Le riz est la principale denrée de plus de la moitié de la population mondiale et joue un rôle particulièrement important dans l'économie mondiale, la sécurité alimentaire, la consommation d'eau, et le changement climatique. L'objectif de cette thèse consistait à développer des méthodes pour le suivi du riz basées sur des données Sentinel-1 ainsi qu'a utiliser les produits de cartographie obtenus dans diverses applications portant sur la sécurité alimentaire et l'environnement mondial. Plus spécifiquement, l'étude a pour but de fournir des outils pour observer la culture du riz, en produisant la cartographie des surfaces cultivées, celle des stades phénologiques de la plante comprenant le début de la saison, celle des deux principales catégories de variétés de riz à cycle court et cycle long, la hauteur de la plante, et la carte annuelle du nombre de récoltes de riz par an. Ces informations sont nécessaires à l'estimation de la production du riz, et à la gestion des cultures à l'échelle régionale. Nous étudions aussi l'intégration des produits ainsi développés dans un modèle de processus destinés à estimer le rendement du riz, et un modèle permettant la dérivation de l'émission du méthane et le volume d'eau nécessaire à la culture. La région test est l'une des régions rizicoles majeures à l'échelle mondiale, qui est le Delta du Mékong, au Vietnam. Cette région est caractérisée par une grande diversité de pratiques agricoles, du nombre de cultures du riz par an, et dans les calendriers des récoltes. La première phase du travail est la compréhension de la variation temporelle des valeurs de rétrodiffusion radar de Sentinel-1, en polarisation VH et VV. Pour cela, des données de terrain ont été collectées sur 60 champs, sur 5 saisons de riz pendant 2 ans. Les variations temporelles des mesures radar ont été interprétées en fonction de la croissance des plantes le long des stades phénologiques. Les mêmes courbes caractéristiques observées lors des 5 saisons ont suggéré l'utilisation d'une courbe 'type' dans le développement des méthodes pour fournir les produits requis. Les résultats obtenus sur le Delta du Mékong ont été validés à l'aide des données terrain de référence, et sont très satisfaisants : 98% de précision pour la carte riz/non riz, une RMSE de 4 jours pour la date de semis, une RMSE de 0.78 cm pour la hauteur de plante, 91,7% de précision pour la distinction entre deux types de riz (cycle court et cycle long), et 98% de précision sur l'estimation du stade phénologique. Enfin, nous avons évalué l'utilisation de ces produits issus de données Sentinel-1 dans le modèle ORYZA2000 destiné à estimer le rendement du riz, et dans le modèle DNDC destiné à estimer le volume d'eau nécessaire à la culture, ainsi que l'émission de méthane par les rizières. Les résultats, préliminaires, montrent le bon potentiel de l'approche pour fournir le rendement, le bilan d'eau, et les taux d'émission de méthane sur les champs de riz considérés. Cette approche permettrait de faire des analyses de sensibilité, par exemple pour optimiser la gestion d'irrigation afin de réduire la consommation d'eau et l'émission de méthane, tout en préservant le rendement du riz. Ces travaux, qui démontrent le potentiel des données Sentinel-1 pour le suivi du riz à large échelle, seront à compléter afin de réaliser des applications effectives opérationnelles. Il s'agira de renforcer les méthodes et de les tester sur différents systèmes rizicoles, et de poursuive l'étude sur l'intégration de ces produits de télédétection dans les modèles permettant d'évaluer la productivité, les besoins en eau et les émissions des gaz à effet de serre des rizières. / Rice is the primary staple food of more than half of world’s population and plays an especially important role in global economy, food security, water use and climate change. The objective of this thesis was to develop methods for rice monitoring based on Sentinel-1 data and to effectively use the mapping products in various applications concerning food security and global environment. Specifically, the study aims at providing tools for observation of the rice cultivation systems, by generating products such as map of rice planted area, map of rice start-of-season and phenological stages, and map of rice crop intensity, together with rice crop parameters such as category of rice varieties (long or short cycle), and plant height. The information to be provided is necessary for the estimation of crop production, and for the management of rice ecosystems at the regional scale. We also investigated on how the products derived