### I N S T I T U T D ' A E R O N O M I E S P A T I A L E D E B E L G I Q U E 3 - A v e n u e Circulaire

### B - 1 1 8 0 B R U X E L L E S

**A E R O N O M I C A A C T A **

### A - N° 1 3 7 - 1 9 7 4

### N e u t r a l a t m o s p h e r e m o d e l i n g by

### G . K O C K A R T S

### FOREWORD

### This paper has been presented at the Summer Advanced Study Institute in Liège (Jul>

### 29 - August 9. 1974). It will be published by D. Reidel Publishing Company, Dordrecht- Holland in a book edited by B.M. McCormac.

### AVANT-PROPOS

### Cet article a été présenté lors du "Summer Advanced Study Institute" à Liège (29 juillet - 9 août. 1974). Il sera publié dans un livre édité par B.M. McCormac chez D Reidel

### Publishing Company. Dordrecht-Holland.

### VOORWOORD

### Deze tekst werd voorgedragen tijdens het "Summer Advanced Study Institute" te Liège (29 juli - 9 augustus 1974) en zal in een boek gepubliceerd worden door D. Reidel Publishing Company, Dordrecht-Holland (B.M. McCormac, ed.).

### VORWORT

### Dieser Text wurde während des "Summer Advanced Study Institute" in Liège

### (29.Juli-9.August 1974) vorgestellt. Er wird in einem Buch durch D. Reidel Publishing

### Company, Dordrecht-Holland (B.M. McCormac, ed.) herausgegeben werden.

### NEUTRAL ATMOSPHERE MODELING

### by

### G. KOCKARTS

**Abstract **

**Abstract**

### This paper is a review of the physical problems associated with the construction of thermospheric models.

**Résumé **

**Résumé**

### Cet article décrit les problèmes physiques associés à l'établissement de modèles thermo- sphériques.

**Samenvatting **

**Samenvatting**

### Deze tekst beschrijft de fysische problemen die verbonden zijn met de constructie van thermosferische modellen.

**Zusammenfassung **

**Zusammenfassung**

### Dieser Text beschreibt die physikalischen Problemen, die mit dem Augbau thermo-

### sphärischen Modellen verbunden sind.

*INTRODUCTION *

### In this introductory paper, no attempt will be made to describe any particular model or to compare different models. Emphasis will be given to the physical phenomena occurring in the terrestrial upper atmosphere.

### A perfect atmospheric model is actually a set of numbers which is intended to represent the physical properties of the atmosphere in the past, in the present and in the future. This is a very ambitious goal which is far from being reached. There are, however, essentially three types of atmospheric model used for our partial understanding of the upper atmosphere.

### The theoretical models result from the solution of some equations with simplifying hypotheses. These models always require a knowledge of boundary conditions which should *come out of experimental data. Most of these models are discussed by Blum et al. (1972). *

### The semi-empirical models are usually based on a certain amount of observational data combined with some theoretical background such as the diffusive equilibrium hypothesis.

### Among the most widely used models are those of Jacchia (1965; 1971) which are based on satellite drag data. These semi-empirical models were initially guided by the theoretical approach developed by Nicolet (1961a; 1961b).

### The purely experimental models are a direct representation of physical parameters measured in situ by satellites. Since these models are by definition limited in time and in space, they lead very often to semi- empirical models by extrapolation and interpolation *processes. As an example, the OGO-6 model (Hedin et al. 1974) gives a worldwide distri-* bution of temperature and concentrations by a spherical harmonic analysis of measured data. The ESRO-IV data described by von Zahn (1975) will probably give the next experi- mental model of the upper atmosphere.

### It is clear, however, that all these models can make a valuable contribution to the knowledge of the upper atmosphere, not only through their internal results, but also by their mutual interaction.

- 2 -

*2: APPLICA TION OF THE CONSER VA TION EQUA TIONS *

### The general conservation equations (Chapman and Cowling, 1970; Landau and Lifshitz, 1959) are usually simplified for atmospheric applications but, unfortunately, the simplifying assumptions are not always clearly stated in the literature. This situation can lead to some confusion when different results are compared.

