The objective of the research reported in this thesis was to develop a technique to monitor the dynamics of sediments and nutrients entering the coastal ocean with river plumes associated with high intensity low frequency events (e.g. floods), using ocean colour remote sensing. To achieve this objective, an inverse bio-optical model was developed, based on analytical and empirical relationships between concentrations of optically significant substances and remote sensing of water-leaving radiance. The model determines concentrations of water-colouring substances such as chlorophyll, suspended sediments, and coloured dissolved organic matter, as well as the values of optical parameters using water-leaving radiances derived from the Sea-viewing Wide Field-of-view Sensor (SeaWiFS). To solve atmospheric correction in coastal waters, the aerosol type over clear waters is transferred to adjacent turbid water pixels. The vicinity of the Herbert River, central Great Barrier Reef zone, Australia, was used as a case study for the application of the algorithm developed. The satellite ocean colour technique was successfully validated using sea-truth measurements of water-colouring constituents acquired in the area during various seasons throughout 2002-2004. A high correlation between chlorophyll and dissolved organic matter was found in the coastal waters of the region, and when the bio-optical model was constrained to make chlorophyll a function of dissolved organic matter, the relationship between in situ and satellite-derived data was substantially improved. With reliable retrieval of the major water-colouring constituents, the technique was subsequently applied to study fluxes of particulate and dissolved organic and inorganic matter following a flood event in the Herbert River during the austral summer of 1999. Extensive field observations covering a seasonal flood in the Herbert River in February 2004 revealed high sediment and nutrient exports from the river to the adjacent coastal waters during the flood event. Due to rapid settling, the bulk of the sediment-rich influx was deposited close inshore, while the majority of nutrients exported from the river were consumed by phytoplankton in a relatively small area of the coastal ocean. With the help of ocean colour remote sensing, it was demonstrated that river-borne sediments and nutrients discharged by a typical flood in the Herbert River are mostly precipitated or consumed within the first 20 km from the coast and therefore are unlikely to reach and possibly affect the midshelf coral reefs of this section of the Great Barrier Reef lagoon.

The objective of the research reported in this thesis was to develop a technique to monitor the dynamics of sediments and nutrients entering the coastal ocean with river plumes associated with high intensity low frequency events (e.g. floods), using ocean colour remote sensing. To achieve this objective, an inverse bio-optical model was developed, based on analytical and empirical relationships between concentrations of optically significant substances and remote sensing of water-leaving radiance. The model determines concentrations of water-colouring substances such as chlorophyll, suspended sediments, and coloured dissolved organic matter, as well as the values of optical parameters using water-leaving radiances derived from the Sea-viewing Wide Field-of-view Sensor (SeaWiFS). To solve atmospheric correction in coastal waters, the aerosol type over clear waters is transferred to adjacent turbid water pixels. The vicinity of the Herbert River, central Great Barrier Reef zone, Australia, was used as a case study for the application of the algorithm developed. The satellite ocean colour technique was successfully validated using sea-truth measurements of water-colouring constituents acquired in the area during various seasons throughout 2002-2004. A high correlation between chlorophyll and dissolved organic matter was found in the coastal waters of the region, and when the bio-optical model was constrained to make chlorophyll a function of dissolved organic matter, the relationship between in situ and satellite-derived data was substantially improved. With reliable retrieval of the major water-colouring constituents, the technique was subsequently applied to study fluxes of particulate and dissolved organic and inorganic matter following a flood event in the Herbert River during the austral summer of 1999. Extensive field observations covering a seasonal flood in the Herbert River in February 2004 revealed high sediment and nutrient exports from the river to the adjacent coastal waters during the flood event. Due to rapid settling, the bulk of the sediment-rich influx was deposited close inshore, while the majority of nutrients exported from the river were consumed by phytoplankton in a relatively small area of the coastal ocean. With the help of ocean colour remote sensing, it was demonstrated that river-borne sediments and nutrients discharged by a typical flood in the Herbert River are mostly precipitated or consumed within the first 20 km from the coast and therefore are unlikely to reach and possibly affect the midshelf coral reefs of this section of the Great Barrier Reef lagoon.

