Effects of atmospheric solar absorption

On the present Earth the idealized picture of climate in which all solar absorption occurs at the ground is useful, but even for the present Earth about 20% of solar radiation is absorbed within the atmosphere. For other atmospheres, the proportion absorbed in the atmosphere could be much greater. The effect of this absorption on climate depends very much on the vertical distribution of the absorption, and that is what we will explore here for selected real gases.

The two key questions we have in mind for this section are the effect of solar absorption on the stratospheric temperature profile and the effect of solar absorption on surface temperature. When is solar absorption strong enough to inhibit convection and chill the surface? Even when the primary effects are stratospheric, it should be kept in mind that indirect feedbacks through stratospheric chemistry can still have an important influence on tropospheric climate. In particular, it takes very little mass of cloud to substantially affect the planet's radiation budget, and the temperature of the stratosphere can influence the kind of clouds that form there. A prime example of this phenomenon in the current Solar System is Titan, whose stratospheric organic haze clouds are a key player in the radiation budget. It has been suggested that such clouds may have played a role in a methane-rich anoxic Early Earth atmosphere, and indeed similar phenomena may be widespread in the universe.

For gas or ice giants, the problem of atmospheric solar absorption is even more critical, since it is the whole story with regard to the external energy supply of the atmosphere, there being no distinct liquid or solid surface at which to absorb solar radiation. The profile of absorption can determine whether the solar driving hinders or helps convection, and (together with heat flux from the interior) determines where the troposphere is located. These planets contain a diverse variety of condensible substances and clouds, many of which are known to absorb near-/R, visible and UV radiation. However, little is known about the vertical distribution of absorbers, or even which ones are dominant in giant planets in our own Solar System. This is a very unsettled area about which we will have little to say; a few pointers into the literature are given in the Further Readings section. The situation with regard to stellar absorption in extrasolar gas or ice giants is of course even more unsettled, but offers wide scope for exploration of hypothetical atmospheres.

5.10.1 Near-IR and visible absorption

We'll begin with an overview of the near-/R and visible absorption characteristics of CO2, CH4 and water vapor, shown in Fig. 5.11 and Fig. 5.12. These are drawn from data in the HITRAN database. As was the case for thermal /R, the absorption coefficients have an intricate line structure leading to fine-scale variations with wavenumber. The figures summarize the properties by showing only the median absorption in bins of width 50 cm-1. This is sufficient to provide a general idea of where the gases absorb strongly and where they are largely transparent. Results are given at a standard pressure of 100 mb, and can be scaled (approximately) linearly to other pressures as discussed in Chapter 4.

CO2 has a dense forest of absorption features in the near-/R below 8000 cm-1, and sparser, weaker absorption features at higher wavenumbers. We see immediately that the spectral class of the planet's star matters very much to the importance of near-/R absorption. For a cool, red M star, the stellar spectrum overlaps very considerably with the absorption features of CO2, whereas a much lower (though not by any means insignificant) proportion of a hotter G star output lead to absorption. The same remark applies to virtually all infrared-active gases.

Which of the CO2 absorption features come into play depends on how much CO2 there is in the planet's atmosphere, and for this we turn to the pressure-adjusted path, defined in Section 4.2.1. For 300 ppmv of CO2 in 1 bar of Earth air under Earth gravity, the adjusted path based on the reference pressure used in Fig. 5.11 is about 20 kg/m2, so it is only the three spikes that rise above an absorption coefficient of .01 m2/kg that contribute significantly to atmospheric heating. (The adjusted path of the present thin Martian atmosphere is similar, at 13 kg/m2, so the near-/R absorption picture for present Mars is rather similar to Earth.) All three spikes overlap significantly with the output of an M-star, but it is only the two higher wavenumber features that contribute significantly for a G star like the Sun. If the CO2 on Earth increases to 20% (mole fraction) of the atmosphere, then the path is over 15000 kg/m2, so all absorption features stronger than 10-5m2/kg come into play, which takes a fairly sizable chunk out of the incoming radiation for a G star and even more so for an M star. CO2 levels of this magnitude are typical of the Faint Young Sun period on Earth, and are also similar to the levels required to enable deglaciation of a Snowball state; in such circumstances, near-IR absorption due to CO2 becomes a significant player in climate. If we go to a 2 bar pure CO2 atmosphere, the adjusted path is about 2 • 105 for Earth gravity, around three times as much for Mars gravity, and about half as much for a typical massive

