Carbon dioxide is taken up by crops during production and released back to the atmosphere at different geographical locations through respiration of consumed crop commodities. In this study, spatially distributed county-level US cropland net primary productivity, harvested biomass, changes in soil carbon, and human and livestock consumption data were integrated into the prior terrestrial biosphere flux generated by the Boreal Ecosystem Productivity Simulator (BEPS). A global time-dependent Bayesian synthesis inversion with a nested focus on North America was carried out based on CO2 observations at 210 stations. Overall, the inverted annual North American CO2 sink weakened by 6.5% over the period from 2002 to 2007 compared to simulations disregarding US crop statistical data. The US Midwest is found to be the major sink of 0.36±0.13 PgC yr-1 whereas the large sink in the US Southeast forests weakened to 0.16±0.12 PgC yr-1 partly due to local CO2 sources from crop consumption.

Carbon dioxide is taken up by crops during production and released back to the atmosphere at different geographical locations through respiration of consumed crop commodities. In this study, spatially distributed county-level US cropland net primary productivity, harvested biomass, changes in soil carbon, and human and livestock consumption data were integrated into the prior terrestrial biosphere flux generated by the Boreal Ecosystem Productivity Simulator (BEPS). A global time-dependent Bayesian synthesis inversion with a nested focus on North America was carried out based on CO2 observations at 210 stations. Overall, the inverted annual North American CO2 sink weakened by 6.5% over the period from 2002 to 2007 compared to simulations disregarding US crop statistical data. The US Midwest is found to be the major sink of 0.36±0.13 PgC yr-1 whereas the large sink in the US Southeast forests weakened to 0.16±0.12 PgC yr-1 partly due to local CO2 sources from crop consumption.

This thesis proposes a nonlinear differential semblance approach to full waveform inversion as an alternative to standard least squares inversion, which cannot guarantee a reliable solution, because of the existence of many spurious local minima of the objective function for typical data that lacks low-frequency energy. Nonlinear differential semblance optimization combines the ability of full waveform inversion to account for nonlinear physical effects, such as multiple reflections, with the tendency of differential semblance migration velocity analysis to avoid local minima. It borrows the gather-flattening concept from migration velocity analysis, and updates the velocity by flattening primaries-only gathers obtained via nonlinear inversion. I describe a general formulation of this algorithm, its main components and implementation. Numerical experiments show for simple layered models, standard least squares inversion fails, whereas nonlinear differential semblance succeeds in constructing a kinematically correct model and fitting the data rather precisely.

