Diante da relevância dos recursos hídricos, é fundamental o seu monitoramento. Considerando as dimensões das bacias brasileiras, principalmente as amazônicas, torna-se viável o uso de ferramentas de sensoriamento remoto, que permitem o acompanhamento de qualidade das águas superficiais, particularmente características de cor e sólidos totais. Os componentes presentes na água alteram a sua reflectância, de acordo com suas características, permitindo um estudo e rastreio das qualidades. O trabalho teve objetivo principal identificar alterações na cor da água em rios da Amazônia Sul Ocidental, nas sub-bacias dos Rio Ji-Paraná, Canumã e Amaná, mapeando as evidências das interações e os impactos da mineração e garimpo. Por meio dos ângulos de refletância e comprimento de onda dominante, para duas sub-bacias do Rio Madeira, Rio Ji-Paraná e Canumã (controle), e sub-bacia do Trombetas Rio Amaná, e seus afluentes com mais de 30 metros de largura. Compreendendo a formação da cor das águas da bacia e a influência de atividades mineiras, e antrópicas. Como metodologia, foram utilizadas as ferramentas de sensoriamento remoto Qgis e Google Earth Engine - GEE, associadas à dados de satélite Landsat 5, 7 e 8. Códigos e rotinas foram elaborados no GEE e RStudio para obtenção das bandas RGB, ângulos e coordenadas cromáticas (hue angle), os dados após tratados no Qgis, foram elaborados cartogramas com os principais resultados de cor em escala Forel-Ule (FU). Os dados fornecidos pelo MapBiomas foram utilizados para compreender o cenário atual da mineração na região. Foram obtidas condições de solo exposto para os anos entre 1994 e 2020, por meio da equação BSI onde as áreas de solo nu podem ser distintas das áreas de floresta. Foi realizada uma análise estatística dos valores médios dos comprimentos de onda dominantes, e nos dados de material em suspensão fornecidos pela ANA, para verificar se as variações são significativas entre 1994 e 2021. Como resultados foram observadas alterações de comprimentos de onda médios (cor), para maior nas bacias do Ji-paraná, e Amaná, as séries históricas elaboradas mostraram as tendencias dos valores médios observadas para as sub- bacias, apresentando crescimento na bacia do Ji-Paraná e Amaná. A espacialização das atividades mineradoras demonstrou presença na sub-bacia do Ji-Paraná, presença substancial na bacia do Amaná, e pouco significativa na sub-bacia do Canumã. Os resultados das equações de solo exposto, demostraram um BSI considerável na bacia do Ji-Paraná que passou por um processo de consolidação da antropização no período analisado entre 1994 e 2021. Para a área do Canumã e Amaná a antropização ainda é pequena, porém presente na bacia. Por meio das análises estatísticas, ficaram demonstradas alterações significativas para maiores comprimentos médios de onda dominante na maioria dos trechos da bacia do Ji-Paraná e Amaná, enquanto que para o Rio Canumã as alterações foram para menores comprimentos e sua maioria. A metodologia e ferramentas empregadas mostraram-se adequadas para atingir o objetivo, foram observadas as alterações ocorridas nas bacias, uso e ocupação, atividades econômicas existentes, e as relações possíveis destas com as alterações das qualidades dos recursos hídricos, principalmente dos volumes de sedimentos inorgânicos presentes nos corpos d’água. | Given the relevance of water resources, their monitoring is fundamental. Considering the dimensions of Brazilian basins, especially the Amazonian ones, the use of remote sensing tools becomes viable, allowing for the monitoring of surface water quality, particularly color and total solids characteristics. The components present in water alter its reflectance according to their characteristics, enabling a study and tracking of qualities. The main objective of the work was to identify changes in water color in rivers of the Southwest Amazon, in the sub-basins of the Ji-Paraná, Canumã, and Amaná Rivers, mapping evidence of interactions and impacts of mining. Through reflectance angles and dominant wavelength, for two sub-basins of the Madeira River, Ji-Paraná and Canumã (control), and the Trombetas River Amaná sub-basin, and its tributaries with widths exceeding 30 meters. Understanding the formation of water color in the basin and the influence of mining and anthropogenic activities. As methodology, the remote sensing tools QGIS and Google Earth Engine (GEE) were used, associated with Landsat 5, 7, and 8 satellite data. Codes and routines were developed in GEE and RStudio to obtain RGB bands, angles, and chromatic coordinates (hue angle). After processing in QGIS, cartograms were developed with the main color results in Forel-Ule (FU) scale. Data provided by MapBiomas were used to understand the current mining scenario in the region. Conditions of exposed soil were obtained for the years between 1994 and 2020, through the BSI equation where bare soil areas can be distinguished from forest areas. A statistical analysis of the mean values of dominant wavelengths and suspended material data provided by ANA was performed to verify if the variations are significant between 1994 and 2021. As results, changes in average wavelengths (color) were observed, mostly increasing in the Ji-Paraná and Amaná basins. Historical series showed trends in average values observed for the sub-basins, presenting growth in the Ji-Paraná and Amaná basins. The spatialization of mining activities demonstrated presence in the Ji-Paraná sub-basin, substantial presence in the Amaná basin, and little significance in the Canumã sub-basin. The results of exposed soil equations showed a considerable BSI in the Ji-Paraná basin, which underwent a process of anthropization consolidation in the analyzed period between 1994 and 2021. For the Canumã and Amaná areas, anthropization is still small but present in the basin. Through statistical analyses, significant changes were demonstrated for higher mean dominant wavelengths in most stretches of the Ji- Paraná and Amaná basins, while for the Canumã River, the changes were mostly towards smaller wavelengths. The methodology and tools employed proved adequate to achieve the objective, observing changes in the basins, land use and occupation, existing economic activities, and their possible relationships with changes in water resource qualities, especially inorganic sediment volumes present in water bodies.

