| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 5.21 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
A humidade do solo é uma variável climática essencial, cuja observação e monitorização se reveste de grande importância em várias áreas científicas e da vida corrente. O efeito da variabilidade da humidade do solo na fase interferométrica tem motivado diversos estudos com incidência quer na resolução do problema inverso de estimativa da humidade do solo a partir da fase interferométrica, quer no contributo que a variabilidade da humidade do solo pode ter na medição da deformação. Esta tese tem como principal objetivo a análise da contribuição da variação temporal da humidade do solo na fase interferométrica e na estimativa do vapor de água integrado da atmosfera, e o seu impacto na estimativa da deformação da superfície. Como segundo objetivo pretende-se desenvolver uma metodologia que permita determinar a estimativa da humidade do solo com elevada resolução espacial a partir da fase interferométrica utilizando dados SAR da banda C das micro-ondas. Para a concretização dos objetivos propostos foram realizados quatro estudos: o primeiro baseado em métodos de regressão por aprendizagem automática, na qual se inclui informação sobre a fase e coerência interferométrica e informação sobre as caraterísticas do solo; o segundo, incidiu no estudo do contributo e da possibilidade de mitigação do efeito atmosférico na estimativa da humidade do solo; o terceiro, foi dedicado a isolar a descorrelação de fase com origem na variação da humidade do solo e o quarto foi dedicado a compreender o impacto da descorrelação de fase na estimativa do valor da deformação de superfície estimado com interferometria SAR, bem como a compreender a possível origem da descorrelação de fase. Relativamente ao primeiro tópico, foi demonstrado que a inclusão de informação sobre o tipo de solo permite melhorar substancialmente os resultados da regressão, tendo também sido demonstrado que se deve utilizar linhas de base temporais até 12 dias para minimizar os efeitos de descorrelação temporal. No segundo estudo foi proposta uma metodologia em que se usa a modelação atmosférica atualizada com base em estações GNSS locais, combinada com informação interferométrica, para remover o efeito atmosférico da fase interferométrica e assim estimar mapas de humidade do solo de elevada resolução espacial. O terceiro estudo incidiu no fecho de fase e foi demonstrada a possibilidade de utilizar a descorrelação de fase interferométrica para determinar a variação da humidade do solo observada entre os momentos de aquisição das imagens SAR usadas no cálculo dos interferogramas. Por último foi possível validar a relação entre a variação da humidade do solo, a variação da água na vegetação e o seu efeito na descorrelação de fase e respetivo impacto nas estimativas de deformação de superfície. Uma vez identificado o efeito da variabilidade da água na vegetação foi proposto um modelo para mitigar o seu impacto na fase interferométrica por forma a melhorar as estimativas de deformação da superfície.
Soil moisture is an essential climatic variable, whose observation and monitoring are of great importance in various scientific areas and everyday life. The effect of soil moisture variability in the interferometric phase has motivated several studies, both in the resolution of the inverse problem of estimating soil moisture from the interferometric phase and in the contribution that soil moisture variability can have on deformation measurement. This thesis aims to analyse the contribution of temporal variation of soil moisture in the interferometric phase and in the integrated water vapour estimation, and its impact on surface deformation estimation. As a second objective, the intention is to develop a methodology for soil moisture retrieval, with high spatial resolution, from the interferometric phase using C-band SAR microwave data. To achieve the proposed objectives, four studies were conducted: the first based on machine learning regression methods, which include information about phase and interferometric coherence, as well as soil characteristics; the second focused on studying the contribution and possibility of mitigating the atmospheric effect in soil moisture estimation; the third was dedicated to isolating the phase decorrelation originating from soil moisture variation, and the fourth was dedicated to understanding the impact of phase decorrelation on the estimation of surface deformation values estimated with SAR interferometry, as well as understanding the possible origin of phase decorrelation. Regarding the first topic, it was demonstrated that the inclusion of information about soil type substantially improves the regression results. It was also shown that a temporal baseline of up to 12 days should be used to minimise the effects of temporal decorrelation. In the second study, a methodology was proposed in which updated atmospheric modelling based on local GNSS stations, combined with interferometric information, is used to remove the atmospheric effect from the interferometric phase and thus estimate high-resolution spatial maps of soil moisture. The third study focused on phase closure, and it was demonstrated that interferometric phase decorrelation can be used to determine the observed soil moisture variation between the acquisition times of the SAR images used in interferogram calculation. Finally, it was possible to validate the relationship between soil moisture variation, vegetation water variation, and their effect on phase decorrelation and its impact on surface deformation estimates. Once the variability effect of vegetation water is identified, a model is proposed to mitigate its impact on the interferometric phase to improve surface deformation estimates.
Soil moisture is an essential climatic variable, whose observation and monitoring are of great importance in various scientific areas and everyday life. The effect of soil moisture variability in the interferometric phase has motivated several studies, both in the resolution of the inverse problem of estimating soil moisture from the interferometric phase and in the contribution that soil moisture variability can have on deformation measurement. This thesis aims to analyse the contribution of temporal variation of soil moisture in the interferometric phase and in the integrated water vapour estimation, and its impact on surface deformation estimation. As a second objective, the intention is to develop a methodology for soil moisture retrieval, with high spatial resolution, from the interferometric phase using C-band SAR microwave data. To achieve the proposed objectives, four studies were conducted: the first based on machine learning regression methods, which include information about phase and interferometric coherence, as well as soil characteristics; the second focused on studying the contribution and possibility of mitigating the atmospheric effect in soil moisture estimation; the third was dedicated to isolating the phase decorrelation originating from soil moisture variation, and the fourth was dedicated to understanding the impact of phase decorrelation on the estimation of surface deformation values estimated with SAR interferometry, as well as understanding the possible origin of phase decorrelation. Regarding the first topic, it was demonstrated that the inclusion of information about soil type substantially improves the regression results. It was also shown that a temporal baseline of up to 12 days should be used to minimise the effects of temporal decorrelation. In the second study, a methodology was proposed in which updated atmospheric modelling based on local GNSS stations, combined with interferometric information, is used to remove the atmospheric effect from the interferometric phase and thus estimate high-resolution spatial maps of soil moisture. The third study focused on phase closure, and it was demonstrated that interferometric phase decorrelation can be used to determine the observed soil moisture variation between the acquisition times of the SAR images used in interferogram calculation. Finally, it was possible to validate the relationship between soil moisture variation, vegetation water variation, and their effect on phase decorrelation and its impact on surface deformation estimates. Once the variability effect of vegetation water is identified, a model is proposed to mitigate its impact on the interferometric phase to improve surface deformation estimates.
Descrição
Palavras-chave
Humidade do Solo Atraso Troposférico Deformação da Superfície Interferometria SAR Soil Moisture Tropospheric Delay Surface Displacement SAR interferometry