from EO Sentinel-1 data can be integrated in process-based models for rice production estimation and methane emission estimation. The test region is one of the world’s major rice regions: the Mekong River Delta, in Vietnam. This region presents a diversity in rice cultivation practices, in cropping density, from single to triple crop a year, and in crop calendar. The first step was to understand the temporal variation of the backscatter Sentinel-1 backscatter of rice fields, at VH and VV polarizations. For this purpose, in-situ data have been collected on 60 fields during 2 years, for the 5 rice seasons. It was found that backscatter time series of rice fields show very specific temporal behavior, with respect to other land use land cover types. The temporal and polarization variations of the rice backscatter have been interpreted with respect to physical interaction mechanisms to relate the backscatter dynamics to the key phenological stages, when the plants change its morphology and biomass. Because the same trend of temporal curves was observed over 5 rice seasons, it was possible to derive a mean curve to be used in the methodology developed for detecting rice phenology, and deriving information such as the date of sowing, the rice varieties of long and short duration cycle, or plant height, at each SAR acquisition date. The methods have been developed and applied to the Mekong delta. Products validation provides a good agreement with the reference data sets: 98% in rice/non-rice accuracy, the sowing dates RMSE of about 4 days, plant height RMSE of 7.8 cm, the long/short variety map has 91.7% accuracy and for phenology, only one season has been processed with good detection rate of 59/60. Finally, the use of the rice monitoring products as inputs in two process-based models was assessed. The models are ORYZA2000 for rice production estimation and DNDC for methane emission and water demand estimation. Sentinel-1 data retrieved information (sowing date, phenology, long/short variety, plant height) were used as model inputs, giving good agreement with the results making use of ground survey only. Based on the two process models with inputs from Sentinel-1 data, it was possible to have an integrated result on rice yield, water use, and methane emissions. The preliminary results show a good potential for the optimization of water management in rice fields in order to reduce water use and GHG emission, without reducing the yield. To achieve the objective which is the effective use of Sentinel-1 data for rice monitoring for food security and global environment, more works need to be done concerning the consolidation of the rice monitoring method development and the integration of Sentinel-1 derived information in models aiming at estimating and predicting rice production, methane emission and water use

The European Water Framework Directive (Directive 2000/60/EC) is a mandatory agreement that guides the member states of the European Union in the field of water policy to fulfil the requirements for reaching the aim of the good ecological status of water bodies. In the last years several work ows and methods were developed to determine and evaluate the haracteristics and the status of the water bodies. Due to their area measurements remote sensing methods are a promising approach to constitute a substantial additional value. With increasing availability of optical and radar remote sensing data the development of new methods to extract information from both types of remote sensing data is still in progress. Since most limitations of these data sets do not agree the fusion of both data sets to gain data with higher spectral resolution features the potential to obtain additional information in contrast to the separate processing of the data. Based thereupon this study shall research the potential of multispectral and radar remote sensing data and the potential of their fusion for the assessment of the parameters of water body structure. Due to the medium spatial resolution of the freely available multispectral Sentinel-2 data sets especially the surroundings of the water bodies and their land use are part of this study. SAR data is provided by the Sentinel-1 satellite. Different image fusion methods are tested and the combined products of both data sets are evaluated afterwards. The evaluation of the single data sets and the fused data sets is performed by means of a maximum-likelihood classification and several statistical measurements. The results indicate that the combined use of different remote sensing data sets can have an added value.