Submitted by STELA ROSA AMARAL GONÇALVES (stelinhaamaral@hotmail.com) on 2016-08-12T03:20:24Z No. of bitstreams: 1 Dissertação_final_imai_v3.pdf: 3555316 bytes, checksum: b4386413378a1a29e513286831a880a2 (MD5) / Approved for entry into archive by Ana Paula Grisoto (grisotoana@reitoria.unesp.br) on 2016-08-15T14:30:47Z (GMT) No. of bitstreams: 1 goncalves_sra_me_prud.pdf: 3555316 bytes, checksum: b4386413378a1a29e513286831a880a2 (MD5) / Made available in DSpace on 2016-08-15T14:30:47Z (GMT). No. of bitstreams: 1 goncalves_sra_me_prud.pdf: 3555316 bytes, checksum: b4386413378a1a29e513286831a880a2 (MD5) Previous issue date: 2016-07-08 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) / Informações quantitativas e qualitativas de sistemas aquáticos podem ser obtidas por meio de dados de sensoriamento remoto. Nesses sistemas, a porção fotoativa do carbono orgânico dissolvido (COD), pode apresentar correlação com o Carbono Orgânico Dissolvido Colorido (CDOM) que é responsável pela alteração da cor da água e consequentemente pode alterar processos de fotossíntese e fotodegradação. Sendo assim o presente trabalho teve como objetivo avaliar a variabilidade espacial das concentrações de CDOM de uma série temporal, de imagens do sensor Operational Land Imager OLI a bordo do satélite Landsat 8, ao longo de um ano no reservatório de Barra Bonita/SP, sendo este o primeiro de uma série de reservatórios em cascata do Rio Tietê, no estado de São Paulo. Foram encontradas correlações entre a concentração laboratorial do COD e o CDOM para o campo de maio o que motivou a modelagem. Foram ajustados modelos inversos, de banda simples e de razão de bandas, para predição de CDOM com dados de dois trabalhos de campo realizados no ano de 2014, sendo o primeiro no mês de maio e o segundo no mês de outubro. Além disso, foram testados modelos encontrados na literatura, também com os dados levantados em campo, para estimativa de CDOM. Os resultados obtidos mostraram que, para os modelos bio-ópticos, de banda simples e razão de bandas, não há diferenças estatistica. O melhor modelo foi selecionado baseado em análise de erro, depois da aplicação dos modelos na imagem OLI coincidente com o trabalho de campo. Uma série de imagens do sensor OLI no período de um ano foi adotada tendo em vista da disponibilidade dos dados gratuitos e resolução radiométrica de 12 bits. A série temporal de imagens do sensor OLI, após ser processada para minimizar os efeitos atmosféricos, permitiu mostrar a dinâmica de CDOM no reservatório em conjunto com dados de precipitação e temperatura do Instituto Nacional de Meteorologia (INMET). Foi possível observar que o reservatório apresenta uma tendência de valores maiores de CDOM após aportes de nutrientes causados por eventos de precipitação. A temperatura apesar de influenciar na velocidade da reação de degradação não foi o fator principal para a variação da concentração de CDOM. O ajuste e avaliação de modelos foi essencial para escolha do modelo mais adequado a realidade local e sazonal do escopo do trabalho. Além disso o modelo aplicado pode contribuir para o cálculo do balanço de carbono, como também em outros estudos que necessitem dessa informação. / Quantitative and qualitative information about aquatic systems can be obtained by using the Remote Sensing techniques. In these aquatic systems, the photoactive portion of Dissolved Organic Carbon (DOC) can be correlated with the Colored Dissolved Organic Matter (CDOM), responsible for the water color alteration and can modify the photosynthesis and photodegradation process. In this sense, this project aimed to evaluate the spatial variability of CDOM concentrations in a one-year temporal analysis from images acquired by the Operational Land Imager OLI sensor onboard the satellite Landsat 8. The study site is the reservoir located in Barra Bonita, State of São Paulo, which is the first in a series of cascading reservoirs from Tietê River in State of São Paulo. The results showed correlations between the COD and CDOM concentrations for the data acquired in May 2014, motivating the CDOM modeling. Inverse models were adjusted, by using simple spectral band and band ratio, for CDOM prediction using data acquired from two field works carried out in 2014, the first in May and the second in October. Moreover, the models found in related literature were tested, also with the data collected in field, for the CDOM estimative. The obtained results showed that for the bio-optical models, from simple spectral band and band ratio, there is no statistic differences. The model who best fitted to the CDOM estimation was selected based on error analysis after their application in OLI images, which coincided with the field work. One year OLI images were used considering their availability, without fees, and their radiometric resolution of 12 bits. This temporal series, after being processed to attenuate the atmospheric effects, allowed showing the CDOM dynamics in the Barra Bonita reservoir combined with temperature and precipitation data from the Instituto Nacional de Metereologia (INMET). It was possible observe that the reservoir shows a bias of higher values of CDOM after nutrients intakes caused by precipitation. The temperature, despite of its influence on degradation reaction velocity, was not the principal factor for the CDOM concentration variation. Finally, the adjustment and models validation was essential for the correct choice of the model which better represents the local and season reality. In addition, the applied model can contribute for the carbon balance calculation, as well for other researches, which use this information.