Esta dissertação avaliou a acurácia de estimativas da temperatura superficial da água (TSA) de rios e lagos utilizando dois sensores remotos, e a aplicabilidade de modelos de temperatura da água de rios em conjunto com estes dados. O trabalho foi dividido em dois capítulos principais, o primeiro focando em lagos e o segundo em rios. No Capítulo 2 foi estudada a acurácia da estimativa da temperatura superficial da Lagoa Mangueira utilizando os sensores Landsat 7 ETM+, este com aplicação de diferentes métodos, e MODIS, utilizando seus produtos MOD11 e MOD28. Verificou-se que o MOD11 e a aplicação da equação de transferência radiativa (RTE) com as imagens Landsat se mostraram as melhores opções, com REQM próximos a 1 C. Também foi verificado que esta equação é bastante sensível aos parâmetros de entrada, mas que os parâmetros de correção obtidos pelo AtmCorr são consistentes, apesar da perda de desempenho quando há alta quantidade de vapor d’água na atmosfera. Além disso, foi analisada a sensibilidade da acurácia a duas variáveis, a emissividade e os parâmetros atmosféricos, e verificados os impactos dos efeitos de superfície nas estimativas da TSA, em especial o efeito de pele fria, que resulta em uma subestimatação média da temperatura, observado na maior parte dos trabalhos neste tema. No Capítulo 3 foi analisada a consistência de séries contínuas da temperatura de um rio geradas por dois modelos de temperatura de rios, o air2stream e um modelo estocástico, calibrados com dados derivados dos sensores Landsat 7 ETM+ e MODIS, este após o uso de uma técnica de downscale. Verificou-se que é possível gerar séries contínuas e consistentes da temperatura de rios utilizando o modelo air2stream calibrado com estes dados. Contudo, esta aplicação está atrelada à acurácia da estimativa da TSA pelo Landsat 7 ETM+, que é inferior à estimativa em lagos, principalmente devido ao tamanho dos pixels e o efeito do entorno do rio na radiância captada pelo sensor devido à reflexão da água. Nesta estimativa, novamente a aplicação da RTE com AtmCorr se mostrou a melhor opção, com REQM próximo a 1,4 C. O desempenho da técnica de fusão de imagens foi bastante fraco, e os critérios de seleção desenvolvidos se mostraram cruciais para uma incorporação apenas dos dados mais acurados. Em virtude deste baixo desempenho, o uso das imagens MODIS junto com as Landsat mostrou pouca resposta pelo modelo, apesar de indicar que mais dados podem melhorar sua performance. Ainda assim, o modelo mostrou um boa performance utilizando apenas os dados do Landsat na sua calibração, e esta abordagem tem grande potencial para ser aplicada em rios sem dados medidos de temperatura, como em análise de tendências das séries temporais da temperatura e na modelagem de rios tributários para aplicação em modelos ecológicos de lagos. | This dissertation assessed the accuracy of estimates of rivers and lakes water surface temperature (WST) using two remote sensors, and the applicability of river water temperature models with these data. This work was divided in two chapters, the first focusing on lakes and the second on ivers. In Chapter 2 the accuracy of the surface temperature estimates was assessed in Lake Mangueira, using Landsat 7 ETM+, by applying different methods, and MODIS, using its products MOD11 and MOD28. Both MOD11 and the application of the radiative transfer equation (RTE) with Landsat images showed the best results, with RMSE close to 1 C. It was also verified that this equation is very sensitive to its input parameters, but the correction parameters provided by AtmCorr are consistent, despite the loss of performance when there is the water vapour in the atmosphere is high. Furthermore, the its sensibility to two variables, the emissivity and the atmospheric parameters, and the impacts of the surface effects on the accuracy of the estimates were evaluated, in particular the cool-skin effect, which results in a mean underestimation of the temperature, and was observed in most studies in this subject. In Chapter 3 the consistency of continuous series of river temperatures generated