NERSCPowering Scientific Discovery for 50 Years

NERSC Initiative for Scientific Exploration (NISE) 2012 Awards

NISE is a mechanism used for allocating the NERSC reserve (10% of the total allocation). It is a competitive allocation administered by NERSC staff and management.  In 2012 we were particularly interested in large scale or data intensive proposals.  We received 61 NISE requests this year and have awarded 108 million hours to 23 projects.

Ocean-Atmosphere Reanalysis for Climate Applications (OARCA) 1850-2013

Gilbert Compo, University of Colorado at Boulder

NISE award: 10,000,000 hours

NERSC Repository: m958

Long-term records of the global weather and climate variations from the 19th to 21st century are urgently needed for assessing the strengths and weaknesses of the next-generation of coupled climate models being used to project the effects of anthropogenic greenhouse gas emissions for the upcoming Fifth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC AR5). Importantly, such records must have quantified uncertainties to allow a quantitative assessment. The Ensemble Kalman Filter method provides the needed uncertainty estimates. This NISE research will make progress towards an Ocean-Atmosphere Reanalysis for Climate Applications system to generate a record of the state of the global atmosphere and ocean back to 1850 at six-hourly resolution in the atmosphere and 5-day resolution in the ocean. This OARCA dataset will be at the resolution to help assess changes in weather extremes, such as severe midlatitude storms and tropical cyclones, which have significant socio-economic consequences.

WeFold: A collaborative effort for protein structure prediction

Silvia Crivelli, Lawrence Berkeley National Laboratory

NISE award: 1,550,000 hours

NERSC repository: m1532

Proteins are present in all living things, including plants, bacteria, and viruses. Knowing how to efficiently determine the three-dimensional structure of a protein is essential for understanding how it works and how it can be synthesized and it will have a tremendous impact on biomedical and biomaterial discoveries. For example, the knowledge of how certain proteins fold will allow scientists to better design drugs to cure diseases such as Alzheimer's and cancer and the knowledge of how certain proteins work in association with other proteins will allow biologists to engineer new kinds of proteins that can be used to efficiently extract energy from grass.

However, predicting protein structure from sequence is computationally very expensive as each protein has a huge conformation space. Determining which of the many, many possible structures corresponds to the actual 3D structure of a protein is regarded as one of the hardest problems in biology today and current methods do not produce consistently good results. That is why a collaborative effort makes perfect sense with respect to the division of a prediction pipeline handled by best methods and people who do different parts of the prediction pipeline. We hope this collaborative effort will speed up the rate at which discoveries are made. 

A multi-decadal reforecast data set to improve weather forecasts for renewable energy applications

Thomas Hamill, National Oceanic & Atmospheric Administration

NISE award: 1,675,000 hours

NERSC repository: refcst

This project will allow NOAA to produce experimental long-lead probabilistic weather forecasts relevant to the renewable energy field, e.g., forecasts for solar, wind, and hydropower.  Currently the emphasis has been on shorter-range forecasts, but there are many decisions that could be made weeks in advance given useful forecast guidance, such as when it may be least problematic to take some wind turbines offline for maintenance, or when it may be necessary to stockpile conventional fossil fuels because the wind power is expected to be marginal a week or two hence.

Excited-State Dynamics and Properties of Complex Interfaces for Energy Conversion

Yosuke Kanai, University of North Carolina at Chapel Hill

NISE award: 2,000,000 hours

NERSC repository: m1029

Energy independence is arguably the most important issue we face in the United States, and it is closely connected to various national security issues because most of known oil reserves in the world are located in the regions that are politically unstable. Solar energy utilization is one of promising strategies in the energy portfolio of the country in the future.

In order to find an economical way to convert abundant solar energy into electricity and/or fuels through improving solar cells, materials innovation plays a crucial role. A significant improvement in the solar cell efficiency requires characterizing, understanding, and designing novel materials for efficient solar energy conversion processes. High-performance supercomputers allow scientists to solve complex equations of quantum mechanics for characterizing important properties and phenomena of nano-materials. The state-of-the-art simulations enable the scientists to also predict and design new materials that are yet to be synthesized in the lab. The field of scientific computation is quickly emerging as one of the most promising avenues in modern science, and this work takes advantage of one of the most powerful supercomputers in the world to simulate and design advanced nano-materials for solar energy conversion.

