AEROSOL EFFECTS

Simulations show that tiny particles can influence climate

Scientists in the Atmospheric Science Division at Lawrence Livermore National Laboratory (LLNL) used NERSC’s Seaborg supercomputer and LLNL’s IMPACT atmospheric chemistry model to perform simulations showing the historic effects of sulfate aerosols at a finer spatial resolution than ever done before. Simulations were carried out for five different decades, from the 1950s through the 1990s (Figure 1).

The results clearly show the effects of the changing global pattern of sulfur emissions. Whereas in 1950 the United States emitted 41 percent of the world’s sulfate aerosols, this figure had dropped to 15 percent by 1990, due to conservation and anti-pollution policies. By contrast, Asian emissions jumped six-fold during the same time, from 7 percent in 1950 to 44 percent in 1990.

Figure 1. Predicted effects on shortwave radiation (watts/m2) due to anthropogenic sulfate aerosols in the 1950s (top left), 1970s (top right), 1980s (bottom left), and 1990s (bottom right). Note the increased regional cooling effects in central Asia, India, and particularly eastern Asia.

Under a special allocation of computing time provided by the Office of Science INCITE program, Dan Bergmann, working with a team of LLNL scientists including Cathy Chuang, Philip Cameron-Smith, and Bala Govindasamy, was able to carry out a large number of calculations on 128 and 256 processors.

The objective was to assess the effects of anthropogenic (man-made) sulfate aerosols. The IMPACT model calculates the rate at which SO₂ (a gas emitted by industrial activity) is oxidized and forms particles known as sulfate aerosols. These particles have a short lifespan in the atmosphere, often washing out in about a week. This means that their effects on climate tend to be more regional, occurring near the area where the SO₂ is emitted.

To accurately study these regional effects, Bergmann needed to run the simulations at a finer horizontal resolution, as the coarser resolution (typically 300 km by 300 km) of other climate models is insufficient for studying changes on a regional scale. Livermore’s use of the Community Atmosphere Model (CAM3), a high-resolution climate model developed at the National Center for Atmospheric Research (NCAR) with collaboration from DOE, allows a 100 km by 100 km grid to be applied. NERSC’s terascale computing capability provided the needed computational horsepower to run the application at the finer level.

Depending on their composition, aerosols can either scatter or absorb sunlight, thereby cooling the Earth and acting as a counter to the warming effects of greenhouse gases. Greenhouse gases such as carbon dioxide are much more long-lived, so they stay in the atmosphere and have more uniform distribution. But since global warming has not increased as much as some computer models predict based on greenhouse gases alone, scientists have become more interested in aerosols and their possible role in countering some effects of greenhouse gases.

The atmospheric concentrations of both aerosols and greenhouse gases have increased over the past century and, because of their effects on shortwave and longwave radiation (which are partially offsetting), have presumably upset to some degree the thermal equilibrium of the climate system. While satellite measurements and field studies provide vital information to document the global and regional climate impacts by aerosols and greenhouse gases, accurate climate model simulations are an equally important tool in quantifying the radiative forcing as well as identifying and analyzing the climate response.

By running numerous calculations at finer resolution, the team was able to see the cooling effects by region, and to see more differences in the effects of varying concentrations.

The simulations also gave the team an opportunity to test new meteorological data generated using LLNL’s version of NCAR’s CAM3. This model calculates the various factors influencing climate, such as winds, temperatures, and precipitation. This data, which bordered on the terabyte scale, was also generated at NERSC.

The LLNL researchers then used the climate data to drive their chemistry model, known as IMPACT. By varying the SO₂ emissions levels while running IMPACT, the team was able to simulate different scenarios for each decade, analyze aerosol distributions from each simulation, and then examine their effects on radiative forcing at regional scales.

Research funding: BER, INCITE, LLNL