主催：JST/CREST，Vrije University, ALTERRA, IBPC

M. Ing. / Digital Surface Models (DSM) are widely used in the earth sciences for research, visu- alizations, construction etc. In order to generate a DSM for a speci c area, specialized equipment and personnel are always required which leads to a costly and time consuming exercise. Image processing has become a viable processing technique to generate terrain models since the improvements of hardware provided adequate processing power to complete such a task. Digital Surface Models (DSM) can be generated from stereo imagery, usually obtained from a remote sensing platform. The core component of a DSM generating system is the image matching algorithm. Even though there are a variety of algorithms to date which can generate DSMs, it is a computationally complex calculation and does tend to take some time to complete. In order to achieve faster DSMs, an investigation into an alternative processing platform for the generation of terrain models has been done. The Graphics Processing Unit (GPU) is usually used in the gaming industry to manipulate display data and then render it to a computer screen. The architecture is designed to manipulate large amounts of oating point data. The scientic community has begun using the GPU processing power available for technical computing, hence the term, General Purpose computing on a Graphics Processing Unit (GPGPU). The GPU is investigated as alternative processing platform for the image matching procedure since the processing capability of the GPU is so much higher than the CPU but only for a conditioned set of input data. A matching algorithm, derived from the GC3 algorithm has been implemented on both a CPU platform and a GPU platform in order to investigate the viability of a GPU processing alternative. The algorithm makes use of a Normalized Cross Correlation similarity measurement and the geometry of the image acquisition contained in the sensor model to obtain conjugate point matches in the two source images. The results of the investigation indicated an improvement of up to 70% on the processing time required to generate a DSM. The improvements varied from 70% to some cases where the GPU has taken longer to generate the DSM. The accuracy of the automatic DSM generating system could not be clearly determined since only poor quality reference data was available. It is however shown the DSMs generated using both the CPU and GPU platforms relate to the reference data and correlate to each other. The discrepancies between the CPU and the GPU results are low enough to prove the GPU processing is bene cial with neglible drawbacks in terms of accuracy. The GPU will definitely provide superior processing capabilites for DSM generation above a CPU implementation if a matching algorithm is speci cally designed to cater for the bene ts and limitations of the GPU.

Accurate simulation of energy, water, and carbon fluxes exchanging between the land surface and the atmosphere is beneficial for improving terrestrial ecohydrological and climate predictions. We systematically assessed the Noah land surface model (LSM) with mutiparameterization options (Noah-MP) in simulating these fluxes and associated variations in terrestrial water storage (TWS) and snow cover fraction (SCF) against various reference products over 18 United States Geological Survey two-digital hydrological unit code regions of the continental United States (CONUS). In general, Noah-MP captures better the observed seasonal and interregional variability of net radiation, SCF, and runoff than other variables. With a dynamic vegetation model, it overestimates gross primary productivity by 40% and evapotranspiration (ET) by 22% over the whole CONUS domain; however, with a prescribed climatology of leaf area index, it greatly improves ET simulation with relative bias dropping to 4%. It accurately simulates regional TWS dynamics in most regions except those with large lakes or severely affected by irrigation and/or impoundments. Incorporating the lake water storage variations into the modeled TWS variations largely reduces the TWS simulation bias more obviously over the Great Lakes with model efficiency increasing from 0.18 to 0.76. Noah-MP simulates runoff well in most regions except an obvious overestimation (underestimation) in the Rio Grande and Lower Colorado (New England). Compared with North American Land Data Assimilation System Phase 2 (NLDAS-2) LSMs, Noah-MP shows a better ability to simulate runoff and a comparable skill in simulating R-n but a worse skill in simulating ET over most regions. This study suggests that future model developments should focus on improving the representations of vegetation dynamics, lake water storage dynamics, and human activities including irrigation and impoundments.

