in REPORTS
season the evergreen forests of tropical South America, Africa, and Indonesia are needed to conclude which source type is responsible for the observed discrepancy and to what extent this discrepancy is caused by an underestimated or hitherto unknown methane source. In addition, investigations of satellite data for longer time periods are necessary to estimate the annual methane excess and to constrain the source attribution. An additional methane source of 30 Tg in 4 months is large but can be accommodated by the uncertainty in the global budget, which is estimated to be 50 to 100 Tg yearj1 (1). Further, surface observations (24) are not in disagreement with a large additional tropical source. Methane emitted in the tropics is generally rapidly uplifted by convection, so that the surface stations, which are located in remote ocean sites, are sensitive to these emissions only to a limited extent. Inverse modeling studies based on ground-based measurements (25, 26) have also indicated higher tropical emissions than those estimated from bottom-up inventories. However, this evidence has been rather indirect and uncertain. The present satellite measurements over the tropical land masses are sensitive to the entire atmospheric column, thus directly reflecting enhancement patterns. Global measurement from space has proven feasible for detection of CH4 emissions. The satellite_s ability to sense sources globally is unique and opens a new window for the analysis of the biogeochemical cycle of methane and anthropogenic impacts. With SCIAMACHY hopefully in operation for several years to come, we should be able to examine the temporal and spatial variations of CH4 over longer time periods. The integration of these measurements with atmospheric models and precise ground-based observations should greatly reduce uncertainties in the methane source strength, helping to draw a consistent picture of the global methane budget. This will be a high priority for climate research.
References and Notes
1. J. Houghton et al., Eds., Climate Change 2001: The Scientific Basis (Cambridge Univ. Press, Cambridge, 2001). 2. J. Lelieveld, P. J. Crutzen, F. J. Dentener, Tellus B50, 128 (1998). 3. J. G. J. Olivier, J. J. M. Berdowski, in The Climate System, J. J. M. Berdowski, R. Guicherit, B. J. Heij, Eds. (A. A. Balkema/Swets and Zeitlinger, Lisse, Netherlands, 2001), pp. 3378. 4. L. A. Morrissey, G. P. Livingston, J. Geophys. Res. 97, 16661 (1992). 5. B. P. Walter, M. Heimann, Global Biogeochem. Cycles 14, 745 (2000). 6. H. Bovensmann et al., J. Atmos. Sci. 56, 127 (1999). 7. J. P. Burrows et al., J. Atmos. Sci. 56, 151 (1999). 8. T. Wagner et al., Monitoring of Trace Gas Emissions from Space: Tropospheric Abundances of BrO, NO2, H2Co, SO2, H2O, O2 and O4 as Measured by GOME in Air Pollution X (WIT Press, Southampton, UK, 2002). 9. U. Platt, Differential Optical Absorption Spectroscopy (DOAS) in Air Monitoring by Spectroscopic Techniques (Wiley, New York, 1994). 10. M. Buchwitz, V. Rozanov, J. Burrows, J. Geophys. Res. 105, 15231 (2000). 11. D. M. OBrien, P. J. Rayner, J. Geophys. Res. 107, 4354 (2002). 12. C. Frankenberg, U. Platt, T. Wagner, Atmos. Chem. Phys. 5, 9 (2005). 13. C. Clerbaux, J. Hadji-Lazaro, S. Turquety, G. Megie, P. Coheur, Atmos. Chem. Phys. 3, 1495 (2003). 14. S. C. Olsen, J. T. Randerson, J. Geophys. Res. 109, 10.1029/2003JD003968 (2004). 15. F. Dentener, M. van Weele, M. Krol, S. Houweling, P. van Velthoven, Atmos. Chem. Phys 3, 73 (2003). 16. S. Houweling et al., J. Geophys. Res. 105, 8981 (2000). 17. J. A. Van Aardenne, F. J. Dentener, J. G. J. Olivier, C. G. M. