Twilights and stratospheric dust before and after the agung eruption.

Many absolute measurements of twilight sky radiance were made in six spectral ranges from lambda377 nm to lambda970 nm at 20 degrees elevation in the solar vertical, and mostly from sunset to 6 degrees -10 degrees sun depression (sd) in southern Germany, from fall 1962 to summer 1967, and at Bedford, Massachusetts, during winter 1967-1968. A strong change of twilight radiance, especially a large increase in the amplitude of the red/green color ratio between 4.5 degrees (main purple light) and 1 degrees sd, was observed in January 1964 following the explosive eruption of the Agung volcano in March 1963, but twilights were still abnormal and variable in early 1968. Dust mixing profiles of some typical twilight groups were derived, and the volcanic twilights show, as do other optical and direct soundings of the same period, a strong maximum near 20 km. A strong secondary red/green maximum (visually observed as secondary purple light) developed at sun depressions greater than 6 degrees of the volcanic twilights. This seems to have been caused by multiple scattering and not by mesospheric dust. A few twilight observations in 1964 from Australia indicate considerably more volcanic dust in accordance with transmission data, while no stratospheric dust is indicated in Arctic twilights of spring 1966. In order to verify the persistence of abnormal twilight conditions for more than five years, the dependence on red wavelength of the red/green color ratio is investigated, and some earlier twilight measurements are discussed. This persistence may have to be related to even higher amounts of dust in the tropical stratosphere.