Determinação da origem dos sedimentos na bacia hidrográfica da Lagoa Ibirité, Minas Gerais, utilizando sensores proximais

Abstract

Soil erosion and the consequent production and transport of sediments represent one of the main challenges for the environmental management of urbanized watersheds, particularly in lacustrine systems and reservoirs subject to siltation. Identifying sediment sources is therefore essential for understanding hydrosedimentological dynamics and for supporting soil and water conservation planning. This study aimed to apply the sediment fingerprinting approach, integrated with proximal sensing techniques and multivariate statistical analyses, to investigate the origin of sediments deposited in Ibirité Lake, Minas Gerais, Brazil. Surface soil samples distributed throughout the watershed and sediment samples collected from the reservoir bottom were analyzed. Elemental composition was determined by X-ray fluorescence spectrometry, magnetic susceptibility was measured using a magnetic susceptibility meter, and particle-size distribution was assessed using the pipette method. The resulting dataset was integrated using cluster analysis techniques applied to both soils and sediments, in order to identify similarity patterns and associations between sediment source areas and deposited materials. The results highlighted the heterogeneity of soils within the watershed and the occurrence of sediment groups with contrasting chemical and magnetic characteristics, reflecting different transport and deposition processes. By integrating clustering results with spatial information on land use and land cover, sectors of the watershed with higher potential for sediment contribution to the reservoir were identified. Overall, the study demonstrates that the combination of proximal sensing data and multivariate analyses represents an efficient, rapid, and non-destructive approach for investigating sediment provenance in urbanized watersheds. Keywords: Sediment fingerprinting; pXRF; Magnetic susceptibility; Multivariate analysis.