic neural network to classify deposits into types on the basis of deposit tonnage and average Cu, Mo, Ag, Au, Zn, and Pb grades is conducted. The purpose is to examine whether this type of system might serve as a basis for integrating geoscience information available in large mineral databases to classify sites by deposit type. Benefits of proper classification of many sites in large regions are relatively rapid identification of terranes permissive for deposit types and recognition of specific sites perhaps worthy of exploring further. Total tonnages and average grades of 1,137 well-explored deposits identified in published grade and tonnage models representing 13 deposit types were used to train and test the network. Tonnages were transformed by logarithms and grades by square roots to reduce effects of skewness. All values were scaled by subtracting the variable's mean and dividing by its standard deviation. Half of the deposits were selected randomly to be used in training the probabilistic neural network and the other half were used for independent testing. Tests were performed with a probabilistic neural network employing a Gaussian kernel and separate sigma weights for each class (type) and each variable (grade or tonnage). Deposit types were selected to challenge the neural network. For many types, tonnages or average grades are significantly different from other types, but individual deposits may plot in the grade and tonnage space of more than one type. Porphyry Cu, porphyry Cu-Au, and porphyry Cu-Mo types have similar tonnages and relatively small differences in grades. Redbed Cu deposits typically have tonnages that could be confused with porphyry Cu deposits, also contain Cu and, in some situations, Ag. Cyprus and kuroko massive sulfide types have about the same tonnages. Cu, Zn, Ag, and Au grades. Polymetallic vein, sedimentary exhalative Zn-Pb, and Zn-Pb skarn types contain many of the same metals. Sediment-hosted Au, Comstock Au-Ag, and low-sulfide Au-quartz vein types are principally Au deposits with differing amounts of Ag. Given the intent to test the neural network under the most difficult conditions, an overall 75% agreement between the experts and the neural network is considered excellent. Among the largestclassification errors are skarn Zn-Pb and Cyprus massive sulfide deposits classed by the neuralnetwork as kuroko massive sulfides—24 and 63% error respectively. Other large errors are the classification of 92% of porphyry Cu-Mo as porphyry Cu deposits. Most of the larger classification errors involve 25 or fewer training deposits, suggesting that some errors might be the result of small sample size. About 91% of the gold deposit types were classed properly and 98% of porphyry Cu deposits were classes as some type of porphyry Cu deposit. An experienced economic geologist would not make many of the classification errors that were made by the neural network because the geologic settings of deposits would be used to reduce errors. In a separate test, the probabilistic neural network correctly classed 93% of 336 deposits in eight deposit types when trained with presence or absence of 58 minerals and six generalized rock types. The overall success rate of the probabilistic neural network when trained on tonnage and average grades would probably be more than 90% with additional information on the presence of a few rock types.

Application of remote sensing and GIS in mineral resource mapping - An overview

Remote sensing, as a direct adjunct to field, lithologic and structural mapping, and more recently, GIS have played an important role in the study of mineralized areas. A review on the application of remote sensing in mineral resource mapping is attempted here. It involves understanding the application of remote sensing in lithologic, structural and alteration mapping. Remote sensing becomes an important tool for locating mineral deposits, in its own right, when the primary and secondary processes of mineralization result in the formation of spectral anomalies. Reconnaissance lithologic mapping is usually the first step of mineral resource mapping. This is complimented with structural mapping, as mineral deposits usually occur along or adjacent to geologic structures, and alteration mapping, as mineral deposits are commonly associated with hydrothermal alteration of the surrounding rocks. In addition to these, understanding the use of hyperspectral remote sensing is crucial as hyperspectral data can help identify and thematically map regions of exploration interest by using the distinct absorption features of most minerals. Finally coming to the exploration stage, GIS forms the perfect tool in integrating and analyzing various georeferenced geoscience data in selecting the best sites of mineral deposits or rather good candidates for further exploration.

Zinc deposits and related mineralization of the Burketown mineral field, including the world-class Century Deposit, Northern Australia: Fluid inclusion and stable isotope evidence for basin fluid sources

