A Numerical Experiment on the Effects of Regional Atmospheric Pollution on Global Climate

Abstract

The 1973 Rand version of the Mintz-Arakawa model of the general circulation of the atmosphere was used to study the consequences of high concentrations of hygroscopic aerosols in a limited geographic region. The experiment was designed to investigate effects of the activity of the aerosol as condensation nuclei and the consequent abnormal production of cloudiness and alteration of precipitation were of primary concern. Values of meteorological parameters generated by the experiment were compared with values simulated by a control (in which the polluted area was absent). Statistical tests of significance were performed in order to judge the certainty that experimental changes, rather than random wandering of the simulations, actually caused changes in certain variables. Cloudiness increased over the polluted area and the temperature at the surface and in the lower atmosphere became warmer. Rainfall, however, was not significantly altered in the polluted region in spite of changes in the parameterization of rain that would make rainfall more difficult with a given atmospheric structure. The data suggest a possible connection between North American pollution and South American rainfall. Comparison of data from the experiment and the control indicate that in the experiment, rainfall decreases gradually and steadily within a large area of South America centered on the equator. The immediate cause is the movement of the region of maximum moisture convergence northward, toward the polluted area, while it is at the same time decreasing in value. The movement is clear and the rainfall changes were found to be statistically significant. Although possible causes are discussed, the model does not properly simulate the hydrological cycle in the tropics, and a definitive explanation of decreasing rainfall in South America is marred by an unrealistic, unstable, moisture-convergence/rain cycle that occurs in the primary control. This behavior of the model precludes a conclusion that North American pollution will cause decreased tropical rainfall, but the data are sufficiently suggestive that this possibility should be examined in future investigations.