NCAR Strategic Capability projects 2017

Forecasts of high-impact weather using convection-allowing models with 1-km horizontal grid spacing

Project lead: Ryan Sobash, MMM
Cheyenne allocation: 8 million core-hours

As computing power has increased, weather forecast models, such as NCAR’s Advanced Research Weather Research and Forecasting (hereafter WRF) model, are now routinely used to produce forecasts with fine horizontal grid spacing over larger domains. Models with horizontal grid spacing of 3-4 km are able to partially resolve the effects of deep convection, such that convective-scale features (e.g., squall lines, convective systems) are explicitly represented in the forecasts. These models are commonly referred to as “convection-permitting” models (CPMs), in contrast to models designed for coarser resolution simulations that simplify the effects of convective processes (e.g., parameterized convection). One- to two-day predictions of high-impact weather phenomena (e.g., severe weather events such as tornadoes, hailstorms, flash flooding, etc.) are markedly improved in CPM depictions of severe weather events. While CPMs with 3-4 km grid spacing have proven skillful in predictions relative to those with parameterized convection, further increases in model resolution (e.g., to 1 km horizontal grid spacing) are necessary to properly represent turbulent processes inherent in clouds and convection. Despite evidence of errors originating from inadequate grid spacing within idealized simulations of convection, the benefit of added resolution on one- to two-day real-data forecasts of severe convection remains uncertain. A few studies have investigated the role of grid spacing in next-day predictions, yet the results have been inconclusive owing to an insufficient number of cases examined and or deficiencies in their methodology. This project will produce a robust collection of full-CONUS WRF forecasts employing horizontal grid spacings of 3-km and 1-km in order to refine our knowledge of the situations where added resolution, and improved realism of convective processes, is beneficial for next-day CPM forecasts of high-impact weather phenomena across the United States.

Evaluating drag parameterizations for CAM using data assimilation techniques

Project lead: Julio Bacmeister, CGD
Cheyenne allocation: 13 million core-hours

This request is to continue work on diagnosing boundary layer drag processes that was begun under a previous NSC project, “Evaluating a new orographic drag parameterization for CAM using data assimilation techniques.” The representation of drag processes in the atmosphere continues to be a major source of uncertainty in both climate modeling and weather forecasting (ECMWF Workshop on Drag Processes in General Circulation Models, Sept 12-15 2016). In addition to conventional PBL surface drag, most atmospheric GCMs parameterize drag arising from two distinct subgrid orographic scale bandwidths. At scales longer than roughly 3 to 5 km, but not yet explicitly resolved by model dynamics, parameterizations based on gravity wave dynamics are used to estimate drag. For shorter scales, parameterizations based on nearly neutral sheared flow dynamics are used. The “handoff” between these processes is poorly constrained and stratification dependent. Furthermore, long-standing issues with stable PBL parameterizations remain, especially related to the diurnal cycle and wind turning. Some of these problems arise because of our incomplete understanding of turbulence in stratified boundary layers and the possible role of gravity waves. Other problems arise when global models are tuned to correct biases in the general circulation, for example by implementing “long-tail” stability functions or enhanced surface roughness. These tunings can be detrimental for the representation of the PBL in models when compared with observed PBL heights and wind-turning angles.

High-resolution tropospheric ensemble chemical data assimilation across the scales

Project lead: Jerome Barre, ACOM
Cheyenne allocation: 10.5 million core-hours

It is of crucial importance to integrate various observation types from satellite measurements for atmospheric composition into models across scales, i.e. from global model (CAM-Chem) to regional model (WRF-Chem). In this proposal, we propose to integrate the CAM-Chem/DART and WRF-Chem/DART systems in the same framework to accomplish the following tasks:

  • Task 1: Provide an integrative and evolutive chemical data assimilation system across scales. 
  • Task 2: Assimilate GEO observations concurrently with LEO observations and assess impacts at variable scales.
  • Task 3: Infer the unobserved species of the atmospheric state that include cross-correlated species and emissions estimation.

Using data assimilation to determine the predictability of the middle and upper atmosphere in WACCMX

Project lead: Nicholas Pedatella, HAO
Cheyenne allocation: 9.1 million core-hours

The variability in Earth’s upper atmosphere has historically been considered to be driven primarily by solar processes, including solar ionizing radiation and geomagnetic storms. However, in recent years it has become increasingly recognized that processes in the troposphere and stratosphere also contribute significantly to the upper-atmosphere variability. We propose to perform extensive simulations in the data assimilation version of WACCMX (WACCMX+DART) in order to study the coupling processes between the lower and upper atmosphere, as well as the upper-atmosphere predictability. Specifically, our objectives are to (1) demonstrate the capabilities of WACCMX+DART for reproducing the short-term variability in the upper atmosphere due to the lower atmosphere; (2) investigate the lower-atmosphere sources of the short-term variability in the upper atmosphere; and (3) quantify the predictability of the upper atmosphere. The proposed simulations will serve to advance present modeling and data assimilation capabilities of the whole atmosphere and significantly improve our understanding of lower-upper atmosphere coupling processes. Furthermore, relatively little is known about the predictability of the upper atmosphere, and we will shed light on how whole-atmosphere model data assimilation can potentially enhance the capabilities for forecasting the upper-atmosphere variability. Understanding the predictability of the upper atmosphere is of increasing importance due to the influence of space weather on our technological society, and our simulations will therefore help to advance the field of space weather predictions.

