Search NASA⌕ Search

SEARCH · Search NASA

Results for “large data”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,621 records · Page 90

Analysis of events with b-jets and a pair of leptons of the same charge in pp collisions at √s = 8 TeV with the ATLAS detector

An analysis is presented of events containing jets including at least one b -tagged jet, sizeable missing transverse momentum, and at least two leptons including a pair of the same electric charge, with the scalar sum of the jet and lepton transverse momenta being large. A data sample with an integrated luminosity of 20.3 fb –1 of pp collisions at √s = 8 TeV recorded by the ATLAS detector at the Large Hadron Collider is used. Standard Model processes rarely produce these final states, but there are several models of physics beyond the Standard Model that predict an enhanced rate of production of such events; the ones considered here are production of vector-like quarks, enhanced four-top-quark production, pair production of chiral b'-quarks, and production of two positively charged top quarks. Eleven signal regions are defined; subsets of these regions are combined when searching for each class of models. In the three signal regions primarily sensitive to positively charged top quark pair production, the data yield is consistent with the background expectation. There are more data events than expected from background in the set of eight signal regions defined for searching for vector-like quarks and chiral b'-quarks, but the significance of the discrepancy is less than two standard deviations. Furthermore, the discrepancy reaches 2.5 standard deviations in the set of five signal regions defined for searching for four-top-quark production. The results are used to set 95% CL limits on various models.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Predicting fault slip via transfer learning

Abstract Data-driven machine-learning for predicting instantaneous and future fault-slip in laboratory experiments has recently progressed markedly, primarily due to large training data sets. In Earth however, earthquake interevent times range from 10’s-100’s of years and geophysical data typically exist for only a portion of an earthquake cycle. Sparse data presents a serious challenge to training machine learning models for predicting fault slip in Earth. Here we describe a transfer learning approach using numerical simulations to train a convolutional encoder-decoder that predicts fault-slip behavior in laboratory experiments. The model learns a mapping between acoustic emission and fault friction histories from numerical simulations, and generalizes to produce accurate predictions of laboratory fault friction. Notably, the predictions improve by further training the model latent space using only a portion of data from a single laboratory earthquake-cycle. The transfer learning results elucidate the potential of using models trained on numerical simulations and fine-tuned with small geophysical data sets for potential applications to faults in Earth.

58 GEOSCIENCES↗

The Living Publication

Within the ICSTI Insights Series we offer three articles on the 'living publication' that is already available to practitioners in the important field of crystal structure determination and analysis. While the specific examples are drawn from this particular field, we invite readers to draw parallels in their own fields of interest. The first article describes the present state of the crystallographic living publication, already recognized by an ALPSP (Association of Learned and Professional Society Publishers) Award for Publishing Innovation in 2006. The second article describes the potential impact on the record of science as greater post-publication analysis becomes more common within currently accepted data deposition practices, using processed diffraction data as the starting point. The third article outlines a vision for the further improvement of crystallographic structure reports within potentially achievable enhanced data deposition practices, based upon raw (unprocessed) diffraction data. The IUCr in its Commissions and Journals has for many years emphasized the importance of publications being accompanied by data and the interpretation of the data in terms of atomic models. This has been followed as policy by numerous other journals in the field and its cognate disciplines. This practice has been well served by databases and archiving institutions such as the Protein Data Bank (PDB), the Cambridge Crystallographic Data Centre (CCDC), and the Inorganic Crystal Structure Database (ICSD). Normally the models that are archived are interpretations of the data, consisting of atomic coordinates with their displacement parameters, along with processed diffraction data from X-ray, neutron or electron diffraction studies. In our current online age, a reader can not only consult the printed word, but can display and explore the results with molecular graphics software of exceptional quality. Furthermore, the routine availability of processed diffraction data allows readers to perform direct calculations of the electron density (using X-rays and electrons as probes) or the nuclear density (using neutrons as probe) on which the molecular models are directly based. This current community practice is described in our first article. There are various ways that these data and tools can be used to further analyze the molecules that have been crystallized. Notably, once a set of results is announced via the publication, the research community can start to interact directly with the data and models. This gives the community the opportunity not only to read about the structure, but to examine it in detail, and even generate subsequent improved models. These improved models could, in principle, be archived along with the original interpretation of the data and can represent a continuously improving set of interpretations of a set of diffraction data. The models could improve both by correction of errors in the original interpretation and by the use of new representations of molecules in crystal structures that more accurately represent the contents of a crystal. These possible developments are described in our second article. A current, significant, thrust for the IUCr is whether it would be advantageous for the crystallographic community to require, rather than only encourage, the archiving of the raw (unprocessed) diffraction data images measured from a crystal, a fibre or a solution. This issue is being evaluated in detail by an IUCr Working Group (see http://forums.iucr.org). Such archived raw data would be linked to and from any associated publications. The archiving of raw diffraction data could allow as yet undeveloped processing methods to have access to the originally measured data. The debate within the community about this much larger proposed archiving effort revolves around the issue of 'cost versus benefit'. Costs can be minimized by preserving the raw data in local repositories, either at centralized synchrotron and neutron research institutes, or at research universities. Archiving raw data is also perceived as being more effective than just archiving processed data in countering scientific fraud, which exists in our field, albeit at a tiny level of occurrences. In parallel developments, sensitivities to avoiding research malpractice are encouraging Universities to establish their own data repositories for research and academic staff. These various 'raw data archives', would complement the existing processed data archives. These archives could however have gaps in their coverage arising from a lack of resources. Nevertheless we believe that a sufficiently large raw data archive, with reasonable global coverage, could be encouraged and have major benefits. These possible developments, costs and benefits, are described in our third and final article on 'The living publication'.

