NWSC science impact

Yellowstone and its associated cyberinfrastructure (CI) and the future generations of systems to be housed at the NCAR-Wyoming Supercomputing Center (NWSC) are being deployed to enable dramatic improvements in the scientific capabilities for a broad spectrum of important Earth System science applications.

NCAR justified the deployment of Yellowstone and the construction of the NWSC with an extensive evaluation of the potential science impact of petascale and exascale computing capabilities on the Earth System sciences. To understand how Yellowstone and future systems deliver on this potential, NCAR wants to encourage and identify computational activities that contribute to or achieve the “Community Science Objectives” in the NWSC Science Justification.

To that end, researchers making large-scale requests for Yellowstone access are encouraged to review the relevant portion of the NWSC Science Justification and include in their allocation requests how their projects’ objectives align with or contribute to the overall Community Science Objectives. Compiled from the much more detailed and complete NWSC Science Justification document, the following table summarizes the types of scientific objectives that will are possible with the NWSC and the computational platforms installed there.

The table is organized according to scientific domains and the six “HPC Dimensions of Earth System Modeling” identified by Figure 2.1 (p. 14) in the NWSC Science Justification (see attached). Each objective listed below includes a reference to a section and page number in the NWSC Science Justification.

To cite these objectives in your allocation request, you can identify them by their location in this chart rather than reproduce the text. Use this format:

[domain code].[dimension number].[objective number]

For example, “A.1.7” would refer to “Hurricane simulations at high resolution to properly represent hurricanes in coarser-resolution models.”

If you feel that an additional objective should be included in this table, please describe it in an email to alloc@ucar.edu and identify its location in the justification document.

Scientific Domain

(1) New Science
(new processes and interactions not previously included)

(A) Atmospheric Science

  1. Climate models integrated with models of evolving natural, anthropogenic systems (§2.4.1, p. 18)
  2. Next-generation climate models bringing in biogeochemical, ecological, and economic model components (§2.4.1, p. 18)
  3. ESM with dynamic global carbon and nitrogen cycles, terrestrial and oceanic biogeochemistry, land cover/use (§2.4.1, p. 18-19)
  4. By 2018 (exascale), multi-year ESM simulations with well characterized precipitation and hydrology (§2.4.1, p. 19)
  5. Improved treatments of sub-grid-scale physical, chemical and biological phenomena in component models of global climate models (§2.4.1, p. 19)
  6. Modeling the carbon cycle, including the ocean ecosystem model (§2.4.1, p. 20)
  7. Hurricane simulations at high resolution to properly represent hurricanes in coarser-resolution models (§2.4.3, p. 25)
  8. Hurricane simulations with full coupling to ocean and representation of ocean waves and detailed processes at the air-water interface and air-sea exchange (§2.4.3, p. 25)
  9. New techniques for explicit sub-grid-scale representations of dynamic, microphysical and radiative processes that make up the interactions of clouds with weather and climate systems. (§2.4.4, p. 28)
  10. Regional climate modeling in complex terrain coupling WRF with land surface and hydrological models covering the state of Wyoming (§2.4.5, p. 30)

(S) Space Weather

  1. First principles modeling of solar convection and its contribution to the 22-year solar cycle (§2.5, p. 33)
  2. Modeling the emergence of the magnetic flux from the solar convection zone and the conditions that lead to solar flares and CMEs (§2.5, p. 33)
  3. Guiding center particle codes modeling the energization and transport of radiation belt electrons, driven by electric and magnetic fields (§2.5.2, p. 40)

(O) Ocean Science

  1. Increased accuracy by replacing ad hoc parameterizations of sub-grid-scale processes with more realistic treatments from computational models, field observations and lab experiments (§2.6, p. 41)
  2. More accurate representation of mesoscale eddies, features of the mean flow (§2.6.1, p. 43)
  3. Adding parameterizations for processes such as inertial wave excitation, gravity currents, mixed layer stratification as the range of explicitly resolved scales increases. (§2.6.1, p. 43)
  4. Simulate regional oceanic ecosystem dynamics within a unified Earth system climate model, including ocean, atmosphere, food web, etc. (§2.6.2, p. 43)
  5. How saltwater turbulence is controlled by temperature, salinity, and velocity between the water layers; inclusion of parameterizations of saltwater mixing in larger-scale ocean models (§2.6.3, p. 46)

(E) Earth Science

  1. First principles simulations of the migration, trapping, and possible leakage of subsurface CO2 plumes (§2.7.1, p. 48)
  2. Very high resolution numerical simulation of the injection of CO2 into brine aquifers (§2.7.2, p. 49)
  3. Preliminary Monte Carlo studies of the injection process, aiming at the quantification and reduction of uncertainty. (§2.7.1, p. 49)
  4. Combine reservoir flow models with advanced seismic imaging techniques to predict post-injection CO2 saturation in aquifers. (§2.7.4, p. 52)
  5. Adaptive, particle-based pore-level modeling of subsurface flows (§2.7.6, p. 55-56)

