BU Today

Science & Tech

BU Profs Brace for Storms from Outer Space

Part two: Predicting solar storms like hurricanes


Click here to watch short videos of computer simulations created by the Center for Integrated Space Weather Modeling.

The solar eruptions that spewed high-energy particles toward Earth in 2003 portend the destructive force of space storms: in three weeks the storms knocked out one satellite and damaged others, triggered power outages, and disrupted air travel. Before the next cycle of solar storms revs up in the next 5 to 10 years, scientists want to be able to forecast solar storms as reliably as meteorologists predict hurricanes.

Warnings issued for the 2003 storms by the Space Weather Prediction Center (SWPC) and the Johnson Space Center were based largely on empirical models — probabilities of certain outcomes based on observations of the sun and solar wind calculated with data from historic solar conditions and the resulting storms.

Such empirical models are handy, because their relative simplicity means they can be run quickly on small computers. But they’re limited, explains Nicholas Gross (GRS’95), education director of the multi-institute Center for Integrated Space Weather Modeling (CISM) at Boston University and BU’s Center for Space Physics, by the sparsity of observation points between the sun and the Earth and by the relative scarcity of historical data on the rare events that are typically the most severe.

“If you run into a situation that you’ve never seen before, which happens a lot in space weather, then it’s very difficult to use empirical models to make forecasts,” Gross says.

“We need to understand the complexity of this huge system, with the sun and the Earth, plasmas and magnetic fields, and how it all interacts,” says Jack Quinn, a College of Arts and Sciences research professor of astronomy. “What are the rules that Mother Nature has given us?”

So CISM, established by the National Science Foundation and headed by Quinn and Jeffrey Hughes, a CAS professor of astronomy, is working on the front lines of a national effort to increase our preparedness for future solar storms. Researchers have been building physics-based models similar to those developed for terrestrial weather forecasting over the past 50 years. Instead of relying on data from past storms, these models create forecasts based on applicable laws of physics at work in the sun’s atmosphere, within the solar wind, and near the Earth.

In their quest to mimic Mother Nature with numbers, CISM researchers didn’t start from scratch. Beginning in the 1980s, scientists at several universities and research labs (many now CISM partner institutions) were developing models that focused on four different portions of the sun-to-Earth system — the sun’s outer atmosphere, or corona, the solar wind, the Earth’s protective magnetosphere, and the ionosphere in our upper atmosphere.

“At the core of each of these models is a grid,” explains Quinn. That is, the region of space or atmosphere being modeled is plotted by a three-dimensional series of points. Computers use variables, such as pressure and temperature, at each point to solve equations that represent the operative laws of physics. Change the density of these points, says Hughes, and you vary the resolution of the model.

“Where conditions are relatively uniform, you can have these points far apart,” he says. “Where things change rapidly, you need a lot of them to describe the change.”

The grids for the CISM models often contain more than a million points, necessitating a degree of number crunching that Hughes says can lead to “computational complications.” For example, a magnetosphere model adopted by CISM was developed on a grid that moved in a uniform direction dictated by the flow of the solar wind. Meanwhile, the grid of an independently developed ionosphere model rotates with the Earth.

“Each model developer did the sensible thing for his own purposes,” says Hughes. “But now you’ve got a rotating grid sitting inside a fixed grid, and they need to talk to each other.”

While six BU faculty members work in some capacity with CISM, much of the work of joining the models, known as coupling, is done by graduate students and postdoctoral researchers, among them Viacheslav “Slava” Merkin, a senior research associate in the BU astronomy department. Merkin has been working for more than three years to join the magnetosphere model with the ionosphere model. One of his tasks is writing code that can, as he says, “do all the dirty work of coupling,” such as interpolating data from one grid to another and converting units used to measure variables that can change from model to model.

Another senior research associate in astronomy, Mathew Owens, is testing a model of the sun’s corona coupled with a model of the solar wind. Owens and his team validate the forecasting capabilities of this combined model by measuring predictions it makes against actual solar observations and the resultant solar wind readings from a satellite stationed about a million miles above Earth called the Solar and Heliospheric Observatory (SOHO).

Despite all the number crunching, Owens says, observational work using SOHO keeps him focused on the big picture of what he’s modeling. “It goes back to being a kid,” he says. “I realized that being unfit and having bad eyesight meant I couldn’t be an astronaut, and so this was the next best thing.”