74 ATOMIC AND MOLECULAR PHYSICS↗

Integrating Data Centers and Grid Technologies at Scale

This presentation focuses on the challenge of integrating AI-driven data centers with the power grid at scale. It examines the AI data center capacity challenge and the role of new Medium Voltage Direct Current (MVDC) and other grid-enhancing technologies in enabling efficient and reliable power delivery. The session will highlight the National Laboratory of the Rockies' ARIES capabilities and planning tools, along with collaborative examples involving Verrus, Compass, and Schneider through the Agora test bed for grid-friendly data center evaluations, and ON. Energy for UPS evaluation. It will showcase the NLR Stable Grid Platform for studying oscillations caused by large-scale data centers, along with planning tools to assess grid security and reliability. Additionally, the presentation covers reconductoring strategies to increase grid capacity and explores innovative data center architectures, including the Advanced DC Architectures with Power-electronic Transformers (ADAPT) platform, which enables testing of complete DC architectures for data centers.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Convection and electrodynamic signatures in the vicinity of a Sun-aligned arc: Results from the Polar Acceleration Regions and Convection Study (Polar ARCS)

An experimental campaign designed to study high-latitude auroral arcs was conducted in Sondre Stromfjord, Greenland, on February 26, 1987. The Polar Acceleration Regions and Convection Study (Polar ARCS) consisted of a coordinated set of ground-based, airborne, and sounding rocket measurements of a weak, sun-aligned arc system within the duskside polar cap. A rocket-borne barium release experiment, two DMSP satellite overflights, all-sky photography, and incoherent scatter radar measurements provided information on the large-scale plasma convection over the polar cap region while a second rocket instrumented with a DC magnetometer, Langmuir and electric field probes, and an electron spectrometer provided measurements of small-scale electrodynamics. The large-scale data indicate that small, sun-aligned precipitation events formed within a region of antisunward convection between the duskside auroral oval and a large sun-aligned arc further poleward. This convection signature, used to assess the relationship of the sun-aligned arc to the large-scale magnetospheric configuration, is found to be consistent with either a model in which the arc formed on open field lines on the dusk side of a bifurcated polar cap or on closed field lines threading an expanded low-latitude boundary layer, but not a model in which the polar cap arc field lines map to an expanded plasma sheet. The antisunward convection signature may also be explained by a model in which the polar cap arc formed on long field lines recently reconnected through a highly skewed plasma sheet. The small-scale measurements indicate the rocket passed through three narrow (less than 20 km) regions of low-energy (less than 100 eV) electron precipitation in which the electric and magnetic field perturbations were well correlated. These precipitation events are shown to be associated with regions of downward Poynting flux and small-scale upward and downward field-aligned currents of 1-2 micro-A/sq m. The paired field-aligned currents are associated with velocity shears (higher and lower speed streams) embedded in the region of antisunward flow.

Weiss, L. A.↗

A simplified regression building thermal modelling method for detached two-floor house in U.S.

