NCAR Strategic Capability projects 2013-2014

The role of coupled climate drift in efficiently initialized forecast simulations: Toward CESM intra-seasonal prediction readiness

Project lead: Julio Bacmeister, NESL/CGD/AMP
Yellowstone allocation: 6.4 million core-hours

Global climate models continue to suffer from equilibrium systematic biases that impact the skill in simulating present and past observed climate states. In CESM certain biases have persisted even in the presence of continued development of Community Atmosphere Model (CAM) parameterized physical processes and increases in model resolution to resolve finer scales. Systematic biases are also thought to play a significant role in initial value prediction problems. Forecast model skill beyond week one is not only hampered by the uncertainty of an inherently chaotic system, but also model drift to tropical equilibrium systematic biases. The tropical mean circulation and predominantly the MJO has a significant impact on week two forecasts over the continental U.S. This predictability is not realized in model predictions due to drift to the mean climate and a degradation in the MJO itself. Quantifying the role of these two types of error in hindcast model simulations will provide valuable insight into which systematic errors should be prioritized. Current methodologies for addressing systematic errors in GCMs tend to be simple and poorly defined. Therefore, we propose to use a nudging framework whereby model prognostic equations contain an extra forcing term to nudge the model state to the observed (reanalyzed) state. This nudging strategy will be applied to spin-up of the fully coupled CESM in order to provide the most accurate coupled state to initialize MJO hindcasts. Analysis will focus on the role of atmospheric resolution, nudging strength and location, and the starting time for the MJO hindcasts.

Fully coupled data assimilation with CESM + DART

Project lead: Gokhan Danabasoglu, NESL/CGD
Yellowstone allocation: 6.3 million core-hours

The scientific challenge of modeling the earth system with sufficient fidelity to inform society of the likelihood of climate change in the future is a formidable one for several reasons. These include the inherent complexity of the coupled earth system, the chaotic nature of its dynamical evolution, and the relative paucity of quality observations of the entire climate system. One of the most powerful methodologies that assist in tackling each of the above-mentioned limitations to scientific progress while providing useful climate information to society on seasonal to decadal time scales is data assimilation. The computer resources requested here will enable us to extend our earlier efforts in ensemble data assimilation for decadal climate predictions and begin the production of a first ever fully coupled reanalysis of CESM. Our science objectives are 1) to build from the current loosely coupled ensemble adjustment Kalman filter data assimilation that we have developed to produce our initial states for the decadal prediction activity in support of the CESM's participation in CMIP5, 2) further develop specific aspects of this system than can be improved, and 3) move all development to a fully coupled atmosphere and ocean assimilation for the purpose of producing a consistent climate system reanalysis for the 1960-2011 period. This final product will serve the dual purpose of producing i) a calibrated reanalysis data set against which we can compare climate simulation and prediction accuracy and ii) a set of ensemble initial conditions that can be used to initialize climate system forecasts of future climate anomalies on seasonal to decadal timescales.

MHD simulations of coronal mass ejections

Project lead: Yuhong Fan, HAO
Yellowstone allocation: 5.7 million core-hours

Coronal mass ejections (CMEs) are large-scale, spontaneous ejections of plasma and magnetic flux from the lower solar corona into interplanetary space and are major drivers of space weather near the Earth. It is generally believed that CMEs and eruptive flares result from a sudden release of the free magnetic energy stored in the previously quasi-equilibrium, twisted/sheared coronal magnetic fields. However, the detailed magnetic field evolution associated with CMEs and the physical mechanisms that cause the sudden eruptions remain fundamental unanswered questions under investigation. In this proposal we apply for Yellowstone allocation to carry out (1) high-resolution magneto-hydrodynamic (MHD) simulations of homologous CMEs driven by a continued, slow flux emergence at the lower boundary, and (2) an improved MHD simulation of the observed Dec. 13, 2006 CME event with a widened simulation domain. These simulations are aimed to understand the nature of the coronal magnetic field evolution that leads to CMEs and the resulting 3D magnetic field structure in the CME ejecta from the lower solar corona. These investigations are critically important for understanding the source and origin of space weather, and directly contribute to the NWSC community science objectives listed under New Science in the "Space Weather" science domain: "Modeling the emergence of the magnetic flux from the solar convection zone and the conditions that lead to solar flares and coronal mass ejections (CMEs)."

High-resolution time-slice experiments for climate-change prediction

Project lead: Richard Neale, NESL/CGD
Yellowstone allocation: 7.7 million core-hours

Our goal is to make climate-change projections that better simulate regional climate and extreme events through higher resolution. In particular, we plan to perform a series of time-slice experiments using the Community Atmosphere Model 5 (CAM5) at 0.25° global resolution. Our models are ready to run, so access through ASD would allow an even earlier glimpse of potential change in the climatology of extreme events. By investigating climate change with our newest atmosphere model, CAM5, with our newest dynamical core, spectral element, at a global resolution that begins to resolve extreme events, 0.25°, our proposal advances along three of the tradeoff dimensions of earth-system modeling defined for the NWSC: new science, better science, and spatial resolution.

A detailed evaluation of the water cycle over the contiguous United States including potential climate change scenarios

Project lead: Roy Rasmussen, RAL
Yellowstone allocation: 15 million core-hours

Recent high-resolution simulations using the WRF model over the Rocky Mountains revealed that proper spatial and temporal depiction of snowfall adequate for water resource and climate change purposes can be achieved with the appropriate choice of model grid spacing and parameterizations. A subsequent sensitivity experiment consistent with expected climate change resulted in an enhanced hydrological cycle with increased fraction of rain versus snow, increased snowfall at high altitudes, earlier melting of snowpack, and increased total runoff. A total increase of precipitation on the order of 10-25% was found in the pseudo-global-warming experiment, but the enhancement was less pronounced in the core of the Colorado headwaters due to a rain-shadow effect from upstream mountains. In order to investigate regional differences between the Rockies and other major mountain barriers, we propose to expand the modeling study to encompass the contiguous U.S. A domain expansion provides the opportunity to assess water cycle issues in the very sensitive western half as well as the very dominant role of convection in the eastern half. The planned simulations will leverage many other RAL and NCAR-wide imperatives and frontiers including: applications to wind and solar energy, surface and air transportation issues, predictability of severe weather, discovery of model biases, and advancement of state of the art numerical models.

