Abstract. Observations of the nighttime atmospheric potassium layer were performed on the German research vessel Polarstem from March to June 1996. K density profiles were obtained between 71øS and 45øN. The nightly mean peak densities ranged from 140 cm -3 in the equatorial region to 10 cm -3 in the Antarctic, and the column abundances decreased from 1.2 x 108 to 1.3 x 107 cm -2 going from low to high latitudes. High peak densities and column abundances were also commonly observed together with sporadic K layers. The global mean peak height of the normal (background) K layer was found to be 88.3 km. After the Polarstern campaign, observations were continued at Kfihlungsborn (54øN). The summer and winter K layers, observed during July 1996 and January 1997, were quite different in shape but had similar peak densities and column abundances. A one-dimensional model of the K layer was developed which includes meteoric deposition, vertical transport through eddy diffusion, and a full chemical scheme. This model was able to reproduce very satisfactorily the seasonal behavior of the K layer at 54øN if the wintertime deposition flux of the metal was reduced by 30% compared to the summer. The midlatitude ratio of K to Na was about 1%, much less than either the chondritic or cosmic ratios of the two metals (•8 or 6%, respectively). The most likely reason is that potassium vaporizes less efficiently from meteoroids than sodium, in agreement with a thermodynamic model of a nonideal chondritic magma and observations in the exosphere of Mercury. Finally, the model was generally very successful in reproducing the latitudinal variations in the K layer.
IntroductionThe deposition of extraterrestrial material in the Earth's upper atmosphere gives rise to layers of free neutral metal atoms in the altitude range from 80 to 110 km [Plane, 1991]. Investigations of the K layer were initiated by Lytle and Hunten [1959], who were able to estimate the ratio of potassium to sodium. Some years later, Sullivan and Hunten [1962, 1964] determined the peak density and column abundance from weak twilight emission data by using a birefringent filter photometer. These measurements were continued by Gault and Rundle [1969] The first model of mesospheric potassium was developed by Swider [1987] from an existing model of sodium chemistry. He proposed that the seasonal differences between Na and K might be explained by the kinetics of key potassium reactions having smaller temperature dependences than their sodium analogues. However, the sodium model that he used was quite 17,173