The residential building sector accounts for approximately 37% of total U.S. electricity consumption. Within the residential building sector, heating and cooling is the main target for peak load shifting/reduction since it is the largest contributor to peak demand. In fact, the flexibility of residential HVAC loads can provide continuous variation of demand to provide grid services by varying their demand over a baseline. The performance of HVAC load control to provide grid services relies heavily on the accuracy of indoor air temperature or cooling/heating demand predictions and therefore the quality of building model. Besides forward models, popular building models are data-driven models which can be broken down into two categories: gray-box model, e.g. Resistance-Capacitance (RC) model and black-box model. RC model, also called lumped capacitance or network model, which is constituted with electrical analogue pattern with resistance (R) and capacitance (C). In general, RC models require considerable computation burden and long periods of data to train limited number of model coefficients. Black-box models have gained increasing interest due to their capability in analyzing large-scale data and flexibility in practical applications. But, the data-mining based (machine learning algorithms/techniques based) models tend to have invisible model structures which poses a problem when trying to use the model for optimal control or model predictive control of the HVAC system. Hence, there is a continuing need for efficient online system identification techniques, which can provide explicit parameters for the model. Traditional regression models fit well for this specific purpose. This paper presents an innovative way to predict average indoor temperature in separate floors of typical detached residential house. A rolling horizon linear regression model, which includes online adaptive correction component, is proposed to predict the temperature difference between downstairs and upstairs. A RC model is used to predict the overall mean indoor air temperature. Since the adaptive algorithm needs to be implemented online, a less computation-demanding polynomial fitting algorithm is adopted. This kind of fitting problem can be cast as linear regression problem with multiple variables, parameters of which can be efficiently obtained by well-known gradient descent method.The validation is conducted by comparing the predicted results with the results from data-mining based models as well as measured data from a real typical detached two-floor house. The results show that the developed method has satisfactory performance in predicting the building indoor temperature in 1st and 2nd floors.

Cui, Borui↗

A Unified Data Infrastructure for Biological and Environmental Research: A Report from the BER Advisory Committee

The Biological and Environmental Research (BER) program within the U.S. Department of Energy (DOE) Office of Science supports large-scale data generation efforts across its two divisions: Biological Systems Science and Earth and Environmental Systems Sciences. These efforts include user facilities in atmospheric radiation measurements, genomics, metabolomics, proteomics, compute, and imaging. In addition, BER supports the development of plant-based fuels; research in biosystems design, environmental microbiomes, and atmospheric systems; energy flux monitoring; climate-based ecosystem experiments; pathogen biopreparedness; and modeling of climate, urban interfaces, and interactions between people and energy resources. For data access, BER supports community data services at its user facilities, along with specialized data initiatives for Earth and environmental science, climate modeling, genomic and microbial analysis, and multisector dynamics modeling.

54 ENVIRONMENTAL SCIENCES↗

Quantitative phase analysis of challenging samples using neutron powder diffraction. Sample #4 from the CPD QPA round robin revisited

Here, quantitative phase analysis (QPA) using neutron powder diffraction more often than not involves non-ambient studies where no sample preparation is possible. The larger samples and penetration of neutrons versus X-rays makes neutron diffraction less susceptible to inhomogeneity and large grain sizes, but most well-characterized QPA standard samples do not have these characteristics. Sample #4 from the International Union of Crystallography Commission on Powder Diffraction QPA round robin was one such sample. Data were collected using the POWGEN time-of-flight (TOF) neutron powder diffractometer and analysed together with historical data from the C2 diffractometer at Chalk River. The presence of magnetic reflections from Fe 3 O 4 (magnetite) in the sample was an additional consideration, and given the frequency at which iron-containing and other magnetic compounds are present during in-operando studies their possible impact on the accuracy of QPA is of interest. Additionally, scattering from thermal diffuse scattering in the high- Q region (<0.6 Å) accessible with TOF data could impact QPA results during least-squares because of the extreme peak overlaps present in this region. Refinement of POWGEN data was largely insensitive to the modification of longer d -spacing reflections by magnetic contributions, but the constant-wavelength data were adversely impacted if the magnetic structure was not included. A robust refinement weighting was found to be effective in reducing quantification errors using the constant-wavelength neutron data both where intensities from magnetic reflections were ignored and included. Results from the TOF data were very sensitive to inadequate modelling of the high- Q (low d -spacing) background using simple polynomials.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

On the selection of a sampling filter bandwidth for a digital data detector.

Consideration of the problem of selecting a low-pass sampling bandwidth for the digital mechanization of a matched-filter bit-synchronizer combination. An attempt is made to determine how large a data rate can be provided for a fixed sampling rate (limited by hardware considerations on the sampling device) without paying an excessive penalty in SNR performance.