5000 1 10'

Wavenumber (cm 1)

5000 1 10'

Wavenumber (cm 1)

5000 1 10'

Wavenumber (cm 1)

5000 1 10'

Wavenumber (cm 1)

Figure 5.11: The lower panels show the median absorption coefficient in 50 cm-1 bins for CO2 (left panel) and CH4 (right panel). The absorption coefficients were computed with T = 260K and p = 100mb. The 75th percentile coefficients are typically about an order of magnitude above the median. The upper panels show the distribution of incoming stellar radiation for a typical cool, red M dwarf, and for a hotter yellow G-dwarf like the Sun.





Super-Earth. At these values, the gaps where the atmosphere is transparent narrow considerably, and the atmosphere becomes optically thick in most of the spectral region below 6000 cm-1. A 100 bar Venus-like atmosphere would absorb nearly all the incoming stellar radiation within all the absorption regions shown in Fig. 5.11, though the net increase in absorbed flux would be modest since the atmosphere will have already absorbed most of what it can absorb in the top 2 bar.

CH4 is a much more potent near-IR absorber than CO2, and has fewer transparent window regions. This limits the potential of CH4 as a greenhouse gas at high concentrations, since the anti-greenhouse effect arising from heating of the upper atmosphere partly offsets the surface warming due to thermal-IR opacity. At present Earth methane concentrations concentrations of around 2 ppmv, the adjusted path is only 0.06 kg/m2, so given the typical magnitude of the absorption coefficients, the near-IR absorption can be neglected. However, if the concentration rises to 1000 ppmv, as it easly could in an anoxic atmosphere, CH4 absorbs a considerable portion of the stellar flux below 10000 cm-1. As before, the implications for climate are even more consequential for an M star than for a G star.

Water vapor has strong absorption features extending well into the visible range, though it is also fairly well supplied with relatively transparent window regions. There are three distinct archetypical planetary situations to be thought about with regard to water vapor. First, in Earthlike conditions, water vapor is a minor and condensible constituent, which is concentrated in the lower atmosphere. For example, in a saturated 100 mb thick near-ground layer at 300K there are 22 kg/m2 of water vapor, yielding a pressure-adjusted path of about 200 kg/m2. There are several absorption peaks below 6500 cm-1 that are effective for a path this large. Because water

Figure 5.12: As for Fig. 5.11, but for water vapor. Note that the spectral range shown is twice that for CO2 and CH4, since water vapor absorbs strongly out to higher wavenumbers.

vapor in the Earthlike regime absorbs largely near the ground, it acts almost the same as reducing the ground albedo. However, if the ground would have absorbed the near-IR anyway, the net effect on climate would be minimal. Over a high albedo surface, the absorption would be more consequential. On a Snowball Earth with 250K near surface air temperatures, the same layer still has an adjusted path of almost 5 kg/m2, and referring to Fig. 5.12 we see that there would still be considerable near-surface absorption.

The second regime to consider is Venus-like, in which water vapor is a noncondensible well-mixed trace gas. 20 ppmv of water vapor in the atmosphere of Venus yields a pressure-adjusted path of over 3000 kg/m2, which would yield strong absorption all the way out to 15000 cm-1, with the exception of a few narrow window regions. Water vapor thus plays a very significant role in solar absorption on Venus, just as it does in the Venus greenhouse effect. By proportion, there is not much water in the atmosphere of Venus, but because the atmosphere is so massive the amount of water adds up to a considerable value, and its absorption is further strengthened by the high-pressure environment.

The final regime to consider is a runaway greenhouse steam atmosphere. The top bar of such an atmosphere under Earth gravity has a pressure-adjusted path of 50000 kg/m2, and so nearly all of the star's near-IR output would be absorbed in that layer, considerably heating it and affecting the planet's energy balance. For an M star, this spectral region contains most of the star's output, and so the near-IR absorption has the potential to play a key role in the climate of a runaway greenhouse atmosphere for planets orbiting M stars.

