PARTICLES, ENVIRONMENTS, ANDPOSSIBLEECOLOGIES IN ...oe reyes 232.6937 LSTOADIS.. No.4,1976...

34en

LSTOADIS..

Tue ASTROPHYSICAL JOURNAL SUPPLEMENT SERIES, 32: 737-755, 1976 December

© 1976. The American Astronomical Society. All rights reserved. Printed in U.S.A.

PARTICLES, ENVIRONMENTS, AND POSSIBLE ECOLOGIESIN THE JOVIAN ATMOSPHERE

CARL SAGAN AND E.E. SALPETERCenter for Radiophysics and Space Research, Cornell University

Received 1975 December 11; revised 1976 June 1

ABSTRACT

The eddy diffusion coefficient is estimated as a function ofaltitude, separately for the Joviantroposphere and mesosphere. The growth-rate and motion of particles is estimated for varioussubstances: the water clouds are probably nucleated by NH.Cl, and sodium compoundsarelikelyto be absent at and abovethe levels of the water clouds. Complex organic molecules produced bythe Le photolysis of methane may possibly be the absorbers in the lower mesosphere whichaccount for the low reflectivity of Jupiter in the near-ultraviolet. The optical frequency chromo-phores are localized at or just below the Jovian tropopause. Candidate chromophore moleculesmust satisfy the condition that they are produced sufficiently rapidly that convective pyrolysismaintains the observed chromophore optical depth. Organic molecules and polymeric sulfurproduced through H.S photolysis at 4 > 2300 A probably fail this test, even if a slow, deepcirculation pattern, driven by latent heat, is present. The condition maybesatisfied if complexorganic chromophores are produced with high quantum yield by NH, photolysis at A < 2300 A.However, Jovian photoautotrophsin the upper tropospheresatisfy this condition well, even withfast circulation, assuming only biochemical properties of comparable terrestrial organisms. Unlessbuoyancy can be achieved, a hypothetical organism drifts downward and is pyrolyzed. Anorganism in the form of a thin, gas-filled balloon can grow fast enough to replicate if (i) it cansurvive at the low mesospheric temperatures, or if (ii) photosynthesis occurs in the troposphere.If hypothetical organisms are capable of slow, powered locomotion and coalescence, they cangrow large enough to achieve buoyancy. Ecological niches for sinkers, floaters, and huntersappearto exist in the Jovian atmosphere.

Subject headings: planets: atmospheres — planets: Jupiter

I. INTRODUCTION

The contemporary Jovian atmosphere has somesimilarities to the primitive terrestrial atmosphere,where organic molecules were produced readily; andexperiments in which a wide variety of organicmolecules are produced under conditions similar tothose of the Jovian atmosphere have been carried out(Sagan and Miller 1960; Sagan et al. 1967; Woellerand Ponnamperuma 1969; Rabinowitz et al. 1969;Sagan and Khare 197la, b; Khare and Sagan 1973;Ferris and Chen 1975). This has led to the hypothesisthat there may be a Jovian biology (Sagan 1961;Shklovskii and Sagan 1966). Observations in thevisible and in the ultraviolet have implied the presenceof particulate matter in the Jovian atmosphere whichmay (but need not necessarily) be connected withquestions of Jovian organic chemistry and biology.Both chemical and biological issues are affected byfluid dynamics in the Jovian atmosphere. Weare hereconcerned largely with these effects.Modern models of the structure of Jupiter (for

references, see, e.g., Stevenson and Salpeter 1976)predict that the bulk constituent, a hydrogen-heliummixture, is fluid throughout, with temperatures in-creasing inward to above 10* K. Plausible organismsrequire temperatures well below 10° K. Oneprincipal

737

problem which must be faced by such hypotheticalorganisms is descent to pyrolytic depths due either tothe acceleration of gravity or to convective down-drafts. The second half of the present paper examines(1) whether such organisms undergosignificant growthand replication in less than the time scale for drift topyrolytic depths, or (2) whether other adaptations toavoid pyrolysis are hydrodynamically and biologicallyfeasible. We shall see that (1) becomes easier forsmaller organisms and (2) for larger ones.That plausible and internally consistent Jovian

ecologies can be described cannot, by itself, demon-strate the likelihood of life on Jupiter. A tenableargumentfor life on Jupiter must also show that theorigin of life on that planet was possible. Since many‘steps, particularly the later ones, which led to theorigin of life on Earth are only incompletely under-stood (see, e.g., Miller and Orgel 1973), a thoroughdiscussion of the origin of life on Jupiter cannot beundertaken at the present time. Most skepticism onthis subject seems to arise from the contention thatsynthesized molecules will be carried convectively topyrolytic depths before the origin of life can occur.The unspoken premise in such an argumentis thatpyrolysis would not have occurred on the primitiveEarth. However, the measured Arrhenius rate con-stants imply, through a regression analysis, that many

© American Astronomical Society * Provided by the NASA Astrophysics Data System

oe

reyes

232.6937

LSTOADIS..

738 SAGAN AND SALPETER

essential organic molecules, for example a number ofbiological amino acids, suffer thermal degradation attypical terrestrial surface temperatures in geologicallyshort time periods. The very limited available experi-mental data show, for example, that the half-life at30° C of the amino acid serine is ~10* yrs and ofthreonine is < 10° yrs; while at the same temperaturethe half-life of the simplest amino acid, alanine, is>10'° yrs (Vallentyne 1964). Because of the ex-ponential Arrhenius kinetics, the highest temperatureson a temperature-variable planet dominate the netmolecular degradations; also there is evidence thataqueoussolution and alkaline pH’s—both expected onthe primitive Earth—enhance the decomposition rates(Vallentyne 1964). It does not seem at all improbablethat some essential precursor molecules for the originof life had half-lives on the primitive Earth of 10? to10° yrs, not greatly dissimilar to the lifetimes againstconvective pyrolysis of organic molecules on con-temporary Jupiter, as calculated below.Many workers on the origin of life have suggested

that the rate-limiting steps were stochastic in nature.The volume available for natural molecular experi-ments on Jupiter is approximately 10° times largerthan on Earth. If phase interfaces play a critical rolein the origin of life (see, e.g., Bernal 1973) Jupiter,with a great concentration of solid and liquid at-mospheric aerosol particles, would be particularlyfavorable.The origin of life on Earth took a period

oe

reyes

232.6937

LSTOADIS..

No.4, 1976

helium abundance ofabout 10% by number,consistentwith separate measurements from the occultation ofBeta Scorpii (Elliot et al. 1974); from Pioneer 10 and 11ultraviolet photometry (Carlson and Judge 1974); andfrom the pressure broadening of the J = 0 and J = 1hydrogenlines (Houcket al. 1975). Assuming a heliumabundance of about 10, (by number), it then followsthat y/(y — 1) © 3, where y is the ratio of specificheats, and that the mean molecular weight p ~ 2.2.For pressure P, pressure scale height H = P(dP/dz)~',density p, and kinematic viscosity v in the tropospherewe adopt

TPX (EKx) bar

THs (Exz)27 km

y P | 6 x 10-* gmcP~\TISK em em-1.5

~w 2e-l1lva (eKx) 0.5cm?s~?. (1)

In the mesosphere, the temperature gradient decreasesrapidly (for P < 0.2 bar, say), and the minimum tem-perature is about 120 K followed by a temperatureinversion whose structure is not yet well known(possibly Tax X 150 K at P ~ 107-3 bar).For a general convective circulation pattern the

vertical convective heat flux F, is given by

F, = c,pATw,, (2)

wherec, is the specific heat at constant pressure, AT isthe typical vertical excess in the temperature drop, andw, the typical vertical convection speed. For verticaldistances larger than the vertical extent / of the cir-culation pattern, one has an eddydiffusion coefficientK, = w-l; and we shall denote by ¢, the eddy diffusiontime, H?/K,, for diffusion through a single pressurescale height H. In the simplest form of Prandtl mixing-length theory (see, e.g., Schwarzschild 1958), thecirculation is characterized by two parameters, themixing length / and the superadiabaticity with w,? andAT/T both proportional to the superadiabaticity. Wereplace F, by 1.4 x 10* ergs cm~? s~*, a typical valuefor the Jovian troposphere, employ equation (), andwrite K., w., and ¢, in the forms

K, = &(T/175 K)*® x 107 cm?s71,

we = &/4(7/175 K)~ 23 x 76cms7!

t, = €-\(T/175 K)°® x 7.5 x 10°s. (3)

In the notation of mixing-length theory our parameter€ equals (//0.05H)*'*. The Jovian tropospheric lapserate is then very close to adiabatic (A7/T ~ 107°), asis true by the same argument (Sagan 1960) for allother planetary tropospheres.

JOVIAN ATMOSPHERE 739

In the simplest version of mixing-length theory1 ~ Hand € ~ 100. The actual circulation patterns inthe troposphere are likely to be much more complex(Hunten 1975). In particular, one should consider theeffects of rotation at great depths and of latent heatrelease in the cloud layers near the tropopause(Gierasch 1973). For a constant value of the heat fluxF, in equation (2), these effects usually increase ATanddecrease w, considerably, leading to smaller values ofK, and € in equation (3). These considerations suggest,as a simple working hypothesis, € = 1, which shouldbe accurate to within a factor of about 200.