Secondary currents are one of the important features that characterize flow in river bends and confluences. Fluid particles follow a helical path instead of moving nearly parallel to the axis of the channel. The local imbalance between the vertically varying centrifugal force and the cross-stream pressure gradient results in generating the secondary flow and raising a typical motion of the helical flow. A number of studies, including experimental or mathematical, have been conducted to examine flow characteristics in curved open channels, river meanders, or confluences. In this research, the influence of secondary currents is studied on the elevation of water surface and the hydraulic structures in channel bends and confluences by employing a 3D OpenFOAM numerical model. The research implements the 3D OpenFOAM numerical model to simulate the horizontal distribution of the flow in curved rivers. In addition, the progress in unraveling and understanding the bend dynamics is considered. The finite volume method in (OpenFOAM) software is used to simulate and examine the behavior of secondary current in channel bends and confluences. Thereafter, a comparison between the experimental data and a numerical model is conducted. Two sets of experimental data are used; the data provided by Rozovskii (1961) for sharply curved channel, and the dataset provided by Shumate (1998) for confluent channel. Two solvers in (OpenFOAM) software were selected to solve the problem regarding the experiment; InterFoam and PisoFoam. The InterFoam is a transient solver for incompressible flow that is used with open channel flow and Free Surface Model. The PisoFoam is a transient solver for incompressible flow that is used with closed channel flow and Rigid-Lid Model. Various turbulence models (i.e. Standard k-ε, Realizable k-ε, LRR, and LES) are applied in the numerical model to assess the accuracy of turbulence models in predicting the behaviour of the flow in channel bends and confluences. The accuracies of various turbulence models are examined and discussed.

L'augmentation probable de la fréquence des sécheresses au cours du 21ème siècle, en réponse au réchauffement climatique, pourrait faire basculer la forêt amazonienne d'un puits à une source de carbone, déclenchant ainsi une rétroaction positive sur l'accroissement de l'effet de serre. La sensibilité des cycles de l'eau et du carbone aux sécheresses représente donc un point critique dans la stratégie d'évaluation des modèles de climat et pour la crédibilité des scénarios climatiques. Cette thèse vise à évaluer la représentation et la réponse aux sécheresses des cycles de l'eau et du carbone en Amazonie par le modèle de surface ISBA-CC mis en oeuvre dans le modèle de climat du CNRM. Pour ce faire, nous avons utilisé, outre des mesures écophysiologiques, des mesures de tours de flux ainsi que deux expériences d'assèchement artificiel. La version originale du modèle n'étant pas satisfaisante, la paramétrisation des processus métaboliques et la sensibilité de la végétation au stress hydrique ont été modifiées, puis validées, sur ces sites, mais également à l'échelle du bassin en utilisant notamment des données de débits, des reconstructions d'évapotranspiration et d'assimilation de carbone, ainsi que des observations spatiales de variations de stocks d'eau et de fluorescence chlorophyllienne. Une analyse succincte des changements climatiques régionaux a été réalisée sur la base des modèles CMIP5, mettant en évidence un certain consensus à l'allongement et au renforcement de la saison sèche au cours du 21ème siècle. Ces scénarios, après calibration des sorties de quelques modèles, nous ont ensuite permis de tester en mode off-line la sensibilité des projections du cycle de l'eau et du carbone aux modifications apportées au modèle ISBA-CC. Les résultats montrent sans surprise l'influence prépondérante du choix du modèle forceur dans ce type de simulation, mais révèle également l'importance des paramétrisations liées à la végétation. / The likely increase in drought frequency as a result of climate change, might switch the Amazon forest from being a sink to a source of carbon, setting off a positive feedback on the increasing greenhouse effect. Hence, the drought sensitivity of the water and carbon cycles in the Amazon is a crucial point in the evaluation of climate models.In this PhD thesis, we evaluate in off-line mode, how the CNRM land surface model ISBACC, represents the water and carbon cycles and their response to drought in the Amazon. For this purpose, we used published ecophysiological data together with five Amazonian flux tower sitesand two artificial drought experiments. Since the standard version of the model was not satisfactory, we modified the parameterizations of the metabolic processes and the sensitivity of vegetation to water stress. We evaluated this new version against the flux tower and artificial drought sites, but also at regional scale using river discharge data, reconstructed evapotranspiration and carbon assimilation fluxes, and remotely sensed total water storage variations and chlorophyll fluorescence data. A brief analysis of the regional climate changes projected by CMIP5 models showed a relative consensus for longer and more intense dry seasons during the 21st century. Two of these climate scenarios were calibrated and used to test the sensitivity of the water and carbon cycle projections to our modifications of the ISBACC model in off-line mode. As expected, the results reveam the importance of the driving climate model, but also highlight the role of the vegetation parameterizations.