### For each atmospheric species with concentration nj, the general continuity equation is

### (dn,/3t) + V . n , ^ + V.) = Pj- Lj (1)

### where P

;### and L

(### are respectively the production and loss rates of the constituent i. is the diffusion velocity and T

**Q**### is the mass average velocity. In a gas mixture, the diffusion velocity is characterized by the relation

### S n ^ V j = 0 (2)

### where n^ is the mass of the particle i. Using this relation, the total density (p =2n

j### m

j### ) continuity equation is obtained from (1) as

### Op/91) + V . ( p ^ ) = 0 (3)

### The term SCP-Lj) vanishes, since the total mass of the system is conserved. It should be noted that the use of Equation (3) for a single constituent is incorrect unless the diffusion velocity V

;### and the production and loss terms are zero. Extensive use has been made of Equation (1) to solve the problem of atomic and molecular oxygen distributions in the lower thermosphere.

### Although the molecular diffusion velocity is given by gas kinetic theory (Chapman and

### Cowling, 1970), the mass average velocity T

0### requires, however, the solution of the

### momentum equation. Considering the viscosity n as constant and neglecting the gradient of

### Vv , the momentum equation for a gas mixture, can be written

### P [(3v„/at) + v V v j + Vp-ju V

^{2}

### ^ . 2 rijX. = 0 ^{(4) } where the total pressure is the hydrostatic pressure and X

^{(4) }

^{t}

### is any external force acting on the i-type particles. In a rotating frame of reference

### where the first three terms respectively represent the gravitational force, the Coriolis force and the centrifugal force at a geocentric distance r*"when the angular velocity is oT The external forces F

^{(}

### are the frictional forces such as ion drag. One sees that it is through these frictional forces that the ionosphere can interact with the neutral atmosphere. Applications of the momentum equation to the determination of winds in the upper atmosphere are described by Kohl (1975).

### Any solution of the momentum equation requires a knowledge of the thermal structure given by the energy equation. When P and L are respectively the heat production and loss, the energy equation is given by

### where T is the temperature and c

^{v}

### is the specific heat at constant volume for the gas mixture. The heat flow E is given by

### E =-XVT+ T E c

^{p j P i}

### Vj (7)

### where X is the heat conductivity of the gas mixture and c

^{p j}

### is the specific heat at constant pressure for the i-type particle.

### = [ g - ^{ 2c} ^{T} ^{xv} ^{^} - oTx(c3xr5] m

^{;}

### + (5)

### p[(3c

^{v}

### T/at)+ vV(c

^{v}

### T)] +

^{ P}

### V.vV.E-Sn

^{i}

### F

^{i}

### V.

^{i}

### = P-L

^{( 6 ) }

### With the addition of an expression for the diffusion velocity, the system of coupled

### equations (3), (4), and (6) is complete and can in principle be solved. Up to now.

### simplifying assumptions have always been used. They are discussed by Banks and Kockarts (1973).

### Solutions of the energy equation are usually based on the assumption that the atmo- spheric constituents are in diffusive equilibrium, i.e. V. = 0. The diffusion terms are, there- fore, omitted in Equations (6) and (7), with the consequence that the diffusion time should be small compared to the heat conduction time. On the other hand, such an assumption, mainly introduced to simplify the mathematical problem, is not necessarily correct. The diffusion time r

D### is usually defined as

### r

D### = H

2### / D <*>

### where H is the atmospheric scale height and D is the molecular diffusion coefficient. The heat conduction time r

^{ H}

### can be written

### r

H### = pc

y### H

2### /X <

9### >

### where all the symbols have been previously defined.

### The comparison between T

d### and T

h### is given in Figure 1 which is computed using the

### parameters of the mean COSPAR international reference atmosphere (Champion and

### Schweinfurth, 1972). The expression for the heat conductivity is taken from Banks and

### Kockarts (1973). It can be seen from Figure 1 that the heat conduction time and the

### molecular diffusion time are of the same order of magnitude over the whole thermospheric

### height range. This means that the diffusion terms are not necessarily negligible in the energy

### equation, but further studies are required to clarify this point. Since the diffusion time is of

### the order of one day around 120 km, variations of the thermal structure will add to the

### difficulty for reaching diffusive equilibrium in that height region. Possible deviations from

### diffusive equilibrium have also been suggested (Stubbe. 1972) as a consequence of hori-

### zontal winds. There is. however, no general agreement for the magnitude of such an effect

### (Blum, 1974; Stubbe, 1974; Rishbeth et al. 1974).

Fig. 1 C omparison between diffusion time and heat conduction t i m e in the mean COSPAR international reference atmosphere

### Up to now, almost all theoretical models were based on fixed boundary conditions at 120 km altitude. As a result of the variations observed in this height region, it seems better now to adopt boundary conditions at lower heights. In this case the transport effects of atomic and molecular oxygen become more important and the energy equation must be solved simultaneously with the continuity and momentum equations. This procedure has been used by Chandra and Sinha (1974) in a one-dimensional case. These authors stressed, however, the importance of downward heat fluxes by eddy transport. Such a flux is to be added to the molecular conduction given in Equation (7). Hunten (1974) pointed out that the introduction of eddy conduction may result in an overestimate of the downward heat transfer since dissipative phenomena such as gravity waves may also contribute to the heating. This shows that lowering the boundary level results not only in more complicated mathematics, but introduces also complex physical problems which are not yet completely understood.