Gierasch (1976) has suggested a specific large-scalecirculation pattern which extendsall the way from thebottom of the water-cloud layer (T ~ 280 K) to thetropopause. This pattern is driven by the latent heatof water which leads to a relatively large value ofAT/T ~ 10-2. This corresponds to € ~ 0.01 withtime scales for complete turnover of the region of upto a few years, even though some cloud patterns(above the water clouds) are observed to change inshorter times (Peek 1958). On Gierasch’s model, theeddy diffusion coefficient increases by a very largefactor just below the cloud bottom to € ~ 10.

In the mesosphere the bulk of the heat flux F,,; iscarried by radiation, but some fraction of the fluxcould drive a circulation pattern working against asubadiabatic temperature gradient. The vertical speedW, Of such a circulation pattern must satisfy an in-equality, based on the fact that the adiabatic coolingrate cannot exceed the total available heating rate:

Hpw,,[c,(dT/dz) + g] < Fro- (4)

Except in the immediate vicinity of the tropopause, thetemperature gradient in equation (4) can be neglectedfor the troposphere. Putting T ~ 150 K and assumingthat the vertical length scale for the circulation patternis of order H, the eddy diffusion coefficient K,, is oforder w,,H, so that

Wm < (1 bar/P)0.016 cm s7?;

Ky, < (1 bar/P)7 x 10*cm?s7?. (5)

Thus typical velocities just below the tropopause aremany thousands of times larger than just above thetropopause. Analysis of the observed La albedo of theupper mesosphere (Wallace and Hunten 1973; Carlsonand Judge 1974) indicates K,, ~ 3 x 108*1cm?s~? atlevels where P ~ 3 x 10~‘ bar; analysis of 6 Scooccultation measurements (Sagan et al. 1974) indicatesK,, > 10° cm? s~? at similar levels. It is theoreticallyplausible (French and Gierasch 1974) that gravity-wavepropagation upward in the mesosphere contributes tothe eddy diffusion. For a single wave mode, K,, has astep-function behavior, but overall it seems plausibleto assume some constant fraction of the maximumexpression in equation (5). The value given by the Laalbedo then suggests

Ky, ~ (i bar/P)100 cm? s~*, (6)

uncertain by factors of order 10 or 100. Note that


232.6937

LSTOADIS..

740 SAGAN AND SALPETER Vol. 32

log K (cm* s"')

S 4 6 7 8| | | q | | J

| T T

-7L

-6bL Ly a absorption

-5;+- @ H4 Photolysis—

® T at \ 4

o = UV Chromophores

2tr _|ropopause Visible0-—. > . Chromophores;1

@ AmmoniaCl aa4g Watér-Clouds «2----._-____.| He

sPhotolysis‘“

‘\

2 o Ammonium Chloride Clouds K ,. 5°o T Pyrolysis

3-6 Silica Clouds oforganic.hua ee es pow re

50 100 200300 500 1000 2000 5000

T (°K)Fic. 1.—A schematic plot of temperature 7 and eddy diffusion coefficient K versus pressure P for the Jovian atmosphere.

The dashed portion of the curve for K corresponds to Gierasch’s (1976) model for a deep circulation pattern driven by latentheat effects. Methane photolysis extends upward from the indicated level.

K,, « e?'", Temperature 7 and K(z) are plotted againstpressure, schematically, in Figure 1. (The dashedcurve for K corresponds to Gierasch’s 1976 model inwhich the latent heat of water condensation is takeninto account.)We shall need continuity equations for n,(z), the

number density as a function of height z of somecomplex molecule (of molecular weight pp, > 2.2) or ofsome particle. Let w,(z) be the steady-state verticaldrift velocity of the molecule or droplet falling undergravity through an atmosphere with pressure scaleheight H, eddy diffusion coefficient K, and meanmolecular weight 2.2. In standard notation, the con-tinuity equation (see, e.g., Strobel 1973) reads

dn, K D;\ _(K + DI)B+ (+ Fn = —9Q,,

where D; is the effective molecular (or particle)diffusion coefficient and H, = H(2.2/u,) « HA is theeffective scale height for the jth particles alone. ®,isthe vertical flux of particles of species 7. H, and thedrift velocity wg for free fall under gravity are relatedby D,; = waH;. We shall be interested only in levelswell below the turbopause where D, « K and giveonly a simplified version of the standard treatment(Strobel 1973; Prinn 1973). The steady-state continuityequation then reduces to

Ky)A + wale)nde) = 0/2).

In § IV we shall apply equation (7) to the numberdensity n(z) of organic molecules or other chromophoreparticles for which the drift velocity can be neglected(w, « K/H). Assume further that the particles aremainly produced at some higherlevel z Y Z,) > 0 anddestroyed at some lowerlevel z = 0, so that the flux ®equals the total production rate (cm~2*s~+) for0

34en

LSTOADIS..

No. 4, 1976

The situation is more complicated if the productionlevel Z, is in the mesosphere and the destruction level(z = 0) in the troposphere, with z = Z, < Zp, repre-senting the tropopause. However, neglecting againderivatives of K = K, (for z < Z,) and assumingthatK = K,, (for Z, < z) is proportional to e?/* as inequation (6), the continuity equation (7) can again besolved analytically. With the fact that Z, > H andK.(Z,) > Kn(Z,) for cases of practical interest, thesolution can be approximated by

n(z) © (OA/K,)[1 — exp(—2/H)] for0

232.6937

LSTOADIS..

742 SAGAN AND SALPETER Vol, 32

TABLE1

EQUILIBRIUM CONDENSATION TEMPERATURE, Joon, AND ESTIMATES OF THE RELEVANT MIXING RATIO, /, THE RADIUS a;, THEPARAMETER yap, AND THE MINIMUM MIXING RATIO f,;, FOR A NUMBER OF SUBSTANCES ABOVE THEIR CONDENSATION LEVELS

Parameter Fe Mg2Si0, SiOz NazeSi2O5 NaCl NH,Cl (NaS)

—lOBi0 fo eee ceceeecaeese 4.1 4.6 4.4 5.7 6.3 6.3 (5.7)Toon/100 K..... eee ee eee 25 23 16 10 4 10ay[10-3 cm... eee ee eee 120 60 20 4 2 8Nuvaprrsecsccccceccscceeees 60 20 0.1 0.03 0.4—lOBi0 fine. cece cece eee 7.9 7.7 7.2 7.8 7.0 6.4 7.0

equation (1), we have If myvap >> 1 for a particular material, most of it willform into large “‘raindrops”’ in a cloud layeratlevels

an, = 0.007 cm (2 175 K\'8 close to the condensation temperature T,,,, and only° Par T small amounts of the material will be found at higher,

9 cooler levels. If ny.) « 1 one might expect only smallw,(a) = 2 a" Par droplets of the material to form, so that wz, « w, and

9 v py, upward convection can maintain an appreciable mix-ing ratio of the droplets at higher levels. In practice,

~ (0.15 cm s~) LT '"\~"? par ( a__\?, however, various complications usually tend toward175K 2 /\10-*cm more rapid formation of larger droplets (Rossow

1976). For nyap ~ 107-1 or 10~?, the circumstances(11) are probably mixed: some rain near the T.,, level,

where pg, is the internal density (in gmcm~°) of aliquid drop. Anothercritical length is a,, the dropletradius for which the drift velocity wz equals the typicalconvection speed w, in equation (3). For the cases ofinterest (JT < 2500 K) we have a, < dz,, so thatequation (11) holds. With py, x2 and 21 inequation (3), we have

aq © 0.002 cm(7/175 K)~ 1/22. (12)

Large drops with a > a, ‘‘rain out” rapidly, whereassmall droplets with a « a, are carried along by theconvection pattern and can move up to higherlevels.

Forall cases in Table 1 (except possibly for NH,Cl)we have a; > az, so that homogeneous nucleationleads to large enough drops for most of the matter tobe rained out. However, if a very large number ofnucleation seeds are present originally, a large numberof much smaller droplets condense out, and the drop-lets must grow before they can fall out as rain. LetQooa, be that radius for which Cp = par/p,:

~ Par — 2/3 T 1/3 4

Qcoa (2) (asxx) x 3 x 10-*cm. (13)

Once @ & Aoa, droplets usually can grow very quicklyby coalescence, but for a < a,., growth must occur bythe slower process of smaller drops evaporating andthe excess vapor diffusing to larger drops (Langmuir1948; Mason 1972; Rossow 1976). An importantparameter 7,,, is the growth rate from this process(from radius 0.5a,o tO Qooa, SAY) EXPressed in units ofH*/K,, a typical convection time. With an uncertaintyof at least a factor of 10, we have

T 7/6 r 2

tran ~ 4 x 104f(575)""(B)°- ag

some small droplets a few scale heights higher up, butwith a mixing ratio which decreasesslightly with in-creasing height. Values for 7,,, from equation (14) aregiven in Table 1. Except possibly for NH,Cl, nyap > |10-2 as well as a; > a,, and mostof the material rainsout and cannot move upto higher levels. However, inupward-moving convection cells the saturation vaporpressure continues to fall; droplets then continue tocondense out, but with decreasing values of the mixingratio f and hence of the parameter 7. Let f,, be that(decreased) value of the mixing ratio, for which thedroplet size a, (due to homogeneous nucleation) inequation (10) equals az. Values off,, are also given inTable 1. The mixing ratio will not drop appreciablybelow f,, 1n upward moving convection cells, sincedroplets are now too small to rain out rapidly enough(and further growth of droplets is also too slow).For the cases with 7,,., % 1600 K the radii satisfy

As > Agony COalescence is rapid, and large drops arerained out immediately. At the high pressures inJupiter’s interior, liquid Mg.SiO, (as well as Ca andAlsilicates and metallic Fe) condenses before MgSiO;and before SiO,. When liquid SiO, begins condensingnear the 1800 K level, most of the metallic elementshave already been rained out, but sodium hasnot. Forpure Na.SiO; our estimates indicate a slightly lowercondensation temperature than for SiO., whichintroduces an uncertainty: [t is conceivable thatall thesilicon has been rained out before sodium condenses,in which case gaseous NaCl and NaOH (since theelemental abundance of Na exceeds that of Cl) surviveto higher, cooler levels. It is more likely that much ofthe sodium condenses out between 1800 K and 1600 Kin the form of some sodium silicate-silica solution, inwhich case gaseous HCl and some NaCl survive tohigher levels. In either case, more of the survivingNaCl condenses out near 1000 K; in the morelikely


232.6937

LSTOADIS..

