Accelerated Scientific Discovery projects on Derecho
The university and NCAR lab projects scheduled for completion in 2023 are described below.
University projects | NCAR projects
A rendering of the skin design for the Derecho supercomputer.
University projects
Dynamically downscaling CMIP6 GCMs across the western United States to quantify climate change uncertainty at landscape scales: WRF experiments conducted on convective-permitting scales
Project lead: Alex Hall, UCLA
Collaborators: Stefan Rahimi, Chad Thackeray, Lei Huang, Jesse Norris (UCLA), Zach Lebo (U Wyoming), David Pierce, Dan Cayan (SIO), Timothy Juliano (RAL)
Science domain: Climate
Derecho allocation: 56.82 million core-hours
We request nearly 57 million core hours on Derecho to dynamically downscale a 15-member GCM ensemble from the 6th Coupled Model Intercomparison Project (CMIP6) across western North America to convective permitting scales. The creation of such an ensemble will complement a pre-existing four-member ensemble and, for the first time, allow for the application of the emergent constraint (EC) technique to reduce uncertainties in future regional climate-change projections. Additionally, these data are to be used to train a state-of-the-art statistical downscaling model, which will then be used to downscale a majority of the best-performing CMIP6 GCMs in a novel hybrid downscaling approach. This dataset is being created with community usability in mind, as we post-process daily output of 36 key variables, selected in collaboration with fire and hydrologic modelers, atmospheric scientists, and stakeholder engagement specialists. With unprecedented water resources shortages, fire weather conditions, and heat extremes, such a dataset is needed immediately in order to reduce climate change uncertainties at landscape scales.
High-resolution integrated hydrologic modeling of the Continental US: a community resource to accelerate hydrologic discovery in changing systems
Project lead: Reed Maxwell, Princeton University
Collaborators: Laura Condon (U Arizona)
Science domain: Hydrology
Derecho allocation: 20 million core-hours and 62,200 GPU hours
We are requesting a combination of CPU and GPU resources to complete a set of novel national integrated hydrologic simulations with particle tracking. The proposed simulations will be the first long term national particle tracking simulations of their kind and will provide unprecedented ability to quantitatively assess whether the hydrologic cycle is accelerating (a major unanswered question in hydrology). Additionally, all of our simulations will be shared as a community resource through the HydroFrame NSF cyber infrastructure project and our tests will provide new GPU workflows for hydrologic modelers. Understanding the interactions between surface water, groundwater, and lower atmosphere systems is crucial to understanding and predicting the dynamics of water availability and extreme events. Our team has completed development and initialization of the next generation ParFlow-CONUS2.0 model. The new model has increased coverage area, additional geologic layering, and new topographic processing and overland flow formulations.
Producing regional climate simulations co-informed by Indigenous guidance to support climate change preparedness and resilience in Alaska and the Yukon
Project lead: Keith Musselman, CU Boulder
Collaborators: J. Hamman (CGD)
Science domain: Hydrology
Derecho allocation: 71.4 million core-hours
Climate change is transforming Arctic hydrology. The impacts of river transformation on Indigenous people, their fisheries, and winter travel corridors remain deeply uncertain. This collaborative project between CU Boulder and NCAR seeks to strengthen understanding of terrestrial hydrologic change in the Arctic and the potential impacts on rivers, fish and Indigenous communities. Guided by an Advisory Council of Indigenous leaders and representatives, our NSF-funded project is conducting regional climate simulations for a 30-year historical period and two future climate scenarios. We are using the Regional Arctic System Model (RASM) with the Community Terrestrial Systems Model (CTSM) applied at high spatial resolution over Alaska and the Yukon. The Accelerated Scientific Discovery (ASD) opportunity presents a unique outcome multiplier that would help us to better characterize the possible range of future climate and river conditions for Indigenous and First Nation decision-makers and the Arctic science community.
Assessing the impact of climate change and solar climate intervention on hazardous convective weather in South America
Project lead: Kristen Rasmussen, Colorado State
Collaborators: Jim Hurrell, Changhai Liu (RAL)
Science domain: Climate
Derecho allocation: 40,700,000 core-hours
To potentially offset anthropogenic warming and its associated impacts, there is a growing body of research on climate intervention (CI) methods. While concerns exist over the potential adverse effects that CI schemes may have if implemented, there is a growing realization of the need to research their impacts. The primary goals of our project are to produce ensemble convection-permitting regional climate simulations of convective storms in South America driven by large climate model ensembles in order to: (1) assess the influence of climate change and quantify the range of uncertainty associated with internal climate variability on the production of convective storms, and (2) determine how the impacts of stratospheric aerosol injection might influence mesoscale processes and convective storms. To accomplish these science goals, existing CESM climate change and solar climate intervention large ensemble simulations will be used to drive high-resolution convection-permitting (4-km) ensemble WRF model simulations to explicitly resolve convection in warm seasons over South America.