Des mesures de résistivité électrique et des campagnes d'observations géologiques ont été menées pour cartographier le contact sédiment/roche basale ainsi que les structures d'impact associées au cratère d'impact Bosumtwi datant de 1,07 millions d'années. Le cratère de 10,5 km de diamètre s'est formé dans des roches métamorphiques du Précambrien (2,1 à 2,2 milliard d'années), d'origine sédimentaire et volcanique. Il est actuellement rempli par le lac Bosumtwi de 8,5 km de diamètre. Ce cratère est la source des tektites et microtektites dispersées jusqu'en Côte d'Ivoire et au large des côtes ouest africaine. La campagne de mesures éophysiques consiste en 16 tomographies de résistivité électrique effectuées radialement des rives du lac vers les bords du cratère d'impact. Chaque profil utilise un système d'acquisition multi - électrodes avec une distance minimum entre les électrodes de 5 m. Les données ont été corrigées des effets topographiques et inversées en utilisant le programme commercial d'inversion Res2DInv, avec la norme L1 considérée plus robuste. La zone comprise entre les rives du lac et le bord externe du cratère de divise en trois formations géologiques principales. Les régions de faibles résistivités (< 64 .m) représentent les sédiments de lac. Les zones de resistivités moyennes (entre 128 et 200 .m) sont interprétées comme des brèches liées à l'impact, en dikes, allochtones ou parautochtones. Les régions de hautes résistivités (> 128 .m) représentent les roches métamorphiques sous-jacentes, d'origine volcanique ou sédimentaire. Les profils de résistivités permettent de retrouver la géométrie et l'extension latérale de ces trois types de roches. Une correspondance directe entre la lithologie observée en surface et les structures mises en évidence par les mesures de résistivité électrique dans le sous-sol est observée à Dwamam au Sud-Est dulac. À Dwamam, les sédiments sont environ à 200 m de la rive du lac et s'étendent environ sur 400 m vers les les bords du cratère, à la différence des autres zones où ont pu être cartographiés les sédiments. La topographie du contact sédiment/roche basale présente une direction particulière NE-SO avec un pendage variant entre 16 degrés au NE et 36 degrés au SO. Une majorité de fractures ont été remarquées dans le SO, remplies par des clasts et des brèches d'impacts. Les failles sont surtout présentent à l'Ouest du lac. En moyenne elles présentent un pendage variant entre 60o à l'Est et 80 degrés à l'Ouest du cratère. Des analyses statistiques ont été effectuées sur les directions et les pendages des failles selon les loi de von Mises et Fisher. Elles démontrent que les failles s'alignent préférentiellement le long de deux directions principales. L'analyse cumulée de la surface de contact sédiment/roche, des failles et de la localisation du champ de dispersion des tektites indique que le bolide d'environ 0,8 à 1 km de diamètre responsable du cratère est arrivé du NE. Les résultats démontrent que les panneaux de résistivité électrique fournissent des informations utiles pour l'étude des cratères d'impact.

Cloud computing systems, like MapReduce and Pregel, provide a scalable and fault tolerant environment for running computations at massive scale. However, these systems are designed primarily for data intensive computational tasks, while a large class of problems in scientific computing and business analytics are computationally intensive (i.e., they require a lot of CPU in addition to I/O). In this thesis, we investigate the use of cloud computing systems, in particular MapReduce, for computationally intensive problems, focusing on two classic problems that arise in scienti c computing and also in analytics: maximum clique and matrix inversion. The key contribution that enables us to e ectively use MapReduce to solve the maximum clique problem on dense graphs is a recursive partitioning method that partitions the graph into several subgraphs of similar size and running time complexity. After partitioning, the maximum cliques of the di erent partitions can be computed independently, and the computation is sped up using a branch and bound method. Our experiments show that our approach leads to good scalability, which is unachievable by other partitioning methods since they result in partitions of di erent sizes and hence lead to load imbalance. Our method is more scalable than an MPI algorithm, and is simpler and more fault tolerant. For the matrix inversion problem, we show that a recursive block LU decomposition allows us to e ectively compute in parallel both the lower triangular (L) and upper triangular (U) matrices using MapReduce. After computing the L and U matrices, their inverses are computed using MapReduce. The inverse of the original matrix, which is the product of the inverses of the L and U matrices, is also obtained using MapReduce. Our technique is the rst matrix inversion technique that uses MapReduce. We show experimentally that our technique has good scalability, and it is simpler and more fault tolerant than MPI implementations such as ScaLAPACK.

