Aerosol detection techniques using infrared wavelengths have a distinct advantage over visible techniques by providing coverage over bright surfaces and during the night. This study investigates detection of volcanic and soil-derived aerosols, two important aerosols in studies of the earth's climate, using infrared observations at the following approximate wavelengths 8.5, 11, and 12 μm. Detection is based on brightness temperature differences among the three channels BT11-BT12 and BT8-BT11. It is demonstrated that these three infrared channels are useful for detecting stratospheric volcanic aerosols over oceans. Theoretical simulations agree with observations from current satellite instruments. Detection of the stratospheric aerosol over land is complicated by spectral variation of surface emissivity. Retrieving aerosol optical depth over land requires defining the surface spectral emittance. Detecting the presence of soil-derived aerosols can also be aided with infrared observations. Increasing the dust optical depth increases BT11-BT12 and BT8-BT11. The effect is opposite to that of an H2SO4 stratospheric aerosol and differs from an increase in atmospheric precipitable water, though addition of ice clouds moves the differences in the same direction. Retrievals of aerosol optical depth over the desert must account for surface emissivity and the vertical distribution of the dust. Negative differences in BT11-BT12 are observed to occur for dust storms over the Arabian Peninsula, Africa, and the southwest United States and is useful for remote sensing source regions of dust outbreaks. These negative differences can be simulated using the theoretical model but requires a specific dust aerosol model. There are inconsistencies between theoretical simulations of the infrared properties of heavy dust loadings and the satellite observations. Negative differences in BT11-BT12 are useful for detecting and tacking dust storms. ¿ 1997 American Geophysical Union