Turbulence and Lagrangian transport in the hurricane boundary layer
Project lead: David Richter, University of Notre Dame
Collaborators: G. Bryan (MMM), J. Dennis, S. Mickelson (CISL)
Science domain: Fluid Dynamics
Derecho allocation: 90,000 GPU-hours
The NSF-funded project associated with this request is focused on better understanding airborne sea spray droplet transport in the marine boundary layer. While the past several decades has seen great improvements in reducing the uncertainty of so-called sea spray generation functions (SSGFs) for small droplets, these estimates remain highly unconstrained for large droplets. More broadly, however, the NSF project is concerned with the issue of quantifying one of the largest sources of uncertainty in understanding spray-related feedback: a near complete lack of knowledge regarding turbulence and the resulting spray transport throughout the entire the high-wind boundary layer, including its vertical distribution in extreme conditions. As such, the Accelerated Scientific Discovery (ASD) program on Derecho provides an unexpected and unique opportunity to enhance the ongoing research on this project by allowing for high-resolution, turbulence-resolving large-eddy simulations (LES) of the high-wind boundary layer, including explicit Lagrangian representation of water droplets for collecting one- and two-point droplet transport statistics which are crucial for parameterizing spray distribution in the boundary layer.
Spontaneous Magnetotail Reconnection
Project lead: Mikhail Sitnov, Johns Hopkins University
Collaborators: John Lyon (Dartmouth College), Viacheslav Merkin, Kareem Sorathia (JHU)
Science domain: Magnetospheric Physics
Derecho allocation: 37,500,000 core-hours
Magnetic reconnection between the dayside magnetosphere and the solar wind magnetic field inserts energy into the magnetosphere. This is transferred to the inner magnetosphere by a process that begins with reconnection of magnetic flux transferred from the dayside to the magnetotail. This proposal seeks to address two questions: how exactly does the plasma and magnetic field transit from the reconnection site to the inner magnetosphere and how is the site of reconnection determined. This latter question is aided by new empirical models of the magnetosphere based on data mining of large volumes of spacecraft data. This proposal will study a sequence of global magnetospheric simulations using ideal MHD, as well as Hall and resistive MHD. Tracer test particles will also be launched using the fields coming out of these simulations to test the consistency of the MHD simulations and to suggest how they might be improved. The sequence of simulation runs will include three idealized cases as well as six that will attempt to model actual magnetospheric events observed by the NASA Magnetosphere Multiscale Mission.
NCAR projects
Benchmark simulations using a Lagrangian microphysics scheme to study cloud-turbulence interactions
Project lead: Hugh Morrison, MMM
Science domain: Fluid Dynamics
Derecho allocation: 50,000,000 core-hours
We propose novel simulations addressing key science questions related to cloud microphysics and dynamics as well as cloud-aerosol interactions. A hierarchy of simulations will use NCAR’s Cloud Model 1 (CM1) with Lagrangian super-droplet microphysics (SDM). SDM tracks Lagrangian trajectories of sampled “super-droplets” that each represent a multitude of real droplets in the flow. The Lagrangian approach represents a major shift compared to traditional Eulerian bulk and bin schemes and provides a significant advancement for representing cloud microphysics in models (see Grabowski et al. 2019, BAMS). Exploring the behavior of SDM across scales and for different cloud regimes is a fundamental part of the transition from the Eulerian to Lagrangian viewpoint.
Enhancing Earth System Predictability by Diagnosing Model Error in Mode Water Regions
Project lead: Ben Johnson, CISL
Science domain: Oceanography
Derecho allocation: 44,800,000 core-hours
This experiment uses an eddy-parameterizing and eddy-resolving model hierarchy to allow for scientists to contrast the behavior of ocean models in regions that are important for major modes of climate variability such as the Pacific Decadal Oscillation. The Pacific Decadal Oscillation exhibits an off-equatorial dipole core that coincides with regions of mode water formation in the ocean. Simulations of mode water formation in eddy-parameterizing models diverge from observations in these regions, suggesting significant model error that could be characterized via the model hierarchy. The project will produce a hierarchical reanalysis spanning seven complete years, 2010-2016.
High-resolution simulations of wildland fires and long-range smoke transport during the 2020 August Complex Fires in California
Project lead: Timothy Juliano, RAL
Science domain: Atmos Chemistry
Derecho allocation: 41,635,000 core-hours
We propose to conduct multiscale and multiphysics simulations of the 2020 California August Complex fires using the Weather Research and Forecasting (WRF)-Fire and WRF-Chem models, in addition to the MUlti-Scale Infrastructure for Chemistry and Aerosols (MUSICA) configuration of the Community Earth System Model (CESM). The coupled fire-atmosphere simulations, using WRF-Fire, will span the mesoscale and microscale. Emissions and plume rise information from WRF-Fire will inform WRF-Chemistry and CESM/MUSICA for a complete picture of wildfire smoke generation, transport, and evolution.