Kinetic Simulations in Laboratory and Space Plasmas

Homa Karimabadi, University of California San Diego

NISE award: 7,000,000 hours

NERSC repository: m1303

Earth and other planets are engulfed in the Sun's atmosphere. The Earth's dipole magnetic field shields us from most of the Sun's effects and its frequent storms. However, the shielding is not perfect and, through a process called magnetic reconnection, the solar wind is able to penetrate ("crack") Earth's magnetosphere.

This cracking, called "space weather," can affect the Earth and its technological systems, and has caused over $4 billion in satellite losses alone. Our goal is to use petascale kinetic simulations to develop a more complete understanding of the cause and conditions for the development of this crack in the magnetosphere.

Turbulent Reacting Flows for Multi-physics Model Development

Colleen Kaul, Stanford University

NISE award: 4,500,000 hours

NERSC repository: m1426

Turbulent reacting flows are important in a wide range of engineering devices. These devices include aircraft engines, automobile engines, and industrial combustors. All of these devices are governed by multi-physics phenomena that interact over a wide range of length and time scales. The existence of these disparate scales presents a tremendous challenge to the accurate numerical simulation of turbulent reactive flows. Using the resources provided by this grant, we will perform high fidelity direct numerical simulation (DNS) of flows governed by turbulent reactive physics. These DNS results will guide large eddy simulation (LES) model development, enabling accurate simulation of realistic combustion-based devices. The resulting modeling framework is expected to assist in both reducing the emissions and improving the efficiency of these technologies.

Systems Biology Knowledge Base

Keith Keller, Lawrence Berkeley National Laboratory

NISE award: 100,000 hours

NERSC repository: kbase

The Systems Biology Knowledge Base will support open community science by serving as a freely available computational environment for sharing and integrating diverse biological data types, accessing and developing software for data analysis, and providing resources for modeling and simulation. It will leverage community-wide capabilities, experimental results, and modeling efforts and bring together research products from many different projects and laboratories to create an extensible, comprehensive cyberinfrastructure focused on DOE scientific objectives related to microbes, plants, and metacommunities (complex communities of organisms).

Electronic Properties of Novel Nitride Nanostructures

Emmanouil Kioupakis, University of Michigan

NISE award: 3,000,000 hours

NERSC repository: m1380

Light-emitting diodes (LEDs) made from nitride materials are efficient sources of light that can replace incandescent and fluorescent light bulbs for indoors lighting. They promise to significantly reduce the electricity cost and carbon-dioxide emissions that lighting generates. Currently, however, the efficiency of LED light bulbs is limited, and this brings up their cost. A promising solution is to use ideas from nanotechnology and fabricate LEDs out of thin nanowires about a millionth of an inch thick. Our theoretical work will study these new nanomaterials and provide a guide for the development of efficient light bulbs.

Integrated Carbon Cycle Data Assimilation with NCAR Carbon-Climate Model

Junjie Liu, Jet Propulsion Laboratory, California Institute of Technology

NISE award: 5,100,000 hours

NERSC repository: m189

The proposed research is to build an integrated carbon cycle data assimilation system to estimate surface carbon dioxide flux and improve our understanding of the terrestrial carbon cycle . The integrated carbon cycle data assimilation system includes an atmosphere component, which assimilates observations of both meteorology and carbon dioxide, and a terrestrial biosphere component, which estimates global terrestrial biosphere productivity and model parameters by assimilating multiple terrestrial biosphere observations.

Environmental Fluctuations and Gating in Bio-Inorganic Proton-Coupled Electron Transfer

Thomas Miller, California Institute of Technology

NISE award: 7,000,000 hours

NERSC repository: m822

Proton-coupled electron transfer (PCET) reactions are central to the chemistry of energy conversion, respiration, and enzyme kinetics. However, key aspects of such reactions remain poorly understood due to the coupling of the intrinsically quantum mechanical motions of the proton and electron to the slower, classical motions of the surrounding environment. We propose to employ the large-scale NERSC computational resources to perform direct simulations of these processes to reveal the detailed mechanisms and nature of the dynamical coupling between the environment and the transferred quantum particles. These simulations will yield breakthroughs in our fundamental understanding of the reaction dynamics of energy-related chemical processes, a key NERSC research goal. In particular, we propose path-integral molecular dynamics simulation studies of PCET dynamics in solvated iron bi-imidazoline complexes, which are a key prototype for bioinorganic catalysis and an important model system for understanding solar photocatalytic water splitting in the chemistries of photosynthesis and respiration.

