Avaliação do sedimento e nutrientes aportados em um reservatório no semiárido

• Ano de defesa: 2019
• Autor(a) principal: Pereira, Erich Celestino Braga
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Assoreamento; Espectroscopia Vis-NIR; Gestão hídrica; Silting; Spectroscopy Vis-NIR; Water management
• Link de acesso: http://www.repositorio.ufc.br/handle/riufc/50824

The surface reservoir is one of the main instruments for management of water resources in Brazilian semi-arid, which retains sediments and nutrients together with water storage. Knowing the sediment and nutrients impacts over water quality, the periodic monitoring is necessary, presenting the remote sensing associated with mathematical models as efficient alternatives for estimating the dredged sediment constituents. Thus, we aimed evaluate the distribution of sediment and nutrients along the hydraulic basin of Desterro Reservoir, evaluating whether this reservoir acts as a sink or disperser of nutrients associated with the reservoir. Another objective was evaluate the performance of the partial least squares regression method (PLSR) in estimating sediment variables. The study was carried out in the Desterro Reservoir, Caridade, Ceará, in March 2018, when the reservoir presented itself in the null storage volume. During the study we collected sediment and riverbed material samples disturbed and undisturbed at intervals of 20 cm intervals in seven trenches (T1 to T7) digged in the reservoir riverbed. We evaluated in this study: total organic carbon (TOC), total nitrogen (NT), available phosphorus (P), total iron (TFe), exchangeable aluminium (Al), pH, sand, clay and silt concentration, particle density and porosity. We applied Pearson's correlation to evaluate the relationships between the studied variables, as well as cluster analysis and parametric and non-parametric means tests. We got the spectral responses using hyperspectral sensor ASD FieldSpec 3, in the range between 350 and 2500 nm, and estimated the COT, NT, sand, silt and clay by the partial least squares regression method (PLSR), with selected bands from Stepwise regression method. The sediment presented high variability along the hydraulic basin, with an increase in the depth and concentration of fine sediments (silt and clay) from the entrance of the watercourses in the reservoir towards the dam. TOC and NT showed a high correlation with the fine sediments (r = 0.848 and r = 0.780, respectively), and the inverse relation with the available P (r = -0.519). Thus, we concluded fine sediments and organic matter retained largely the nutrients contributed, serving as a sink. The C/N ratio shows an average value of 8.77 demonstrate the predominance of non-vascularized photosynthetic organisms in the mineralization of these nutrients to the sediment. The sediment reflectance showed and average value 10% lower than the reflectance of the riverbed samples. The models were evaluated as suitable for TOC (r 0.97, Ra2 0.94 and RPD de 6.70), for NT (r 0.98, Ra2 0.96 and RPD 2.59), for clay (r 0.98, Ra2 0.96 and RPD 10.79), for silt (r 0.97, Ra2 0.94 and RPD 6.29) and for sand (r 0.97, Ra2 0.93 and RPD 11.25). For the P (r 0.96, Ra2 0.91 and RPD de 0.17) and for TFe (r 0.96, Ra2 0.91 and RPD 0.76), the models were classified as unreliable.