Optimización de la estimación de concentraciones de cobre (Cu) y molibdeno (Mo) mediante kriging co-regionalizado con cokriging en contextos de geoquímica de superficie

Abstract

Abstract: The spatial estimation of geochemical concentrations is a fundamental tool in the early stages of mineral exploration, where limited sampling density and high natural variability in the data lead to uncertainty in the delineation of mineralised anomalies. The aim of this degree project was to optimise the spatial estimation of copper (Cu) and molybdenum (Mo) concentrations by applying a co-regionalised kriging (Cokriging) model and to compare its performance with that of Ordinary Kriging. To this end, 1,782 surface geochemical samples were analysed using exploratory data analysis techniques, logarithmic transformation, variographic modelling and the fitting of a Linear Co-regionalisation Model (LMC). The models were validated using cross-validation, with the root mean square error (RMSE) and the coefficient of determination (R²) as performance metrics. The results demonstrated a positive spatial correlation between Cu and Mo, allowing for the application of co-kriging. Furthermore, this method showed slightly better predictive performance than Ordinary Kriging for both variables, yielding lower errors and higher coefficients of determination.