Turbulence over Complex Terrain: a Wind Energy Perspective

Edward Patton, National Center for Atmospheric Research

NISE award: 3,200,000 hours

NERSC repository: m917

Wind turbines are frequently deployed in regions of undulating topography to take advantage of the expected speed-up of wind as the atmosphere is forced up over the hill. A substantial portion of future wind farm deployments may also be offshore where turbines can be located close to consumers. Atmospheric interactions with these complex underlying surfaces produce highly variable and potentially damaging environments for turbines. Proper characterization of the connections between turbines and their environment is essential for wind turbine deployment strategies and for designing turbines capable of withstanding these environments. The project's goals include using numerical simulations to develop a fundamental understanding of the interconnections between wind turbines, vegetation, orography, water waves, heterogeneity, and stratification on turbulence within the Planetary Boundary Layer for improved turbine design, wind farm siting, and forecast skill.

The Materials Genome

Kristin Persson, Lawrence Berkeley National Laboratory

NISE award: 11,500,000 hours

NERSC repository: matgen

The energy and climate problem facing the world has highlighted the urgent need to accelerate the search and development of new materials. Many technologies to create, transfer, save, or store energy are critically dependent on materials innovation. We will accelerate the design of new materials needed for these technologies by 1) using scalable high-throughput ab-initio computations at an unprecedented scale, to rapidly predict and mine data on all inorganic materials in nature, in order to more rapidly and efficiently design new materials in the energy field. 2) make that data available in an organized way to the larger materials community, so that informed and effective choices can be made in materials research and development programs focused on energy.

Global Full-Waveform Seismic Tomography

Barbara Romanowicz, University of California Berkeley

NISE award: 1,500,000 hours

NERSC repository: m554

Seismic imaging provides our most powerful tool for examining structure of the Earth's interior. We can gain considerable insight into temperature, composition, and dynamics by accurately mapping physical properties at depth, as seen by seismic waves. While previous global-scale studies have relied extensively upon various approximations, our use of the SEM allows us to treat the physics of seismic waves traveling through the Earth "exactly". By treating wave propagation in this manner, we stand to resolve critical open questions regarding the origins and ongoing evolution of our planet.

Quantum Transport Simulation of Nano Scale Electronic Devices for Ultra Low Power Computing

Sayeef Salahuddin, University of California Berkeley

NISE award: 300,000 hours

NERSC repository: m946

We propose to build up a parallel simulation platform for the so-called Spin torque transfer (STT) devices which, by using exotic material properties at the nanoscale, promises to significantly reduce energy dissipation in electronic devices. Our transport calculation will be based on the Non Equilibrium Green's Function (NEGF) formalism which is currently regarded to be the state of the art for quantum transport simulation within the single electron picture. In a latter section we shall briefly describe the NEGF formalism and what is involved from a computational point of view. While this proposal is focused on the STT devices, the same transport platform can be used to simulate transport in other nanostructures where quantum effects are of significant importance, for example, nanoscale transistors and memory devices, electrical nano sensors and molecular electronics. Hence the simulation platform built during this work, the algorithms developed and the insights gained, will be generally applicable to a large array of different nano-structures.

Carbon Dioxide Gas Separation Using Nanoporous Graphene

Joshua Schrier, Haverford University

NISE award: 700,000 hours

NERSC repository: m1031

Separating mixtures of gases plays an important role in energy technology. For example, separating carbon dioxide from the methane in landfill gas is necessary to improve the heating-value and capability for power generation of this renewable resource. Another example, the separation of carbon dioxide from nitrogen, is necessary for reducing the greenhouse gas emissions from fossil-fuel based energy production. Both of these are challenging separations, due to the similarity of the weak intermolecular interactions of these molecules. Existing strategies are too energy-intensive to be feasible, so a breakthrough would be of tremendous environmental and economic importance. Our approach is based on using nanoporous forms of graphene, and carefully tuning both the size of the pores through which the molecules pass and the adsorption of molecules on the surface. Our preliminary calculations indicate that this has the potential to drop the energy costs for CO2/CH4 and CO2/N2 separations by a factor of 10 or more.  The calculations we propose here will allow us to obtain a precise figure of this value.

Attribute-Based Unified Data Access Service

Arie Shoshani, Lawrence Berkeley National Laboratory

NISE award: 1,525,000 hours

NERSC repository: udas

The proposed work will make data from astronomical observations easier to access and more convenient to analyze. This project intends to work with SDSS and Pan-STARRS data sets. Both of them involve a large number of partners from US with a number of DOE Labs. Improving data accesses could increase the scientific output from the investment already made in building the astronomical observations.