Storm chasers

In January 2004, President George W. Bush announced a new agenda for NASA: the agency would pursue “new journeys to worlds beyond our own.” Bush proposed that humans return to the lunar surface as early as 2015, create a permanent outpost there for research and mineral exploration, and plan for a manned mission to Mars.

To accomplish this without putting astronauts at serious risk, NASA must predict solar storms days in advance. That’s particularly important for a Mars mission, during which astronauts would spend a lot of time beyond the protection of the Earth’s magnetosphere.

Nathan Schwadron, a CAS associate professor of astronomy, says the astronauts of the Apollo missions in the late 1960s and early 1970s were lucky to have escaped serious injury from space storms. “If we had had more missions, eventually there would have been a flare or some very large radiation event from the sun that would have wreaked havoc,” he says. “The goal is to be able to give days and days of warning, so astronauts will be able to explore those environments.”

CISM’s work is not the only BU-led effort to help achieve that goal. Schwadron, for instance, is partnering with radiation biologists on a project known as the Earth-Moon-Mars Radiation Exposure Module (EMMREM), a three-year NASA-funded initiative that began in 2006.

“We have a couple of different models that we’re putting together,” he says. “One is a model of how charged particles are generated and move through space. The other is a model that predicts how that radiation will interact with atmospheres, with shielding on spacecraft, and with human tissue.”

Schwadron’s team will use data generated by CISM’s solar wind models. They will also refine their own model with measurements from a probe designed by Harlan Spence (CAS’83), a CAS professor of astronomy, and scheduled for launch later this year on board NASA’s Lunar Reconnaissance Orbiter. The instrument, the Cosmic Ray Telescope for the Effects of Radiation, or CRaTER, is made of tissue-equivalent plastic and will measure the effects of cosmic ray radiation on the human body, which could include radiation sickness and cancer.

Of course, the true test for both EMMREM and the CISM models is their usefulness in real-time forecasting at places such as the Johnson Space Center, where members of the Space Radiation Analysis Group wear pagers that alert them to major changes in the 24-hour data feed from solar observatories and satellites.

“If the pager messages tell us we’ve gone above certain thresholds, then we come in and stay around the clock until things have calmed back down,” explains Steven Johnson, a senior member of the Space Radiation Analysis Group.

“You start to work with these guys, and immediately you realize what they’re facing,” Schwadron says. “When we give them a very sophisticated model that has beautiful pictures of coronal mass ejections and can illustrate solar shock waves in 3-D, they’re not terribly impressed if it doesn’t give them any extra warning capability.” Coronal mass ejections are shock waves of solar plasma made up of charged protons and electrons that can slam into the Earth’s magnetosphere and cause geomagnetic disturbances that can knock out satellites and overload electric power grids.

Consequently, CISM recently turned over two coupled models to SWPC for testing and refinement (essentially, both halves of the sun-to-Earth model CISM plans to produce).

“Making a model operational is a huge challenge,” says Hughes. “It can’t crash. It’s got to work under a whole range of inputs. You can’t tune it after the fact to make yesterday work better.”

He points out that while researchers in lab conditions can clean up the data used in their models, operational models need to assimilate live feeds from observatories and satellites, “and that data has glitches and gaps that the model has to deal with on the fly and in a reasonable time.”

After each round of validating by CISM researchers and further testing by forecasters, the model’s physics and the code are tweaked. The goal is to show that physics-based models are superior to the simpler empirical models, says Hughes, not only for understanding the science of space weather, but for the forecasters safeguarding the world’s communications, infrastructure, and astronauts. It’s not about achieving a perfect model, he adds. “We’re not going to get there in our lifetime,” he says, “or in our graduate students’ lifetimes.” But even an imperfect model could save billions of dollars in avoided space weather disruption of communications satellites, power grids, and airlines, according to the 2006 federal Report of the Assessment Committee for the National Space Weather Program. The BU researchers hope to have a model that will allow at least three days of warning by 2012, just about the time the sun will be entering its next peak of storming.

For now, the race is on.

Click here to read part one of “BU Profs Brace for Storms from Outer Space.”

Chris Berdik can be reached at cberdik@bu.edu.

+ Comments

Post Your Comment

(never shown)