218 



SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS 



all the satellites analyzed in Jacchia and 

 Slowey, 1963) we obtain temperature varia- 

 tions which are, on the average, smaller by 

 6 percent. 



It should be remembered, of course, that a 

 comparison between temperatures becomes 

 impossible in atmospheric regions where the 

 density is nearly independent of temperature. 

 This situation occurs for heights lower than 



+ 200° 



+ 100' 



AT 



-100° 



2000° 



500° 1000° 1500° 



Figure 2. — Correction to the exospheric temperatures ob- 

 tained from densities by use of the Nicolet II models to 

 reduce them to temperatures obtained using the present 

 models. 



200 km. at sunspot minimum; at sunspot 

 maximum, however, the nearly isopycnic layer 

 extends much higher, to about 300 km. At 

 these heights and in these conditions even a 

 minuscule difference in density corresponds 

 to enormous temperature differences. 



5. Formulae for the systematic temperature 

 variations 



Formulae for the variation of the exospheric 

 temperature for use with Nicolet's models 

 were given by Jacchia (1964). These formulae 

 necessitate some revision if we want to use 

 the present atmospheric models. 

 a. Variation with the solar cycle. — The relation 

 between the exospheric temperature T«, and the 

 10.7 cm. solar flux F l07 , both smoothed over 

 two or three solar rotations, shows practically 

 no departure from linearity in the new tem- 

 perature system. In figure 3 we have plotted 

 revised values of the nighttime minimum and 

 daytime maximum temperature from satellite 

 drag data covering the years 1958-1964. As 



can be seen, the smoothed nighttime minima 

 T can be represented by 



7' =418 +3?60F I o.7- 



(6) 



The bar indicates averages over two or three 

 solar rotations. The daytime maxima are 

 represented by 



T A/ = 1.287V (7) 



The smaller range of the diurnal variation 

 (by a factor of 1.28 instead of 1.30) reflects the 

 overall smaller temperature ranges explained 

 in section 4. It should be recalled that the 

 same diurnal density variation requires a much 

 larger temperature oscillation according to the 

 time-dependent models of Harris and Priester. 

 Although the latter are probably closer to 

 reality, the density variations are represented 

 equally well with the present static models. 



Equation (6) is valid for average quiet geo- 

 magnetic conditions (K p =2, a p =7). To re- 

 duce it to a p =0 the absolute term should read 

 357° instead of 418°. 

 6. Variation within one solar rotation. — We can 



use 



To = T + l°8(F i0 . 7— F io. 7) , 



(8) 



i.e., the same equation as given by Jacchia 

 (1964), but with the numerical coefficient 

 changed from 1?9 to 1?8. There is some indi- 

 cation that this coefficient might be somewhat 

 smaller (1?5 or so) near sunspot minimum and 

 larger (possibly 2?4) near sunspot maximum. 

 c. Semiannual variation. — We can use the for- 

 mula of Jacchia (1964), with a 6 percent 

 reduction in the amplitudes: 



T a =T, 



o+(o. 



37+0.14 sin 2tt 



rf— 151 



365 

 F w . 7 sin 4tt 



) 0) 



d-59 



365 



(d in days counted from January 1). 

 d. Diurnal variation. — The same parameters as 

 those found in Jacchia (1964) can be used, 

 except for R, which should be changed from 

 0.30 to 0.28. For convenience we shall repeat 

 the equations with their explanations. 



Let the temperature maximum occur at a 

 point on the globe which has the same latitude 

 as the subsolar point, and let the minimum 