京都大学 / 新制・課程博士 / 博士(工学) / 甲第23165号 / 工博第4809号 / 新制||工||1752(附属図書館) / 京都大学大学院工学研究科社会基盤工学専攻 / (主査)教授 立川 康人, 准教授 市川 温, 教授 田中 茂信 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DGAM

Water is undeniably among the most important natural resources and the most critical in semi-arid regions like the Intermountain West of the United States. Such regions are characterized by low precipitation, the majority of which is transferred to the atmosphere from the soil and vegetation as evapotranspiration (ET). Quantification of ET is thus crucial for understanding the balance of water within the region, which is important for efficiently planning the available water resources. This study was motivated towards advancing the estimation of actual ET (ETA) in mountain ecosystems, where the variation in different types of vegetation and non-uniformity of soil including considerable stone content creates challenges for estimating water use as ET. With the aim of addressing the effect of stone content in controlling soil moisture and ET, this study examined the influence of stone content on bulk soil hydraulic properties. An averaging model referred to as a binary mixing model was used to describe the way in which water is held and released in stony soil. This approach was based on the individual hydraulic behavior of the background soil and of the stones within the soil. The effect of soil stone content on ETA was evaluated by accounting for the water retention properties of stones in the soil using a numerical simulation model (HYDRUS-1D). The results revealed overestimation of simulated ETA when effects of stone content were not accounted for in comparison to ETA measured by the state-of-the-art “eddy covariance” measurement method for ETA. An even larger-scale model was evaluated, named the Noah-Multiphysics (Noah-MP) land surface model. The land surface model was run using different arrangements of complexity to determine the importance of stone content information on simulation results. The version of the model with information about stone content along with detailed soil properties was able to provide the best Noah-MP prediction of ET. The study suggests that improvement in representation of soil properties including stone content information, can substantially advance the ability of numerical and land surface models to more accurately simulate soil water flow and ETA.

<p> </p> <p>As a result of Shadow NeRF and Sat-NeRF, it is possible to take the solar angle into account in a NeRF-based framework for rendering a scene from a novel viewpoint using satellite images for training. Our work extends those contributions and shows how one can make the renderings season-specific. Our main challenge was creating a Neural Radiance Field (NeRF) that could render seasonal features independently of viewing angle and solar angle</p> <p>while still being able to render shadows. We teach our network to render seasonal features by introducing one more input variable — time of the year. However, the small training datasets typical of satellite imagery can introduce ambiguities in cases where shadows are present in the same location for every image of a particular season. We add additional terms to the loss function to discourage the network from using seasonal features for accounting for shadows. We show the performance of our network on eight Areas of Interest containing images captured by the Maxar WorldView-3 satellite. This evaluation includes tests measuring the ability of our framework to accurately render novel views, generate height maps, predict shadows, and specify seasonal features independently from shadows. Our ablation</p> <p>studies justify the choices made for network design parameters. Also included in this work is a novel approach to space carving which merges multiple features and consistency metrics</p> <p>at different spatial scales to create higher quality digital surface map than is possible using standard RGB features.</p>

No description available.

No description available.