No.4, 1976

case, NH,Cl condenses near 400 K;in the less likelycase, Na.S condenses near 1000 K.Wehave sometentative implications for levels with

T

232.6937

LSTOADIS..

744

mass-production rate R,,, is ~5 x 107** gcm=?s7},and equation (16) would give a column density forcomplex organics of No,~ ~ 3 x 107* gcm~?.Complex organic molecules may or may not be the

mesospheric chromophores which have been inferredfrom the UV albedos: The measured albedo ofJupiter (Wallace et al. 1972; Savage and Caldwell1974) to solar photons in the near-UV is quite low,with a minimum ~0.25 near A 2800 A. Unit opticaldepth for pure (conservative) Rayleigh scattering byH, for 2800A photons is reached at a level withP x 0.3 bar in the lower mesosphere. No simplemolecules present near or above this level absorb inthis wavelength region (NHszis effective only for\ < 2300 A), and complex molecules or very smallabsorbing dust grains have been invoked to providethe absorption (Axel 1972). Absorption in this regioncan be provided by a range of not implausible organicmolecules (Sagan 1968; Khare and Sagan 1973). Amass absorption coefficient near 2800 A of ~(2Nore)~?~ 10° cm? g~! would be required if organic molecules(beyond ethane) produced by Lea photons are re-sponsible. This is within the range of measuredabsorption coefficients for complex organic moleculesproduced by ultraviolet light in crude Jupiter simula-tion experiments (see, e.g., Khare and Sagan 1973).

b) Cloud Models

The total thermal infrared flux emerging fromJupiter gives T.;; ~ 127 K. Most of this flux emergesat wavelengths A ~ 20 um to 60ym; the opacitysource is mainly H, (with a small contribution fromgaseous NH3); and most of the radiation emanatesfrom levels with P ~ 0.7 bar to 0.2 bar (J ~ 150 K to~115K). Only a small fraction of the thermal fluxemerges at wavelengths A < 15 um, but the opacitiesof the main gaseous constituents of the Jovianatmosphere are low at some of these shorter infraredwavelengths. In the absence of any particulate matter(or complex molecules), radiation with A ~ 8.2 um to9.5 «zm would emerge from levels with T ~ 155 K, andradiation with A ~ 5 wm from even deeperlevels withtemperatures up to 7 ~ 300 K.Assuming solar abundancesof all elements in the

Jovian interior and assuming thermochemical equi-librium, one can calculate condensation temperaturesToon for various species (cf. Lewis 1969a; Weiden-schilling and Lewis 1973; Prinn and Owen 1976). Toondetermines the level of each ‘“‘cloud bottom,” andabovethis level the gas-phase abundanceis given bythe saturation vapor pressure. The abundance andsize distribution of droplets (or crystals) in the upperlayers of a cloud cannotbe calculated without detailedknowledge of the dynamics (see § III). The cloudbottoms for ice (and aqueous ammonia) are at levelswith T ~ 280 K, for NH,SH at T ~ 200K, and forNHsgat T ~ 150 K (P wz 0.7 bar).Measurements of spectral reflectivities for solar

radiation in the visible range are now availableseparately for the bright zones and for the darkerbelts in the Jovian equatorial and temperate latitudes

SAGAN AND SALPETER Vol. 32

(Orton 1975c). The fraction of the sunlight that isabsorbed varies from only ~0.2 in zones to ~0.3 inbelts for red light, but varies from ~0.3 in zonesto~ (0.5 in belts for blue light. The data at A ~ 20 um and45 wm, separately for zones and belts (Orton 1975c),are compatible with a model which includes anoptically opaque cloud top at the level (T ~ 145 K,P ~ 0.6 bar) for zones, but not for belts. This modelis not unique, but the postulated cloud top occurs closeto the level at which solid NHg crystals are expected toprecipitate out if the zones represent the upward draftin the overall circulation. The spectral data for A & 8to 14m (averaged over some zones and belts),together with spatially resolved data at A = 8.11 umand 8.45 um (Orton 19755), indicate the presence ofsome obscuring matter at altitudes above the T =150 K level. The absorption spectrum of solid NH,crystals fits these data quite well, with more absorbingmaterial over zones than over belts suggested by thedata. Nevertheless, a small amount of absorptionseems to be required over the belts as well, with anextent of only a few kilometers at levels with T ~145 K to 150 K indicated for this thin haze.One question of primary concern is the nature and

location of the red chromophoreswhichare responsiblefor the absorption ofvisible sunlight over the belts (orat least for the preferential absorption in the blue).Upper limits to the location altitude can be inferredfrom the common observation that the red chromo-phores are sometimes overlain by time-variable whiteclouds. The relative heights are apparent because theinterface between white and red material is frequentlyconvex outward from the white clouds. As discussed,the highest lying clouds which are possible on Jupiterare ammonia cirrus clouds at levels with T x 145 to150 K. The red chromophores must therefore residemainly at deeper levels with T > 145 K and P > 0.6bar. Emission at A Y 5 um reveals infrared hotspotsin only someregions of the belts (but in none of thezones) where the radiation emerges from deep levelswith temperatures up to T ~ 300 K. Thered colora-tion is anticorrelated with these infrared hotspots(Keay et al. 1973; Westphal, Matthews, and Terrile1974), which implies either that (a) the red chromo-phores reside above the level of some intermittentclouds which are opaque to 5 um radiation and whichreside somewhere above the 7 © 300 K level; or (db)that the red chromophoresare themselves intermittentand opaque at Sym. The infrared hotspots areassociated with visually dark patches of bluish hue,probably connected with Rayleigh scattering ratherthan with blue chromophores, placing the blue regionsat pressures of several bars, and T < 280 K (see Sagan1971). To summarize: The red chromophoresresidemainly in the troposphere (7 > 145 K), extendingdownward from the tropopause (T ~ 145 K) to atmost the T ~ 300 K level.The condensation temperature of solid NH,is only

slightly higher than the temperature (T ~ 145 K)expected for the tropopause, so that the solid NH,clouds reside in the upper layers of the troposphere.The pattern of patches of white clouds overlying the


oe

reyes

232.6937

LSTOADIS..

No. 4, 1976

red chromophores changes on a time scale of a fewweeks. This is comparable to the total troposphericeddy diffusion time ¢, in equation (3) if € ~ 1, but isalso compatible with the longertotal circulation timeter in Gierasch’s model, since time scales for patchesof horizontal extent d, « D,, (the separation betweenzones and belts) are only of order (d,/D,4)tar. Thepostulated intermittent 5 1m clouds underlying the redchromophores could in principle reside anywherebetween the T ~ 145 K and T ~ 300 levels, but onGierasch’s modelthese levels in belt regions representa.continuous downdraft of warming material whichcould not form NH,SH or H,0 crystals. Solid NH,clouds can be present in belts to varying degrees,because the upper horizontal circulation path fromzones to belts can include layers just below the tropo-pause which carry NH,crystals. On this picture thered coloration would have to be provided by redchromophores which more orless coexist with NH;clouds (P ~ 0.5 to 0.7 bar), but with varying amountsof (i) high-level clouds with small NH; crystals, (ii) redchromophores, and (iii) clouds at lower levels withlarger NH,crystals.

c) Radiative Transfer Excursion

Besides the red chromophoresweare also interestedin inferred near-UV chromophores: The measuredalbedo of Jupiter (Wallace, Prather, and Belton 1974;Savage and Caldwell 1974) to solar photons in thenear-UV (A x 2000A to 3500A, say) is quite low,with a minimum albedo ~0.25 near X ~ 2800 A (afterrising slowly toward shorter A, the albedo again seemsto be lower for A < 2300 A). Absorption by H,S,residing in the NH,SH cloudsat lowerlevels has alsobeen considered (Prinn 1970, 1974; Lewis and Prinn1970; Sagan and Khare 19715; Lewis 1976). Acomputational problem is often presented by absorberswhose abundance decreases very rapidly with increas-ing altitude (because of condensation); we brieflyreview the relevant radiative transfer theory.