The spectral overview we have just provided does not tell us very precisely how much incoming flux is absorbed and how the absorption is distributed in the vertical. For this, we need to compute the flux profile taking into account the full variability of the absorption coefficient.

We shall do this by using the exponential sum method to compute the transmission function from the top of the atmosphere to level p in each of a set of bands covering the spectrum, and then summing up the transmission weighted by the incoming stellar flux in each band. This is in essence a simplified subset of the calculation done in the homebrew radiation code discussed in Chapter 4. In this calculation, the thermal emission from the atmosphere is negligible, so one only needs to compute the transmission of radiation entering from the top. To focus on the most essential features, we'll also assume that all flux reaching the bottom boundary is absorbed there, so we need not consider the transmission of upward-reflected stellar radiation. In most circumstances, this is a minor influence, since the parts of the spectrum where the atmosphere absorbs well are mostly depleted by the time the ground is reached. As a further simplification, we'll assume the atmosphere to be transparent outside the spectral region covered in the spectral survey, and neglect Rayleigh scattering. Rayleigh scattering is fairly weak in the near-/R, but Rayleigh scattering of visible and ultraviolet light would keep some of the incoming radiation from being absorbed at the ground. Finally, the calculations are carried out with a fixed zenith angle having cos Z = 2, rather than averaging the zenith angle over day and season at some given latitude. The profiles were all computed subject to a net downward stellar radiation of 350 W/m2 coming in at the top of the atmosphere, but as the problem is linear, the flux can be readily scaled to any other value. In each case, we did the calculation for an M star with photospheric temperature of 3000K and for a G star at 6000K. The calculations were carried out with a gravitational acceleration of 10m/s2.

Results for a pure CO2 atmosphere, a CO2-air mixture, and a pure H2O atmosphere are shown in Fig. 5.13. These were computed for an isothermal 260K atmosphere, but the transmission is not terribly sensitive to the temperature profile. When examining fluxes plotted against a logarithmic pressure axis, it is good to keep in mind that if F is the flux, then the heating rate per unit mass is proportional to the slope dF/dp, which is p-1dF/d ln p. Thus, a given slope in log coordinates corresponds to a greater heating rate at low pressures than it does at high pressures. For well-mixed greenhouse gases, one typically finds high heating rates aloft, since the upper atmosphere gets the first chance to absorb the part of the spectrum that is absorbed very strongly.

In all cases, the absorption for the M star case is substantially greater than that for the G star case, as expected. A 2 bar CO2 atmosphere absorbs 30 W/m2 of the incident 350 W/m2 for the G star, but 100 W/m2 for the M star. For both kinds of stars, about a third of the total flux is deposited at pressures lower than 100 mb, which will lead to intense heating in the upper atmosphere. Still, a great deal of the flux is absorbed in the lower atmosphere. Over a highly absorbing surface like an ocean or dark land, it matters little to the tropospheric temperature whether the energy is absorbed in the troposphere or at the surface. Over a highly reflective surface, as in a Snowball state, the effect of the lower atmospheric absorption is to decrease the effective albedo, which will lead to a warming of the troposphere. In any event, the vertical distribution of absorption for CO2 does not suggest a pronounced anti-greenhouse effect.

The situation for a mixture of 20% CO2 in air is quite similar, and in fact leads to only slightly less total absorption, because the additional CO2 in the pure CO2 case is only able to absorb parts of the spectrum where CO2 is a relatively poor absorber. This is another instance of the typical logarithmic dependence of radiative properties on greenhouse gas concentrations. The general implication is that stellar absorption by CO2 should have only a minor effect on tropospheric climate when the surface has low albedo. Over a high albedo surface, as in a snowball, we expect some modest tropospheric warming, which can help in deglaciation. For example, assuming a surface pressure of 1000 mb, the atmospheric absorption is 6 W/m2 for a G star between the 400 mb level and the ground, or about twice as much for an M star. If the near-/R albedo of the surface is 50%, then half of the amount absorbed in the troposphere would have been absorbed at the surface anyway, so the additional radiative forcing due to tropospheric absorption is only half the stated values. Assuming a 60% surface albedo averaged over the entire solar spectrum, the surface solar absorption for a transparent atmosphere would be 140 W/m2. Hence, the additional radiative forcing amounts to between 2.5% and 5% of the energy budget. This is not overwhelming, and would be partially offset by the cooling due to stratospheric absorption. Still, it is a factor that should be taken into consideration in determining the conditions for deglaciating a Snowball state.

