Fuzzy K-Means-Based Geochemical Modeling at the Catchment Basin Scale

Geochemical anomaly mapping is essential in mineral exploration, but point-based analyses often fail to capture spatial dispersion and uncertainty in surface geochemical data. This study applies a catchment basin-based approach to identify geochemical anomalies using an unsupervised fuzzy clustering framework. Principal component analysis was employed to reduce data dimensionality and isolate exploration-relevant elements from lithogenic background signals. A selected geochemical matrix was then classified using fuzzy k-means clustering, allowing each catchment basin to belong to multiple anomaly classes with varying membership degrees. This approach better reflects the gradual transition between background and anomalous geochemical signatures. The resulting anomaly map classifies catchment basins into background, weak anomaly, and strong anomaly zones. Strong anomaly basins show clear spatial correspondence with known mineral occurrences, indicating the effectiveness of the proposed method in highlighting mineralized environments. Basin-based modeling also suppresses local noise and enhances regional geochemical patterns.

کلیدواژه ها:

Geochemical Anomaly Detection ، Fuzzy K-Means ، Principal Component Analysis (PCA) ، Catchment-Based Geochemical Analysis