The Forest Inventory and Analysis (FIA) program of the United States Department of Agriculture Forest Service collects forest inventory data that provide estimates with reasonable accuracy at the national scale. However, for smaller domains, these estimates are often not as accurate due to the small sample size. Small area estimation improves the accuracy of the estimates at smaller domains by relying on auxiliary information. This study compared direct (FIA estimates), indirect (multiple linear regression), and composite estimators (Fay-Herriot) using auxiliary information derived from Landsat and Global Ecosystem Dynamics Investigation (GEDI) to obtain county-level estimates of forest attributes namely total and merchantable volume (m3 ha-1), aboveground biomass (Mg ha-1), basal area (m2 ha-1), and Lorey’s mean height (m). Compared with FIA estimates, the composite estimator reduced error by 75-78% for all the variables of interest. This shows that a reasonable amount of precision can be achieved with auxiliary information from Landsat and GEDI, improving FIA estimates at the county level.

This thesis explores the nature of contemporary sculptural practices in relation to the broader field of installed sculpture (which deploy articulated, interrelated, but autonomous components) and in the context of recent approaches to both curation and display. The artistic work and attendant commentary constitute a response to the issues of sculptural agency and display raised by both the practice-based outcomes and key works of several contemporary artists: Gabriel Kuri, Gedi Sibony, Melanie Counsell, Marc Camille Chaimowicz and Michael Dean. In a number of exhibitions ‘post-installation’ practices and the function of ‘montage’ sculpture is examined. Through outlining the current landscape of sculptural production and medium specificity a progressive notion of the monument is established. The sculptural artwork is seen to retain a political resistance, as both art-object and thing in the world. An assessment is made of how sculptures produce space within and through their exhibition context, directly related to the production of space as a whole (a social morphology posited by Henri Lefebvre). Applying a conception of time in reference to spatial production opens up the artwork’s potential to draw on complex codes of mnemonic function, which can potentially generate emancipatory agency from ideological issues in late-capitalism. Re-readings of key installed works by Marc Camille Chaimowicz and Mark Dean, through contexts derived from Nietzsche and Mark Fisher, reveal how sculptures can activate specific mnemonic codes, or collective memory. Such art works utilise a ‘forgetful memory’ – a reflexive process of positing, junking and reimagining relationships to cultural information. The body of artistic work produced for this research, intertwined with its critical reflection, makes an original contribution to knowledge by interrogating theoretically and experientially the potentials of ‘the sculptural’, as part of the plural production of art and exhibition-making. By means of practice and its outcomes, the research engages the current dynamics of spatial production and radicality of sculptural objecthood. The work examines the complex relationships between social memory and historicity, with which sculpture in an exhibition environment can engage.

A análise da expressão gênica através de dados gerados em experimentos de microarrays de DNA vem possibilitando uma melhor compreensão da dinâmica e dos mecanismos envolvidos nos processos celulares ao nível molecular. O aprimoramento desta análise é crucial para o avanço do conhecimento sobre as bases moleculares das neoplasias e para a identificação de marcadores moleculares para uso em diagnóstico, desenho de novos medicamentos em terapias anti-tumorais. Este trabalho tem como objetivos o desenvolvimento de modelos de análise desses dados, propondo uma nova forma de normalização de dados provenientes de microarrays e dois modelos para a construção de redes regulatórias de expressão gênica, sendo uma baseada na conectividade dinâmica entre diversos genes ao longo do ciclo celular e a outra que resolve o problema da dimensionalidade, em que o número de experimentos de microarrays é menor que o número de genes. Apresenta-se, ainda, um pacote de ferramentas com uma interface gráfica de fácil uso contendo diversas técnicas de análise de dados já conhecidas como também as abordagens propostas neste trabalho. / The analyses of DNA microarrays gene expression data are allowing a better comprehension of the dynamics and mechanisms involved in cellular processes at the molecular level. In the cancer field, the improvement of gene expression interpretation is crucial to better understand the molecular basis of the neoplasias and to identify molecular markers to be used in diagnosis and in the design of new anti-tumoral drugs. The main goals of this work were to develop a new method to normalize DNA microarray data and two models to construct gene expression regulatory networks. One method analyses the dynamic connectivity between genes through the cell cycle and the other solves the dimensionality problem in regulatory networks, meaning that the number of experiments is lower than the number of genes. We also developed a toolbox with a user-friendly interface, displaying several established statistical methods implemented to analyze gene expression data as well as the new approaches presented in this work.