Orientador: Nilton Nobuhiro Imai / Abstract: The acquisition of data using Remote Sensing and in-situ sampling allows several data sources to be integrated for the analysis and observations of environmental characteristics and may require computational methods to support the data processing, exploration and analysis. The need to integrate data from different sources is highlighted in studies of dynamic and complex environments that frequently change, such as hydroelectric reservoirs. Reservoirs are artificial ecosystems, which influence directly the regional characteristics, mainly because of their multipurpose use. The interactions of the electromagnetic energy with the optically active components occur along the entire water column, so that the behavior of the light field reflects the changes applied along the entire euphotic zone. However, the values taken from images are used accordingly to a plane and associated with the respective point or area of surface. The calibration of bio-optical models considering only the surface sampling data can not deliver fully effective results because the electromagnetic radiation interacts with the components located along the water column and the response captured by the sensors does not only represent the value associated with the surface. Considering this scenario, this work proposes an investigation on the influence of the vertical distribution of the optical properties along the water column, in order to contemplate records about the interaction in different levels of depth, b... (Complete abstract click electronic access below) / Resumo: A aquisição de dados por meio da combinação de Sensoriamento Remoto e amostragens in-situ permite que várias fontes de dados sejam integradas para a análise e observação de características do alvo de interesse e pode exigir métodos computacionais para apoiar o processamento, exploração e análise de dados. A necessidade de integrar dados de diferentes fontes é destacada em estudos de ambientes dinâmicos e complexos que se alteram frequentemente, como os reservatórios hidrelétricos. Os reservatórios são ecossistemas artificiais, que influenciam diretamente nas características regionais, principalmente devido ao seu uso múltiplo uso. As interações da energia eletromagnética com os componentes opticamente ativos ocorrem ao longo de toda a coluna d’água, de modo que o comportamento do campo de luz reflete as mudanças aplicadas ao longo da zona eufótica. No entanto, as grandezas registradas nas imagens são usados de acordo com o plano e limitadas ao respectivo ponto ou área da superfície. A calibração de modelos bio-ópticos, considerando apenas os dados de amostragem da superfície, pode não fornecer resultados totalmente eficazes, porque a radiação eletromagnética interage com os componentes localizados ao longo da coluna de água e, consequentemente, a resposta capturada pelos sensores não representa apenas o valor associado à superfície. Este trabalho propõe uma investigação sobre a influência da distribuição vertical das propriedades ópticas ao longo da coluna d’água, a fim de co... (Resumo completo, clicar acesso eletrônico abaixo) / Doutor

Indiana University-Purdue University Indianapolis (IUPUI) / Cyanobacterial blooms are currently one of the most important issues faced by environmental agencies, water authorities and public health organizations. Remote sensing provides an advanced approach to monitor cyanobacteria by detecting and quantifying chlorophyll-a (Chl-a) and phycocaynin (PC). In this thesis, an analytical bio-optical model, more typically applied to ocean waters, was modified to accommodate the complexity of inland waters. The newly developed models work well to estimate inherent optical properties, including absorption and backscattering coefficients, in eight different study sites distributed around the globe. Based on derived absorption coefficients, Chl-a and PC concentrations were accurately retrieved for data sets collected annually from 2006 to 2010, and the estimation accuracy exceeded that of currently used algorithms. An important advantage of the model is that low concentrations of Chl-a and PC can be predicted more accurately, enabling early warning of cyanobacterial blooms. In addition, the results also indicated good spatial and temporal transferability of the algorithms, since no specific calibration procedures were required for data sets collected in a different sites and seasons. The compatibility of the newly developed algorithm with MERIS spectra provides the possibility for routine surveillance of cyanobacterial growth in inland waters.