[pt] Imagens de sensoriamento remoto permitem o monitoramento e mapeamento de culturas de maneira precisa, apoiando práticas de agriculturaeficientes e sustentáveis com o objetivo de garantir a segurança alimentar.No entanto, a identificação do tipo de cultura a partir de dados de sensoriamento remoto em regiões tropicais ainda são consideradas tarefas comalto grau de dificuldade. As favoráveis condições climáticas permitem o uso,planejamento e o manejo da terra com maior flexibilidade, o que implica emculturas com dinâmicas mais complexas. Além disso, a presença constantede nuvens dificulta o uso de imagens ópticas, tornando as imagens de radar uma alternativa interessante para o mapeamento de culturas em regiõestropicais. Os modelos de campos aleatórios condicionais (CRFs) têm sidousados satisfatoriamente para explorar o contexto temporal e espacial naclassificação de imagens de sensoriamento remoto. Estes modelos oferecemuma alta precisão na classificação, no entanto, dependem de atributos extraídos manualmente com base em conhecimento especializado do domínio.Neste contexto, os métodos de aprendizado profundo, tais como as redesneurais convolucionais (CNNs), provaram ser uma alternativa robusta paraa classificação de imagens de sensoriamento, pois podem aprender atributosótimos diretamente dos dados. Este trabalho apresenta um modelo híbridobaseado em aprendizado profundo e CRF para o reconhecimento de culturas em áreas de regiões tropicais caracterizadas por ter uma dinâmicaespaço–temporal complexa. O framework proposto consiste em dois módulos: uma CNNs que modela o contexto espacial e temporal dos dados deentrada, e o CRF que modela a dinâmica temporal considerando a dependência entre rótulos para datas adjacentes. Estas dependências podem seraprendidas ou desenhadas por um especialista nas práticas de agriculturalocal. Comparações entre diferentes variantes de como modelar as transiçõestemporais são apresentadas usando sequências de imagens SAR de duas municipalidades no Brasil. Os experimentos mostraram melhorias significativasatingindo ate 30 por cento no F1 score por classe e ate 12 por cento no F1 score medio em relação ao modelo de base que não inclui dependências temporais duranteo processo de aprendizagem. / [en] Remote sensing imagery enables accurate crop mapping and monitoring, supporting efficient and sustainable agricultural practices to ensure food security. However, accurate crop type identification and crop area estimation from remote sensing data in tropical regions are still challenging tasks. Compared to the characteristic conditions of temperate regions, the more favorable weather conditions in tropical regions permit higher flexibility in land use, planning, and management, which implies complex crop dynamics. Moreover, the frequent cloud cover prevents the use of optical data during large periods of the year, making SAR data an attractive alternative for crop mapping in tropical regions. To exploit both spatial and temporal contex, conditional random fields (CRFs) models have been used successfully in the classification of RS imagery. These approaches deliver high accuracies; however, they rely on features engineering manually designed based on domain-specific knowledge. In this context, deep learning methods such as convolutional neural networks (CNNs) proved to be a robust alternative for remote sensing image classification, as they can learn optimal features and classification parameters directly from raw data. This work introduces a novel end-to-end hybrid model based on deep learning and conditional random fields for crop recognition in areas characterized by complex spatio-temporal dynamics typical of tropical regions. The proposed framework consists of two modules: a CNN that models spatial and temporal contexts from the input data and a CRF that models temporal dynamics considering label dependencies between adjacent epochs. These dependencies can be learned or designed by an expert in local agricultural practices. Comparisons between data-driven and prior-knowledge temporal constraints are presented for two municipalities in Brazil, using multi-temporal SAR image sequences. The experiments showed significant improvements in per class F1 score of up to 30 percent and up to 12 percent in average F1 score against a baseline model that doesn t include temporal dependencies during the learning process.

