Water bodies pollution sources apportionment and their spatiotemporal profiling captured by a chemometrics scalar field theory framework

Pollution in aquatic environments arises from multiple sources and involves a wide spectrum of contaminants, with emerging organic compounds being of particular concern. Here, we introduce a three-step chemometrics modelling workflow for processing discrete spatiotemporal monitoring data. First, LC-HRMS data is analyzed using the Regions of Interest–Multivariate Curve Resolution (ROIMCR) approach and annotated with the MSident framework. Second, pollution source apportionment is achieved, and their spatiotemporal profiles are resolved through Multivariate Curve Resolution–Alternating Least Squares (MCR-ALS) trilinear factorization. Finally, source dynamics are modeled within a matrix framework analogous to the Klein–Gordon scalar field equation. This formulation, consistent with network topology, enables decomposition of the field energy into kinetic (temporal), gradient (spatial), and internal (chemical) components. Furthermore, Fourier spectral analysis provides an alternative representation of the pollution sources in the frequency domain as a superposition of spatial and temporal modes. The approach is demonstrated on a contaminated alluvial aquifer in the Besòs River Delta (Barcelona, NE Spain) and is broadly applicable to discrete monitoring datasets.