Consider plane-parallel geometry and assume thatthe scattering is isotropic. Let 7 be the total extinctionoptical depth at some level (scattering plus absorp-tion), &, the single scattering albedo(ratio of scatteringto extinction cross section) at the same level, andx = [3(1 — @)]’”. If « is independent of 7, theEddington approximation is known even forslabs offinite thickness (Chandrasekhar 1960; Irvine 1975) andis discussed in our Appendix. The meanintensity J(7)as a function of depth is particularly simple for asemi-infinite slab of constant « (see eqs. [A3] to [A5]and eq. [A17]). Unfortunately, « can be a very steepfunction of 7 as in the example in Figure 2, whichcorresponds very roughly to absorption by H.S andRayleigh scattering by H. of 2500A wavelengthphotons in the work by Prinn (1970). Prinn usedequation (A3) for J(7) in the integrand of an integralwhich evaluates the absorbed energy layer by layer.Unfortunately, this use of equation (A3) can lead togross errors when « is varying rapidly, and a moreaccurate numerical evaluation is given in our Appendix


a NS

0.7 —

0.5-- J =

0.37- =

0.2 _

015 T\ —

0.10K Vs

0.07 _

0.0 S-- _

0.03 ~

0.021— po} a || 2 3 4 #5 6 7 8

Fic. 2.—A model radiative transfer calculation for anassumed form of the parameter «, as a function of totaloptical depth 7. Numerical results for the mean intensity J aregiven in the solid curve. The dashed curve is an analyticapproximation, discussed in the text.

for the example in Figure 2. Fortunately, the correctresult for the fractionf of the incident energy which isabsorbed is fairly close to the following simpleapproximation, fo: Let 7) be the total optical depth atthe level where x equals some predetermined number,say « = 0.5 (in our example 7, ~ 7.5). Let fo be thefractional energy which would be absorbed accordingto the Eddington approximation if one had noabsorption at all for 7 < 7, and a perfect absorberattT). For diffuse incident illumination (instead ofaveraging over different angles of incidence) one hasfo = (0.7579 + 1)~1 if to > 1 (see eq. [A16]). For thenumerical example in the Appendix the value of focorresponding to the choice « = 0.5 for 7p fits thecorrectf quite well.

d) Near-Ultraviolet

Solar photons in the near-UV (A x 2000A to3000 A, say) are much more abundant than Lephotons. It is of interest to speculate about materialswhich absorb in the near-UVto see(i) if their absorp-tion can explain the low observed albedo in the near-UV, and/or (ii) if the end products of reactionsinitiated by such absorption can provide the redchromophoresfor the troposphere, and(iti) how muchcomplex organic material can be so produced.The absorption of solar radiation with A ~ 2300 A

to 2700 A by HSin the upperlayers ofthe tropospherehas been discussed by the authors listed above (forA > 2700 A the absorption by H,S is too weak andphotons with ’ < 2300A are already absorbed byNH,at higher altitudes). The total flux in this wave-length range (diurnal mean) is about 1 x 10** photons


oe

reyes

232.6937

LSTOADIS..


cm~*s~*, and somefraction f, of these photonsareabsorbed by H.S. Wefirst present estimates if an adhoc value off, = 0.1 is assumed (and € = 1 in eq.[3]).

Laboratory simulation experiments by near-ultra-violet irradiation of CH,, C,Hs, NH3, H.O, and H.Smixtures at ~1 bar pressures have been performed,but without initial excess hydrogen (Sagan 1971;Sagan and Khare 197la, b; Khare and Sagan 1973,1975). Hydrogen sulfide is the primary photon acceptorin these experiments. Amino acids were produced inthese experiments with a quantum yield of ~10~* formolecular weights ~ 100 (~ 10-2 amu photon ~*). Thequantum yield of all organics in such experimentsis ofthe order of 100 times larger (of the order of 1 amuphoton~*). Some red chromophores were abundantlyproduced, consisting mainly of polymeric sulfur, butincluding a substantial fraction of organic compounds.The overall mass absorption coefficient for blue lightof these chromophores is k ~ 10% cm? g~?}. If suchparticles are produced in Jupiter at the rate of oneS-bond per absorbed photon (R ~ 10'7S cm7?s7? ~5 x 10-71 gcm~?s~?), their column density, NV, overa few scale heights downto the thick water clouds,isneeded. According to equations (3) and (8), N x2RH?/K, ~ 10-*® gcm~? and the optical depth toabsorption in the blue is 7 = kN ~ 0.03. This isalmost two orders of magnitude smaller than theobserved optical depth. Moreover, the reaction ofeight sulfhydryl radicals to form Sg, is poisoned bycompeting reactions, and the Sz quantum yield must besignificantly less than 1.

In the laboratory experiments cited, organic mole-cules, but not polymeric sulfur, were produced fromsuperthermal hydrogen atoms. In the presence of athousand times more H, (by number), the quantumyield might be expected to decrease by about the samefactor, if energy loss of the fast atoms is mainly byelastic collisions, to ~10~° amu photon~?. Thetotalproduction rate would then be quite small, R ~ 5 x10-1* gcm~*s~?* (which is still 10 times larger thanour estimated R,,, for the mesosphere). The corre-sponding optical depth for organic chromophoresis,then, many orders of magnitude below the observedvalue. Many uncertainties remain, however, includingindirect effects of scattering. In the laboratory ex-periments cited, large quantities of molecular hydrogenare produced during the irradiation because the endproducts are unsaturated, and it may be that thecorrection factor from laboratory to Jovian situationis Closer to 10~? than to 107°.It is also possible thatsomewhatlonger wavelength ultraviolet light—perhapsas long as 2900 A—isstill effective in these syntheses.Recent experiments (Ferris and Chen 1975) haveshown that 1849 A irradiation of mixtures of CH, andNHg in a 10:1 excess of Hz produce a variety oforganic molecules in very high yields, presumably bythe generation of hot hydrogen atoms from ammoniaphotodissociation. These results suggest that the effectsof H, dilution on poisoning organic photochemistrymay be slower than linear. Hence the contribution oforganic polymers to near-UV photoproduced opticalfrequency absorbers on Jupiter may be significant.

Vol. 32

Polymeric sulfur is not produced by hot hydrogenatoms in such experiments, and so the scaling fromlaboratory to Jupiter is more direct in this case. How-ever, even here, and even underthe excessively optimis-tic assumption that every H.S photodissociation eventleads to polymeric S, the implied optical depth fallsshort by two orders of magnitude from matching theobserved values; moreover, pure polymeric sulfurfitsthe observed optical properties of the Jovian redchromophores poorly at best (Rages and Sagan,in preparation).

Wehave seen that a slow, deep circulation patternis likely to extend all the way from the tropopausedown to the level with T ~ 300 K (Gierasch 1976)with € ~ 0.01 in equation (3). Abundancesof reactionproducts are proportional to €~* and also to ({,/0.1),which was claimed to be as large as 4 by Prinn (1970).One might then expect H,S photolysis to dominatechromophore production, but this still seems unlikelyfor three reasons: (1) As discussed in § [Vc and in ourAppendix, Prinn’s (1970) estimate off, ~ 0.4 is likelyto be an overestimate since fo = (0.7575 + 1)~? issmaller. Thus f, < 0.2 is more likely even in the ab-sence of any scatterers (other than H.) or otherabsorbers above the H,S levels. The smallness off, isalso suggested by the fact that the observed albedo ofJupiter does not increase from A ~ 2800 A (whereH.S is ineffective) to A ~ 2600 A (where HS absorp-tion should already be strong). (2) Since H.S cannotitself be the cause for the low albedo in this wave-length range, whatever absorbers are the cause mustlower f, further. (3) Most important, the systematicdeep circulation pattern mentioned before depressesH.S photolysis both in zones andbelts: In the updraft,represented by zones, H.S is abundant in the NH,SHclouds, but NH, clouds are above these layers andprevent most of the solar radiation from penetratingdown. In the downdraft, represented by belts, theNH, clouds are absent (or at least diminished), butthe material moving downwardis practically devoid ofH.O and H.S (which condensed out on the previousupward journey). H,O and H.Sare again abundantindeep levels with T > 300 K,butlittle of the near-UVpenetrates to such depths.NH, absorbs solar photons of wavelength A <

2300 A, which are somewhat less abundant thanphotons between 2300 A and 2700 A. On the otherhand, NH, should not suffer the shielding difficultiesexperienced by H,S becauseit resides in the highestcloud layer which extends up to the tropopause. NH,should therefore be abundant in both zones and beltsand even flows into the lower mesosphere at a non-negligible rate. It is therefore likely that most of thephotons with A ~ 2300 A are absorbed by ammonia(Strobel 1973b), which is compatible with the observeddecline in albedo from A > 2300A to A < 2300A.Hydrazine particles are also produced by ammoniaphotolysis and may provide a suitable explanation forthe low albedo near A ~ 2800 A (Prinn 1974). More-over, organic molecules produced by NHg3 photo-dissociation at A < 2300A could conceivably be theprincipal source of red chromophores,if the quantum


oe

reyes

232.6937

LSTOADIS..

No.4, 1976

yields of such reactions with 10°:1 H, dilutions are ashigh as the experiments of Ferris and Chen (1975)suggest.

e) Optical Frequency Photoproduction ofChromophores

We here consider two possibilities: If the circula-tion in the upper part of the troposphereis as slow asGierasch (1976) suggests, then € < 0.01 in equation(3) and column densities obtained from equation .(8)are increased by two orders of magnitude; thuspolymeric sulfur or complex organics might providethe red chromophores. But, as we have seen, there areother problems with the contention that the chromo-phores are produced by H.S photolysis. These problemsare avoided if shorter wavelength photolysis of NH,is an adequate source of complex organic chromo-phores, but this is a questionable proposition consider-ing the lower photonflux at A < 2300 A, and whateveris the correct value of the poisoning of organic photo-chemical synthesis in a 10?:1 H, dilution. If, on theother hand, é turns out to be of order unity, thenultraviolet photochemistry cannot in any case providethe observed chromophoresat a sufficient rate. Thusin both cases we are led to the possibility that photo-chemistry at optical frequencies, where more totalenergy is available, is involved in chromophore pro-duction. But no chemical bonds of even moderatelyabundant molecules can be broken by photonsof wavelengths longer than 3500 A on Jupiter, fromwhichit follows that optical frequency photochemistrymust be at least a two-photon process. However,such processes appear to be exclusively biological.We therefore find ourselves led unexpectedly to thehypothesis that the Jovian chromophoresare biologicalin origin and that there is an abundant biota in theJovian clouds.