Now let's turn to the pure water vapor case for a layer extending to a pressure of 2 bar. This can be thought of as the top 2 bar of an atmosphere undergoing a runaway greenhouse, or alternately the whole atmosphere for a world with a substantial ocean having a surface temperature of around 395 K. Since the absorption coefficients in this example are computed with a fixed temperature of 280 K, the opacity of the lower part of the atmosphere has been underestimated, but the example nonetheless suffices to demonstrate just how powerfully water vapor absorbs in the near-IR. Even on a logarithmic plot the slope at 50 mb is greater than the slope at 500 mb, indicating an extremely intense upper level heating which should lead to pronounced warming of the upper atmosphere. Rather little stellar flux makes it to the 2 bar level - only 150 W/m2 in the G star case and 40 W/m2 in the M star case. Calculations with thicker atmospheres (not shown) reveal that the attenuation continues as the surface pressure is further increased. For example, in the M star case, only 15 W/m2 reaches the ground for a 20 bar atmosphere and 5 W/m2 for a 200 bar atmosphere. For a hot atmosphere, temperature scaling of line strengths would further reduce the penetration, though the precise effect takes us into unknown territory with regard to temperature scaling of the water vapor continua at high temperatures. Further, the pressure scaling of absorption coefficients in the exponential sum code probably underestimates the true influence of line broadening at very high pressures, so most probably the flux would be further reduced in a precise calculation. For a G star there would be less absorption, but this would be offset by greater Rayleigh scattering owing to the shorter wavelengths in the incident stellar radiation. A runaway greenhouse would be a very dark place at the surface, with hardly any radiation penetrating to the ground.

With regard to the climate implications of the limited flux penetrating the atmosphere, however, it should be kept in mind that this atmosphere is also very optically thick in the thermal infrared, so that even in the 2 bar M star case, the lower atmosphere would have to achieve a very high temperature in order to be able to lose 40 W/m2 by radiative diffusion through the highly opaque atmosphere. If the tropopause rises so high as a result of this heating that it engulfs the stellar heating region aloft, the anti-greenhouse effect will be suppressed. Even in pure radiative equilibrium one must reckon with the fact that if H2O is a good absorber of stellar near-IR, it is also a good emitter of thermal IR, and the outcome of the competition between these two factors is not easy to resolve a priori. The most that can be said at this point is that near-IR stellar absorption must be considered as a serious factor in steam atmospheres, all the more so in the M star case.

Now let's take a look at the Earthlike case of saturated water vapor mixed with air, shown in Fig 5.14. In this case, the temperature profile matters a great deal, since it determines the vertical distribution of water vapor. These calculations were done for an atmosphere on the moist adiabat. Results are shown for a tropical case with surface temperature of 300K and a Snowball case with a surface temperature of 250 K. In contrast to the case of a pure steam atmosphere, the absorption is trapped in the lower atmosphere for the Earthlike case. This adds significantly to the heating of the troposphere, but again, over a strongly absorbing surface the absorption is only acting to radiatively deposit energy directly in the troposphere, which otherwise would have been absorbed at the surface and communicated to the troposphere by convection. As we

Downward Stellar Flux (W/m2)

Downward Stellar Flux (W/m2)

Downward Stellar Flux (W/m2)

Downward Stellar Flux (W/m2)

Downward Stellar Flux (W/m2)

Downward Stellar Flux (W/m2)

Figure 5.13: Profiles of incoming stellar flux computed using an exponential-sum transmission function for three different atmospheres: 2bar pure CO2 (left), 20% molar CO2 in Earth air (middle) and 2bar pure H2O (right). The incoming vertical flux at the top of the atmosphere is 350 W/m2 in all cases, and results are shown for both a G star and M star spectrum.