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. Klein Goldewijk, J. Lelieveld, Global Biogeochem. Cycles 15, 909 (2001). K. Gurney et al., Nature 415, 626 (2002). R. S. DeFries, J. R. G. Townshend, Int. J. Remote Sens. 15, 3567 (1994). J. M. Melack et al., Global Change Biol. 10, 530 (2004). A. H. Devol, J. E. Richey, B. R. Forsberg, L. A. Martinelli, J. Geophys. Res. 95, 16417 (1990). M. O. Andreae, P. Merlet, Global Biogeochem. Cycles 15, 955 (2001). M. G. Sanderson, Global Biogeochem. Cycles 10, 543 (1996). Information about the Climate Monitoring and Diagnostics Network (CMDL) can be found online at www.cmdl.noaa.gov/ccgg. R. Hein, P. J. Crutzen, M. Heimann, Global Biogeochem. Cycles 11, 43 (1997). S. Houweling, T. Kaminski, F. J. Dentener, J. Lelieveld, M. Heimann, J. Geophys. Res. 104, 26,137 (1999). Most importantly the authors thank all scientists and engineers involved in the ESAs ENVISAT/SCIAMACHY mission, especially J. P. Burrows and H. Bovensmann from the University of Bremen, as well as the space agencies of Germany [German Aerospace Center (DLR)], Netherlands [Netherlands Agency for Aerospace Programmes (NIVR)], and Belgium [Belgian User Support and Operations Center (BUSOC)]. We also thank R. Washenfelder and G. Toon from Jet Propulsion Laboratory for the updated methane spectral database; S. Sanghavi for proofreading the manuscript; and the Netherlands SCIAMACHY Data Center, in particular A. J. M. Piters and J. van de Vegte from [Royal Netherlands Meteorological Institut (RNMI)], for invaluable assistance in handling the SCIAMACHY data format. We acknowledge the European Commission for supporting the 5th Framework Programme RTD project EVERGREEN (contract number EVG1-CT-2002-00079) and A. P. H. Goede, EVERGREEN project coordinator.
Supporting Online Material www.sciencemag.org/cgi/content/full/1106644/DC1 Materials and Methods SOM Text Figs. S1 to S4 Movie S1 22 October 2004; accepted 9 March 2005 Published online 17 March 2005; 10.1126/science.1106644 Include this information when citing this paper.
A Hydrogen-Rich Early Earth Atmosphere
Feng Tian,1,2* Owen B. Toon,2,3 Alexander A. Pavlov,2 H. De Sterck4
We show that the escape of hydrogen from early Earths atmosphere likely occurred at rates slower by two orders of magnitude than previously thought. The balance between slow hydrogen escape and volcanic outgassing could have maintained a hydrogen mixing ratio of more than 30%. The production of prebiotic organic compounds in such an atmosphere would have been more efficient than either exogenous delivery or synthesis in hydrothermal systems. The organic soup in the oceans and ponds on early Earth would have been a more favorable place for the origin of life than previously thought. It is generally believed that the existence of prebiotic organic compounds on early Earth was a necessary step toward the origin of life. Biologically important molecules can be formed efficiently in a highly reducing atmosphere (CH4- and/or NH3-rich) (1, 2). They can also be produced efficiently in a weakly reducing atmosphere (35), where one important factor influencing the efficiency of production is the ratio of hydrogen to carbon (68). However, our current understanding of the composition of early Earth_s atmosphere suggests it was neither strongly reducing nor hydrogen-rich. The concentrations of CH4 and NH3 are thought to have been low (9), and the hydrogen mixing ratio is believed to be of the SCIENCE
order of 10j3 or smaller (10). Because it is difficult to produce organics in the atmosphere, two directions of research into the origin of life on Earth have become dominant: synthesis of organic compounds in hydrothermal systems, and exogenous delivery of organic compounds to early Earth (11). Here we reexamine the theory of diffusion-limited escape of hydrogen and show that hydrogen escape from early Earth_s atmosphere was not as rapid as previously assumed. Hydrodynamic escape should be the dominant mechanism of escape, implying a hydrogen-rich early Earth atmosphere in which organic molecules could be produced efficiently. The assumption that the escape of hydrogen is limited by diffusion into the heteroAstrophysical and Planetary Science Department, Laboratory for Atmospheric and Space Physics, Program in Atmospheric and Oceanic Science, University of Colorado, Boulder, CO 80309, USA. 4 Department of Applied Mathematics, University of Waterloo, Waterloo, Ontario, N2L 3G1, Canada.