Eventos de inundação são fenômenos geralmente associados a eventos de chuvas intensas. Nesses eventos a cobertura de nuvens, normalmente, prejudica o mapeamento com uso de imagens ópticas. Assim, este trabalho tem como objetivo avaliar os resultados de mapeamento de áreas de inundação utilizando imagens SAR e SRTM. Para aplicação dos métodos foram analisadas as áreas de inundação nas cidades de Santa Fe e Parana, na Argentina. Embora a maior inundação registrada tenha sido no ano de 2003, registros de inundação são frequentemente observados nas províncias de Santa Fé e Entre Ríos. Foi utilizado imagens do satélite Sentinel-1, equipado com sensor C-SAR com dupla polarização (VV/VH). As imagens obtidas são do tipo Interferométrico (IW) Ground Range Detected (GRDH) com resolução espacial de 10 m. Foram utilizadas imagens em períodos com e sem eventos de inundação entre 2016 e 2017, calibradas e coregistradas. Sobre as imagens foram aplicadas técnicas de limiarização e de análise temporal para mapear a mancha de inundação. Também foi elaborado mapa a partir do Modelo Digital de Elevação (MDE) utilizando como referência estações de medição de nível da água dos rios. A validação de todos os métodos foi totalmente remota, baseando-se em um mapeamento da inundação de abril de 2003 na cidade de Santa Fe. Além disso, imagens publicadas de eventos de inundação complementaram a validação e foi possível comparar os resultados com uma imagem óptica Landsat – 8 com resolução de 15 m do dia 22 de fevereiro de 2016, quando o nível do rio Paraná estava acima do nível de alerta Os resultados dos três mapeamentos foram somados para formar uma única imagem com a mancha de inundação em comum. Entre as melhores acurácias, o método de análise do MDE atingiu o melhor resultado, 82% da área de inundação, no entanto, considerando os três métodos, a acurácia atinge mais de 91% de precisão. A técnica de limiarização foi mais eficiente em áreas sem alvos verticais, como áreas urbanas por exemplo. O MDE foi eficiente para simular a inundação em todos os alvos, no entanto em modelos de elevação com melhor resolução, o resultado final do mapeamento será mais preciso. A análise temporal mostrou ser uma técnica promissora para mapeamentos de inundação, no entanto um mapa detalhado de uso de solo é fundamental para aprimorar o resultado desta análise. Todos os processos foram feitos remotamente, possibilitando o desenvolvimento no futuro de um sistema automático para detecção de evento de inundação que pode ser aplicado em áreas com características similares. / Flood events usually go hand in hand with intensive rainfall during which clouds compromise any mapping attempts with optical imagery. Thus, this thesis aims at evaluate the results of mapping flood areas using SAR and SRTM images. For this purpose, flood areas in the cities Santa Fe and Parana in Argentina were analyzed. While the worst flood was registered in 2003, flood events frequently occur in both provinces Santa Fé and Entre Ríos. The employed Sentinel-1 satellite carrying a C-SAR sensor with dual polarization (VV/VH) provided interferometric (IW) Ground Range Detected (GRDH) imagery with a spatial resolution of 10 meters. Images from periods with and without flood events between 2016 and 2017 were calibrated and co-registered. Subsequently on the images were applied threshold and time analysis techniques, as well as a Digital Elevation Model (DEM) analysis with data from stations which measure the rivers’ water levels. The validation of all methods was totally remote, based on a flood mapping of April 2003 in the city of Santa Fe. In addition, published images of flood events complemented the validation and it was possible to compare the results with an optical image Landsat - 8 with 15 m resolution of February 22, 2016, when the level of the Paraná River was above the alert level The three maps were summed to form a single image with the flood spot in common. Among the best accuracy, the MDE analysis method achieved the best result, 82% of the flood area, however, considering all three methods, the accuracy reaches more than 91% accuracy. The thresholding technique was more efficient in areas with no vertical targets, such as urban areas. The DEM was efficient to simulate flooding on all targets, however using elevation models with better resolution, the final result of the mapping will be more accurate. The temporal analysis showed to be a promising technique for flood mapping, however a detailed map of land use is fundamental to improve the results of this analysis. All processes were done remotely, allowing the future development of an automatic flood event detection system that can be applied in areas with similar characteristics.