discussed for the CO2 case, the absorption is more consequential if the surface is highly reflective. In this regard, it is important to note that the low level absorption is substantial even when the surface temperature is only 250K. In both the G and M star cases, the absorption is markedly greater than the corresponding low level absorption in the CO2 case. Since tropical temperatures during a Snowball can easily reach 250K, solar absorption by water vapor can significantly assist deglaciation, by lowering the effective surface albedo. Any process which warms the atmosphere will further increase the atmospheric water content and thus further increase the solar absorption. This constitutes a novel sort of water vapor feedback, which operates via the effect of water vapor on the solar spectrum instead of the thermal infrared effect.

For a complete appreciation of the effect of stellar absorption on the temperature profile, one must compute the radiative-convective equilibrium in the presence of absorption. We will discuss a few such calculations now, for the case of pure CO2 atmospheres. In Section 4.8 we carried out thermal infrared radiative-convective solutions by fixing the ground temperature Tg and finding the atmosphere that was in equilibrium with the upwelling radiation from the ground. If one then wanted to know what Tg would be supported by a given amount of absorbed stellar radiation, it was only necessary to vary Tg until the desired OLR was achieved. This procedure will not do in the presence of atmospheric stellar absorption, since the incoming flux must be known in order to compute the temperature profile. Hence, in the calculations to follow, we adopt a somewhat different procedure, specifying the incoming stellar radiation, time-stepping the temperature profile as before, but this time adjusting Tg until the top-of-atmosphere energy balance is satisfied. Where the atmosphere is statically unstable relative to the ground temperature, the temperature profile is reset to the adiabat as before.

There are various ways to approach the problem of adjusting Tg. If one were interested in reproducing the actual time evolution of the system, it would be necessary to use the surface downwelling stellar radiation and thermal infrared in order to determing the surface temperature change; then, turbulent and radiative heat fluxes would heat the low lying air and if instability

Downward Stellar Flux (W/m2) Downward Stellar Flux (W/m2)

Figure 5.14: Profiles of incoming stellar flux computed using an exponential-sum transmission function for saturated water vapor in Earth air on the moist adiabat. The temperature labels indicate surface temperature. The left panel shows results for a G star spectrum, while the right shows M star results. These calculations take into account the effect of the temperature profile on water vapor, but do not incorporate the temperature-scaling of absorption coefficients.

results, the heat would be mixed upward by convection. In our case, we are only interested in getting the equilibrium state, so any procedure that converges to the equilibrium will do. The following simple iteration works quite well in practice. The general idea is that if the net top-of-atmosphere radiation ( incoming stellar minus OLR) is downward, then Tg needs to be increased in order to bring the atmosphere closer to balance; conversely, if the net is upwards, Tg needs to be increased. The stellar absorption is not very sensitive to temperature, so the change in Tg mainly affects the OLR. Thus, the main problem is to figure out the climate sensitivity, dOLR/dTg. If the atmosphere is optically thick, then increasing Tg with the atmospheric temperature fixed would not change OLR, but the process we envision is that increasing Tg warms the rest of the atmosphere by convection and radiation, and this ultimately leads to an increase in OLR. The key simplification in the iteration is to estimate dOLR/dTg as if the atmosphere were a grey gas. Specifically, we compute the radiating temperature Trad from <jT4ad = OLR, since we know the OLR from the radiation calculation. Then, the climate sensitivity is estimated as dOLR/dTg « 4aT4ad. Adopting the convention that a net downward flux is negative, the iteration to be performed at each time step is

T rad where Ftoa is the net top-of-atmosphere flux and e is an under-relaxation factor that adjusts the ground temperature just part of the way towards the target, which improves the stability of the iteration. The following calculations were performed with e = .05, which provides a reasonable compromise between stability and rate of convergence. This iteration procedure is rather ad hoc, and no doubt there are more sophisticated schemes which rest on firmer ground. However, we have found it to serve quite well over a range of situations.