2 3 1
*To whom correspondence should be addressed. E-mail: [e-mail]
1014
13 MAY 2005 VOL 308
www.sciencemag.org
REPORTS
sphere from below is applicable only when the escape is efficient. For Jeans escape to be efficient, the exobase temperature must be high. This condition is satisfied in the current Earth_s atmosphere, where exobase temperatures exceed 1000 K as a result of the efficient absorption of solar ultraviolet (UV) radiation by atomic oxygen. The prebiotic Earth_s atmosphere was anoxic and probably contained substantial amounts of CO2 (12), similar to modern Venus or Mars. CO2 absorbs UV, but unlike oxygen, it can effectively radiate energy back to space and keep the exobase temperatures low. The CO2-rich venusian and martian exobases have temperatures of 275 K and 350 K, respectively (13). In the anoxic early Earth_s atmosphere with low exobase temperature (14), Jeans escape of hydrogen would have been inhibited. For present-day Earth and Mars atmospheres, which have low hydrogen concentrations, the relatively low hydrogen escape rates caused by nonthermal processes are comparable to the Jeans escape rate (15, 16). It is important to consider how the nonthermal escape rate might increase in an atmosphere with large hydrogen abundance. For an H2O-rich early venusian atmosphere, the maximum limit of the nonthermal hydrogen escape rate (caused by the saturation of ionization of hydrogen atoms, which occurs when the homopause hydrogen mixing ratio reaches 2 103) is 1010 cmj2 sj1 (17). The maximum nonthermal escape rate on early Earth should be similar to that on Venus (18). Given the low Jeans and nonthermal escape rates, the total hydrogen escape rate would not have been in balance with the volcanic H2 outgassing rate on early Earth (10) at the previously suggested hydrogen mixing ratio of 10j3. Instead, hydrogen would have been one of the major constituents in the ancient atmosphere. H2 can absorb extreme ultraviolet (EUV), as can O2 and CO2, but cannot effectively radiate energy back to space. However, H2 can escape because of its low molecular weight and thereby carry energy away to space in a hydrodynamic hydrogen escape flow. When hydrogen is the major gas in the heterosphere and the major absorber of EUV, the escape of hydrogen would not be diffusion-limited but would be controlled by the solar EUV flux available to drive the escape flow (energylimited), which would produce an escape rate smaller than the diffusion-limited escape rate. In this report, we describe the application of a hydrodynamic escape model, which was recently developed to study transonic hydrogen hydrodynamic escape (14), to a hydrogenrich early Earth_s atmosphere. Because the solar EUV radiation level could have been much stronger during the Archean era than today (19), EUV radiation levels 1, 2.5, and 5 that of today were used in the simulations for sensitivity studies. The velocity distributions in the high-energy input cases (2.5 and 5) (Fig. 1A) level off and converge near the upper boundary of the model where the sound speed is exceeded (supersonic flow). Transonic points in the higher-energy input cases (2.5 and 5) are near 10 Earth radii (14). Near the upper boundary of the model, the flow velocity is comparable to the escape velocity from the planet. Escape velocity is exceeded by the combination of the flow velocity and the thermal velocity at an altitude below the transonic points. In the 1 EUV level case, the energy absorbed is not adequate to drive supersonic flow, but escape still occurred. Figure 1B shows the temperature profiles in the correspond-