(C) Computational Science

  1. Development of advanced algorithms to allow more detailed physical parameterization (§2.4.1, p. 18)
  2. Numerical techniques to solve large-scale compositional flows in fine computational grids (§2.7.1, p. 48)
  3. Distinct multiscale domain decomposition strategies in the parallel solution of elliptic, parabolic, and hyperbolic systems (§2.7.1, p. 48)
  4. Enhanced understanding of rotation, stratification, and magnetic fields in turbulent flows, as well as interactions of nonlinear eddies and waves (§2.8.3, p. 61)
  5. Petascale DNS to study intermittency and its effects in reactive flows, convective plumes, combustion, the solar corona, and space weather (§2.8.3, p. 62)

 

Scientific Domain

(2) Better Science
(from parameterizations to explicit models)

(A) Atmospheric Science

  1. Explicitly resolve all climate processes (rather than using parameterization or compensating for errors) (§2.4.1, p. 21)
  2. Nested Regional Climate Modeling to give better ideas about how weather patterns are likely to shift from decade to decade (§2.4.2, p. 24)

(S) Space Weather

  1. Coupled or integrated models of radiative MHD processes and magnetic flux emergence (§2.5.1, p. 39)
  2. Coupled magnetospheric model with ionospheric model that includes modeling chemical interactions (§2.5.2, p. 40)

(O) Ocean Science

  1. More comprehensive predictions by coupling physical, biological and chemical processes and human interactions (§2.6, p. 41)

(E) Earth Science

  1. New simulation tools for multiphase, multicomponent flow in multiscale heterogeneous porous media (§2.7.1, p. 48)

(C) Computational Science

  1. Better petascale algorithms to enable existing models to scale to greater domain sizes and compute more efficiently (§2.8.1, p. 59-60)
  2. DNS at increased Reynolds number to provide improvements and better parameterizations for complex flows in other Earth System models (§2.8.3, p. 62)

 

Scientific Domain

(3) Spatial Resolution
(simulate finer details, regions, and transients)

(A) Atmospheric Science

  1. Increase spatial resolution to allow the formation of fine-scale features and phenomena that were previously unresolved. E.g., CCSM with 0.1˚ ocean coupled to 0.25˚ atmosphere. (§2.4.1, p. 19)
  2. By 2018 (exascale), multi-year ESM simulations at cloud-resolving scales globally (§2.4.1, p. 19)
  3. In 10 years, 1-km resolution for cloud-resolving atmospheric components (§2.4.1, p. 20)
  4. Hurricane simulations at resolution <1-2 km horizontal, 50 m vertical, to predict the structure, wind intensity and rainfall (§2.4.3, p. 24)
  5. Severe storms, convective cluster simulations at 100-m resolution for thunderstorm convection and ~20-m resolution for tornadoes (§2.4.3, p. 26)
  6. Boundary layer modeling at 10-m to 20-m resolution across 100-km domains for studies of wind energy (§2.4.3, p. 26)
  7. Cloud dynamics simulations with DNS of microphysics within a domain of ~1 m3 and LES at grid spacing of ~1 m. (§2.4.4, p. 29-30)

(S) Space Weather

  1. Increased resolution to allow for resolving turbulent flows within the elements of sunspot fine structure and comparing with high-res observations from ATST (§2.5.1, p. 37)
  2. Simulating the formation of an active region in a domain 200k x 100k x 24k km at 48x24 km resolution (§2.5.1, p. 37)
  3. Simulating the evolution of umbral dots in a 6k x 6k x 2k km domain at 4 km resolution (§2.5.1, p. 38)
  4. MHD simulations of flux emergence in a vertical domain from 20 Mm below to 100 Mm above photosphere and 100 Mm x 100 Mm horizontal domain (§2.5.1, p. 38)
  5. Refine grid resolutions to accurately track interactions between solar wind plasma and the plasma in the CME (§2.5.2, p. 39-40)
  6. Global scale MHD simulations to model magnetosphere 20 Mm sunward and 200 Mm anti-sunward with a resolution of ~100 km in the inner magnetosphere (§2.5.2, p. 39-40)

(O) Ocean Science

  1. In 10 years, 10 km resolution for eddy-resolving oceanic components (§2.6.1, p. 43)
  2. Improved precision with increased spatial resolution of global- and regional-scale models(§2.6, p. 41)
  3. Fully coupled climate simulations with ocean component models with resolution of 10 km or better (§2.6.1, p. 43)