Figures 5.15 and 5.16 show radiative-convective equilibrium results for pure CO2 atmospheres, carried out using this procedure. As in the flux profiles shown previously, the surface was assumed to be completely absorbing. The results in Fig. 5.15 are in equilibrium with 350 W/m2

Figure 5.15: Radiative-convective equilibrium subject to an incoming stellar flux of 350 W/m2 for pure CO2 atmospheres, for a planet with 10 m/s2 surface gravity. The first two panels have surface pressure of 2 bar and the rightmost panel has surface pressure of 20 bar. The surface is assumed completely absorbing. In the 2 bar cases, calculations without atmospheric stellar absorption are shown for comparison. The G star cases assume a blackbody spectrum of incoming radiation with a temperature of 6000K, whereas the M star cases assume 3000K.

Figure 5.15: Radiative-convective equilibrium subject to an incoming stellar flux of 350 W/m2 for pure CO2 atmospheres, for a planet with 10 m/s2 surface gravity. The first two panels have surface pressure of 2 bar and the rightmost panel has surface pressure of 20 bar. The surface is assumed completely absorbing. In the 2 bar cases, calculations without atmospheric stellar absorption are shown for comparison. The G star cases assume a blackbody spectrum of incoming radiation with a temperature of 6000K, whereas the M star cases assume 3000K.

of incoming stellar radiation, on a planet having a surface gravity of 10m/s2. In the 2 bar G star case, the near-IR absorption moderately warms the stratosphere relative to the no-absorption case. The effect of absorption on surface temperature is hardly detectable in the figure. It amounts to a cooling of about 4K. For the 2 bar M star case, the warming aloft is much more pronounced, and there is a significant lowering of the tropopause. In this case the surface cooling caused by absorption is 15 K, though given the high surface temperature this is hardly a very consequential effect. When the surface pressure is increased to 20 bar in the M star case, the surface temperature increases dramatically, but the stratospheric temperature changes little, with the main effect being a slight warming of the highest stratosphere, which leaves the stratosphere quite isothermal. The no-absorption case for the 20 bar atmosphere (not shown) very closely follows the adiabat and has a very high tropopause. The surface cooling caused by absorption in this case increases to 22 K, but this offsets little of the additional greenhouse warming which raises the surface temperature to 460 K.

Finally, to give an example of the situation for thin atmospheres, we show a calculation carried out in the Present Mars regime in Fig. 5.16. This calculation was done with Mars gravity, and subject to G star illumination. We see that, as in the dense atmosphere cases, the G stellar absorption causes only a moderate warming of the stratosphere. The observed stratospheric temperature is notably warmer than the calculation, which suggests that near-IR absorption alone is not able to fully account for the Martian stratospheric temperature seen in this sounding. Absorption due to dust is a likely culprit, but effects due to the global scale stratospheric circulation may be playing a role as well.

The summary situation for pure CO2 atmospheres is that stellar near-IR absorption causes moderate stratospheric warming for planets about G stars case and more pronounced stratospheric warming for M stars, but in neither case does the stellar absorption lead to a stratospheric temper-

Present Mars Case

Present Mars Case

Figure 5.16: Radiative-convective equilibrium for Present Mars. The incoming Solar flux is 250 W/m2 and has been chosen to yield a tropospheric temperature similar to the observation shown in the figure. The observation shown for comparison is from a summer afternoon tropical sounding, from the Mars Global Surveyor radio occultation dataset.

ature inversion; the temperature is monotonically decreasing everywhere. The effect of absorption on the tropospheric temperature is a modest cooling. Thus, the anti-greenhouse effect is not terribly consequential for pure CO2 atmospheres. We'll note also that all of the stratospheres in the above results are optically thin in comparison with the tropospheres, which implies the happy result that reasonable estimates of surface temperature can be made on the basis of the simple and swift all-troposphere OLR model.

The radiative-convective behavior of thick water vapor atmospheres presents considerably more challenge because of the extreme optical thickness of water vapor both in the near-IR and thermal-IR. This is a very rich problem which entails consideration of the delicate balance between very slow radiative cooling and the small amount of stellar radiation reaching the surface, as well as the effects of condensation on the adiabat and the increase of surface pressure with temperature for an atmosphere in equilibrium with an ocean. For the hot steam atmospheres which are of greatest interest, one also comes up against the largely unknown behavior of the water vapor continua at high temperatures and pressures. We will be content to leave this deep and interesting problem as a subject for research. It is an important problem because the stellar absorption has the potential to significantly increase the threshold illumination needed to trigger a runaway greenhouse. The reader is now in posession of all the tools necessary to carry out an inquiry of this sort.