A
B
Fig. 1. (A) Flow velocity profiles in hydrodynamic escape of hydrogen under solar EUV levels 1, 2.5, and 5 that of today. r0 is the distance between the lower boundary and the center of Earth. The homopause hydrogen density is 5 1012 cm3 , corresponding to a mixing ratio of 50%. The dashed curve represents the escape velocity from Earth. The transonic point is marked approximately by a circle. The dash-dotted curve shows the difference between the escape velocity and the thermal velocity. The exobases are marked by crosses. (B) Temperature profiles in the corresponding cases. Although the peak temperatures are in the range of 700 to 800 K, the temperature at the exobases (marked by crosses) is low (500 to 600 K). Fig. 2. Calculated hydrogen escape rate from early Earths atmosphere as a function of the homopause hydrogen mixing ratio. The homopause air composition is assumed to be the same as that of today, except for the higher hydrogen concentration and lack of oxygen. The diffusionlimited escape rates, previously assumed to apply, are one or two orders of magnitude greater than the hydrodynamic escape rates because of overestimated exobase temperature. The dotted curve shows the Jeans escape rate as a function of homopause hydrogen mixing ratio under the 2.5 energy input level for the exobase temperatures that are likely to have been present for early Earth. The lower horizontal line represents the volcanic outgassing rate of hydrogen from the interior of Earth today (1.8 1010 hydrogen molecules cmj2 sj1) (29). The upper horizontal line is the estimated outgassing rate of hydrogen from the interior of early Earth (5 times the outgassing rate today) (30).
www.sciencemag.org
SCIENCE
VOL 308
13 MAY 2005
1015
REPORTS
ing cases. Although the peak temperatures are 700 to 800 K in the high-energy input cases, the temperatures at the exobases (marked by crosses) are in the range of 500 to 600 K because of adiabatic cooling associated with the hydrodynamic escape. Figure 2 illustrates our calculated escape rates for varying hydrogen homopause mixing ratios. The Jeans escape rates computed for the exobases are more than one order of magnitude smaller than the corresponding hydrodynamic escape rates because of the low exobase temperatures. If the solar EUV radiation level was 2.5 times that of today and the volcanic hydrogen outgassing rate was 5 times that of today (9.25 1010 hydrogen molecules cmj2 sj1), a hydrogen mixing ratio of more than 30% could have been maintained everywhere below the homopause by balancing the volcanic hydrogen outgassing with the hydrodynamic escape of hydrogen (Fig. 2). By increasing the solar EUV radiation level to 5 times that of today, the hydrogen mixing ratio could still have been maintained at 10%. These mixing ratios are two orders of magnitude greater than the 10j3 hydrogen concentration, considering the diffusion-limited hydrogen escape rate. A hydrogen-rich early Earth_s atmosphere could be maintained even for the modest hydrogen outgassing rates appropriate if the oxidation state of Earth_s mantle 3.9 billion years ago were the same as it is today.
Fig. 3. The hydrogen escape rate increases nearly as linearly the solar EUV level increases. The slope is regulated by the homopause hydrogen density n0.