Let us assume that the optical chromophores arecontained primarily within Jovian organisms in theupper troposphere—organisms driven byvisible lightphotosynthesis, perhaps primarily in the blue, andutilizing the abundant methane, water, and ammonia.Terrestrial algae (using CO, instead of CH,) underlaboratory conditions (Lehninger 1972) have a quan-tum yield for synthesizing glucose of about 4 amuphoton™? (in the visible). The equivalent maximumproduction rate of biological material in the Joviantroposphere would be about R = 3 x 10-®gcm~?s-1. With a mass absorption coefficient of k ~ 10°cm? g~1, such organisms wouldyield an optical depthof + ~ 2, slightly more than required by the observa-tions. If the parameter é turns out to be of order 107°,and the red chromophores are due to Jovian photo-synthetic organisms, R and/or K would have to besmaller.Thus Jovian organisms having metabolic and photo-

synthetic parameters typical of terrestrial algae areable to account for the optical chromophores onJupiter. Such organisms must, however, be adapted tothe Jovian environment, and in particular mustmaintain a steady-state population in the face ofconvective pyrolysis. Likely ecological niches for such


organisms, as determined by the hydrodynamics onJupiter, are discussed in the following sections. Thefollowing discussion can, however, be treated inde-pendently of the hypothesis that Jovian chromophoresare biogenic.

V. SINKERS AND FLOATERS

If there is a Jovian biology, we would expect it tofill a rich variety of ecological niches. The bestterrestrial analogy seems to be the surface of the sea.Oceanic phytoplankton inhabit a euphotic zone nearthe ocean surface where photosynthesis is possible.They are slightly denser than seawater and passivelysink out of the euphotic zone and die. But suchorganisms reproduce as they sink, return somedaughter cells to the euphotic zone through turbulentmixing, and in this way maintain a steady-statepopulation. (The early stages of sinking move thephytoplankton from a region of depleted resourcesinto a region of abundant nutrients, thus stimulatingreplication.)A more elaborate adaptation is provided by fish and

other organisms with float bladders which use meta-bolic energy to maintain a habitat at suitable pressurelevels. They are generally not photosynthetic auto-trophs, but rather heterotrophs living on organicmolecules produced by autotrophs. A third ecologicalnicheis filled by marine predators, one step further upthe food chain, which hunt heterotrophs.

In the following discussion we will consider threecomparable ecological niches on Jupiter: The primaryphotosynthetic autotrophs, which must replicatebefore they are pyrolyzed, will be described as sinkers.A second category of larger organism, which may beeither autotrophs or heterotrophs but which activelymaintain their pressure level, will be described asfloaters. A third category of organisms actively seekout other organisms; wecall these hunters, although,as we shall see, the distinction between hunting andmating under these conditions is not sharp. Finally,there is a category of organisms which live almost atpyrolytic depth. They are scavengers, metabolizing theproducts of thermal degradation of other organisms.For our purposes, these pyrolytic scavengers areidentical with floaters. In the following two sections weestimate the growth time ¢,, required for an organismto double its mass.Wefirst consider passive organisms with positive

excess density Ap, above the gas density p,. Suchorganismsfalling under gravity through the atmosphererapidly acquire a vertical drift velocity, wz, whichdepends on the location and size of the organism andincreases with increasing Ap,. There is a premium onkeeping Ap, and hence w, small, and we consider onlyorganisms in the shape of thin, gas-filled balloons.Such organisms have been briefly proposed before,both for Jupiter (Shklovskii and Sagan 1966) and forVenus (Morowitz and Sagan 1967).For simplicity we treat a spherical shell of outer

radius a and skin thickness d « a. In general we con-sider organismsfilled with gas of the same temperature


232.6937

LSTOADIS..


and mean molecular weight (2.2) as the surroundingambient atmosphere. The lack of buoyancy is thencontrolled by the excess averaged density due to theskin, Ap, = (3d/a)p,, where we assumethe biologicalskin material to have unit density (1 g cm~§).For sufficiently small radius a of an organism, the

Stokes formula in equation (11) holds (with pg, —3d/a) for the drift velocity w, and the drag coefficientCp & 12v/aw,g is large. Let ag, be a critical radius forwhich this relation gives Cp = 1 (Reynolds number =24), so that

2Wag = 2 gad fora XK ap, ,

3 vp,

Ape = (6v*p,/gd)'”. (17)

For large organisms with a > dp, (large Reynoldsnumber), the flow around the body becomesturbulentand the drag coefficient Cp (=drag force/0.5aa?p,w)is close to 0.5 (see, e.g., Batchelor 1970). In this regime

Wa & 4(gd/p,)"" . (18)

Weshall have to consider separately organismsliving below and above the tropopause. The tropo-sphere provides conditions closest to terrestrial ones(similar temperatures and abundant waterin liquid orvapor form). We assume that the organisms cansurvive there over a range L ofvertical height (L ~ 5H,say), centered approximately on the level with T ~300 K. An important time scale is then a nominaldrift time t; = L/w,. We consider the skin thickness das an unknown parameter, with d ~ 10~* cm possiblya typical value (biological unit membranes on Earthhave d < 107° cm,but mechanical strength is requiredfor the skin). Using equations (1) and (11) andevaluating coefficients at the level with T = 300 K, wend

13 x 10*s ; ~ T \351cm10-*cm

¢™ 5H \300K a d

for

(oem 300 Ka< ap & 7 7

For @ > dg,, on the other hand, equation (18) gives adrift time which depends on d but not directly onradius a,

—4 2

f ( r "(Gy)" 2.5 x 108s. (20)

1/20.039 cm. (19)

la © Si \300 K d

The eddy diffusion coefficient K, in the convectivetroposphere is fairly large, and we also have to con-sider the contribution of eddy diffusion to the organ-ism’s downward motion overa vertical distance L. Weare interested in cases where L is larger than H andthe effective circulation time ty, for downwarddiffusion is then not given by L?/K,: With L > H andK, a slowly varying function of height, the first

Vol. 32

derivative term in equation (7) can be neglected andeddy diffusion has the sameeffect as increasing w, to(wz + H~1K). The relevant time scale is then linearin L and is approximated by the smaller of t, and thequantity

--, ffi)” 6lor = Fhe = K ~ (sox) 3.7 x 108s,

(21)

where we have used equation (3) and assumed é ~ 1.If d ~ 10-* cm, then eddy diffusion competes withgravitational fall only for microorganisms witha < 30um. Thus Jovian organisms the size of smallterrestrial protozoa and prokaryotes have typicaltimes for falling through the troposphere to pyrolyticdepths of 1 to 2 months. To maintain a steady-statepopulation, they must only replicate in that time scale.

In the Jovian mesosphere the temperature is low (weassume T ~ 150K) and water is absent, but somebiologies have nevertheless been envisioned for suchenvironments (see, e.g., Pimentel et al. 1966; Sagan1970). Weshall see that the lower, denser layers of themesosphere are the most advantageousfor the growthof such organisms; eddy diffusion is negligibly slowhere compared with the downward drift velocity wz.For organisms operating between a lower pressure Pand the tropopause (pressure ~1 bar), the drift-timeL/wg becomes

x 930s += In 1 bar\ 1 cm 107-* cm

q™ P a d

for

Ga cm 1 bara< ape ~ aas

For larger organisms, a > dg., the expression forlarge Reynolds numbergives

1 a(S cm P

1/20.044cem. (22)

ta & In (42 d lbar1/2

16x 10*s. (23)

To obtain the pyrolysis time scale this number must beadded to the appropriate tropospheric time scale givenby equations(19), (20), and (21).The expressions given above for tz hold only for

balloon organisms with ambient atmosphere inside.For organisms capable of pumping gas, buoyancy canbe achieved while maintaining pressure equilibrium byKeeping the interior gas pure hydrogen with molecularweight » = 2.0 instead of the ambient hydrogen-helium mixture with » = 2.2. A pumped organism canthus float if Ap, = (3d/a)p, < 0.1p,. This requiresradii a larger than a threshold radius a,, for floating,given by

d T 1 barAn & (iota) (53x) ( P x l6cm. (24)


232.6937

LSTOADIS..

No. 4, 1976

Instead of maintaining a hydrogeninterior, an organ-ism could use a fraction of its metabolic energy releaseto heat its interior gas. The temperature difference ATbetween interior and exterior is determined by themetabolic rate and the thermal diffusion coefficient ofthe gas. Except at very high pressures, the thermaldiffusion is fast enough to keep A7/T ty, for the dispersules, they can becirculated upward, grow to d,,;x, and complete thelife cycle.

Wesaw in § IV that the most favorable region fordirect photosynthesis is probably the upper tropo-sphere, just above the water clouds (with T ~ 300 K,P = 5S bar). Methane, ammonia, and water buildingblocks are abundantly available there to a photo-synthesizing organism, and the growth rate is con-trolled by the deposition of free energy from solarphotons. Furthermore, the observed optical chromo-phores can be explained by such organisms at such alevel. We saw that a value for the production rate ofRx 3x 10-®gcm~2s~! seems reasonable, and we

JOVIAN ATMOSPHERE

749

find for the growth time

al2 4d d 3x 107° 5‘er = Galt R~ 10-%cm R 1.3 x 10°s.