5.10.2 Ultraviolet absorption

Because of its importance to near-surface life on Earth, ozone (O3) is probably the most familiar of all ultraviolet-absorbing gases. The interest stems from the fact that the shorter wave and more energetic forms of UV radiation wreak havoc with key biological molecules of life as we know it, and in particular genetic information encoded in DNA. It is a highly Earth-centric view to think that an ozone shield is necessary to protect complex life in general from deady UV radiation, but notwithstanding that issue, ozone has some very profound effects on the stratospheric temperature structure that play a key role in the prospects for detecting O2 (and presumably oxygenic photosynthetic life) on extrasolar planets.

UV wavelengths are customarily measured in nanometers (nm, or 10-9m). The radiation begins to become harmful to Earth life at 320 nm, and wavelengths shorter than 300 nm cause extreme damage. Ozone plays a distinguished role in shielding Earth life from UVB (320-280 nm) and UVC (280-100 nm) radiation. Shorter UV wavelengths, to say nothing of Solar X-rays, are even more deadly, but there are many molecules which efficiently absorb wavelengths shorter than 100 nm. For example, CO2, which is far more abundant than O3 in Earth's atmosphere, is every bit as absorbent as O3 in the vicinity of 140 nm. In contrast O3 is a potent absorber between 200 and 300 nm, whereas CO2 and most other reasonably abundant atmospheric gases are nearly transparent there. Ozone also absorbs significantly in the 400 to 700 nm range. These wavelengths are not particularly damaging to life, but because the stellar output is abundant in this range for G-class and hotter stars, the effect on atmospheric heating is significant.

Ozone is a feature of atmospheres rich in free O2, bombarded by UV radiation. So far, Earth's is the only known example of such an atmosphere. Ozone is a highly reactive substance with a short lifetime. Therefore, it is very inhomogeneous. In particular, in Earth's present atmosphere ozone is concentrated in a stratospheric layer near the altitude where its production rate is strongest. At earlier times, when there was less O2 around, the ozone layer was probably found at a lower altitude. At present, maximum ozone values occur at about the 20 mb level in the tropics, and reach values on the order of 2 ppmv. The concentration drops by two orders of magnitude as the tropopause is approached. Even at such low concentrations, ozone is a very effective absorber. A concentration of 1ppmv in the layer of atmosphere above 20 mb gives this layer an optical thickness exceeding 4.0 at 250 nm, which is sufficient to exhaust virtually all of the UV flux at that wavelength. Detailed data on the UV absorption spectrum of O3 and other gases can be found in the resources listed in the Further Readings section.

Heating due to UV absorption by ozone has an important effect on the stratospheric temperature profiel, but ozone is also a very powerful absorber in the thermal infrared range. It is the only abundant infrared absorber which is concentrated in the stratosphere, and that that sense provides a counterpoint to water vapor, which is concentrated in the troposphere.

We will now carry out some calculations with the ccm radiation model which illustrate the key effects of ozone in an Earthlike setting. We adopt an idealized ozone profile of the form no3 = exp[-(p - 20mb)2/(50mb)2] (5.59)

where no3 is the molar mixing ratio of ozone and is the peak value, which we take to be 2 ppmv in the following calculations. The left panel of Fig. 5.17 shows the net downward flux for an atmosphere that contains Earth air and ozone, but no other gases. The atmosphere is illuminated with 400 W/m2 of incoming radiation, 25 W/m2 of which is reflected back by Rayleigh scattering. The ground is assumed to be perfectly absorbing. The temperature profile is immaterial, since UV absorption is nearly independent of temperature for typical planetary temperatures. Of the sunlight that is not scattered, about 15 W/m2 is absorbed within the ozone layer. This includes nearly all of the harmful UVB and UVC radiation. Given the low mass of this region of the atmosphere, the absorption gives rise to a considerable heating, which should substantially warm the stratosphere. There is some weak absorption within the troposphere, which is due to O2.