Mean-Field Solutions for Heavy Nuclei: Structure and Dynamics

Ionel Stetcu, Los Alamos National Laboratory

NISE award: 1,100,000 hours

NERSC repository: m1451

Atomic nuclei are complex quantum-mechanical systems whose description requires considerable computational resources. However, the nuclei are never isolated so the complexity increases even more when one considers the large number of processes in which they are involved. Such processes, from the nuclear response to electroweak probes, to nuclear reactions, or nuclear fusion and fission are of fundamental interest and have important applications in energy production, global security (e.g., nuclear forensics, nuclear detection, non-proliferation), medical applications, etc. Most of the approaches to describe nuclear systems, especially those involving nuclei with more than 20 nucleons, rely on phenomenological models which involve parameters fitted to reproduce experimental data. The extrapolation of the phenomenological models is unreliable and many times fails. We propose the use of a quantum-mechanical approach, the density functional theory and its extension to time-dependent phenomena, to describe nuclear systems and processes of interest. Such an approach has been widely applied in atomic physics and chemistry, but it was computationally too intensive to be applied to nuclei. We have taken advantage of the recent advances in computational power, developing the theoretical and software tools to attack the difficult description of nuclear systems.

Attribution of Extreme Weather Risk to Anthropogenic Emissions

Daithi Stone, Lawrence Berkeley National Laboratory

NISE award: 8,650,000 hours

NERSC repository: m1517

Whenever a damaging weather event occurs these days people ask "Was this caused by our emissions?" Currently climate scientists lack a resource that would allow them to respond both accurately and promptly. This project would generate a dataset that would allow researchers to quickly assess the degree to which our emissions have changed the odds of certain extreme weather events.

This work will contribute to two projects: the Weather Risk Attribution Forecast (WRAF), and the Attribution of Climate-related Events (ACE) activity of the Climate of the 20th Century project (a collaboration of about 30 climate modelling centres around the world). Both projects seek to examine the degree to which the probabilities of extreme weather events have been altered because of anthropogenic emissions of greenhouse gases and aerosols. This project would contribute simulations of the CAM5.1 atmospheric model. For the ACE project, 50 simulations at 1-degree horizontal resolution, each with slightly different initial conditions, would be run with evolving ocean surface temperatures, greenhouse gas concentrations, and aerosol concentrations over the last 50 years. Two more sets of 50 simulations would then be run under two estimates of the conditions of the world that might have been had humans never interfered with the climate (i.e. with greenhouse gas and aerosol concentrations at pre-industrial levels, and with ocean surface temperatures cooled according to two estimates of the warming pattern attributable to anthropogenic emissions). For the WRAF project these simulations would be incrementally advanced each month as observations of surface temperatures become available for the previous month, thus producing a near-real-time dataset for attribution study. Comparison of the frequency of extreme weather events in these different ensembles will provide a much-needed quantitative estimate of the degree to which past emissions are affecting our current weather.

Advanced Simulation of Pore Scale Reactive Transport Processes Associated with Carbon Sequestration

David Trebotich, Lawrence Berkeley National Laboratory

NISE award: 7,500,000 hours

NERSC repository: m1516

The injection of CO2 into the Earth's subsurface forces the subsurface system far from equilibrium, where a range of self-organizing processes can lead to emergent, time-dependent structures.  Current focus is on the structures that emerge due to process coupling at the pore scale, since this is ultimately the framework within which the fluids migrate and/or reside, and minerals dissolve and precipitate.  Research within the DOE Energy Frontier Research Center (EFRC) for Nanoscale Control of Geologic CO2 aims to establish the rules governing emergent behavior at the pore scale under far from equilibrium conditions, and to do so, is bringing to bear a new generation of experimental, imaging and modeling tools specifically designed to address this scale as it pertains to carbon sequestration. The intent of modeling and simulation is to not only inform the experiments a priori, but to also help interpret the experimental results, and to generalize the computational results to the larger (porous-continuum) scales for the purpose of more accurate macroscopic field scale modeling, which is ultimately the scale of interest in carbon sequestration. The primary objective of this project is to perform high resolution simulations of pore scale reactive transport processes to support and validate the experimental effort of EFRC while providing the basis for upscaling to the continuum (reservoir) scale.

Computational Prediction and Discovery of Magnet Materials

Cai-Zhuang Wang, Ames Laboratory, Iowa State University

NISE award: 13,100,000 hours

NERSC repository: m1515

This proposal covers two projects.