Simon, M. K.↗

Recent HARVIST Results: Classifying Crops from Remote Sensing Data

In this paper, we report on recent results from the Heterogeneous Agricultural Research Via Interactive, Scalable Technology (HARVIST) project. HARVIST seeks to provide the tools and scalability required to enable practicioners to analyze large, diverse data sets that may come from different data sources. We have focused on agricultural applications, and our current results demonstrate the ability of the system to train a crop type classifier that operates on orbital remote sensing images. We find that this classifier can label crops with an accuracy of 82%, comparable to other published results.

support vector machines↗

Advances in Entry Modeling for Impact Risk Assessment

A summary of recent advancements in the detailed modeling of asteroid atmospheric entry processes made through NASA’s Asteroid Threat Assessment Project (ATAP) is presented. Understanding, and accurately modeling these processes and their associated uncertainties is critical to predicting all downstream impact effects, such as blast wave and thermal damage footprints. Furthermore, there is (perhaps thankfully) a dearth of empirical data for large impactors of the kind that would pose a threat to human populations, on which to anchor and/or validate models used in risk assessments. Therefore, we must rely heavily on detailed theoretical and numerical modeling to develop robust assessments for decision makers. To that end, ATAP has made some significant progress in advancing the capabilities in this area. Two areas in particular are highlighted in the present work: meteoroid ablation mechanisms, and bolide luminosity. The first research area – meteoroid ablation mechanisms – has focused on performing novel high-enthalpy wind tunnel experiments on meteorites and meteorite analogs, and utilizing the resulting data to develop high-fidelity models for impactor mass loss at scale. These efforts have resulted in several insights. Of note, these data suggest a differential vaporization process where volatiles are liberated preferentially when the asteroidal material is subject to high heat, while refractory components remain in the molten layer on the surface. A numerical model has been developed which considers this phenomena, and its effect on the bulk impactors effective heat of ablation is examined. The second research thrust that is discussed is focused on accurate modeling of impactor radiation phenomena. While another submission to the conference will discuss the application of our approach to thermal ground damage modeling, here, we present an overview of our extensive efforts to utilize available ground- and space-based observations of large bolides (~1m diameter, and above) to inform and validate our detailed modeling approaches. These methods have been shown to accurately reconstruct the detailed spectra for the Benesov bolide, as well as approximate the burn footprint for the Tunguska event. Recent effort has focused on reconciling light curve data from multiple sources (all-sky camera networks, GLM, US government sensors), using our validated model, and providing a calibrated model for luminous efficiency which can then be utilized to infer impactor properties such as shape, mass, and composition. This work will be demonstrated through an exemplar case study focusing on a large bolide event with observational data (e.g. Chelyabinsk, Flensburg, etc.). Finally, our team’s assessment on the current maturity of atmospheric entry models, and priorities for future research will be provided.

ATAP↗

Effects of Improved 17 O Correction on Interlaboratory Agreement in Clumped Isotope Calibrations, Estimates of Mineral-Specific Offsets, and Temperature Dependence of Acid Digestion Fractionation

The clumped isotopic composition of carbonate-CO 2 (denoted Δ 47 ) is a function of carbonate formation temperature and in natural samples can act as a recorder of paleoclimate, burial, or diagenetic conditions. The absolute abundance of heavy isotopes in the universal standards VPDB and VSMOW (defined by four parameters: R 13 VPDB , R 17 VSMOW , R 18 VSMOW , and λ ) impact calculated Δ 47 values. In this work, we investigate whether use of updated and more accurate values for these parameters can remove observed interlaboratory differences in the measured T-Δ 47 relationship. Using the updated parameters, we reprocess 14 published calibration data sets measured in 11 different laboratories, representing many mineralogies, bulk compositions, sample types, reaction temperatures, and sample preparation and analysis methods. Exploiting this large composite data set ( n = 1,253 sample replicates), we investigate the possibility for a “universal” clumped isotope calibration. We find that applying updated parameters improves the T-Δ 47 relationship (reduces residuals) within most labs and improves overall agreement but does not eliminate all interlaboratory differences. We reaffirm earlier findings that different mineralogies do not require different calibration equations and that cleaning procedures, method of pressure baseline correction, and mass spectrometer type do not affect interlaboratory agreement. We also present new estimates of the temperature dependence of the acid digestion fractionation for Δ 47 (Δ* 25-X ), based on combining reprocessed data from four studies, and new theoretical equilibrium values to be used in calculation of the empirical transfer function. Overall, we have ruled out a number of possible causes of interlaboratory disagreement in the T-Δ 47 relationship, but many more remain to be investigated.