(25)

Equating this time with t, in equation (19) then givesfor Qyax, the maximum radius to which an-organismcan grow, |

Amax ~~ (10~* cm/d)70.1 cm. (26)

Values of dni, < 107% seem perfectly feasible, so thatone organism can produce more than 10* dispersulesfor which tg «4x 10®°s ty,. The steady-statepopulation of such organisms would then be con-trolled only by competition for sunlight, and not bygravitational fallout to pyrolytic depths.

Consider next the diffusion onto the skin of anorganism of some organic molecule with molecularweight » and abundance in the surrounding atmosphereof po (in g cm~* of atmosphere). Ethane is the lowestmass (w% = 30) organic molecule which carries freeenergy (relative to methane, the prevalent carbonmolecule at thermal equilibrium). The diffusion co-efficient D of ethane in the Jovian H—He mixture isabout 0.4 times the viscosity v of this mixture (Strobel1973), and for heavier molecules D scales reciprocallyas the cross section of the molecule. For approxi-mately spherical organic molecules we adopt

D & 0.4v(p/30)- 78 , (27)

with v given by equation (11). If vy > D 3 (aw,/12),the diffusion is almost independent of the motion ofthe sphere, and the steady-state rate for diffusion ontothe surface is 47a7(p,)D/a). Let « be the ratio of in-creased body mass of the organism to the mass ofingested metabolites. Assuming unit density for theskin material, the growth time is then

ter = ad/2epoD . (28)

At a mesospheric level with pressure P (in bars), formolecules produced at a mass rate of R = p® atsome higher level, equations (6) and (9) give po(u®H/100)P In P-!. The growth time in the meso-sphere is then

~ 4 d (u/30)?"°sr 1 cm 10-4 cm In (1 bar/P)

(? x | 0.03x< ———______

t

Ve 6s —1.8 x 10%s. (29)

Forethane an efficiency factor « ~ 0.03 seems reason-able and R ~ 5 x 107-14 gcm~?s~? has already beendiscussed; for UV photoproduction of more complexorganic molecules we may leave ec ~ 0.3 and R ~ 5 x10-7 gm cm~?s7?, so that Re is the same, but thefactor »?/> makes heavier molecules less favorable.Replacing the slowly varying factor In (1/P) by 3 and


232.6937

LSTOADIS..


Re by 1.5 x 107-15, we find from equations (22) and(29),

Amax ~ (107 * cm/d)(u/30)~-7/2 x 0.07cm. (30)

If ®e is particularly large and/or d particularlysmall, an organism can grow beyondsize dp,, and thediffusion rate now dependsonthe drift speed wg, givenby equation (18). We omit the complicated transitionregion @ XY dp, and make only order-of-magnitudeestimates: There is a boundary layer in the flowaround the sphere of thickness b ~ a(4v/aw,)1! « a,such that the flow speed a normal distance y from thebody is ~(y/2b)w,. With D < v the typical stand-offdistance for diffusion onto the body surface is bp =2b(D/v)"* and the diffusion rate is ~471a7(p,)D/bp), sothat

ter ~ (2dlepo)(a/Dwa)"(v/Dy” . (31)

For the mesospherethis gives

a \#2/ d_— \3/4/1 bar\1/4 (u/30)*"9a (; ‘.) (iota) (5 In (1/P)

5 x 10-24 0.03 .« PPE) 8 x 10 S. (32)

The same basic formulae apply in the troposphere,but po is now much smaller for the same value of Dbecause ofthe larger eddy diffusion coefficient, K,. Onthe other hand, we have seen that photosynthesis maybe going on there, and this may release complexorganic molecules at a faster rate ©. SubstitutingPo = RH/K, and using equation (3), we find (fora < Ape)

a d1. (T p\2/8

I cm 10-4 cm 300 K 30

3 x 10-°0.3 ,x (=2)8 x 10‘s. (33)

Comparison with equation (19) shows dma, to be oforder (107* cm/d)(30/1)/30.013 cm, which is justmacroscopic.

If small tropospheric photosynthetic organisms haveradii a, = 0.01 cm, their drift speeds are even smallerthan the eddy circulation speed and their abundanceis still given by po = RA/K,. A larger organism ofradius a, drifting downward with speed w,, can growby coalescing with such small organisms if the en-counter satisfies certain conditions. These conditionsare similar to those governing the coalescence ofrain-drops (Mason 1972; Rossow 1976). In analogy withequation (13) one finds, independent of a,, and pro-vided d

oe

reyes

232.6937

LSTOADIS..

No.4, 1976

and grow as they drift downward, so that thesizedistribution changes with height z. Let n(a,z) andp(a, Z) = (47a*d)n(a, z) be their steady-state numberdensity and mass density, respectively, per logarithmicinterval of radius a. In principle we havea relation likeequation (7) separately for each value of a, with theright-hand side representing the flux from one sizerange to another. We shall make only order-of-magnitude estimates for a(z), the mean radius oforganismsatlevel z, and for n(z) = { dIn an(a,z).

Weassumethat the organismsaresufficiently largethat w,(z), the drift velocity for radius @(z) at heightz,is much larger than K/H in equation (7). Multiplyingby mass (assuming as usual unit density for the skinmaterial) and integrating over sizes, we find

[4ra~2(z)d]w,(z)n(z) = €(Z)R. (35)

In this relation R is the constant rate in g cm~2 s~?+ atwhich mass in the form of organismsenters the upperlevel z = Z,; ap., we have, independent of d both for thetroposphere and the mesosphere,

3 ( a )' 175K P 3x ea 23slor © ~~ 20s.

6? \1 cm T 1bar R

(39)

For such organisms the troposphereis quite favorableeven in the absence of photoautotrophs as prey(R ~ 5 x 10°14 gcm~*s~?); for instance, with T x300 K and d~ 10-*cm, we have a, ~ 6cm and


TABLE 2

MAXIMUM HUNTER SIZES

SOURCE OF METABOLITES

Abiological8 Organic Matter Photoautotrophs

2 x 10-7? radians..... 6 cm 40 m10-2... eee 3m 2 km| 30m 20 km

Amax ~ (6/0.002)6 cm, so that buoyancy can beachieved with quite small values of 0. For R ~ 3 x10-°gcm~?s71, dy ~ 6cm requires only dpax ~(6/3 x 10~°)6cm; microradian maneuverability suf-fices. The maximum sizes of organisms embracing thehunter doctrine as a function of the maneuverability,6, and the source of food are given in Table 2. We seethat very large hunters are permitted, within theresolution limits of the Mariner Jupiter/Saturn imagingsystem.To derive an explicit relation between typical radius

a and vertical level Z(a), we must integrate equation(36) with t¢,, given by equation (39). The result dependson whether @ is constant or a function of a. Neglectingthe variations of 0, w,, and « gives (for a muchlargerthan the original size a, at level Z,)

Z, — 2(a) = 3walZ)ter(@) « (40)

Westill have to calculate the mechanical energyexpended by an organism in maintaining its horizontalhunting speed @w,. In pure downwarddrift the rate ofenergy dissipation equals the rate of gravitationalenergy release mgw,. The equivalentrate for maintain-ing the horizontal speed (with 8 < 1)is at most 6mgw,.(With judicious sailing and soaring it could be appreci-ably less.) The relevant quality factor is the ratioAE/Am of mechanical energy expended in finding apartner to the increase in mass of biological material.(Biochemical energy storage is of order 107’ ergs g~?and a mechanical quality factor ~10° ergs g™! is avery modest assumption.) The heat of combustion(full oxidation) of molecular hydrogen is about 10times that of glucose and more than 3 timesthat ofsuch fats as stearic acid. It is not the unavailabilityof reduced compoundsbutrather the unavailability ofoxidized compounds on Jupiter which probably setslimits on the efficiency of internal metabolism ofhypothetical Jovian organisms. We have

AEAm S OgWater 9 (41)

which depends quite weakly on temperature andpressure. For the upper troposphere

AE < a 2/3 d 1/2

Am ~ (; =) (s5-toa)

-9 1/3

x (oo) 5x 10°28 (42)R € g


34en

LSTOADIS..

752

represents quite a modest power expenditure, exceptfor the largest organisms. A 100m radius hunter ofphotoautotrophs with 6=0.0l radian excursioncapability and a 1 cm skin thickness requires a qualityfactor of 10" ergs g-1. The same quality factor isrequired by a | m radius hunterof abiological organicmatter with 6 = 0.03 radians and d = 1 cm.Even smaller power expenditures will be requiredif,

as would be likely, evolution selects organisms withsensory systems which increase the efficiency of hunt-ing food—such as, e.g., optical sensors for detectingchromophores. Likewise, acceleration sensors capableof detecting convective “‘thermals’’ would ease therestrictions on floater sizes and energy expenditures;and habitats near the boundaries between upward anddownward convective regimes would improve the costof Jovian biology.The path from sinkers to hunters to floaters is an

evolutionary path. It provides a possible sequence forthe evolution of floaters which, as we have seen,would be difficult to understand directly from sinkers.Once floaters are established, they may be able toreplicate without passing through the sinker stage atany point in their life cycles. For this to occur, ofcourse, the minimum radius of their dispersules mustbe larger than apy.While we have distinguished passive sinkers,

hunters, and floaters in this discussion, it is clear thata single highly evolved Jovian organism might partakeof qualities of all of these life styles—for example, atdifferent stages in the life cycle. Indeed, it may be thatonly after the hunting doctrine is adopted by sinkerscan growth to sizes large enough for the evolution offloaters become possible. We have in this discussionmadenodistinction among variouslocales on Jupiter;but it is clear that some locales—the Great Red Spot,for example—may be more favored than others be-cause of higher abundances of organic molecules,prevailing updrafts, or other reasons.Amongotherobjects in the outer solar system which

possess atmospheres, only Titan and Saturn exhibitcolorations similar to those on Jupiter. Because of theabsence of convective pyrolysis on Titan, the chromo-phores there can quite readily be ultraviolet photo-produced organic molecules (Sagan 1973). However,the physical environment of Saturn is very similar tothat of Jupiter in the respects relevant to the argumentsof this paper, and we therefore tentatively consider anairborne biota on Saturn as well.The test of these ideas is, of course, observational.

