The National Institute for Computational Sciences

Mapping Layered Flows

Studying Atmospheric Turbulence to Better Understand Weather and Climate

By Jennifer Bailey


Weather and its aspects of temperature, wind, precipitation, clouds, and storms is something that crosses most people's minds at some point during the day, and recently many in the U.S. have wondered when warm weather would return.

This winter, America has faced bitter cold from the polar vortex that blew frigid temperatures into many states in the Midwest, Northeast, and Southeast regions, and it has endured ice and snow storms that have swept those same regions with a fury, crippling the South by stranding multitudes of motorists and leaving hundreds of thousands of homes and businesses without power.

February continued the atypical weather experienced in January. A storm blanketed the South with snow and left numerous areas along the East Coast without power. This particular storm made it necessary to cancel more than 6,500 flights across the U.S. and caused a wide variety of delays. Shortly before the storm, National Weather Service forecaster Jason Deese said the storm "certainly looks like it could be of historic proportions, especially in the last 10 to 20 years."

The Northeast, especially along I-95, received substantial snowfall, with totals of 12.5 inches in New York City to 11.5 inches at the Baltimore Washington International Thurgood Marshall Airport and 11.1 inches at the Philadelphia International Airport. Various areas in Maryland had well over 20 inches of snow, with Glyndon, MD, reaching 26 inches of snow by the morning of Friday, Feb. 14.

On the other end of the thermometer, warm-weather situations have also historically led to some dramatic weather conditions, as well.

In August 2005, Hurricane Katrina wreaked havoc. In May of 2009, the Southern Midwest Derecho—a widespread, long-lived, straight-line windstorm—brought massive damage by way of tornadoes and flooding. The Mid-Atlantic and Midwest Derecho came in June of 2012 and affected more than 700 miles across those regions of the U.S., taking the lives of 13 people. A few months later in October, Hurricane Sandy assaulted the East Coast and left in her wake at least 117 deaths, according to the Red Cross.

Probing Various Scales

Severe weather raises questions as to the causes of the phenomena that characterize it. The answer to all the questions is atmospheric conditions. The atmosphere, one of Earth's physical systems, consists of varying layers of gases or fluid structures—movable or deformable structures with an internal or surrounding fluid flow. When these different fluid structures interact, nonlinearities, or relationships that cannot be fully explained by the combination of their inputs, occur and influence weather conditions; and mapping the nonlinearities will help determine how weather is going to behave. Nonlinearity without explanation can lead to random, unforecasted outcomes such as chaos. Accurately predicting or modeling these occurrences, however, involves examining numerous structures or modes across a variety of scales, says Duane Rosenberg, a researcher at Oak Ridge National Laboratory.

"It is crucial in such applications as numerical weather prediction and climate modeling to understand the effects of scale interactions and the role of waves, because in these applications, the small scales are not computed directly, but rather, modeled," Rosenberg says. "Thus, the effects on the energy-containing scales that are computed are required for accurate modeling of these fluid systems."

Rosenberg and Annick Pouquet of the University of Colorado, Boulder, are currently investigating where energy introduced to a system travels. Among the questions they are asking are whether the energy goes to a larger scale or a smaller scale, and what the implications of these energy flows are.

In the types of turbulence traditionally studied, interactions typically occur between structures of similar scale. But rotating, stratified (layered) turbulence shows many indications that structures of various sizes interact and energy transfers occur. Rosenberg points out that devising theories from the observation of phenomena characterized by rotating, stratified flows of various scales is more difficult than doing so with traditional turbulence and its structures of similar scale. This means that relating observations to a set framework that can be used for future speculation is difficult. However, understanding these flows is key to interpreting current weather and weather patterns. The turbulence systems Rosenberg and Pouquet are modeling are aimed at clarifying the nature of the energy interactions between structures of different scales, with the goal of solving the relevant fluid system equations numerically as exactly as possible at each different scale. This, they believe, will make mapping flows easier in the future.

Rosenberg says another objective of the project "is to achieve the largest separation of scales that reflect those found in nature, such as in weather and climate systems, while including increasingly complex physical processes that are found in the system." Separation of scales, he explains, is what makes simple descriptions of a variety of properties of the world possible.

Toward the goal of optimal separation of scales, Rosenberg and Pouquet are compiling data and information that describe the system to create the most accurate and complex model of the system. Their main approach is not to model the physical processes but to incorporate them precisely at each scale. To accomplish this, they use direct numerical simulation, in which they solve certain fluid dynamics equations without turbulence models. This simpler form allows the solutions to be analyzed in detail and clearly exposes specific interactions that are also present in the more complex systems. "The level of detail realizable in our multiscale computations cannot be extracted from more physically complex integrated system models," Rosenberg says.

Characterizing Energy Fluxes with Kraken

Rosenberg explains that the Kraken supercomputer affords access to large-scale resolutions of the necessary number of nodes and the ability to solve the equations describing the interactions. The system's computing capabilities made it possible to see that the energy fluxes had a bi-directionality to them in the rotating, stratified turbulence runs, and the research will lead to even better models of such intricate flows and better understanding of them, he says.

Currently, Rosenberg and Pouquet, in conjunction with a couple of colleagues, are preparing a new paper titled "The effect of moderate rotation on stratified turbulence," which they anticipate submitting for publication in about a month to Physics of Fluids. In addition, they have a second paper, on small-scale properties of the flows, that is undergoing data analysis.

Article posting date: 18 February 2014

About NICS: The National Institute for Computational Sciences (NICS) operates the University of Tennessee supercomputing center, funded in part by the National Science Foundation. NICS is a major partner in NSF’s Extreme Science and Engineering Discovery Environment, known as XSEDE.