Eventos de inundação são fenômenos geralmente associados a eventos de chuvas intensas. Nesses eventos a cobertura de nuvens, normalmente, prejudica o mapeamento com uso de imagens ópticas. Assim, este trabalho tem como objetivo avaliar os resultados de mapeamento de áreas de inundação utilizando imagens SAR e SRTM. Para aplicação dos métodos foram analisadas as áreas de inundação nas cidades de Santa Fe e Parana, na Argentina. Embora a maior inundação registrada tenha sido no ano de 2003, registros de inundação são frequentemente observados nas províncias de Santa Fé e Entre Ríos. Foi utilizado imagens do satélite Sentinel-1, equipado com sensor C-SAR com dupla polarização (VV/VH). As imagens obtidas são do tipo Interferométrico (IW) Ground Range Detected (GRDH) com resolução espacial de 10 m. Foram utilizadas imagens em períodos com e sem eventos de inundação entre 2016 e 2017, calibradas e coregistradas. Sobre as imagens foram aplicadas técnicas de limiarização e de análise temporal para mapear a mancha de inundação. Também foi elaborado mapa a partir do Modelo Digital de Elevação (MDE) utilizando como referência estações de medição de nível da água dos rios. A validação de todos os métodos foi totalmente remota, baseando-se em um mapeamento da inundação de abril de 2003 na cidade de Santa Fe. Além disso, imagens publicadas de eventos de inundação complementaram a validação e foi possível comparar os resultados com uma imagem óptica Landsat – 8 com resolução de 15 m do dia 22 de fevereiro de 2016, quando o nível do rio Paraná estava acima do nível de alerta Os resultados dos três mapeamentos foram somados para formar uma única imagem com a mancha de inundação em comum. Entre as melhores acurácias, o método de análise do MDE atingiu o melhor resultado, 82% da área de inundação, no entanto, considerando os três métodos, a acurácia atinge mais de 91% de precisão. A técnica de limiarização foi mais eficiente em áreas sem alvos verticais, como áreas urbanas por exemplo. O MDE foi eficiente para simular a inundação em todos os alvos, no entanto em modelos de elevação com melhor resolução, o resultado final do mapeamento será mais preciso. A análise temporal mostrou ser uma técnica promissora para mapeamentos de inundação, no entanto um mapa detalhado de uso de solo é fundamental para aprimorar o resultado desta análise. Todos os processos foram feitos remotamente, possibilitando o desenvolvimento no futuro de um sistema automático para detecção de evento de inundação que pode ser aplicado em áreas com características similares. / Flood events usually go hand in hand with intensive rainfall during which clouds compromise any mapping attempts with optical imagery. Thus, this thesis aims at evaluate the results of mapping flood areas using SAR and SRTM images. For this purpose, flood areas in the cities Santa Fe and Parana in Argentina were analyzed. While the worst flood was registered in 2003, flood events frequently occur in both provinces Santa Fé and Entre Ríos. The employed Sentinel-1 satellite carrying a C-SAR sensor with dual polarization (VV/VH) provided interferometric (IW) Ground Range Detected (GRDH) imagery with a spatial resolution of 10 meters. Images from periods with and without flood events between 2016 and 2017 were calibrated and co-registered. Subsequently on the images were applied threshold and time analysis techniques, as well as a Digital Elevation Model (DEM) analysis with data from stations which measure the rivers’ water levels. The validation of all methods was totally remote, based on a flood mapping of April 2003 in the city of Santa Fe. In addition, published images of flood events complemented the validation and it was possible to compare the results with an optical image Landsat - 8 with 15 m resolution of February 22, 2016, when the level of the Paraná River was above the alert level The three maps were summed to form a single image with the flood spot in common. Among the best accuracy, the MDE analysis method achieved the best result, 82% of the flood area, however, considering all three methods, the accuracy reaches more than 91% accuracy. The thresholding technique was more efficient in areas with no vertical targets, such as urban areas. The DEM was efficient to simulate flooding on all targets, however using elevation models with better resolution, the final result of the mapping will be more accurate. The temporal analysis showed to be a promising technique for flood mapping, however a detailed map of land use is fundamental to improve the results of this analysis. All processes were done remotely, allowing the future development of an automatic flood event detection system that can be applied in areas with similar characteristics.