### Even if the mathematical techniques were available to solve the three-dimensional time-dependent conservation equations, one has still to determine the different production and loss terms as well as the external forces acting on the system.

**3. ENERGY SOURCES AND SINKS **

**3. ENERGY SOURCES AND SINKS**

### The terms involving the mass average velocity in the energy equation are sometimes

### considered as compressional heat sources during the night and expansive heat sinks during

### the day. Since the total contribution is zero over a 24h period, these terms do not constitute

### a real global source or sink for the system. They must, however, be included in the energy

### equation, since they strongly influence the diurnal temperature distribution. When the

### momentum and energy equations are not solved simultaneously it is difficult to estimate the

### effect of the compression and expansion terms. It has been shown (Kockarts, 1973) how

### such an effect can be simulated in a one-dimensional calculation by introducing a very small

### heat source during the night and a compensating sink during the day. Although the absolute

### magnitude of these sources or sinks is negligible compared with the total UV absorbed

### energy, their effect is important since they play a role in a height region above 200 km

### where not very much UV radiation is absorbed.

### The major heat sources of the thermosphere are solar UV radiation, Joule dissipation from ion-neutral interactions and atmospheric waves such as gravity waves and tidal oscil- lations. Although the physical mechanisms are fairly well understood, there is still some uncertainty with regard to the amount of energy deposited in the upper atmosphere.

### Most of the theoretical models are based on solar UV heating alone. For practical purposes the solar spectrum is often divided into two parts: the Schumann-Runge continuum below 1750 A and the wavelength region below Lyman-0 (1026A) down to 80A.

### Longer and shorter wavelengths are not considered, since they are not significantly absorbed above 100 km altitude. The amount of energy available in the Schumann-Runge continuum is of the order of 15 erg cm'

2### s"

1### according to the tabulation presented by Ackerman (1971). Between 1026A and 80A the UV flux is more variable with solar activity and for medium solar activity the measured flux is of the order of 2.3 erg cm"

2### s"

1### (Hinteregger, 1970). The validity of Hinteregger's fluxes has been questioned several times in relation with the neutral and ionospheric heat balance. Recently, Prasad and Furman (1974) concluded that the arguments advanced for doubling the solar UV fluxes below 1300A are not compelling, since too many uncertain parameters are involved in such studies. The effects of changing solar UV fluxes from 1.7 to 4.5 erg cm"

2### s"

1### are indicated by Banks and Kockarts (1973) and by Kockarts (1973). New measured solar UV fluxes are now available below 1000A (Schmidtke et al. 1974). Figure 2 shows a comparison between the fluxes measured on board AEROS-A (Schmidtke et al. 1974) on March 2, 1973 and the values given by Hinteregger (1970). With the exception of an Si XII, an Fe XV line, and the 176-155A wavelength interval, Hinteregger's values are always smaller than the fluxes given by Schmidtke et al. (1974). The ratio of these measured values is presented in the lower part of Figure 2. The total flux below 1000A measured on board AEROS-A is 3.8 erg cm'

2### s'

1### . More data are, however, required before a definite conclusion can be reached with regard to the absolute UV fluxes. Furthermore, long term variations of the UV radiation emitted by the whole sun are not yet available, despite their fundamental importance in aeronomic studies.

- 8 -

### 10

### 10 -1 E u Ol |B>

### <u

X

*Z> *

*>• *

### o a:

### 111

7### z i O UJ

### LyP **cm **

### NIII LyV

### Ly6

### ,SVI

### Oil III

### OV He 1

### 0 III N . V I H

^{0 V }

### MgX

### X X OIV Si XII X

### t ; ^{NeVIl }

### ^Hell

### Fe XV X

### Mg IX

v#

### FeXVL

### X I

### FeXVI

### I;

### 10

3**2 **

### O 1 5 1

### 1000 900 800 700 600 500 400 300 200

### W A V E L E N G T H (A)

### Another important mechanism is the Joule heating resulting from ionospheric currents (Cole, 1971. 1972). Since Joule heating is proportional to the electron concentration and to the square of the electric field, its effect is predominant in the auroral region during disturbed conditions (see Wickwar, 1975). Joule dissipation is actually a linking mechanism between the neutral and the ionized atmospheres, as can be seen from the momentum and energy equations. Neutral winds related to the convection electric fields have been measured in the E-region using the incoherent radar facility at Chatanika, Alaska. At nighttime, during storm conditions, neutral wind velocities up to 200 m s'

1### are not uncommon (Brekke et al.