by two river temperature models, air2stream and a stochastic model, and calibrated with data derived by Landsat 7 ETM+ and MODIS, in this case after the application of a downscaling technique, were analysed. It was observed that it is possible to generate continuous and consistent series of river temperature with air2stream calibrated with these data. However, this application is dependant on the accuracy of the WST estimates using Landsat, which is inferior in rivers when compared to lakes, mainly due to the pixel size and the effect of the surrounding surfaces on the radiance measured by the sensor due to the reflection by the water surface. In this case, again the the application of the RTE with AtmCorr showed the best results, with RMSE close to 1.4 C. The image fusion technique underperformed, and the data selection criteria were crucial for the incorporation of the accurate data only. Due to this weak performance, the use of MODIS images along with Landsat showed little response by the model, although there is indication that more data can improve its performance. Still, the model showed a good performance when calibrated with Landsat-derived data, and this approach has great potential for studies in rivers with no measured water temperature data, such as trend analysis of temperature and in modelling tributary rivers temperatures as input data for lake ecological modelling.

En el presente trabajo de grado se presentan los resultados relacionados con el análisis de algunas variables asociadas a la calidad del agua como aporte en la búsqueda tecnológica de sensores y su parametrización en el marco del proyecto IoT para el análisis de calidad del agua en el humedal rio blanco para el proyecto IoT en el humedal rio blanco dentro del proyecto de investigación: “IoT para el análisis de la calidad del agua en el humedal del nacimiento del Río Blanco”. Inicialmente, se realizó una búsqueda bibliográfica sobre experiencias de análisis de parámetros fisicoquímicos en diferentes humedales, que permitiera establecer las tecnologías utilizadas para determinar la calidad del agua en humedales. Se evidencio que la tecnología más utilizada en cuanto a sensores remotos, es el uso de imágenes satelitales proporcionadas por satélite landsat esté dando mayor variabilidad de resultados. Por otra parte, se evidenciaron pocos estudios realizados con sensores conectados a Arduino para obtener valores in situ sobre la calidad del agua respecto a los parámetros fisicoquímicos. Posteriormente, se identificaron los principales indicadores fisicoquímicos asociados a los humedales para la conservación de la biodiversidad como lo son: OD, Solidos, pH, Temperatura, DBO, DQO, Nitritos y nitratos, conductividad eléctrica, turbiedad y fosforo total | In this degree work, the results related to the analysis of some variables associated with water quality are presented as a contribution in the technological search for sensors and their parameterization within the framework of the IoT project for the analysis of water quality in the wetland. Rio Blanco for the IoT project in the Rio Blanco wetland within the research project: “IoT for the analysis of water quality in the wetland at the source of the Río Blanco”. Initially, a bibliographic search was carried out on experiences of analysis of physicochemical parameters in different wetlands, which allowed to establish the technologies used to determine the quality of water in wetlands. It was evidenced that the most widely used technology in terms of remote sensors, is the use of satellite images provided by the Landsat satellite, which is giving greater variability of results. On the other hand, there were few studies carried out with sensors connected to Arduino to obtain in situ values on the quality of the water with respect to the physicochemical parameters. Subsequently, the main physicochemical indicators associated with wetlands for the conservation of biodiversity were identified, such as: DO, Solids, pH, Temperature, BOD, COD, Nitrites and nitrates, electrical conductivity, turbidity and total phosphorus. Humedale.