The stratiform Century Zn-Pb deposit and the discordant Zn-Pb lode deposits of the Burketown mineral field, northern Australia, host ore and gangue minerals with primary fluid inclusions that have not been affected by the Isan orogeny, thus providing a unique opportunity to investigate the nature of the ore-forming brines. All of the deposits are hosted in shales and siltstones belonging to the Isa superbasin and comprise sphalerite, pyrite, carbonate, quartz, galena, minor chalcopyrite, and minor illite. According to Pb model ages, the main ore stage of mineralization at Century formed at I575 Ma, some 20 m.y. after deposition of the host shale sequence. Microthermometry on undeformed, primary fluid inclusions hosted in porous sphalerite shows that the Zn at Century was transported to the deposit by a homogeneous, Ca2+- and Na+-bearing brine with a salinity of 21.6 wt percent NaCl equiv. delta D-fluid of the fluid inclusion water ranges from -89 to -83 per mil, consistent with a basinal brine that evolved from meteoric water. Fluid inclusion homogenization temperatures range between 74 degrees and 125 degrees C, which are lower than the 120 degrees to 160 degrees C range calculated from vitrinite reflectance and illite crystallinity data from the deposit. This discrepancy indicates that mineralization likely formed at 50 to 85 Mpa, corresponding to a depth of 1,900 to 3,100 m. Transgressive galena-sphalerite veins that cut stratiform mineralization at Century and breccia-filled quartz-dolomite-sphalerite-galena veins in the discordant Zn-Pb lodes have Pb model ages between 1575 and 1485 Ma. Raman spectroscopy and microthermometry reveal that the primary fluid inclusions in these veins contain Ca2+, Na+. but they have lower salinities between 23 and 10 wt percent NaCl equiv and higher delta D-fluid values ranging from -89 to -61 per mil than fluid inclusions in porous sphalerite from Century. Fluid inclusion water from sphalerite in one of the lode deposits has delta O-18(fluid) values of 1.6 and 2.4 per mil, indistinguishable from delta O-18(fluid) values between -0.3 to +7.4 per mil calculated from the isotopic composition of coexisting quartz, dolomite, and illite. The trend toward lower salinities and higher delta D-fluid values relative to the earlier mineralizing fluids is attributed to mixing between the fluid that formed Century and a seawater-derived fluid from a different source. Based on seismic data from the Lawn Hill platform and paragenetic and geochemical results from the Leichhardt River fault trough to the south, diagenetic aquifers in the Underlying Calvert superbasin appear to have been the most likely sources for the fluids that formed Century and the discordant lode deposits. Paragenetically late sphalerite and calcite cut sphalerite, quartz, and dolomite in the lode deposits and contain Na+-dominated fluid inclusions with much lower salinities than their older counterparts. The isotopic composition of calcite also indicates delta O-18(fluid) from 3.3 to 10.7 per mil, which is larger than the range obtained from synmineralization minerals, supporting the idea that a unique fluid source was involved. The absolute timing of this event is unclear, but a plethora of Pb model, K-Ar, and Ar-40/Ar-39 ages between 1440 and 1300 Ma indicate that a significant volume of fluid was mobilized at this time. The deposition of the Roper superbasin from ca. 1492 +/- 4 Ma suggests that these late veins formed from fluids that may have been derived from aquifers in overlying sediments of the Roper superbasin. Clear, buck, and drusy quartz in veins unrelated to any form of Pb-Zn mineralization record the last major fluid event in the Burketown mineral field and form distinct outcrops and ridges in the district. Fluid inclusions in these veins indicate formation from a low-salinity, 300 degrees +/- 80 degrees C fluid. Temperatures approaching 300 degrees C recorded in organic matter adjacent to faults and at sequence boundaries correspond to K-Ar ages spanning 1300 to 1100 Ma, which coincides with regional hydrothermal activity in the northern Lawn Hill platform and the emplacement of the Lakeview Dolerite at the time of assemblage of the Rodinia supercontinent.

Data-driven fuzzy analysis in quantitative mineral resource assessment

The integration of geo-information from multiple sources and of diverse nature in developing mineral favourability indexes (MFIs) is a well-known problem in mineral exploration and mineral resource assessment. Fuzzy set theory provides a convenient framework to combine and analyse qualitative and quantitative data independently of their source or characteristics. A novel, data-driven formulation for calculating MFIs based on fuzzy analysis is developed in this paper. Different geo-variables are considered fuzzy sets and their appropriate membership functions are defined and modelled. A new weighted average-type aggregation operator is then introduced to generate a new fuzzy set representing mineral favourability. The membership grades of the new fuzzy set are considered as the MFI. The weights for the aggregation operation combine the individual membership functions of the geo-variables, and are derived using information from training areas and L, regression. The technique is demonstrated in a case study of skarn tin deposits and is used to integrate geological, geochemical and magnetic data. The study area covers a total of 22.5 km(2) and is divided into 349 cells, which include nine control cells. Nine geo-variables are considered in this study. Depending on the nature of the various geo-variables, four different types of membership functions are used to model the fuzzy membership of the geo-variables involved. (C) 2002 Elsevier Science Ltd. All rights reserved.