by J. William Bell, NCSA Senior Science Writer
Champaign, IL–Large-scale features of the sun have been known for centuries. Galileo and his contemporaries began observing and recording the existence of sunspots not long after the invention of the telescope in the early 1600s. And ancient Aztec creation describe a sun god with a pockmarked face, giving us what may be the first image of a sun with a mottled complexion. Sunspots, some of the most noticeable features on the sun’s face, measure as large as 50,000 miles in diameter, survive several months, and result from magnetic activity deep within the sun.
A face, however, is more than just its most distinguishing features. The upper regions of the sun create and are littered with smaller but no less significant details. Patches of hot gas called granules create bright patches about 1,200 miles across that last approximately ten minutes and are surrounded by lanes of cool, darker gas. Other tiny pores and bright spots leave their minute marks. Even magnetic fields too small to be observed from earth contribute to the sun’s countenance.
Robert Stein, a physics and astronomy professor at Michigan State University, and Aake Nordlund of Copenhagen University Observatory in Denmark are using NCSA’s SGI Origin2000 supercomputer to simulate the processes behind the sun’s smaller-scale features. Creating massive models of portions of the sun, their research team is focused on understanding convection and magnetic flux near the solar surface.
Although the sun’s energy is produced at its core, the ways that energy travels through different layers of the sun influence the nature of those layers. At the sun’s core, nuclear fusion creates helium out of hydrogen, and temperatures reach as high as 29 million F. The fusion-produced energy radiates most of the way to the surface through a region of the sun that is about 4.5 million F with a density similar to that of water.
In the outer third of the sun’s radius, however, a dramatic change takes place. At about 1 million F, this outer layer is cooler and has only about one-tenth the density of water. Energy is carried through this layer by the churning of the sun’s gases. This churning, called convection, is created as energy radiates into space at the surface, thus cooling the gases there. Because of their higher densities, the cooled gases fall while hot gases constantly rise.
“Looking at the sun in layers is somewhat arbitrary,” says Stein, “The lines get pretty fuzzy. But the convection zone is the region that drives much of what goes on at the surface. Most dynamic things happen there.”
Though most of the magnetic activity is born deep within the sun near the base of the convection zone, the near-surface region in which Stein and Nordlund are interested also sports small-scale magnetic activity of its own. This magnetism is inextricably tied to convection. The ionization level of the gas is related to its temperature and density, and temperature and density are determined by the vagaries of the ever-stirring sun. The ionized gases, in turn, serve as a great conductor, leaving convection and the magnetic field to constantly influence one another.
Stein and Nordlund have used supercomputers to study the nature of the sun for more than 15 years, and a fair amount of their work has been done at NCSA. Their newest simulation on the Alliance’s SGI Origin2000 at NCSA, however, is one of their largest undertakings to date. The simulation covers a swath of the sun that’s 18,000 km square and runs from the surface to 9,000 km deep.
Rather than creating the simulation through a variety of modeling applications, the team uses a single, integrated code. Their code is based on the laws of conservation of mass, momentum, and energy as well as the forces of pressure, gravity, and the magnetic field. Solving the equations that represent these laws and forces allow the researchers to see the essential physics at work, the radiative cooling at the surface that drives convection and the turbulent motions that generate small-scale magnetic fields and shuffle them around, for example.
On four to six Origin2000 processors, it takes about one day to simulate 30 seconds of time on the sun. Eventually, they hope to simulate about one day on the sun.
The team has been using this code for many years, constantly updating and improving it. Past simulations have already helped the team make a great deal of sense out of the region near the surface of the sun. They have discovered that moderately powerful magnetic fields influence the patterns of convection-created granules, making them smaller and more irregular. They’ve also given other researchers a whole new way to look at the nature of solar convection in general. Previously, most solar researchers characterized convection patterns as a hierarchy of eddies of decreasing size. Through computer simulations, however, the team found that convection actually flows in cool, turbulent downdrafts that plunge through hot, smooth upflows.
The newest simulations promise similar discoveries. The team hopes to discover how much small scale magnetic flux is generated by convection, how larger-scale structures are related to granulation, and how to calibrate sound waves to observe phenomena near the solar surface.
Trying to outdo themselves even as they complete their huge simulation of a day on the sun, the team is improving on an different, idealized version of the code that runs more efficiently on large numbers of processors. They are using better representations of the real sun instead of the simplified physics and boundary conditions that are used. They’re also completing a series of runs that looks at a smaller section of the sun in higher resolution. This high-resolution picture will couple nicely with the current massive-scale run, offering two different views of the same processes.
Computing time is relatively scarce, even after 15 years in the business. But, according to Stein, the scope of the projects gives researchers a lot to work with.
“Because of the size of the computations we do, we can’t do that many runs at the same time. But there are always many different ways to look at any of our simulations-many different things to be considered and understood,” he says. “The beauty of it is we’ve only got one sun.”