58 GEOSCIENCES↗

Training data selection for event classification in a highly variable environment

A problem of interest for nuclear nonproliferation is monitoring activities at nuclear facilities, where proliferation events may only take place a few times and often under variable conditions. Machine learning has revolutionized data analytics by enabling the use of measurable signatures to generate predictive models of facility operations. However, traditional methods for training these models require large, reliable data sets with labeled observations, a challenge for nonproliferation. Highly variable conditions further complicate this as events from training data may have occurred in conditions quite different from the event of interest. Our hypothesis is that when events occur in a highly variable environment, careful training data selection for each test event could outperform the standard approach of using all available training data. We developed a method to optimize training data selection for the given test event and applied it to predicting the power level of the High Flux Isotope Reactor (HFIR) at Oak Ridge National Laboratory. In this study, the reactor startup exhibits variability between occurrences due to natural variability in environmental conditions and operational procedures. Using a combination of analysis techniques, a similitude assessment was performed on data collected from HFIR to isolate clusters that were optimal for training a predictive model. Concepts such as dynamic time warping and Jaccard similarity were used in conjunction with clustering analysis. In order to validate this approach, the model was trained on every combination of unique training events and the predictive performance was compared to the performance using a subset of the training data selected by isolated clusters found through the similitude assessment.

Iyer, A↗

Dimensionality Reduction of SDSS Spectra with Variational Autoencoders

High-resolution galaxy spectra contain much information about galactic physics, but the high dimensionality of these spectra makes it difficult to fully utilize the information they contain. We apply variational autoencoders (VAEs), a nonlinear dimensionality reduction technique, to a sample of spectra from the Sloan Digital Sky Survey (SDSS). In contrast to principal component analysis (PCA), a widely used technique, VAEs can capture nonlinear relationships between latent parameters and the data. We find that a VAE can reconstruct the SDSS spectra well with only six latent parameters, outperforming PCA with the same number of components. Different galaxy classes are naturally separated in this latent space, without class labels having been given to the VAE. The VAE latent space is interpretable because the VAE can be used to make synthetic spectra at any point in latent space. For example, making synthetic spectra along tracks in latent space yields sequences of realistic spectra that interpolate between two different types of galaxies. Using the latent space to find outliers may yield interesting spectra: in our small sample, we immediately find unusual data artifacts and stars misclassified as galaxies. In this exploratory work, we show that VAEs create compact, interpretable latent spaces that capture nonlinear features of the data. While a VAE takes substantial time to train (≈1 day for 48,000 spectra), once trained, VAEs can enable the fast exploration of large astronomical data sets.

79 ASTRONOMY AND ASTROPHYSICS↗

Improved navigation by combining VOR/DME information with air or inertial data

The improvement was determined in navigational accuracy obtainable by combining VOR/DME information (from one or two stations) with air data (airspeed and heading) or with data from an inertial navigation system (INS) by means of a maximum-likelihood filter. It was found that the addition of air data to the information from one VOR/DME station reduces the RMS position error by a factor of about 2, whereas the addition of inertial data from a low-quality INS reduces the RMS position error by a factor of about 3. The use of information from two VOR/DME stations with air or inertial data yields large factors of improvement in RMS position accuracy over the use of a single VOR/DME station, roughly 15 to 20 for the air-data case and 25 to 35 for the inertial-data case. As far as position accuracy is concerned, at most one VOR station need be used. When continuously updating an INS with VOR/DME information, the use of a high-quality INS (0.01 deg/hr gyro drift) instead of a low-quality INS (1.0 deg/hr gyro drift) does not substantially improve position accuracy.

Bobick, J. C.↗

Life Sciences Data Archive (LSDA) in the Post-Shuttle Era

Now, more than ever before, NASA is realizing the value and importance of their intellectual assets. Principles of knowledge management, the systematic use and reuse of information/experience/expertise to achieve a specific goal, are being applied throughout the agency. LSDA is also applying these solutions, which rely on a combination of content and collaboration technologies, to enable research teams to create, capture, share, and harness knowledge to do the things they do well, even better. In the early days of spaceflight, space life sciences data were been collected and stored in numerous databases, formats, media-types and geographical locations. These data were largely unknown/unavailable to the research community. The Biomedical Informatics and Health Care Systems Branch of the Space Life Sciences Directorate at JSC and the Data Archive Project at ARC, with funding from the Human Research Program through the Exploration Medical Capability Element, are fulfilling these requirements through the systematic population of the Life Sciences Data Archive. This project constitutes a formal system for the acquisition, archival and distribution of data for HRP-related experiments and investigations. The general goal of the archive is to acquire, preserve, and distribute these data and be responsive to inquiries from the science communities.