Trabajo presentado en el III Simposio Internacional en Ciencias del Mar (ISMS12), celebrado en Cádiz del 24 al 26 de enero de 2012. | Digital elevation models (DEM) of intertidal areas are important in a number of applications (ecological modelling, geomorphodynamics, sea-level change). However, field measurements of height profiles can be difficult to collect (either on foot or using a boat) and advanced airborne methods (such as LIDAR) can be rather expensive. We evaluate the use of freely available remote sensing (RS) images (LANDSAT 5TM) to derive a DEM for intertidal areas within Cádiz Bay. Waterlines were extracted from 12 images over a period of 18 months (2009 to 2011), representing different stages of the tide (measured in Cádiz Port). These contour lines were used to construct a DEM via kriging. RS-derived topography was compared to transects of shore height measured using a dGPS. The waterline method applied to LANDSAT 5TM images offers a cheap and quick way to derive a reasonably accurate DEM within intertidal areas. The method can be semi-automated and is generally suitable for low-energy zones, where large waves are infrequent, the topography is relatively stable and not very abrupt. | Peer reviewed

Groundwater recharge is an important metric for sustainable water management, particularly in semi-arid regions. Hard-rock aquifers underlie two-thirds of India and appropriate techniques for estimating groundwater recharge are needed, but the accuracy of such values is highly uncertain. The chloride mass balance (CMB) method was employed to estimate annual groundwater recharge rates in a monsoon-dependent area of Jaisamand Lake basin in Rajasthan, which contains the Gangeshwar watershed. A monitoring program was established within the watershed during summer 2009, with local participation for the collection of rainfall and groundwater samples. Groundwater recharge was estimated spatially over a 3-year period with pre-monsoon and post-monsoon datasets. Recharge rates estimated using the CMB method were then compared to those estimated using the water-table fluctuation (WTF) method. Specific yield was 0.63 % and assumed to be homogenous across the watershed. The average recharge rate derived from the WTF method (31 mm/year) was higher than that derived from the CMB method (24.3 mm/year). CMB recharge rates were also applied to obtain a water balance for the watershed. CMB recharge rates were used to estimate annual groundwater replenishment and were compared with estimates of groundwater withdrawal using Landsat imagery. Over the 2009–2011 study period, groundwater demand was about seven times greater than the estimated groundwater renewal of 5.6 million cubic meters. This analysis highlights the challenges associated with estimating groundwater recharge in fractured hard-rock aquifers, and how renewable groundwater-resource estimates can be used as a metric to promote sustainable water use.

The feasibility of a hydrogeological modeling approach to simulate several thousand shallow groundwater-fed lakes and wetlands without explicitly considering their connection with groundwater is investigated at the regional scale (~40,000 km²) through an application in the semi-arid Nebraska Sand Hills (NSH), USA. Hydraulic heads are compared to local land-surface elevations from a digital elevation model (DEM) within a geographic information system to assess locations of lakes and wetlands. The water bodies are inferred where hydraulic heads exceed, or are above a certain depth below, the land surface. Numbers of lakes and/or wetlands are determined via image cluster analysis applied to the same 30-m grid as the DEM after interpolating both simulated and estimated heads. The regional water-table map was used for groundwater model calibration, considering MODIS-based net groundwater recharge data. Resulting values of simulated total baseflow to interior streams are within 1% of observed values. Locations, areas, and numbers of simulated lakes and wetlands are compared with Landsat 2005 survey data and with areas of lakes from a 1979–1980 Landsat survey and the National Hydrography Dataset. This simplified process-based modeling approach avoids the need for field-based morphology or water-budget data from individual lakes or wetlands, or determination of lake-groundwater exchanges, yet it reproduces observed lake-wetland characteristics at regional groundwater management scales. A better understanding of the NSH hydrogeology is attained, and the approach shows promise for use in simulations of groundwater-fed lake and wetland characteristics in other large groundwater systems.

Vertical leakage (discharge to upper aquifers) is an important but poorly constrained component of water balance in the Great Artesian Basin (GAB), Australia. It ranges from negligible discharge where the GAB is overlain by aquitards, to high discharge where artesian water feeds the shallow unconfined aquifer (thereby raising the water table) causing elevated surface soil moisture and extensive surface salinisation. Adequately representing the temporal and spatial variability of vertical leakage is difficult due to the large scale over which the discharge occurs. An innovative method is presented that integrates a supervised classification of high-discharge zones using time-series Landsat data with landform mapping information to improve classification results. ‘Wetness persistence’ and ‘salt persistence’ classes, determined from the time series data, are related to groundwater discharge processes through a discharge framework that allows scaling up of field-based discharge estimates. The results show that using multi-image classification integrated with landform data will significantly reduce uncertainty by reducing false positives. No significant temporal trends were found in a time series assessment, with results featuring high variability, most likely due to image normalisation issues. The lack of a clear temporal signal suggests that an assumption of steady-state discharge is valid for estimating annual fluxes of vertical leakage. Supervised classification and landform outputs provide updated knowledge on GAB vertical leakage rates by providing useful lower and upper bounds of discharge rates respectively. Additionally, groundwater-dependent ecosystem classification, covering the full extent of the basin margins, is a new source of information resulting from the work.