### 1974). Horizontal winds as high as 1000 m s'

1### have been deduced at 140 km (Chang et al.

### 1974) from geomagnetic variations recorded at College, Alaska. As a result of the involved physical mechanism, Joule heating is strongly dependent on the latitude. A global estimation of the amount of heat available through this process requires simultaneous know- ledge of the worldwide electric field and the electron concentration distributions. Using a global ionospheric model and electric field data for high and middle latitudes, Ching and Chiu (1973) computed the Joule dissipation in a given neutral model and concluded that it was similar in magnitude and height profile to the global solar UV absorption. Although Joule heating is mainly concentrated at high latitudes, a redistribution of the heat in the whole thermosphere occurs. A quantitative analysis of this transport phenomenon can only be made with a three-dimensional model. Furthermore, the Joule heating mechanism indicates that the complete self-consistent construction of a neutral atmosphere model would require a simultaneous computation of the ionospheric structure, i.e. the three neutral conservation equations should be coupled to the corresponding ionic and electronic equations.

### The potential significance of atmospheric waves as contributors to the thermospheric

### heat budget has been pointed out by Hines (1965, 1973) who estimated an average energy

### heat input of the order of 0.1 erg cm'

2### s"

1### above 120 km altitude. The variations of the

### actual supply of energy are, however, not yet known and it is, therefore, very difficult to

### introduce this effect in theoretical models. Furthermore, wave dissipation can result from

### medium-scale waves generated in the lower atmosphere, or from high- altitude large-scale

### wave generated in the auroral regions and traveling over large horizontal distances toward

### the equator. This means that atmospheric waves can introduce energy into the thermosphere from below and from above. Klostermeyer (1973) has shown that gravity waves generated in the auroral regions during geomagnetic disturbances can lead to temperature increases which are in agreement with the empirical relation deduced from satellite drag (Jacchia et al.

### 1967).

### The energy sources previously described can, however, not explain all the observational facts. Using observed temperatures, densities and winds, Barlier et al. (1974) have shown that a definite north-south asymmetry exists in the thermosphere. As an example. Figure 3 gives the ratio of the observed densities and Jacchia's 1971 model. It is seen that, even at the equinoxes, there is an asymmetry when the solar illumination is identical for both hemi spheres. The total density is greater in the south, A similar asymmetry has also been observed in the global temperature distribution. From the large volume of data accumulated during the last solar cycle maximum, Barlier et al. (1974) conclude that the temperature is.

### on the average, higher in the southern thermosphere. These authors suggest that this fact could be related to the geomagnetic field asymmetry and to the tidal wave dissipation which is linked to the asymmetrical worldwide ozone distribution in the stratosphere and meso- sphere.

### Since the construction of the first atmospheric models, heat sources other than the solar UV radiation have been discovered. Up to now, no thermospheric model has been published taking into account all these sources. The former problem of the lack of sufficient sources of energy is now perhaps replaced by the problem of how to lose sufficient energy to avoid high temperatures which are not observed. Besides the downward heat conduction mechanism, the infrared emission of atomic oxygen at 63 fim is the sole loss process usually introduced in thermospheric models. It has been shown (Kockarts and Peetermans, 1970) that, below 150 km. radiative transfer strongly reduces the 63 nm volume emission rate and.

### therefore, decreases the cooling rate resulting from atomic oxygen. At 100 km altitude the

### 63 Mm cooling rate is negligible. Molecular heat conduction can then lead to large tempe-

### rature gradients in the 100 to 120 km region. Such a situation can. however, be avoided if

### other loss processes are introduced.

**G E O G R A P H I C L A T I T U D E **

### • i i i i • • • • • i c - i m u - t -> w o< <o o o o o o o o o_ o_ o_

wo o o o m m m w® o o

**G E O G R A P H I C L A T I T U D E ** I I I I

### ( O - J U i U - t - ^ U U i - x i t t ) O . O . o„ o . o „ o„ o „ o . O

o### o .