Fitts, Mary A.↗

Life Sciences Data Archives (LSDA) in the Post-Shuttle Era

Now, more than ever before, NASA is realizing the value and importance of their intellectual assets. Principles of knowledge management-the systematic use and reuse of information, experience, and expertise to achieve a specific goal-are being applied throughout the agency. LSDA is also applying these solutions, which rely on a combination of content and collaboration technologies, to enable research teams to create, capture, share, and harness knowledge to do the things they do well, even better. In the early days of spaceflight, space life sciences data were collected and stored in numerous databases, formats, media-types and geographical locations. These data were largely unknown/unavailable to the research community. The Biomedical Informatics and Health Care Systems Branch of the Space Life Sciences Directorate at JSC and the Data Archive Project at ARC, with funding from the Human Research Program through the Exploration Medical Capability Element, are fulfilling these requirements through the systematic population of the Life Sciences Data Archive. This project constitutes a formal system for the acquisition, archival and distribution of data for HRP-related experiments and investigations. The general goal of the archive is to acquire, preserve, and distribute these data and be responsive to inquiries for the science communities. Information about experiments and data, as well as non-attributable human data and data from other species' are available on our public Web site http://lsda.jsc.nasa.gov. The Web site also includes a repository for biospecimens, and a utilization process. NASA has undertaken an initiative to develop a Shuttle Data Archive repository. The Shuttle program is nearing its end in 2010 and it is critical that the medical and research data related to the Shuttle program be captured, retained, and usable for research, lessons learned, and future mission planning. Communities of practice are groups of people who share a concern or a passion for something they do, and learn how to do it better as they interact regularly. LSDA works with the HRP community of practice to ensure that we are preserving the relevant research and data they need in the LSDA repository. An evidence-based approach to risk management is required in space life sciences. Evidence changes over time. LSDA has a pilot project with Collexis, a new type of Web-based search engine. Collexis differentiates itself from full-text search engines by making use of thesauri for information retrieval. The high-quality search is based on semantics that have been defined in a life sciences ontology. Additionally, Collexis' matching technology is unique, allowing discovery of partially matching dicuments. Users do not have to construct a complicated (Boolean) search query, but can simply enter a free text search without the risk of getting "no results". Collexis may address these issues by virtue of its retrieval and discovery capabilities across multiple repositories.

Fitts, Mary A.↗

Scalable and efficient algorithms for the propagation of uncertainty from data through inference to prediction for large-scale problems, with application to flow of the Antarctic ice sheet

The majority of research on efficient and scalable algorithms in computational science and engineering has focused on the forward problem : given parameter inputs, solve the governing equations to determine output quantities of interest. In contrast, in this paper, we consider the broader question: given a (large-scale) model containing uncertain parameters, (possibly) noisy observational data, and a prediction quantity of interest, how do we construct efficient and scalable algorithms to (1) infer the model parameters from the data (the deterministic inverse problem ), (2) quantify the uncertainty in the inferred parameters (the Bayesian inference problem), and (3) propagate the resulting uncertain parameters through the model to issue predictions with quantified uncertainties (the forward uncertainty propagation problem )? We present efficient and scalable algorithms for this end-to-end, data-to-prediction process under the Gaussian approximation and in the context of modeling the flow of the Antarctic ice sheet and its effect on loss of grounded ice to the ocean. The ice is modeled as a viscous, incompressible, creeping, shear-thinning fluid. The observational data come from satellite measurements of surface ice flow velocity, and the uncertain parameter field to be inferred is the basal sliding parameter, represented by a heterogeneous coefficient in a Robin boundary condition at the base of the ice sheet. The prediction quantity of interest is the present-day ice mass flux from the Antarctic continent to the ocean. We show that the work required for executing this data-to-prediction process—measured in number of forward (and adjoint) ice sheet model solves—is independent of the state dimension, parameter dimension, data dimension, and the number of processor cores. The key to achieving this dimension independence is to exploit the fact that, despite their large size, the observational data typically provide only sparse information on model parameters. This property can be exploited to construct a low rank approximation of the linearized parameter-to-observable map via randomized SVD methods and adjoint-based actions of Hessians of the data misfit functional.

54 ENVIRONMENTAL SCIENCES↗