A key question is whether ozone causes the stratospheric temperature to increase with height in the stratosphere, as is seen in observations of the Earth's atmosphere. The right panel of Fig.

Downward Solar Flux (W/m )

-Ozone SW + LW

-•■-Ozone LW

---Dry Adiabat

l4O l6O lSO 2OO 22O 24O 26O 2SO 3OO

lOOO ri i i I i i i I i i i I i i i I i i i I i i i I i i l4O l6O lSO 2OO 22O 24O 26O 2SO 3OO

Figure 5.17: Left panel: Net downward solar flux for an atmosphere with the ozone profile described in the text. The calculation was performed for an isothermal 300K atmosphere, but the results are essentially insensitive to temperature. The incoming solar radiation is 400W/m2, some of which is scattered back by Rayleigh scattering. Right panel: Radiative-convective equilibrium for a dry atmosphere containing 300 ppmv of CO2, computed for three cases as follows. Thin solid curve with open circles - no ozone or solar absorption; Dashed curve with filled circles - ozone thermal infrared effects incorporated; Thick solid curve - ozone infrared and solar absorption incorporated. The plain dashed curve gives the adiabat for the third case. The ground temperatures differ slightly between the cases, but the difference is not visible in the figure. All calculations were performed with the ccm radiation model.

5.17 shows radiative-convective equilibrium calculations using the ccm code, computed with a dry atmosphere containing 300 ppmv of CO2, in which convectively unstable layers are adjusted to the dry air adiabat. The profiles shown are in equilibrium with 400 W/m2 of incident solar radiation. Results with ozone are compared with a control case for a solar-transparent CO2/air mixture. In the control case, the radiative-convective equilibrium temperature decreases monotonically with height, just as seen in the simulations of Chapter 4.

Because of the dual role of ozone as an infrared and UV absorber, its effect on the temperature profile is complex. In the circumstances of this particular simulation, introducing just the effect of ozone on infrared absorption introduces a temperature increase with height in the lower stratosphere. This arises because the ozone is concentrated aloft, and absorbs at wavelengths that escape the CO2 effects in the troposphere. This leads to an intense heating layer, which must warm up until it comes into equilibrium. Without solar absorption by ozone, however, the upper stratospheric temperature still declines sharply with height. Introducing the solar absorption warms the upper stratosphere considerably, and causes it to increase with height. It also results in a pronounced lowering of the tropopause.

The effect of ozone on stratospheric temperature is profound, but its effect on surface temperature is modest and largely invisible in Fig. 5.17. For the control case, the surface temperature is 295.48K. It rises to 297.8K when ozone infrared effects are introduced, owing to the greenhouse effect of ozone. However, when the ozone solar absorption is brought into the picture, the warming of the stratosphere allows the upper atmosphere to radiate better to space, and this brings the surface temperature back down to 295.22K. Thus, the main climatic effects of ozone are in the stratosphere, though it is quite possible that the lowering of the tropopause would have repercussions for tropospheric climate. Further, the absence of ozone in the anoxic Early Earth atmosphere would have led to a much colder stratosphere, which is important to take into account in working out the chemistry of Titan-like stratospheric haze clouds.

Water vapor, CO2 and CH4 all absorb ultraviolet quite strongly for wavelengths shorter than 180 nm, but the only common atmospheric constituent that competes with ozone at longer wavelengths is SO2. This gas is abundant in volcanic outgassing, but in oxygenated atmospheres it forms sulfates which are removed by rainout if the planet supports liquid water. On dry planets or planets without oxygen, SO2 can build up to higher concentrations, but its status as an ultraviolet shield still hinges on atmospheric chemistry. The very fact that it absorbs ultraviolet so well tends to dissociate the molecule, and the question then is whether there are chemical pathways that can restore it. There is no question, however, that SO2 is a molecule that holds very interesting prospects as a mediator of planetary climate evolution, especially in view of the fact that it is also a potent greenhouse gas.

Was this article helpful?

0 0

Post a comment