High-resolution seismic images of lithospheric structures allow us to infer the tectonics that modified the lithosphere. We apply such methods to understand Cenozoic modification of the lithosphere by tectonic and magmatic processes in the tectonically active western United States. Using USArray Transportable and Flexible Array data, we present high-resolution images for three regions in this thesis. (1) In the Mendocino triple junction, we use a joint inversion of Rayleigh-wave dispersion data and receiver functions to obtain a new crust and upper Vs model to ~150km depth. The model shows four distinct, young lithosphere-asthenosphere boundary systems. A low-Vs anomaly beneath the Great Valley-Sierra Nevada reconciles existing slab window models with the mantle-wedge geochemical signatures in Coast Range volcanics, and explains the ~3 Myr delay of the onset of volcanism after slab removal. Uppermost mantle low velocities provide evidence for forearc mantle serpentinization extending along the Cascadia margin. (2) In the Colorado Plateau, a Rayleigh wave tomography model sheds light on the volcanism along the margins and plateau uplift. Strong upper mantle heterogeneity across the plateau edge results from the combined effect of a ~200-400 K temperature difference and ~1% partial melt. A ring of low velocities under the plateau periphery suggests that the rehydrated Proterozoic lithosphere is progressively removed by convective processes. Particularly, a high-Vs anomaly imaged beneath the western plateau adds evidence for a downwelling/delamination hypothesis [Levander et al., 2011]. Thermo-chemical edge-driven convection causing localized lithospheric downwelling provides uplift along the margins and magmatic encroachment into the plateau center. (3) In the final study, we developed a 3-D teleseismic scattering wave imaging technique based on the Kirchhoff approximation and 3-D inverse Generalized Radon Transform. Synthetic tests demonstrate higher resolution imaging for continuous, irregular interfaces or localized scatterers, in comparison to conventional methods. Applied to the High Lava Plains dataset, the transmission coefficient structure shows a deepening Moho near 117.6°W and three negative events that correlate well with the Rayleigh wave low-Vs zones. Images made with the Mendocino data clearly show rapidly decreasing lithosphere-asthenosphere boundary depths from the subduction to transform regime.

Systolic architectures are capable of achieving high throughput by maximizing pipelining and by eliminating global data interconnects. Recursive algorithms with regular data flows are suitable for systolization. The computation of multiplicative inversion using algorithms based on EEA (Extended Euclidean Algorithm) are particularly suitable for systolization. Implementations based on EEA present a high degree of parallelism and pipelinability at bit level which can be easily optimized to achieve local data flow and to eliminate the global interconnects which represent most important bottleneck in todays sub-micron design process. The net result is to have high clock rate and performance based on efficient systolic architectures. This thesis examines high performance but also scalable implementations of multiplicative inversion or field division over Galois fields <i>GF</i>(2<i>^m</i>) in the specific case of cryptographic applications where field dimension <i>m</i> may be very large (greater than 400) and either <i>m</i> or defining irreducible polynomial may vary. For this purpose, many inversion schemes with different basis representation are studied and most importantly variants of EEA and binary (Stein's) GCD computation implementations are reviewed. A set of common as well as contrasting characteristics of these variants are discussed. As a result a generalized and optimized variant of EEA is proposed which can compute division, and multiplicative inversion as its subset, with divisor in either <i>polynomial</i> or <i>triangular</i> basis representation. Further results regarding Hankel matrix formation for double-basis inversion is provided. The validity of using the same architecture to compute field division with polynomial or triangular basis representation is proved. Next, a scalable unidirectional bit serial systolic array implementation of this proposed variant of EEA is implemented. Its complexity measures are defined and these are compared against the best known architectures. It is shown that assuming the requirements specified above, this proposed architecture may achieve a higher clock rate performance w. r. t. other designs while being more flexible, reliable and with minimum number of inter-cell interconnects. The main contribution at system level architecture is the substitution of all counter or adder/subtractor elements with a simpler distributed and free of carry propagation delays structure. Further a novel restoring mechanism for result sequences of EEA is proposed using a double delay element implementation. Finally, using this systolic architecture a CMD (Combined Multiplier Divider) datapath is designed which is used as the core of a novel systolic elliptic curve processor. This EC processor uses affine coordinates to compute scalar point multiplication which results in having a very small control unit and negligible with respect to the datapath for all practical values of <i>m</i>. The throughput of this EC based on this bit serial systolic architecture is comparable with designs many times larger than itself reported previously.