Extreme Weather Events Under a Wide Range of Climates in High-Resolution Coupled CESM
Project lead: Bette Otto-Bliesner, CGD
Science domain: Paleoclimate
Derecho allocation: 39,400,000 core-hours
We propose an unprecedented, landmark set of fully coupled high-resolution (HR) climate simulations for past greenhouse and icehouse climates to study the dynamics that govern the characteristics of extreme weather events in both atmosphere and ocean under altered climate states. We target well-studied paleoclimate intervals with higher and lower atmospheric CO2, including the preindustrial, the Last Glacial Maximum, the Pliocene, and the Early Eocene. We employ scientifically validated and extensively tested CESM code and configuration, the iHESP high-resolution CESM1.3 (~0.25° atmosphere/land and ~0.1° ocean/sea ice) with water isotopes. The unique water isotope capability enables unprecedented integration of information from model and paleoclimate observational data.
Response of tropical cyclone rainfall to thermal forcing in long-term convection-permitting simulations
Project lead: Rosimar Rios-Berrios, MMM
Science domain: Meteorology
Derecho allocation: 36,000,000 core-hours
Recent Hurricanes Maria, Harvey, Lane, and Florence had something in common: they brought record-breaking, catastrophic rainfall to their landfall locations. Their associated rainfall amounts were considered “extreme” in our current climate, but those amounts may become more common if our planet continues to warm at the projected rates. At the same time, an increasing body of literature suggests that interactions between clouds and radiation—known as cloud-radiative feedbacks (CRFs)—impact several processes in the tropical atmosphere, including convective organization, precipitation extremes, and tropical cyclone formation. Given the impactful nature of tropical cyclone rainfall, this proposed study will use convection-permitting idealized simulations to investigate if tropical cyclone rainfall will increase under a projected 4-K warming while also investigating if CRFs affect extreme rainfall in tropical cyclones.
Urban Air Quality Across the Globe with MUSICA
Project lead: Louisa Emmons, ACOM
Science domain: Atmos Chemistry
Derecho allocation: 34,000,000 core-hours
Air quality is primarily driven by local anthropogenic emissions sources, but it can also be strongly influenced by long-range transport of pollutants and regional influences. In turn, local air quality can have impacts that extend all the way to the global scale. Global models including CESM2(CAM-chem) with comprehensive chemistry in the troposphere and stratosphere usually perform well in reproducing distributions of important air pollutants, including those of ozone and particulate matter (PM2.5). So far however, global models are not able to increase the horizontal resolution sufficiently because of needed large computer resources for transporting 200-300 chemical tracers. This project will perform simulations of MUSICAv0 with a custom variable resolution mesh with 3 refined regions of special interest targeting the United States, Europe and southern and eastern Asia. This grid has a base ne60 (0.5-degree) resolution for most of the globe, zooming into ne240 (~1/8 degree) resolution in the 3 regions.
Global Convection-Permitting Simulations with GPU-MPAS
Project lead: Falko Judt, MMM
Science domain: Climate
Derecho allocation: 110,000 GPU-hours
We propose a series of global convection-permitting simulations using GPU-MPAS to better understand the dynamics of tropical convection, and the predictability of the atmosphere in different climate zones. Furthermore, we plan to assess the “added value” of convection-permitting resolution in simulating structure and life cycle of mesoscale convective systems across different climate zones, capturing the diurnal cycle and the duration, frequency, and intermittency of precipitation, predicting extreme weather from local to global scales, and representing orographic precipitation.
Data-inspired MURaM simulations of flares resulting from sunspot collisions
Project lead: Matthias Rempel, HAO
Science domain: Solar Physics
Derecho allocation: 87,400 GPU-hours
Major solar eruptions often originate from complex active regions, specifically in active regions that are composed of several bipolar spot groups that interact with each other. AR 11158 (Feb 2011) is a well-studied example in which two opposite polarities collide and form a flare productive collisional polarity inversion line (cPIL). We propose a data inspired simulation of AR 11158 with the MURaM radiative MHD code in which the observed spot motions will be imposed at the lower sub-photospheric boundary condition. Synthetic observables covering visible to EUV observations will be computed and compared to the available observations from NASA/SDO. The investigation will focus on connecting changes in the magnetic topology prior to flares to available observables, specifically constraining the build-up and release of magnetic energy. Unlike earlier MURaM simulations, this simulation aims for the first time at reproducing processes in a specific observed active region through data-constrained boundary driving.
Deep Learning-based Large Ensemble for Subseasonal Prediction of Global Precipitation
Project lead: Maria Molina, CGD
Science domain: Machine Learning
Derecho allocation: 50,000 GPU-hours
Our proposal aims to improve subseasonal prediction of precipitation using a data-driven, deep learning approach. NCAR’s next supercomputer, Derecho, will be designed with capabilities that enable state-of-the-art deep learning applications for Earth system science. With capabilities provided by Derecho, and as part of the Accelerated Scientific Discovery opportunity, we will use a data-driven, deep learning approach to create a 100-member ensemble of subseasonal forecasts of global precipitation, by leveraging deep learning with observational products to improve upon an already existing set of CESM2 subseasonal reforecasts.