The role of magnetic field in supergranular scale selection

Project lead: Matthias Rempel, HAO
Yellowstone allocation: 12.5 million core-hours

There are two confirmed length scales of convection on the surface of the sun: granulation (about 1,000 km diameter) and supergranulation (about 30,000 km diameter). Supergranulation has implications for our understanding of flux transport in the solar dynamo, chromospheric and coronal heating, and the total solar irradiance variation near solar minimum. More than 50 years after it was first observed there is no confirmed physical mechanism that selects the length scale of supergranules. Recent work has suggested that the network magnetic field may play a role in determining the length scale of supergranules, in part due to the change in supergranular size scale over the solar cycle. We study two possible mechanisms: 1. magnetic stresses may imprint larger convective scales from deeper in the domain up to the photosphere; 2. the increased radiative losses through the magnetic network elements due to decreased opacity may result in longer-lived convective downflows that organize into larger convective structures. We investigate these two mechanisms in a series of runs with different initial magnetic field strength, all in domains 98x98 Mm in width and 32 Mm in depth. Simulations of this size allow us to directly compare our simulations to observations and give us the resources necessary to understand the role of magnetic field in supergranular scale selection. This will yield a new understanding of the physics of radiative-magnetohydrodynamic convection, which is aligned with the NWSC Community Science Objective: First principles modeling of solar convection and its contribution to the 22-year solar cycle.

New frontiers in applying NCAR's WRF-DART ensemble data assimilation to probabilistic forecasts of severe convective weather

Project lead: Glen Romine, NESL/MMM
Yellowstone allocation: 8.63 million core-hours

Data assimilation and ensemble forecasting for severe convective storms are two frontier problems in weather prediction. Real-time ensemble data assimilation within MMM has produced exciting successes for both tropical cyclones and springtime severe weather and provides support for field programs such as the Deep Convective Clouds and Chemistry experiment. Existing computing limits the resolution of the data-assimilation system, the ensemble size, and the number and length of probabilistic forecasts. Because ensemble forecasts are highly scalable, the NWSC capabilities are an important opportunity to advance this research. Our proposed work will explore refinements of the ensemble predictions and increase the statistical significance of results by increasing the number and length of forecasts. It will also test, for the first time anywhere in the world, ensemble data assimilation that both resolves (or at least "permits") moist convection and employs a domain encompassing most of the continental United States. The tools for this work are the Weather Research and Forecasting model and the Data Assimilation Research Testbed, which together form a system we term WRF/DART. WRF/DART represents a long-term and active collaboration between NESL and CISL, which will continue under this proposal.

The CAM cloud perturbed physics experiment

Project lead: Ben Sanderson, NESL/CGD
Yellowstone allocation: 6 million core-hours

This experiment sets out to explore the parameter space in the new Community Atmosphere Model version 5.0 and Community Land Model version 4.0. The ensemble will contain approximately 100 50-year climate simulations, each with unique perturbations to the parameters in the cloud, microphysics and land-surface schemes. There are two stages to the experiment, the first being "exploratory" – a Monte-Carlo type exploration of the parameter space – and the second being an "optimal" ensemble of likely simulations in parts of the parameter space which are predicted to be the best performing. The final product will be a small (~10) number of complete parameter settings for the model which could be used by the community to produce a range of climate projections for a given scenario, thus sampling to some degree the uncertainty in that projection.

Meso- to planetary-scale processes in a global ultra-high-resolution climate model

Project lead: R. Justin Small, NESL/CGD
Yellowstone allocation: 21.9 million core-hours

This is a collaborative multi-institution request responding to the NWSC science justification of using global high resolution to explore interactions between different scales, from mesoscale to planetary. The main computational objective is to perform and assess Community Earth System Model (CESM) simulations with 1/8-degree atmosphere and land models and 1/10-degree ocean and ice models. The climate science objectives are (1) to investigate the climate response to the coupling of ocean and atmosphere mesoscale features, (2) to assess the ability of a high-resolution and frequently coupled (two-hour) ocean and atmosphere simulation to represent near-inertial waves in the ocean, and (3) to investigate the role of small-scale ice features such as polynyas in the climate system.

Large eddy simulations of high-wind atmospheric boundary layers above a spectrum of resolved moving wind waves

Project lead: Peter Sullivan, NESL/MMM
Yellowstone allocation: 12 million core-hours

The momentum and scalar fluxes at the atmosphere-ocean interface depend on numerous small and large scale processes that couple atmospheric turbulence and the underlying surface gravity (water) wave field. These fluxes impact winds in the marine atmospheric boundary layer and are crucial to high wind storm dynamics. The proposed computational project will use newly developed large-eddy simulation technology to examine the couplings between stratified atmospheric turbulence at high winds and a three-dimensional time-dependent spectrum of phase resolved moving surface waves. The scales of interest vary from 1 m to 1000m. The very high spatial and temporal resolution of the proposed simulations will provide information about the impacts of the wave field on sea-surface drag, turbulence fluxes and variances, mean wind profiles, and coherent structures in the marine atmospheric boundary layer. The turbulence simulation code to be used is highly parallel and runs on thousands of processors.