L'étude des noyaux légers riches en neutrons à la limite de la stabilité et au-delà suscite un intérêt marqué depuis quelques années car elle offre des tests approfondis de notre compréhension de la structure nucléaire. Expérimentalement, la disponibilité récente de faisceaux radioactifs avec des intensités suffisantes et le développement de nouveaux systèmes de détection de grande acceptance ont permis de sonder la structure de ces noyaux. Cette thèse porte sur l'étude de la structure de l'9He, système N = 7 non lié, via la réaction de transfert d'un neutron d(8He, p)9He avec l'aide d'un faisceau SPIRAL1. L'objectif principal de l'expérience était de clarifier la structure des états à basse énergie d'excitation de l'9He, en particulier l'inversion de parité possible de l'état fondamental. L'expérience a été une des premières à utiliser le nouvel ensemble de détection MUST2. Les statistiques relativement faibles dans la région du seuil d'émission de neutron et les divers fonds physiques, ne permettent pas de conclusion définitive concernant la nature de l'état de plus basse énergie. Des états à des énergies d'excitation plus élevées ont été identifiés. Une comparaison entre les distributions angulaires pour chacune des résonances avec des calculs DWBA et CRC a été effectuée. Les facteurs spectroscopiques ont été estimées pour chaque spin possible.

The transformed rejection method, a combination of the inversion and the rejection method, which is used to generate non-uniform random numbers from a variety of continuous distributions can be applied to discrete distributions as well. For the Poisson distribution a short and simple algorithm is obtained which is well suited for large values of the Poisson parameter $\mu$, even when $\mu$ may vary from call to call. The average number of uniform deviates required is lower than for any of the known uniformly fast algorithms. Timings for a C implementation show that the algorithm needs only half of the code but is - for $\mu$ not too small - at least as fast as the current state-of-the-art algorithms. (author's abstract) / Series: Preprint Series / Department of Applied Statistics and Data Processing

Systolic architectures are capable of achieving high throughput by maximizing pipelining and by eliminating global data interconnects. Recursive algorithms with regular data flows are suitable for systolization. The computation of multiplicative inversion using algorithms based on EEA (Extended Euclidean Algorithm) are particularly suitable for systolization. Implementations based on EEA present a high degree of parallelism and pipelinability at bit level which can be easily optimized to achieve local data flow and to eliminate the global interconnects which represent most important bottleneck in todays sub-micron design process. The net result is to have high clock rate and performance based on efficient systolic architectures. This thesis examines high performance but also scalable implementations of multiplicative inversion or field division over Galois fields <i>GF</i>(2<i>^m</i>) in the specific case of cryptographic applications where field dimension <i>m</i> may be very large (greater than 400) and either <i>m</i> or defining irreducible polynomial may vary. For this purpose, many inversion schemes with different basis representation are studied and most importantly variants of EEA and binary (Stein's) GCD computation implementations are reviewed. A set of common as well as contrasting characteristics of these variants are discussed. As a result a generalized and optimized variant of EEA is proposed which can compute division, and multiplicative inversion as its subset, with divisor in either <i>polynomial</i> or <i>triangular</i> basis representation. Further results regarding Hankel matrix formation for double-basis inversion is provided. The validity of using the same architecture to compute field division with polynomial or triangular basis representation is proved. Next, a scalable unidirectional bit serial systolic array implementation of this proposed variant of EEA is implemented. Its complexity measures are defined and these are compared against the best known architectures. It is shown that assuming the requirements specified above, this proposed architecture may achieve a higher clock rate performance w. r. t. other designs while being more flexible, reliable and with minimum number of inter-cell interconnects. The main contribution at system level architecture is the substitution of all counter or adder/subtractor elements with a simpler distributed and free of carry propagation delays structure. Further a novel restoring mechanism for result sequences of EEA is proposed using a double delay element implementation. Finally, using this systolic architecture a CMD (Combined Multiplier Divider) datapath is designed which is used as the core of a novel systolic elliptic curve processor. This EC processor uses affine coordinates to compute scalar point multiplication which results in having a very small control unit and negligible with respect to the datapath for all practical values of <i>m</i>. The throughput of this EC based on this bit serial systolic architecture is comparable with designs many times larger than itself reported previously.