Multilevel and structural equation modeling approach to identify spatiotemporal patterns and source characterization of metals and metalloids in surface water and sediment of the Ctalamochita River in Pampa region, Argentina

Abstract

Simple models capable of describing spatial and temporal patterns of pollutants are needed to understand surface water quality and to identify the main processes that influence the fate and transport of pollutants. This study combines exploratory factor analysis with multilevel and structural equation modeling to: find spatial and temporal patterns of metal and metalloids in the Ctalamochita River Basin (Argentina); identify the processes responsible for the occurrence of these elements; and, establish relationships between latent variables of different aquatic compartments. For this purpose, the metal and metalloid concentration in different matrices was analyzed at sampling sites along 330 km of the river course, including seasonal variations. Our results indicate that the concentration of natural occurring elements in surface water increased downstream the catchment and during the warm-rainy season, as a consequence of surface runoff and groundwater discharge. In contrast, the spatial and seasonal variation in sediment showed an inverse pattern (negative correlation) compared to surface water. On the other hand, copper, zinc and lead were derived from anthropogenic sources, such as old mining areas and metal industries. Moreover, the input of these metals into the river evidenced a direct effect of surface water on sediment. Overall, the combined use of multilevel model and SEM allowed for a proper description of spatial and temporal patterns of metals and metalloids in the catchment, and successfully highlighted sources and processes influencing metal transport. The proposed methodological framework could be extended to wider spatial scales including nested regions, and could be applied to environmental studies dealing with repeated observations and multiple matrices.