Firstly, permanent magnetic materials are essential in electrical generators using wind, water, and even carbon based fuels. Permanent magnets are also essential in electric motors for vehicles and other electro-mechanical devices, including levitators. Because of the role of such devices in new energy economies there is a great increased demand for strong permanent magnet materials. Currently, most widely used permanent magnetic materials use rare earth materials. Due to the potential shortage in resources and supply of rear earth materials, There is a strategic national need to find replacement materials beyond rare earth to meet the performance and cost goals for advanced electric drive motors.

Secondly, he urgent demand for new energy technologies has greatly exceeded the capabilities of today's materials and chemical processes. New materials and processes are critical pacing elements for progress in advanced energy systems and virtually all industrial technologies. The ability to predict the crystal and interface structures for any given stoichiometry by using computational algorithms is crucial for aiding the material design and accelerating the pace of technological advances. Accurate theoretical structure/property determinations will complement the traditional experimental efforts in material searches if they can be done rapidly. The algorithms, code, and experience created in this effort will be directly applicable to other accelerated materials discovery efforts. The database created will also become public and be available for other materials discovery efforts aimed at optimizing different material properties.

Spin-lattice Coupling in Magnetic Phase Transition

Yi Wang, Pennsylvania State University

NISE award: 6,000,000 hours

NERSC repository: m891

Quantitatively considering the role of the freedom of spin in thermodynamics has to be the key in understanding most advanced materials and phenomena. Solution of this can reveal the microscopic origin of the intriguing properties of many materials. The fundamental examples are the elemental metals Fe, Co, and Ni, each of which undergoes the well-documented ferromagnetic-paramagnetic transition at its Curie temperature.

However, since the discovery of the electron spin, it has been an enduring problem on how to precisely formulate the interplay among spin fluctuation, electromagnetism, and thermodynamics near the intervening critical regime in solid-state phase transitions. The current frontier is how to account for the effects of spin fluctuations at finite temperature on the Helmholtz energy within the framework of first-principles calculation.   We have proposed a general theoretical framework for a magnetic system at finite temperature. This is accomplished by reaching the partition function with specifying explicitly the microscopic Hamiltonian.

Understanding Multiple Exciton Generation and Charge Extraction in Nanoparticle-based Solar Cells

Stefan Wippermann, University of California Davis

NISE award: 5,500,000 hours

NERSC repository: m1518

Semiconductor nanoparticles (NPs) are remarkably promising for increasing solar cell efficiency and are expected to play an important role in upcoming 3rd generation solar cell architectures. Ideally, this project will 1) contribute to the fundamental understanding of nanoparticle-based solar cells and 2) will allow for qualitative predictions about how to optimize both the efficiency of charge extraction and generation of multiple electron-hole pairs per photon in realistic solar cell applications. This project will thus directly help to increase the energy conversion efficiency of next generation solar cells.

Guest-Host Interactions in Hydrate Lattices: Implications for Hydrogen Storage and Carbon Dioxide Sequestration

Sotiris Xantheas, Pacific Northwest National Laboratory

NISE award: 7,500,000 hours

NERSC repository: m1513

Hydrogen hydrate, a clathrate hydrate generated from hydrogen guest molecules inside hydrate host lattices, is one of the promising hydrogen storage materials. Its applicability as a low-cost hydrogen storage alternative providing a higher storage capacity is still debated because of issues related to its thermal stability near ambient conditions. The maximum hydrogen capacity of the hydrogen hydrate is 5.3 wt.% (mass H2 per mass H2O) for cubic structure II (sII) at pressures of P = 300 MPa and temperatures of T = 250 K, a value that is close to the US DOE's 2015 goal of 5.5 wt.%. Experiments have previously indicated that up to four H2 molecules can occupy the large cages and up to two H2 molecules the smaller cages. However, the high-pressure requirement for the stability of the hydrogen hydrate places a limiting constraint on its practical application. An alternative approach to reduce the storage conditions near ambient conditions (e.g. at P = 5 MPa and T = 280 K) is to use a binary hydrate with hydrogen and tetrahydrofuran (THF) with the expense of dramatically reducing the hydrogen storage capacity down to ?2 wt.%. A scientific challenge is therefore associated with devising ways to increase the hydrogen storage capacity in hydrogen hydrates near ambient conditions. We propose to perform first principles electronic structure simulations of hydrogen accommodation inside mixed hydrate lattices and suggest possible scenarios for enhancing the hydrogen storage capacity of those molecular scaffolds.