DETERMINATION OF THE ORIGIN OF SEDIMENTS IN IBIRITÉ LAKE WATERSHED, MINAS GERAIS, USING PROXIMAL SENSING
Soil erosion, fingerprinting, soil spatial variability
The identification of sediment source areas in urbanized watersheds is a challenge for the environmental management of reservoirs subject to siltation, due to the complexity of erosion and transport processes. In this study, the sediment fingerprinting approach was applied to determine the origin of sediments deposited in the Ibirité Lagoon, Minas Gerais, Brazil, integrating data obtained from proximal sensing, particle-size analyses, and statistical clustering techniques. A total of 82 soil samples from the 0–2 cm layer, distributed across the watershed, and 15 sediment samples collected from the reservoir bottom were analyzed. Chemical composition was determined by portable X-ray fluorescence spectrometry (pXRF), magnetic susceptibility was measured at low and high frequencies using a magnetic susceptibilimeter, and texture was assessed by the pipette method. Principal component and cluster analyses allowed the identification of three distinct sediment groups, associated with different hydrosedimentological processes: (i) sandy inputs related to higher-energy hydrological events, (ii) predominantly silty sediments resulting from mixing processes during transport, and (iii) fine sediments enriched in clay, Fe, and associated elements, dominant in the depositional environment. The spatial extrapolation of the clusters, integrated with land use and land cover information and hydrosedimentological connectivity, enabled the delineation of watershed areas with higher sediment contribution potential. The results demonstrate that the integration of proximal sensing and multivariate analyses constitutes a satisfactory, rapid, and non-destructive approach for sediment tracing in urbanized watersheds, providing support for the planning of siltation mitigation actions and the environmental management of reservoirs.