220 



SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS 



magnetic storms the temperature variations lag 

 about 6 hours behind the variations in a p 

 (Jacchia and Slowey, 1964b). There is evidence 

 that AT/a p is somewhat larger in high geo- 

 magnetic latitudes (Jacchia and Slowey, 1964c). 



6. Limitations of the present models 



As we stated in section 1, atmospheric models 

 must suffer from the oversimplified assumptions 

 one is obliged to make to construct them. Our 

 models share with those of Nicolet the limita- 

 tions imposed by the invariance of the tempera- 

 ture profiles and of the boundary conditions; 

 this latter limitation is common also to the 

 Harris-Priester models. 



A consequence of the fixed boundary condi- 

 tions is a nearly isopycnic layer at 200 km. at 

 times of moderate to high solar activity. At 

 such times, according to the models (ours, 

 Nicolet's, and the Harris-Priester models), the 

 density at 200 km. should not show appreciable 

 variations when the exospheric temperature 

 varies. This condition is nearly fulfilled by the 

 diurnal variation which practically disappears 

 at heights lower than 200 km. On the other 

 hand, density variations at the 200 km. level 

 have been observed at times of high solar 

 activity in correspondence with geomagnetic 

 storms, and also of the erratic ("27 day") 

 component of the 10.7 cm. flux (Jacchia, 1959). 



The different response of the density at 200 

 km. to different types of heating could be ex- 

 plained by assuming that the temperature at 

 120 km. is not subject to a diurnal variation, 

 but increases in correspondence with geomag- 

 netic storms and transient enhancements of 

 solar EUV radiation. If we increase the tem- 

 perature at 120 km. by 50° without changing 

 the composition, the density at 200 km. will 

 increase, according to our models, by a little 

 over 30 percent when the exospheric tempera- 

 ture is about 1400° K. This is just about the 

 order of magnitude of the erratic density 

 changes observed in Sputnik 2 and 3. At 

 greater heights the density change is more or 

 less the same, decreasing only slightly with 

 height, but its relative importance becomes 

 smaller because of the increased response of the 

 density to changes in the exospheric tempera- 

 ture (or, to be more accurate, to changes in the 

 corresponding temperature gradient above 120 

 km.). 



Satellites at heights as low as 160 km. have 

 recently shown that the density changes during 

 magnetic storms are in phase with those at 

 greater heights (Zirm, 1964). This indicates 

 that most of the heating during these storms 

 must occur at heights considerably lower than 

 160 km. It therefore looks highly probable 

 that the temperature at 120 km. must undergo 

 changes during a magnetic storm. 



If we assume that also the erratic changes in 

 solar EUV affect the temperature at 120 km., 

 it is difficult to see how the much larger varia- 

 tions of EUV in the course of the 11 year solar 

 cycle could leave the temperature at 120 km. 

 undisturbed. Perhaps there is such a change 

 and the construction of better models will be 

 possible when this change becomes known. 



7. Comparison with recent satellite-drag 

 data at heights below 200 km. 



A valuable collection of drag data on satellites 

 with low perigee heights has been recently 

 presented by Small (1964). These data extend 

 in an unbroken series to heights as low as 160 

 km., and for one satellite (1962 pa) to 126 km. 

 Apart from the assumed boundary conditions, 

 our atmospheric models are based on drag data 

 from satellites with perigee mainly above 250 

 km. and were completed before we had knowl- 

 edge of Small's densities. It was gratifying to 

 find that the agreement of these densities with 

 our models is excellent, as can be seen from 

 figure 4. In this plot we divided the data into 

 three groups according to the mean exospheric 

 temperature prevalent at the pertinent time, 

 in addition we have separately marked the 

 points derived from Sputnik 3 (1958 52), which 

 are particularly numerous and may be affected 

 by a small systematic error. 



According to our models log p (p= density) 

 at 180 km. varies by about 0.2 from sunspot 

 maximum to sunspot minimum. Since the 

 residuals in log p for the three temperature 

 groups do not show any clear evidence of 

 systematic differences, we must conclude that 

 our models represent rather well not only the 

 average densities, but also their variations. 

 Since, however, the density variations below 

 200 km. are relatively small, the agreement with 

 observations in this region must be ascribed 

 mainly to the boundary conditions, which are 

 obviously satisfactory. The increase in scatter 