The hydrodynamic escape rate increases nearly linearly as the solar EUV radiation level increases (Fig. 3), which reflects the energylimited nature of the hydrodynamic escape. The slope depends on the hydrogen density at the homopause (20). An early Earth_s atmosphere with high hydrogen concentration has important consequences for the origin and evolution of life. Endogenous sources of prebiotic organics, such as production by lightning or photochemistry, are dominant in a reducing early Earth_s atmosphere, whereas exogenous sources, such as delivery from space or production in hydrothermal systems, become major contributors in an atmosphere of an intermediate oxidation state (1). To provide more specific examples of the influence of high hydrogen concentration, we consider two organic molecule formation mechanisms, realizing that these are not the only prebiotic organic molecule formation mechanisms that are affected by the high hydrogen concentration. Both H2 and CO2 are uniformly mixed below the homopause (21, 22), where we have shown that the hydrogen mixing ratio could have been greater than 30%. Most photochemistry of interest occurs well below the homopause. Because the CO2 concentration is likely to be less than 30%, H2/C ratios in early Earth_s atmosphere could have been greater than 1 throughout the chemically interesting part of the atmosphere. The formation of cer-
Fig. 4. Hydrocarbon production rate increases rapidly as the hydrogen concentration increases, and exceeds the delivery of organics by interplanetary dust particles (IDPs) for H2 mixing ratios above 10%. pCO2 is the partial pressure of CO2 and fCH4 is the mixing ratio of CH4.
1011 1010 10 9 10 8 10 7 10 6 10 -3 10 -2 10 -1 H 2 mixing ratio 10 0 upper limit on IDP organic delivery 4 Gyr ago pCO2 = 0.01 bars fCH 4 = 0.001
tain prebiotic organic compounds in an atmosphere of CO2 or CO by electric discharge is almost as productive as that in an atmosphere of CH4, when H2/C Q 1 (7, 8). The conservative estimate of amino acid production rate by electric discharge is 107 kg/year when H2/C Q 1 (23). Although early Earth_s atmosphere might have been dominated by CO2 immediately after the heavy bombardment period, as continents formed on early Earth, the atmospheric CO2 concentration would decline because of weathering, and the H2/C ratio would become suitable for efficient formation of prebiotic organic compounds through electric discharge. Formation of prebiotic organic compounds by electric discharge at this conservative rate in a hydrogen-rich early Earth_s atmosphere would have created an ocean with a steady-state amino acid concentration 10j6 mol/liter (24), which is orders of magnitude greater than the amino acid concentration estimated for a hydrogenpoor early Earth_s atmosphere (25). This amino acid concentration is highly uncertain because neither the production rate nor the destruction rate is well known. In addition, organic films may have formed at the ocean surface, leading to higher concentrations of organic compounds than in the bulk sea water (26). Because the magnitude of energy deposition from electric discharge in the ancient atmosphere is poorly understood, it is difficult to predict the exact production rate of organic materials from these sources. Alternatively, organics can be formed through photolysis of methane by Lyman-a (Ly-a) photons with subsequent polymerization. The rate of photochemical haze production is critically dependent on the CH4/CO2 ratio (27). A similar dependence on the CH4/CO2 ratio has been found for hydrogen cyanide (HCN) formation in an atmosphere with N2, CO2, and CH4 (5). We used a one-dimensional photochemical model to study the organic production rate in a hydrogen-rich early Earth_s atmosphere (14). Figure 4 shows that the production rate of hydrocarbons is enhanced by about three orders of magnitude (from G107 kg/year to 1010 kg/year) when the hydrogen concentration in early Earth_s atmosphere changes from 10 3 to 30%. Hence, the atmospheric production rate of organics through UV photolysis would have been orders of magnitude greater than the rate of either the synthesis of organic compounds in hydrothermal systems or the exogenous delivery of organic compounds to early Earth (28). On the basis of our new model of hydrodynamic hydrogen escape, we conclude that diffusion-limited escape theory does not apply to a hydrogen-rich early Earth atmosphere. Rather, the escape of hydrogen was energy limited. Hydrogen mixing ratios greater than 30% could have been maintained in the atmosphere of prebiotic Earth without either invok-
1016
13 MAY 2005 VOL 308
Hydrocarbon Production (kg C/year)
SCIENCE
www.sciencemag.org
REPORTS
ing huge volcanic hydrogen outgassing rates or assuming a reduced mantle. The efficient production of organics in a hydrogen-rich early Earth_s atmosphere would have led to an organic soup in the oceans and ponds on the early Earth. The world ocean could have been the birthplace of life (14).
