<p>Calibrating airborne geophysical data with in situ petrophysical measurements can significantly enhance the accuracy and interpretability of these datasets. In situ petrophysical measurements act as a ground truth process for airborne geophysical data, facilitating the correction of discrepancies, reducing noise, and resolution mismatches. This paper introduces an innovative framework for calibrating airborne radiometric and apparent magnetic susceptibility data using in situ petrophysical measurements, followed by integrating these calibrated datasets into a predictive mineral prospectivity mapping workflow. The framework leverages XGBoost random forest, a hybrid machine learning algorithm, to perform dual tasks: (1) regression modeling for calibrating airborne data using in situ measurements, and (2) classification modeling to generate mineral prospectivity maps. Beyond calibration, a suite of complementary geophysical predictor layers was developed and integrated into a predictive model for gold mineralization in southwestern New Brunswick. The resultant prospectivity map exhibits a strong spatial correlation between high-probability zones and known mineral occurrences, identifying several promising exploration targets. Model validation metrics underscore the framework’s robustness, with an accuracy of 0.973 and an area under the curve of 0.995, confirming its predictive reliability. Analysis of feature importance revealed that structural complexity and radiometric maps were the most influential predictors in the model. Also, a confidence-weighted prospectivity map was proposed using the bootstrap variance of predictions in order to prioritize the high-potential zones. This integrated methodology offers a robust, scalable, and data-driven approach to mineral exploration which supports the identification of prospective zones.</p>