(E) Earth Science

  1. Increase multiphase flow model resolution beyond 200 m x 200 m to allow more accurate predictions of C02 sequestration capacity (§2.7.5, p. 55)
  2. Adaptive, particle-based pore-level modeling of subsurface flows, increasing the domain modeled from 0.065 mm3 to the scale of core samples, ~144k mm3 with ~1x1012 particles (§2.7.6, p. 58)

(C) Computational Science

  1. Development of advanced algorithms to allow increased spatial resolution (§2.8, p. 59)
  2. By 2018 (exascale), radical redesigns of “legacy” climate codes in use today (§2.4.1, p. 19)
  3. Code development to take advantage of adaptive or nested grids in simulations of sunspot fine structure or large-eddy simulation of oceans (§2.5.1, p. 38; §2.6.1, p. 43)

 

Scientific Domain

(4) Timescale
(more simulated time, shorter time steps)

(A) Atmospheric Science

  1. Modeling the carbon cycle, including the ocean ecosystem model over 3,000 years (§2.4.1, p. 20)

(S) Space Weather

  1. Extending sunspot simulations to 8 hours to study how fine structure changes with further thermal relaxation (§2.5.1, p. 37)

(O) Ocean Science

  1. Paleoclimate studies requiring integrations of many millennia to investigate events like ice age initiation and termination (§2.6, p. 41)

(C) Computational Science

  1. Development of advanced algorithms to allow finer-grained time steps (§2.8, p. 59)

 

Scientific Domain

(5) Data Assimilation
(decadal prediction, initial value forecasts)

(A) Atmospheric Science

  1. Advances in decadal predictions involving ocean initialization and integration (§2.4.1, p. 20)
  2. Assimilation into climate system models of the ever-expanding observational data sets (§2.4.1, p. 21)
  3. Hurricane prediction using data assimilation with ~100 ensemble members (§2.4.3, p. 26)

(O) Ocean Science

  1. Data assimilation for state estimation and initialization of ocean and coupled models, used for predictions from days to decades (§2.6, p. 41)

(C) Computational Science

  1. Algorithms for assimilation of observational datasets from remote sensing systems into climate system models (§2.8.2, p. 60)

 

Scientific Domain

(6) Ensemble Size
(quantifying statistical properties of simulation)

(A) Atmospheric Science

  1. Advances in decadal predictions involving ocean initialization and integration (§2.4.1, p. 20)
  2. Hurricane prediction using data assimilation with ~100 ensemble members (§2.4.3, p. 26)
  3. 30-100 ensemble members for accurate short-range prediction of convective storm at 1 km horizontal and 300 m vertical resolution (§2.4.3, p. 26)

(S) Space Weather

  1. Ensemble forecasting of Earth-Sun interactions driven by the variable solar wind (§2.5.2, p. 40)

(O) Ocean Science

  1. Expanded ensemble sizes to provide more robust statistical quantification of forced and natural climate variability (§2.6, p. 41)

(C) Computational Science

  1. Ensembles of ~100 members for data assimilation in Earth System models (§2.8.2, p. 61)

 

Scientific Domain

(7) Societal Impacts

(A) Atmospheric Science

  1. Climate change decision support, assessing and predicting the impact of policies on climate (§2.4.1, p. 18)
  2. Details in wind, wind variability from land-atmosphere interaction in landfalling hurricanes and incorporation into damage models — extremely high resolutions, 10s of meters (§2.4.3, p. 26)
  3. Better statistics from longer simulations and larger ensembles to study extreme (rare) events (§2.4.1, p. 19)
  4. Improved understanding of climate change issues and ability to investigate a broader range of scenarios in climate change impact and assessment research (§2.6, p. 41)
  5. Global ESM to study bioenergy and food production in a changing climate (§2.4.1, p. 18-19)
  6. NRCM to understand how high-impact weather events such as hurricanes may change in frequency, intensity, size and rainfall — needed to manage resources, protect lives, deal with societal changes (§2.4.2, p. 24)

(S) Space Weather

  1. Improved space weather predictions and their impacts on satellites, global positioning systems, and power grids—brining space weather forecasting to practical use (§2.5, p. 33)

(O) Ocean Science

  1. Ocean modeling and prediction system covering the US continental shelves to meet the needs of environmental assessment, forecasting, management, navigation and national security (§2.6.2, p. 45)
  2. Applying understanding of saltwater mixing to dispersion of hypersaline plumes from desalinization facilities (§2.6.3, p. 47)

(E) Earth Science

  1. Prediction of the effects and impacts of underground carbon sequestration used to mitigate global climate change (§2.7, p. 47)
  2. High-resolution reservoir simulations with sensitivity analysis or uncertainty management techniques to evaluate CO2 injection, determine optimal injection rates, and assess CO2 storage capacity and potential leakage. (§2.7, p. 47)

(C) Computational Science

  1. Enhanced remote sensing and aviation safety through detailed studies of stratified turbulence allowing for scale separation between different phenomena (§2.8.3, p. 63)