Eventos de inundação são fenômenos geralmente associados a eventos de chuvas intensas. Nesses eventos a cobertura de nuvens, normalmente, prejudica o mapeamento com uso de imagens ópticas. Assim, este trabalho tem como objetivo avaliar os resultados de mapeamento de áreas de inundação utilizando imagens SAR e SRTM. Para aplicação dos métodos foram analisadas as áreas de inundação nas cidades de Santa Fe e Parana, na Argentina. Embora a maior inundação registrada tenha sido no ano de 2003, registros de inundação são frequentemente observados nas províncias de Santa Fé e Entre Ríos. Foi utilizado imagens do satélite Sentinel-1, equipado com sensor C-SAR com dupla polarização (VV/VH). As imagens obtidas são do tipo Interferométrico (IW) Ground Range Detected (GRDH) com resolução espacial de 10 m. Foram utilizadas imagens em períodos com e sem eventos de inundação entre 2016 e 2017, calibradas e coregistradas. Sobre as imagens foram aplicadas técnicas de limiarização e de análise temporal para mapear a mancha de inundação. Também foi elaborado mapa a partir do Modelo Digital de Elevação (MDE) utilizando como referência estações de medição de nível da água dos rios. A validação de todos os métodos foi totalmente remota, baseando-se em um mapeamento da inundação de abril de 2003 na cidade de Santa Fe. Além disso, imagens publicadas de eventos de inundação complementaram a validação e foi possível comparar os resultados com uma imagem óptica Landsat – 8 com resolução de 15 m do dia 22 de fevereiro de 2016, quando o nível do rio Paraná estava acima do nível de alerta Os resultados dos três mapeamentos foram somados para formar uma única imagem com a mancha de inundação em comum. Entre as melhores acurácias, o método de análise do MDE atingiu o melhor resultado, 82% da área de inundação, no entanto, considerando os três métodos, a acurácia atinge mais de 91% de precisão. A técnica de limiarização foi mais eficiente em áreas sem alvos verticais, como áreas urbanas por exemplo. O MDE foi eficiente para simular a inundação em todos os alvos, no entanto em modelos de elevação com melhor resolução, o resultado final do mapeamento será mais preciso. A análise temporal mostrou ser uma técnica promissora para mapeamentos de inundação, no entanto um mapa detalhado de uso de solo é fundamental para aprimorar o resultado desta análise. Todos os processos foram feitos remotamente, possibilitando o desenvolvimento no futuro de um sistema automático para detecção de evento de inundação que pode ser aplicado em áreas com características similares. / Flood events usually go hand in hand with intensive rainfall during which clouds compromise any mapping attempts with optical imagery. Thus, this thesis aims at evaluate the results of mapping flood areas using SAR and SRTM images. For this purpose, flood areas in the cities Santa Fe and Parana in Argentina were analyzed. While the worst flood was registered in 2003, flood events frequently occur in both provinces Santa Fé and Entre Ríos. The employed Sentinel-1 satellite carrying a C-SAR sensor with dual polarization (VV/VH) provided interferometric (IW) Ground Range Detected (GRDH) imagery with a spatial resolution of 10 meters. Images from periods with and without flood events between 2016 and 2017 were calibrated and co-registered. Subsequently on the images were applied threshold and time analysis techniques, as well as a Digital Elevation Model (DEM) analysis with data from stations which measure the rivers’ water levels. The validation of all methods was totally remote, based on a flood mapping of April 2003 in the city of Santa Fe. In addition, published images of flood events complemented the validation and it was possible to compare the results with an optical image Landsat - 8 with 15 m resolution of February 22, 2016, when the level of the Paraná River was above the alert level The three maps were summed to form a single image with the flood spot in common. Among the best accuracy, the MDE analysis method achieved the best result, 82% of the flood area, however, considering all three methods, the accuracy reaches more than 91% accuracy. The thresholding technique was more efficient in areas with no vertical targets, such as urban areas. The DEM was efficient to simulate flooding on all targets, however using elevation models with better resolution, the final result of the mapping will be more accurate. The temporal analysis showed to be a promising technique for flood mapping, however a detailed map of land use is fundamental to improve the results of this analysis. All processes were done remotely, allowing the future development of an automatic flood event detection system that can be applied in areas with similar characteristics.