A análise da expressão gênica através de dados gerados em experimentos de microarrays de DNA vem possibilitando uma melhor compreensão da dinâmica e dos mecanismos envolvidos nos processos celulares ao nível molecular. O aprimoramento desta análise é crucial para o avanço do conhecimento sobre as bases moleculares das neoplasias e para a identificação de marcadores moleculares para uso em diagnóstico, desenho de novos medicamentos em terapias anti-tumorais. Este trabalho tem como objetivos o desenvolvimento de modelos de análise desses dados, propondo uma nova forma de normalização de dados provenientes de microarrays e dois modelos para a construção de redes regulatórias de expressão gênica, sendo uma baseada na conectividade dinâmica entre diversos genes ao longo do ciclo celular e a outra que resolve o problema da dimensionalidade, em que o número de experimentos de microarrays é menor que o número de genes. Apresenta-se, ainda, um pacote de ferramentas com uma interface gráfica de fácil uso contendo diversas técnicas de análise de dados já conhecidas como também as abordagens propostas neste trabalho. / The analyses of DNA microarrays gene expression data are allowing a better comprehension of the dynamics and mechanisms involved in cellular processes at the molecular level. In the cancer field, the improvement of gene expression interpretation is crucial to better understand the molecular basis of the neoplasias and to identify molecular markers to be used in diagnosis and in the design of new anti-tumoral drugs. The main goals of this work were to develop a new method to normalize DNA microarray data and two models to construct gene expression regulatory networks. One method analyses the dynamic connectivity between genes through the cell cycle and the other solves the dimensionality problem in regulatory networks, meaning that the number of experiments is lower than the number of genes. We also developed a toolbox with a user-friendly interface, displaying several established statistical methods implemented to analyze gene expression data as well as the new approaches presented in this work.

<p dir="ltr">Canopy height is a fundamental metric for extracting valuable information about forested areas. Over the past decade, Lidar technology has provided a straightforward approach to measuring canopy height using various platforms such as terrestrial, unmanned aerial vehicle (UAV), airborne, and satellite sensors. However, satellite Lidar data, even with its global coverage, has a sparse sampling pattern that doesn’t provide continuous coverage over the globe. In contrast, satellites like LANDSAT offer seamless and widespread coverage of the Earth's surface through spectral data. Can we exploit the abundant spectral information from satellites like LANDSAT and ECOSTRESS to infer structural information obtained from Lidar satellites like Global Ecosystem Dynamic Investigation (GEDI)? This study aims to develop a deep learning model that can infer canopy height derived from sparsely observed Lidar waveforms using multi-sensor spectral data from spaceborne platforms. Specifically designed for localized site, the model focuses on county-level canopy height estimation, taking advantage of the relationship between canopy height and spectral reflectance that can be established in a local setting – something which might not exist universally. The study hopes to achieve a framework that can be easily replicable as height is a dynamic metric which changes with time and thus requires repeated computation for different time periods.</p><p dir="ltr">The thesis presents a series of experiments designed to comprehensively understand the influence of different spectral datasets on the model’s performance and its effectiveness in different types of test sites. Experiment 1 and 2 utilize Landsat spectral band values to extrapolate canopy height, while Experiment 3 and 4 incorporate ECOSTRESS land surface temperature and emissivity band values in addition to Landsat data. Tippecanoe County, predominantly composed of cropland, serves as the test site for Experiment 1 and 3, while Monroe County, primarily covered by forests, serves as the test site for Experiment 2 and 4. When compared to the Airborne Lidar dataset from the United States Geological Survey (USGS) – 3D Elevation Program (3DEP), the model achieves a Root Mean Square Error (RMSE) of 4.604m for Tippecanoe County using Landsat features while 5.479m for Monroe County. After integrating Landsat and ECOSTRESS features, the RMSE improves to 4.582m for Tippecanoe County but deteriorates to 5.860m for Monroe County. Overall, the study demonstrates comparable results to previous research without requiring feature engineering or extensive pre-processing. Furthermore, it successfully introduces a novel methodology for integrating multiple sources of satellite data to address this problem.</p>