References and Notes
1. C. F. Chyba, C. Sagan, in Comets and the Origin and Evolution of Life (Springer, New York, 1997), pp. 147173. 2. S. L. Miller, Science 117, 528 (1953). 3. J. P. Pinto, G. R. Gladstone, Y. L. Yung, Science 210, 183 (1980). 4. P. H. Abelson, Proc. Natl. Acad. Sci. U.S.A. 55, 1365 (1966). 5. K. J. Zahnle, J. Geophys. Res. 91, 2819 (1986). 6. V. M. Canuto, J. S. Levine, T. R. Augustsson, C. L. Imhoff, M. S. Giampapa, Nature 305, 281 (1983). 7. G. Schlesinger, S. L. Miller, J. Mol. Evol. 19, 376 (1983). 8. R. Stribling, S. L. Miller, Orig. Life 17, 261 (1987). 9. Both CH4 and NH3 in the atmosphere of early Earth would have been subject to rapid loss driven by solar UV radiation. It is unlikely that the volcanic outgassing rate of CH4 or NH3 could have been adequate to maintain high concentrations of these gases. 10. The hydrogen concentration is determined by the balance between the volcanic outgassing rate of hydrogen and the escape of hydrogen to space. The modern volcanic hydrogen outgassing rate is 1.8 1010 hydrogen molecules cmj2 sj1 (29). Because of higher heat flow in the past, the overall outgassing rate of gases, and hydrogen in particular, might have been 5 times greater on ancient Earth (30). In the present Earths atmosphere, oxygen is dominant at the exobase level (defined as the boundary beyond which rapidly moving molecules may escape without collision). The current exospheric temperature is high (1000 to 2500 K) because of the efficient absorption of solar UV radiation by oxygen. If the exospheric temperature on the early Earth were as high as it is today, Jeans escape of hydrogen from the atmosphere would have been efficient, and the diffusion of hydrogen through the background gases to the homopause level would have been the limiting process. The rate of diffusion-limited escape can be expressed as F(H2) 0 2.5 1013 ftotal molecules cm2 s1, with ftotal defined as the total mixing ratio of hydrogen (in all chemical forms) at the homopause (31). By balancing the diffusion-limited escape rate of hydrogen with the hydrogen outgassing rate, the hydrogen mixing ratio up to the homopause in early Earths atmosphere should be of the order of 10j3 or below (21, 31), unless the oxidation state of Earths mantle was more reduced than its current oxidation state. But Earths mantle has been suggested to be in a similar oxidation state to that of today for the past 3.96 billion years (32). The common consensus among planetary scientists for the past 30 years has been that early Earths atmosphere had a low hydrogen concentration. 11. However, experiments to date generate only methane or formate in realistic hydrothermal-like systems (33). The exogenous flux of organic materials at about 4 billion years ago (Ga), primarily interplanetary dust particles (IDPs), may be less than 150 times the present value (34), although the interpretations of the Akilia rocks are debatable (35). 12. J. C. G. Walker, Orig. Life 16, 117 (1985). 13. I. de Pater, J. J. Lissauer, in Planetary Sciences (Cambridge Univ. Press, Cambridge, 2001), pp 6671. 14. Materials and methods are available as supporting material on Science Online. 15. H. Lammer et al., Icarus 165, 9 (2003). 16. D. M. Hunten, Planet. Space Sci. 30, 773 (1982). 17. S. Kumar, D. M. Hunten, J. B. Pollack, Icarus 55, 369 (1983). 18. Nonthermal escape from a hydrogen-rich early Earth has not been studied in detail. Although a similar upper limit of nonthermal hydrogen escape rate should apply to early Earth, it is important to note that nonthermal hydrogen escape processes may also be rather different for Earth than for Venus because of the presence of a strong magnetic field on Earth. Future work should include these processes in escape models to make more accurate estimates. J. D. Dorren, E. ...