Using of Sentinel-1 data for radar interferometry Abstract The diploma thesis deals with extraction of a digital surface model (DSM) using synthetic aperture radar interferometry (InSAR) and Sentinel-1 data in selected locations of the Czech Republic. The InSAR technique, the Sentinel-1 data, their parameters and possibilities of their usage are described in the theoretical part of the thesis. The specification of the model areas and used data follows. The practical part is focused on creating a methodology of deriving a digital surface model and its extracting in the three tested locations. These locations differ from each other in their geomorphological features and land cover. At the end of this part the comparison of the extracted model with the reference model DMP 1G using statistical methods is carried out. At the end of this thesis the results are evaluated and discussed. Key words: InSAR, Sentinel-1, SAR, DSM

Monitoring carbon stock has emerged as a critical environmental problem among several worldwide organizations and collaborations in the context of global warming and climate change. This study seeks to provide a remote sensing solution based on three types of data, to explore the feasibility and reliability of estimating aboveground biomass (AGB) in order to improve the efficiency of monitoring carbon stock. The study attempted to investigate the potential of using Google Earth Engine (GEE), and the combinations of different datasets from Sentinel-1 (SAR), Sentinel-2 multispectral imagery, and LiDAR data to estimate AGB, by using the random forest algorithm (RF). Two models were proposed: the first one (Model 1) detected the AGB temporal changes from 2016 to 2021 in Southern Sweden; while the second one (Model 2) focused on Hultsfred municipality and studied the influence of different variables including the canopy height. Besides, six experimental groups of variables were tested to determine the performance of using different types of remote sensing data. We validated these two models with the observed AGB, and the findings showed that the combination of SAR polarization, multisprectral bands, vegetation indices able to estimate AGB for Model 1. In addition, Model 2 showed that further using the canopy height data can further improve the estimation.  We also found out that the spectral bands from Sentinel-2 contributed the most to AGB estimation for Model 1 in terms of: bands B3 (Green), B4 (Red), B5 (Red edge), B11 (SWIR), B12 (SWIR); and, vegetation indices of RVI, DVI, and EVI. On the other hand, for Model 2, B1(Ultra blue), B4 (Red), EVI, SAVI, and the canopy height are the most crucial variables for estimating AGB. Besides, the radar backscatter values using VV and VH modes from Sentienl-1 were both important for Models 1 and 2. For Model 1, the experimental group with the best accuracy was the group that used all variable combinations from Sentinel-1 and 2, and its   was 0.33~0.74. For Model 2, the group that used all the variables, in addition to the canopy height performed the best, where its   is 0.91. These therefore showed the benefit of integrating different remote sensing data sources.  In conclusion, this study showed the potential of using RF and GEE to estimate AGB in Southern Sweden. Furthermore, this study also shows the possibility of handling large dataset for a large scale area, at the resolution of 10 m, and producing time series AGB maps from 2016 to 2021. This can help enhance our understanding of AGB temporal changes and carbon stock detection in Southern Sweden, that can provide valuable insights for forest management and carbon monitoring.

