Remote sensing as a surface water quality monitoring support in the semiarid region of Brazil

• Autor(a) principal: Fernando Bezerra Lopes
• Data de Publicação: 2013
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da UFC
• Texto Completo: http://www.teses.ufc.br/tde_busca/arquivo.php?codArquivo=11302
• topic: Ãguas continentais no semiÃrido; Sensoriamento de sistemas aquÃticos; Monitoramento dos recursos hÃdricos; Continental waters in the semiarid; Sensing of aquatic systems; Water resources monitoring; ENGENHARIA AGRICOLA

The contamination of surface water bodies due to antropic action has made water ever more scarce. Knowledge of the water quality is essential to determine instruments for it's management . Monitoring water quality in huge areas requires a high number of saimples for water quality control. This fact, allied to the high costs of water analysis, limits the evaluation that can be made of continental waters. Even though in later years geoprocessing and remote sensing techniques have been developed with important results in water quality studies, these techniques have yet to be applied in the study of aquatic systems in semiarid regions. Therefore, it was attempted to develop a methodology based on the spectral characteristics of water as a support method for the evaluation of water quality in semiarid regions. Water samples were collected in seven points from 2008 to 2010 (every two months) and in 20 points from 2011 and 2012 (every three months) . The campaigns for radiometric data acquisition occurred in 2011 and 2012, alongside the water sampling. The determinant factors and water similarity were identified by the multivariable analysis. Data from orbital and insitu remote sensing correlated with limnologic data were used. The determinant indicators for water quality in the Oros were defined mainly by the following factors: the geologic components of the soil; sediment transport by surface flow and organic pollution. The Cluster Analysis formed three distinct groups. The water similarity was defined by natural conditions and by the land use around the reservoir and along the basin. The models developed for limnologic variables, suspended inorganic solids, turbidity, transparency and electric conductivity, showed themselves to be trustworthy, indicating that these variables can be quantified remotely through remote sensing data. The models developed for the clorophile - a fitted well, indicating that this variable can be quantified accurately through in situand orbital remote sensing data. The general three banded model presented a better efficiency to that of the two band model. According to the developed model and the image use of the MERIS sensor, the Oros waters, as far as trophic state is concerned, presented 61.15% of it's water bater classified as eutrophic for the month of February, 2010. For the MERIS image of August 2011, 95.77% of it's waters were classified as eutrophic. Therefore, through remote sensing data it is possible to elaborate a water resources management at a lower cost, generating a usefull information for decision making by managers, vital for the implementation of public policies at county, regional, state and federal levels.