Search

Geoscientists have extensively used machine learning for geological mapping and exploring the mineral prospect of a province. However, the interpretation of results becomes challenging due to the complexity of machine learning models. This study uses a convolutional neural network (CNN) and Shapley additive explanation (SHAP) to estimate potential locations for gold mineralisation in Rengali Province, a tectonised mosaic of volcano-sedimentary sequences juxtaposed at the interface of the Archaean cratonic segment in the north and the Proterozoic granulite provinces of the Eastern Ghats Belt in Eastern India. The objective is to integrate multi-thematic data involving geological, geophysical, mineralogical and geochemical surveys on a 1:50 K scale with the aim of prognosticating gold mineralisation. The available data utilised during the integration include aero-geophysical (aeromagnetic and aerospectrometric), geochemical (national geochemical mapping), ground geophysical (gravity), satellite gravity, remote sensing (multispectral) and National Geomorphology and Lineament Project structural lineament maps obtained from the Geological Survey of India Database. The CNN model has an overall accuracy of 90%. The SHAP values demonstrate that the major contributing factors are, in sequential order, antimony, clay, lead, arsenic content and a magnetic anomaly in CNN modelling. Geochemical pathfinders, including geophysical factors, have high importance, followed by the shear zones in mineralisation mapping. According to the results, the central parts of the study area, including the river valley, have higher gold prospects than the surrounding areas. Gold mineralisation is possibly associated with intermediate metavolcanics along the shear zone, which is later intruded by quartz veins in the northern part of the Rengali Province. This work intends to model known occurrences with respect to multiple themes so that the results can be replicated in surrounding areas.

The irregularly spaced assay values of lead and zinc in different boreholes of Rampura-Agucha Pb-Zn deposit (Rajasthan), have been interpolated at equal intervals of one meter using cubic splines to obtain a constant lag, discrete spatial (time) series. The time series has been used subsequently for stochastic modelling of assay values along the drill holes. Univariate stationary models under the integrated autoregressive moving average scheme - ARIMA (p,d,q) - were established for each assay series along four borehole based on statistical analyses of their auto (acf) and their partial (pacf) correlation functions. These models can be used for univariate forecasting of assay values at different distances. Among the candidate ARIMA (p,d,q) models, ARIMA (1,1,1) and ARIMA (1,1,0) were found to fit spatial distributions of Zn and Pb, respectively, using the principle of parsimony. Though linear stochastic models were developed for Rampura - Agucha Pb-Zn deposit, a brief review of the scope and use of non-linear stochastic model for modelling of complex ore mineralisation has also been attempted. Geochemical interpretation has been carried out on the petrochemical data of the amphibolites interlayered with the footwall and the hanging wall rocks using partial correlation coefficients between various oxides. This has led to comprehensive characterisation of the geochemistry and mineralogy of the amphibolites.

Multivariate canonical correlation analysis has been carried out taking physical variables (mean, standard deviation, skewness, kurtosis) as predictors and chemical variables (Cd, Co, Cr, Cu, Fe, Mn, Ni, Pb, and Zn) as responses from the soils and sediments of Thane Creek and Ulhas River areas in Bombay, India. Soil samples provide the first canonical correlation to the presence of more clay fractions, which controls the concentration of trace elements such as Co, Fe, and Ni. In sediment samples, the first canonical correlation explains the role of mean particle size in controlling the concentration of Pb and Ni. The second correlation shows the role of clay minerals in controlling the concentration of the trace elements such as Fe and Zn. The plot of transformed scores of first canonical correlation for soil illustrate the high correlation between sets of variables as all points are grouped closely within an ellipsoidal field. The plot of transformed scores of first canonical correlation illustrate that there is a clear distinction between the type of sediments collected from Thane creek and the Ulhas river region.