Permafrost deformation is expected to increase due to climatic perturbations such as amplified air and soil temperatures, resulting in permafrost thawing and subsequent subsidence. Palsas and peat plateaus are uplifted ice-rich peat mounds that experience permafrost subsidence. This is due to the uppermost layer of permafrost, known as the Active Layer (AL), that seasonally thaws and freezes. Interferometric Synthetic Aperture Radar (InSAR) is an interferometric stacking technique successfully applied over permafrost regions when monitoring ground subsidence. The Small Baseline Subset (SBAS) technique is based on interferograms produced by stacking Synthetic Aperture Radar (SAR) acquisitions with small normal baselines. In this study, seasonal Sentinel-1 SAR C-band data obtained during June, July, August and September (JJAS) was used to generate seasonal Line of Sight (LoS) deformation time series of palsas and peat plateaus in Tavvavuoma (Sweden) by using the SBAS technique. Chersky (Russia) has documented permafrost subsidence and was used as a reference site. Findings include that seasonal stacks with short normal baselines generated more robust results than inter-annual stacks with longer normal baselines and temporal data gaps. No instances of pronounced subsidence were reported during JJAS. Nevertheless, minor subsidence during the early season and negative development trends were identified in the Tavvavuoma 2020 andChersky 2020-2021 stacks, respectively. Increased subsidence during the mid-and late thaw season was detected. The SBAS technique performed better and resulted in less temporal and seasonal decorrelation in areas above the tree line (Tavvavuoma) compared to the lowlands in the forest-tundra (Chersky). The challenge lies in whether surface subsidence of palsas and peat plateaus in sporadic permafrost regions experience irreversible long-term changes or seasonally cyclic changes in the permafrost ground regime. Future studies are recommended to implement annual intervals, including winter images over Tavvavuoma.

The launch of Sentinel-1A/B satellites in 2014 and 2016 marked a pivotal moment in Synthetic Aperture Radar (SAR) technology, ushering in a golden era for SAR. With a revisit time of 6–12 days, these satellites facilitated the acquisition of extensive stacks of high-resolution SAR images, enabling advanced time series analysis. However, processing these stacks posed challenges like interferometric phase degradation and tropospheric phase delay. This study introduces an advanced Small Baseline Subset (SBAS) algorithm that optimizes interferometric pairs, addressing systematic errors through dyadic downsampling and Delaunay Triangulation. A novel statistical framework is developed for elite pixel selection, considering distributed and permanent scatterers, and a tropospheric error correction method using smooth 2D splines effectively identifies and removes error components with fractal-like structures. Beyond geodetic technique advancements, the research explores geological phenomena, detecting five significant slow slip events (SSEs) along the Southern San Andreas Fault using multitemporal SAR interferometric time series from 2015-2021. These SSEs govern aseismic slip dynamics, manifesting as avalanche-like creep rate variations. The study further investigates Managed Aquifer Recharge (MAR) as a nature-engineering-based solution in the Santa Ana Basin. Analyzing surface deformation from 2004 to 2022 demonstrates MAR's effectiveness in curbing land subsidence within Orange County, CA. Additionally, MAR has the potential to stabilize nearby faults by inducing a negative Coulomb stress change. Projecting into the future, a suggested 2% annual increase in recharge volume through 2050 could mitigate land subsidence and reduce seismic hazards in coastal cities vulnerable to relative sea level rise. This integrated approach offers a comprehensive understanding of geological processes and proposes solutions to associated risks. / Doctor of Philosophy / The launch of Sentinel-1A/B satellites in 2014 and 2016 marked a big step forward in radar technology, especially Synthetic Aperture Radar (SAR). These satellites, which revisit the same area every 6-12 days, allowed us to collect many high-quality radar images. This helped us study changes over time in a more advanced way. However, there were challenges in handling all these images, like errors in the radar signals and delays caused by the Earth's atmosphere. We devised a smart algorithm based on the Small Baseline Subset (SBAS) to tackle these challenges. It helps optimize how we use pairs of radar images, reducing errors. We also developed a new method to pick the best pixels in the images and corrected errors caused by the atmosphere using mathematical methods. Moving beyond just technology, our research also looked at interesting Earth events. We found five major slow slip events along the Southern San Andreas Fault by studying radar data from 2015 to 2021. These events are like slow-motion slips along the fault, influencing how the ground moves. We also explored Managed Aquifer Recharge (MAR) as a solution in the Santa Ana Basin. By studying ground movement from 2004 to 2022, we found that MAR helped prevent the land from sinking in Orange County, California. It even has the potential to make nearby faults more stable. Looking ahead, increasing MAR activities by 2% each year until 2050 could protect against land sinking and reduce earthquake risks in coastal cities facing rising sea levels. This combined approach gives us a better understanding of Earth's processes and suggests ways to tackle related problems.