A Geostatistical Study of a Fuzzy-Based Dataset from Airborne Magnetic Particle Biomonitoring

Abstract

Airborne magnetic particles (AMPs) are associated with potentially toxic elements, and their size, mineralogy, and concentration can significantly impact both the environment and human health. However, their spatial analysis is often limited by small datasets, non-normality, and pronounced local variability. In this work, two sites with distinct demographic and geographic characteristics, the city of Mar del Plata (Argentina) and the Aburrá Valley region (Colombia), were analyzed using the fuzzy Magnetic Pollution Index (IMC) as an indicator of the concentration of AMPs. Moreover, an original methodological framework that explicitly incorporates measurement uncertainty through fuzzy numbers, combined with an approach modeling fuzzy semivariances via α-cuts, performs spatial prediction via ordinary kriging. This study produces maps that simultaneously reflect the magnitude of IMC and its associated uncertainty. Unlike classical geostatistics, the fuzzy-based model captures the inherent imprecision of magnetic measurements and reveals spatial patterns where uncertainty becomes informative about the type and origin of pollution. In particular, this approach demonstrates that areas with higher IMC levels are associated with high anthropic activity (near industrial zones, main avenues, slow traffic). In contrast, lower values were found in residential areas. Overall, the fuzzy-driven approach provides an additional layer of information not accessible through traditional methods, improving spatial interpretation and supporting the identification of priority areas for environmental monitoring.