The efficient and timely monitoring of forest dynamics is of paramount importance and requires accurate, high-resolution and time-tagged predictions at global scale. Despite numerous methodologies have been proposed in the literature, existing approaches often compromise on accuracy, resolution, temporal fidelity or coverage. To tackle these challenges and limitations, the main objective of this doctoral thesis is the investigation of the potential of artificial intelligence (AI) for the regression of bio-physical forest parameters from spaceborne Earth Observation (EO) data. This work explores for the first time the combined use of TanDEM-X single-pass interferometric products and convolutional neural networks for canopy height estimation at country scale. To achieve this, a novel deep learning framework is proposed, leveraging the capability of deep neural networks to effectively capture the complex spatial relationships between forest properties and satellite data, as well as ensuring the adaptability to different environmental conditions. The design and the understanding of the model is driven by explainable AI principles and by considerations on large-scale forest dynamics, with a great emphasis set on the challenges related to the variable acquisition geometry of the TanDEM-X mission, and by relying on the use of LVIS-derived LiDAR measurements as reference data. Moreover, several investigations are conducted on the adaptability of the developed framework for transferring knowledge to related domains, such as digital terrain model regression and above-ground biomass density estimation. Finally, the capability of the proposed approach to be extended to the use of other EO sensors is also evaluated, with a particular emphasis on the ESA Sentinel-1 and Sentinel-2 missions. The developed deep learning framework sets a solid groundwork for the generation of large-scale products of bio-physical forest parameters from spaceborne EO data. The approach achieves cutting-edge performance, significantly advancing the current state of forest assessment and monitoring technologies.

京都大学 / 新制・課程博士 / 博士(農学) / 甲第25318号 / 農博第2584号 / 新制||農||1104(附属図書館) / 京都大学大学院農学研究科森林科学専攻 / (主査)教授 北島 薫, 教授 小野田 雄介, 教授 Daniel Epron / 学位規則第4条第1項該当 / Doctor of Agricultural Science / Kyoto University / DGAM

<p dir="ltr">As a new spaceborne laser remote sensing system, the Global Ecosystem Dynamics Investigation, or GEDI, is being widely used for monitoring forest ecosystems. However, its measurements are subject to uncertainties that will affect the calculation of ground elevation and vegetation height. This research intends to investigate the quality of the GEDI elevation data and its relevance to topography and land cover.</p><p dir="ltr">In this study, the elevation of the GEDI data is compared to 3DEP DEM, which has a higher resolution and accuracy. All the experiments in this study are conducted for two locations with vastly different terrain and land cover conditions, namely Tippecanoe County in Indiana and Mendocino County in California. Through this investigation we expect to gain a comprehensive understanding of GEDI’s elevation quality in various terrain and land cover conditions.</p><p dir="ltr">The results show that GEDI data in Tippecanoe County has better elevation accuracy than the GEDI data in Mendocino County. GEDI in Tippecanoe County is almost four times more accurate than in Mendocino County. Regarding land cover, GEDI have better accuracy in low vegetation areas than in forest areas. The ratio can be around three times better in Tippecanoe County and around one and half times better in Mendocino County. In terms of slope, GEDI data shows a clear positive correlation between RMSE and slope. The trend indicates as slope increases, the RMSE increases concurrently. In other words, slope and GEDI elevation accuracy are inversely related. In the experiment involving slope and land cover, the results show that slope is the most influential factor to GEDI elevation accuracy.</p><p dir="ltr">This study informs GEDI users of the factors they must consider for forest biomass calculation and topographic mapping applications. When high terrain slope and/or high vegetation is present, the GEDI data should be checked with other data sources like 3DEP DEM or any ground truth measurements to assure its quality. We expect these findings can help worldwide users understand that the quality of GEDI data is variable and dependent on terrain relief and land cover.</p>