Role of stratospheric processes in predicting the NAO on subseasonal timescale

Project lead: Jadwiga (Yaga) Richter, CGD
Cheyenne allocation: 6.3 million core-hours

The North Atlantic Oscillation (NAO) is a phenomenon in the North Atlantic Ocean of varying differences in sea level pressure between the Icelandic low and the Azores high. During the positive phase, the Icelandic low is very low and the Azores high is very high, creating a large pressure gradient. Although the NAO is very important for weather in Europe and eastern U.S., its predictability is currently low. This proposal aims to improve the understanding of the role of the stratosphere on the predictability of the NAO and related weather extremes on a subseasonal (S2S), 2- to 8-week timescale. We plan to systematically explore the role of the stratosphere in the ability to predict the NAO by comparing numerous ensembles of S2S forecasts in a standard 30-level configuration of the Community Atmosphere Model and a 46-level (with a better-resolved stratosphere) configuration of the model.

Tools for CESM verification at multiple timescales

Project lead: Allison Baker, CISL
Cheyenne allocation: 12.6 million core-hours

We will generate an ensemble of 75 100-year runs with CESM 2.0 on Cheyenne. This ensemble will establish a baseline climate state for CESM 2.0 and allow an investigation of how different variables are affected by software and hardware modifications at different timescales. In particular, we will now look at the spectrum of variability beyond one year. This investigation will improve the feedback to climate scientists from the CESM Ensemble Consistency Test and will be valuable in terms of verification of future modifications.

High vertical resolution studies of the quasibiennial oscillation

Project lead: Rolando Garcia, ACOM
Cheyenne allocation: 8.4 million core-hours

Under a previous NSC project (NCGD0025), we used a high-vertical resolution version of WACCM to simulate the QBO as part of the international QBO inter-comparison project (QBOi; http://users.ox.ac.uk/~astr0092/QBOi.html). These simulations showed that (1) WACCM produces a realistic QBO under present-day climate; (2) the period of the QBO shortens as sea-surface temperatures (SST) warm with increasing CO2; and (3) for the warmest SST, WACCM produces interrupted west QBO cycles, similar to the unprecedented behavior observed in spring and summer of 2016 following the major ENSO event of 2015-2016. The proposed project will focus on the behavior of the QBO in a warming climate, addressing the following questions: (1) What is the momentum budget of the interrupted cycle found in WACCM simulations under warm SST conditions and its predictability? (2) Can WACCM reproduce the interrupted west cycle of the QBO observed in 2016? (3) What is the behavior of the QBO in the warming climate of the 21st century as simulated by WACCM? (4) What is the relationship between the phase of the QBO and the behavior of the winter polar stratospheric vortex? The results will be made available for studies at NCAR and by the university community. The proposed studies will also continue to support the QBO inter-comparison (QBOi) project.

An Earth system understanding of the end-Cretaceous mass extinction

Project lead: Clay Tabor, CGD
Cheyenne allocation: 8 million core-hours

Explanations for the end-Cretaceous mass extinction remain divided between an asteroid impact and volcanism. Both events overlap temporally and likely caused significant stress to flora and fauna. Currently, fossil assemblages and theoretical modeling work cannot rule out either hypothesis. With recent advancements in the Community Earth System Model, we are now able to simulate the chemical, physical, and dynamical responses to these abrupt forcings. Starting from a previously generated, fully coupled end-Cretaceous simulation, we will explore a wide range of emission scenarios from the Chicxulub impactor and Deccan Traps volcanism to better interpret the patterns of extinction across the Cretaceous-Paleogene boundary. Here, we will use the Whole Atmosphere Community Climate Model with active chemistry and the Community Aerosol and Radiation Model for Atmospheres for explicit aerosol calculations. With the addition of the ocean Biogeochemisty Element Model, we will better determine the consequences of these extreme perturbations on life. Our results will improve understanding of the mechanisms responsible for the K-Pg extinction and help settle this long-standing debate in the geological sciences.

Magneto-hydrodynamic (MHD) simulations of solar prominence eruptions

Project lead: Yuhong Fan, HAO
Cheyenne allocation: 5.6 million core-hours

Solar prominences/filaments (elongated large-scale structures of cool and dense plasma suspended in the much hotter and rarefied solar corona) are major precursors or source regions of coronal mass ejections (CMEs). The hosting magnetic structure of the prominence is likely a magnetic flux rope with helical field lines twisting about its center, supporting the dense prominence plasma at the dips of the field lines. In this project, we will carry out MHD simulations of the formation of prominences in coronal magnetic flux ropes under coronal streamers and their destabilization and eruption. Previous simulations have either (1) focused on the mechanisms for the destabilization and eruption of coronal flux ropes using simplified thermodynamics without the formation and the effects of the prominence condensations, or (2) modeled the formation and fine scale structure of the prominence condensations on an equilibrium flux rope. Here we will simulate both the formation of prominence condensations due to radiative instability in a large-scale coronal flux rope confined under a coronal streamer during the quasi-static phase as well as its evolution towards destabilization and eruption as the flux rope twist continually increases driven by the flux emergence.