Abstract
Continental brine deposits hosted in salar-type systems of the Central Andean Lithium Triangle (Argentina, Bolivia, and Chile) account for more than 60% of known global lithium resources, making the accurate spatial characterization of lithium concentration a matter of paramount importance for resource evaluation and investment decision-making. However, the pronounced geometric anisotropy of these deposits, characterized by large lateral extents at the kilometer scale and limited vertical thicknesses on the order of tens of meters, imposes severe constraints on the reliable inference of spatial continuity, and consequently accurate estimation of lithium concentration in these deposits under traditional geostatistical frameworks. This study evaluates a novel methodological framework based on transitive kriging as an alternative to conventional ordinary kriging for the estimation of lithium concentration in a brine deposit of the Lithium Triangle. Applied to a dataset of 332 composite brine samples from 60 irregularly distributed boreholes, the transitive approach yielded a markedly more stable experimental covariogram in the vertical direction than does the conventional approach. Leave-one-out cross-validation demonstrated that the transitive approach reduced the root mean square error by approximately 13% and achieved a higher coefficient of determination. Furthermore, the smoothing effect, an inherent limitation of kriging-based estimators, was significantly less pronounced under the transitive approach, yielding a more accurate assessment of the recoverable resources above economically relevant cut-off concentrations. These results demonstrate that transitive kriging constitutes a technically sound and practically valuable framework for lithium resource evaluation in geometrically constrained salar-type deposits.