Flyby, orbital, and entry probe spacecraft each canperform significant tests of these ideas. For example,entry gas chromatograph/mass spectrometers for thenear future are anticipated to have detectivities oforganic matter as good as about 10-?° gcm~%. Thepredicted steady-state density of complex organic

SAGAN AND SALPETER

Vol. 32

TABLE 3

STEADY-STATE DENSITY OF ORGANIC MOLECULES (g cm~?)

Abiological Biological

Lower Mesosphere........... 2x 1071! 9x 1078Upper Troposphere.......... 2 x 10735 9 x 107?°

molecules on Jupiter is RH/K, and is roughly estimated,using the numerical values of this paper, both for thelower mesosphere and upper troposphere and forabiological and biological sources of organic matter,in Table 3. While most biological organic matter willbe in organisms and notin free molecules, the densitiespredicted in Table 3 are sufficiently large that a high-sensitivity entry gas chromatograph maybea significanttest of the biological hypotheses of the present paper.Because of the slow convective velocities of themesosphere, such instruments should be capable ofworking at such pressure levels. For comparison, thedensity of dissolved organic matter at 2000 m depth inthe north-central Pacific is equal to 5 x 107-’ gcm~8,and of particulate organic matter, 3 x 10-9 gcm™=3(Williams and Carlucci 1976).

VII. FINAL REMARKS

The possible existence of indigenous Jovian organ-isms is also relevant to the question of sterilization ofspacecraft intended for entry into the atmosphere ofJupiter. Even if the ambient environment is incon-sistent with the growth of microbial contaminantsfrom Earth, the internal environment of Jovianorganisms may be much more hospitable. Thereplication of terrestrial contaminants in the Jovianclouds also depends on the availability of traceelements. We have seen (§ III) that indigenous Naisprobably missing from the upper troposphere, andthat indigenous Mg must certainly be missing.Magnesium ions are essential for nucleic acid replica-tion and a wide rangeof other biological functions; ifMg is missing from the clouds, the likelihood ofterrestrial biological contamination of Jupiter becomesnil. However, exogenous sources of sodium, magnes-ium, and other trace elements, particularly frommicrometeorites, while small, cannot be neglected(§ IIL). Consequently, it seems judicious not to excludeprematurely the possibility of biological contamina-tion of Jupiter by terrestrial microorganisms.

This research was supported by NASA grants NGR33-010-101 and NGR 33-010-082 and NSF grantsAST 75-21153 and MPS 74-17838. We are gratefulto Drs. R. Danielson, B. Draine, F. Dyson, M. Eigen,P. Gierasch, K. Hare, N. H. Horowitz, D. Hunten, A.Ingersoll, W. Irvine, J. Lederberg, J. Lewis, W. Press,R. Prinn, W. Rossow,and D. Stevenson forinterestingdiscussions.

APPENDIXConsider a plane-parallel slab of total optical thickness 7, and let J(7, ~) be the intensity of radiation at depth 7

with pw the cosine of the angle between the propagation direction and the forward normal. Assume isotropicscattering and let @, be the single-scattering albedo, which may be a function of 7. The equation of radiative


34

rar

ewe

IG

LSTOADIS..

No. 4, 1976 JOVIAN ATMOSPHERE 753

transfer (Chandrasekhar 1960; Irvine 1975) then reads

d , a ~ t arr S ~

iep) = —I'(7, pw) + ak| du'I'(r, p') + Go exp (—7/pX) — (Al)

for an incident beam of unit intensity with direction cosine jo, where I’ is the intensity of the diffuse radiation andexp (— 7/49) that of the attenuated incident beam. The Eddington approximation can be obtainedbysubstituting afunction of a single variable /,.'(r) for I'(7, ») when 0 < y» < 1 and another function /_'(r) when —1 < p < 0.Let

x = [3(1 — &)]*” (A2)

and let J'(r) be the average over all » of (7, 4). For constant « and a semi-infinite medium (7) — ©0), one finds

_ 1 + (2/3)Ho™* 5 b _ 3@o .

i+ (Q)3e -” on? —

Theeffective reflection coefficient R = 1 — f, wherefis the fraction of the energy flux that is absorbed,is given by

J'(7) = aexp (—«1) — bexp(—p"*7), (A3)

”= OF mull + OP)Generalizations for anisotropic scattering are discussed by Chandrasekhar (1960), Prinn (1970), and Irvine (1975).For isotropic scattering Prinn used equation (A3) for normal incidence (uw) = 1) in his numerical work. Forgeneral p> but with « « 1, equation (A4) reduces to

F(Ho) = 1 — R(t) © «(2/3 + po). (A5)

For many purposes oneis interested only in results averaged over a whole daylight hemisphere, i.e., averagedover }4p from 0 to 1. Such averages are represented by the single problem weshall consider here, where the incidentintensity Z,,.(@) equals unity for all 1 between 0 and1. In the spirit of the Eddington approximation wereplace theproblem by one for intensity J(7, ») without explicitly isolating the attenuated incident beam: By analogy withequation (A1) we have

(A4)

d af a. ' |mF Mew) =Geo) + Gok| dn'TCs, w’), (A6)which leads to the coupled equations

dK(r) _

dF(r) _ .OO = -In -a@), S=-Fe, (A7)

where |

i) = [duly), FG) = [. dupl(t,p), K(x) =4[. dup2T(z, 1). (A8)

We again make the approximation (but here for ai/ the radiation) that J(7, ») = J_(r) for » > 0 and J(7, p) =I_(7) for p < 0, so that

J=3K=370,+TL.), F=47, —J_). (A9)

In this problem, any explicit consideration of the incident beam is simply replaced by the boundary condition thatI,(7 = 0) = I and theeffective reflection coefficient R = 1 — fis simply given by J_(7 = 0).For general #(r) and with «(7) defined by equation (A2), the use of the Eddington approximation (A9) in

equation (A7) leads to the eikonal equation

2

E — eo)|I(7) = 0 (A10)

and to the further relation

2(1, + 1) = 320-1). (A11)

© American Astronomical Society ¢ Provided by the NASA Astrophysics Data System

34

rar

ewe

IG

LSTOADIS..

154 SAGAN AND SALPETER Vol. 32

Wefirst solve these equations for the case of constant « for a slab of total thickness 7): We assumethat the effectivereflection coefficient Rp = I_(t0)/I,(70) 18 given as one boundary condition, the other boundary condition for theincident beam being J, (0) = 1. The general solution of the coupled equations (A10) and (A11)is

I(r) = a{(3 + 2k) exp (—x«7) — ¥(3 F 2x) exp [k(r — 275)]}, (A12)

with a and y disposable constants. The requirement of J_/7, = Ro at the second boundary (rt = 79) leads to

_ 3 — 2k) — Rof3 + 2x)

~ (3 — 2) — Ro(3 — 2k)’(A13)

and the condition /,(0) = 1 then gives a. Thefinal expression for the intensity J (averaged over direction cosines)is then

_ 3{exp (—«7) — y exp [k(r — 270)}J(7) = (3 + 2x) — (3 — 2x)y exp (—27) (A14)

Another quantity of interest is the effective reflection coefficient R, of the whole slab at the surface of incidence,whichis ocJ_(0),

_ 3B — 2k) — vB + 2k) exp (—2«79)

R= (3 + 2x) — »(3 — 2k) exp (—2«79) (Al>)

Two limiting cases are of interest for the solutions given by equations (A12) to (A15). Considerfirst the casewith Ry = 0 in the limit of « > 0 with 7, kept finite. In this case equations (A14) and (A15) reduce to

4 _ 30 — +2,fel-R=ssq JO=“aRea (Al6)

The other case is the limiting case of 7) —> oo for any constant, nonzero x. Independently of the value of Ro, thiscase gives

po x < 1, the value offin equation (A17) agrees with the average over 0 < po < 1 of equation (A5) whichis= (4x/3).Noexact solutions of the eikonal equation (A10) exist for general «(7), but if « varies sufficiently slowly, oneis

tempted to use the WKB approximation. For a semi-infinite medium, only one of the two WKB solutionssurvives,and we have

I(t) 0 [x(2)]-2? exp |-I dr'x()| . (A18)

Note that substituting a varying «(7r) into equation (A17) (or into eq. [5] of Prinn 1970, which is the equivalent ofour eq. [A3] for normal incidence instead of diffuse illumination) is not a good approximation to equation (A18)if x varies appreciably, even if the variation is slow. Furthermore, the requirement for the validity of the WKBapproximationis that (dx/dr) « «*, which is not satisfied by practical cases for x(r), such as the example in Figure2. An inelegant but reliable numerical method proceedsas follows: Subdivide the total slab into a large but finitenumberof thin slabs (with Ar ~ 0.3, say) and replace «(r7) for each thin slab by its average value overthis slab.Start with the /ast thin slab, assume Ry = (3 — 2x)/(3 + 2x), and evaluate equations (A13) to (A15) with the slabthickness 7, replaced by Ar. The value for R,; obtained from equation (A15) for the last slab is then used as Ry forthe penultimate thin slab, and J and R, are again evaluated, and so on iteratively until the first thin slab. The solidcurve labeled J is the result of such a numerical solution for J/(7), and the effective reflection coefficient R; for thewhole medium (obtained at sr = 0) is R,; ~ 0.85. The dashed curveis the result for J(7) in a different problem—noabsorption for 7 < 7, and complete absorption for + > 79, with tr) ~ 7.5 (the value of 7 for which « = 0.5 in theactual problem). Equation (A16) gives the solution for J/(7) for this problem and also yields R; ~ 0.85. Equation(A16) for R; should be a fairly good solution for any «(7), with 7) the value where « ~ 0.5, as long as 7) > 1 andx is small for 7 appreciably less than 7.The accuracy of the numerical solution described aboveis limited (apart from computational errors) by errors

inherent in the Eddington approximation, including the use of a constant coefficient in equation (A11). For asingle slab this approximation is excellent when « « 1, although other two-stream approximations are better


LSTOADIS..