**G E O G R A P H I C L A T I T U D E ** l | I

### (O -J Ul U U O) ~J <0 O . o . o_ o „ o „ o . o . o_ o_ o_

### u Ol r - -4 O

**O**

*to*

*> *

### l ! "

### m -

### <0 ro

1

### V

3### -

1### "

^{1}

### L i ^ H Z

## d

### - l V s k t r

### i •

## P

### i

*Ot *
*o *

II

### Fig. 3 - Ratio at 280 km between observed densities and the nighttime minimum density of Jacchia (1971) as a

### function of locar solar time and latitude. The solar declination is respectively - 20°, 0° and + 20°

### Some experimental results seem to indicate that other infrared emissions could play a role in the upper atmosphere heat budget. Very high infrared emission in the 3 to 8 Mm spectral region has been observed at altitudes above 150 km (Markov, 1969) but no complete physical explanation has been given. Recently the infrared emission of 0

^{3}

### at 9.6 nm and of C0

^{2}

### at 15 pm has been observed in the mesosphere and lower thermosphere *(Stair et al. 1974). According to these data, the 15 Mm C0*

^{2}

### emission at 120 km and 100 km corresponds respectively to 1 and 10 erg cm'

^{2}

### s'

^{1}

### . It is, therefore, not excluded that carbon dioxide could be the cooling agent in the 100 to 120 km height range.

*4. CONCLUSION *

### It appears from the above discussion that there is a need for more investigations

### designed to develop the numerical techniques used for the solution of the three-dimensional

### conservation equations. Furthermore, longterm observations are still needed for a better under-

### standing of the time and space evolution of all heat sources and sinks which are the

### fundamental input parameters of any atmospheric model.

*REFERENCES *

### ACKERMAN, M. : 1971, in G. Fiocco (ed.) Mesospheric Models and Related Experiments D. Reidel Publishing Company, Dordrecht-Holland, p. 149.

### BANKS, P.M. and KOCKARTS, G. : \913\Aeronomy, Part B, Academic Press, New York BARLIER, F., BAUER, P., JAECK, C„ THUILLIER, G. and KOCKARTS, G„ : 1974, /

*Geophys. Res. 79, to be published. *

### BARLIER, F., FALIN, J.L., ILL, M. and JAECK, C. : 1973, Space Research 13, 349.

### BLUM, P.W. : 1974, Planetary Space Sei. 22, 183.

### BLUM, P., HARRIS, I. and PRIESTER, W. : 1972, in Cospar International Reference *Atmosphere 1972 Akademie-Verlag, Berlin, p. 399. *

### BREKKE, A., DOUPNIK, J.R. and BANKS, P.M. : 1974,7. Geophys. Res. 79, 2448.

### CHAMPION, K.S.W. and SCHWEINFURTH, R.A. : 1972, in Cospar International Reference *Atmosphere 1972 Akademie-Verlag, Berlin, p. 3. *

### CHANDRA, S. and SINHA, A.K. : 1974,/. Geophys. Res. 79, 1916.

### CHANG, S.-C., WU, S.T. and SMITH, R.E. : 1974,/. Atmospheric Terrest. Phys. 36, 889.

### CHAPMAN, S. and COWLING, T.G. : 1970, The Mathematical Theory of Non-Uniform *Gases 3rd ed. Cambridge Univ. Press, London. *

### CHING, B.K. and CHIU, Y.T. : 1973, Planetary Space SQL 21, 1633.

### COLE, K.D. : 1971, Planetary Space Sei. 19,59.

### COLE, K.D. : 1972, Planetary Space Sei. 20, 2205.

### HEDIN, A.E., MAYR, H.G., REBER, C.A., SPENCER, N.W. and CARIGNAN. G.R. : 1974.

*J. Geophys. Res. 79, 215. *

### HINES, C.O. : 1965,/. Geophys. Res. 70, 177.

### HINES, C.O. : 1973, Planetary Space Sei. 21, 2238.

### HINTEREGGER, H.E. : \910,Ann. Geophys. 26, 547.

### HUNTEN, D.M. : 1974,/. Geophvs. Res. 79, 2533.

### JACCHIA, L.G. : 1965, Smithsonian Contr. to Astrophys. 8. 215.

### JACCHIA, L.G. : 1971, Smithsonian Inst. Astrophys. Obs. Spec. Rep. 332.

### JACCHIA, L.G., SLOWEY, J. and VERNIANI, F. : 1967./. Geophys. Res. 72. 1423.

### KLOSTERMEYER, J. : 1973,/. Atmospheric Terrestr. Phys. 35, 2267.

**- 1 4 -**