No.4, 1976 JOVIAN ATMOSPHERE 755

(Sagan and Pollack 1967; Irvine 1975) when #, « 1. For the problems with 7) > 1, discussed here, « is small nearthe incident surface and increases only slowly with 7 (at first). The results for J(7r) and R,; are most sensitive tolayers with moderately small 7 and are insensitive to inaccuracies in the Eddington approximation at large r.

REFERENCES

Aitken, D. K., and Jones, B. 1972, Nature, 240, 230.Allen, C. W. 1973, Astrophysical Quantities (3d ed.; London:

Athlone).Aumann, H. H., Gillespie, C. M., and Low, F. J. 1969, Ap. J.

(Letters), 157, L69.Axel, L. 1972, Ap. J., 173, 451.Batchelor, G. K. 1970, An Introduction to Fluid Dynamics

(New York: Cambridge University Press).Bernal, J. D. 1967, The Origin of Life (London: Weidenfeld

and Nicolson).Brown, R. 1976, Ap. J. (Letters), 206, L179.Carlson, R. W., and Judge, D. L. 1974, J. Geophys. Res., 79,

3623.Chandrasekhar, S. 1960, Radiative Transfer (New York:

Dover).Elliot, J. L., Wasserman, L. H., Veverka, J., Sagan, C., and

Liller, W. 1974, Ap. J., 190, 719.Ferris, J. P., and Chen, C. T. 1975, Nature, 258, 587.French, R. G., and Gierasch, P. J. i974, J. Atm. Sci., 31, 1707.Gierasch, P. J. 1973, Icarus, 19, 482.

. 1976, Icarus, in press.Gillett, F. C., Low, F. J., and Stein, W. A. 1969, Ap. J., 157,

925.Hogan, J. S., Rasool, S. I., and Encrenaz, P. J. 1969, J. Atm.

Sci., 26, 898.Houck, J. R., Pollack, J. B., Schaak, D., Reed, R. A., and

Summers, A. 1975, Science, 189, 720.Hunten, D. M. 1975, in Atmospheres ofEarth and the Planets,

ed. B. McCormick (Dordrecht: Reidel).Ingersoll, A. P., Munch, G., Neugebauer, G., and Orton, G.S.

1976, in Jupiter, the Giant Planet, ed. T. Gehrels (Tucson:University of Arizona Press).

Irvine, W. M. 1975, Icarus, 25, 175.JANAF. 1971, Thermochemical Tables, 2d ed. (NSRDS-

NGS37).Keay, C. S., Low, F. J., Rieke, G. H., and Minton, R. B. 1973,

Ap. J., 183, 1063.Khare, B. N., and Sagan, C. 1973, Icarus, 20, 311.

. 1975, Science, 189, 722.Kliore, A. J., and Woiceshyn, P. M. 1976, in Jupiter, the Giant

panels ed. T. Gehrels (Tucson: University of ArizonaPress).

Lacy, J. H., Larrabee, A. I., Wollman, E. R., Geballe, T. R.,Townes, C. H., Bregman, J. D., and Rank, D. M. 1975,Ap. J. (Letters), 198, L145.

Langmuir, I. 1948, J. Meteor., 5, 175.Lehninger, A. L. 1972, Bioenergetics (2d ed.; Menlo Park:

Benjamin).Lewis, J. S. 1969a, Icarus, 10, 365.

. 1969b, Icarus, 10, 393.

. 1976, Origins ofLife, in press.Lewis, J. S., and Prinn, R. G. 1970, Science, 169, 472.iiatre J. 1971, The Physics of Clouds (Oxford: Clarendon

Press).Miller, S. L., and Orgel, L. 1973, The Origins of Life (Engle-wood Cliffs: Prentice-Hall).

Morowitz, H., and Sagan, C. 1967, Nature, 215, 1259.Noyes, W. A., Jr., and Leighton, P. A. 1941, The Photo-

chemistry of Gases (New York: Reinhold).

Orton, G. S. 1975a, Icarus, 26, 125.. 19756, Icarus, 26, 142.. 1975c, Icarus, 26, 159.

Peek, B. M. 1958, The Planet Jupiter (London: Faber & Faber).Pimentel, G. C., Atwood, K. C., Gaffron, H., Hartline, H. K.,

Jukes, T. H., Pollard, E. C., and Sagan, C. 1966, Biologyand the Exploration of Mars, ed. C. S. Pittendrigh, W.Vishniac, and J. P. T. Pearman (NAS/NRC Publication1296), chap.12.

Prinn, R. G. 1970, Icarus, 13, 424.. 1973, Science, 182, 1132.. 1974, Bull. AAS, 6, 375.

Prinn, R. G., and Owen, T. 1976, in Jupiter, the Giant Planet,ed. T. Gehrels (Tucson: University of Arizona Press).

Rabinowitz, J., Woeller, F., Flores, J., and Krebsbach, R.1969, Nature, 224, 796.

Ridgway, S. T. 1974, Ap. J. (Letters), 187, L41.Rossow, W. 1976, Ph.D. thesis, Cornell University.Sagan, C. 1960, Jet Propulsion Lab. Tech. Rept. 32-34.

. 1961, Radiation Res., 15, 174.

. 1968, Science, 159, 448.

. 1970, Encyclopaedia Britannica, s.v. “‘Life.”’

. 1971, Space Sci. Rev., 11, 73.———. 1973, Icarus, 18, 649.

. 1975, Origins ofLife, 5, 497.Sagan, C., Dayhoff, M. O., Lippincott, E. R., and Eck, R.

1967, Nature, 213, 273.Sagan, C., and Khare, B. N. 197la, Science, 173, 417.

. 1971b, Ap. J., 168, 563.Sagan, C., and Miller, S. L. 1960, A.J., 65, 499.Sagan, C., and Pollack, J. B. 1967, J. Geophys. Res., 72, 469.Sagan, C., Veverka, J., Wasserman, L., Elliott, J., and Liller,W. 1974, Science, 184, 901.

Salpeter, E. E. 1973, Ap. J. (Letters), 181, L83.. 1974, Ap. J., 193, 579.

Savage, B. D., and Caldwell, J. J. 1974, Ap. J., 187, 197.Schwarzschild, M.1958, Structure and Evolution ofthe Stars,

(Princeton: Princeton University Press).Shklovskii, I. S., and Sagan, C. 1966, Intelligent Life in the

Universe (San Francisco: Holden-Day).Smoluchowski, R. 1973, Ap. J. (Letters), 185, L95.Stauffer, D., and Kiang, C. S. 1974, Icarus, 21, 129.Stevenson, D. J., and Salpeter, E. E. 1976, in Jupiter, The

plantPlanet, ed. T. Gehrels (Tucson: University of ArizonaPress).

Strobel, D. F. 1973a, J. Atm. Sci., 30, 489.. 19736, J. Atm. Sci., 30, 1205.. 1975, Rev. Geophys. Space Phys., 13, 372.

Trafton, L. M., and Munch, G. 1969, J. Atm. Sci., 26, 813.Vallentyne, J. R. 1964, Geoch. Cosmoch. Acta, 28, 157.Wallace, L., and Hunten, D. M. 1973, Ap. J., 182, 1013.Wallace, L., Prather, M., and Belton, M. J. 1974, Ap. J., 193,

481.Weidenschilling, S. J., and Lewis, J. S. 1973, Icarus, 20, 465.Westphal, J. A., Matthews, K., and Terrile, R. J. 1974, Ap. J.

(Letters), 188, L111.Williams, P. M., and Carlucci, A. F. 1976, Nature, 262, 810.Woeller, F., and Ponnamperuma, C. 1969, Icarus, 10, 386.

CARL SAGAN: Laboratory for Planetary Studies, Cornell University, Ithaca, NY 14853

E. E. SALPETER: Newman Laboratory for Nuclear Studies, Cornell University, Ithaca, NY 14853


PARTICLES, ENVIRONMENTS, ANDPOSSIBLEECOLOGIES IN ...oe reyes 232.6937 LSTOADIS.. No.4,1976...

Documents

Transcript of PARTICLES, ENVIRONMENTS, ANDPOSSIBLEECOLOGIES IN ...oe reyes 232.6937 LSTOADIS.. No.4,1976...