Oxygen escape from the Earth during geomagnetic reversals: Implications to mass extinction
Yong Weia, Zuyin Pub, Qiugang Zongb, Weixing Wana, Zhipeng Rena, Markus Fraenzc, Eduard Dubininc, Feng Tiand, Quanqi Shif, Suiyan Fub, Minghua Honga
Earth and Planetary Science Letters
Volume 394, 15 May 2014, Pages 94–98
Geomagnetic field reversal substantially weakens the protection for the atmosphere.
Solar wind energizes more oxygen ions to escape when geomagnetic field is weakened.
Oxygen escape may explain the drop of atmospheric level during mass extinction.
The causal relation between reversal and mass extinction should be “many-to-one”.
The simulated oxygen escape rate based on knowledge of Mars support our hypothesis.
The evolution of life is affected by variations of atmospheric oxygen level and geomagnetic field intensity. Oxygen can escape into interplanetary space as ions after gaining momentum from solar wind, but Earth’s strong dipole field reduces the momentum transfer efficiency and the ion outflow rate, except for the time of geomagnetic polarity reversals when the field is significantly weakened in strength and becomes Mars-like in morphology. The newest databases available for the Phanerozoic era illustrate that the reversal rate increased and the atmospheric oxygen level decreased when the marine diversity showed a gradual pattern of mass extinctions lasting millions of years. We propose that accumulated oxygen escape during an interval of increased reversal rate could have led to the catastrophic drop of oxygen level, which is known to be a cause of mass extinction. We simulated the oxygen ion escape rate for the Triassic–Jurassic event, using a modified Martian ion escape model with an input of quiet solar wind inferred from Sun-like stars. The results show that geomagnetic reversal could enhance the oxygen escape rate by 3–4 orders only if the magnetic field was extremely weak, even without consideration of space weather effects. This suggests that our hypothesis could be a possible explanation of a correlation between geomagnetic reversals and mass extinction. Therefore, if this causal relation indeed exists, it should be a “many-to-one” scenario rather the previously considered “one-to-one”, and planetary magnetic field should be much more important than previously thought for planetary habitability.
Keywords: atmospheric ion escape; geomagnetic reversal; mass extinction
In the past half a century, many efforts have been devoted to seeking the links between geomagnetic reversals and mass extinctions, but no consensus has been reached (Glassmeier and Vogt, 2010). An important test of any successful mechanism is to explain the patterns of mass extinctions. Fossil records reveal that a mass extinction has a gradual pattern persisting for millions of years, during which a stepwise pattern is manifested by a series of impulsive extinctions (Jin et al., 2000). These patterns suggest that the main cause should be continual environmental degradation. The drop of atmospheric O2 level has been verified to be able to induce environmental degradation because reducing the supply of O2 is lethal for most species (Huey and Ward, 2005). However, it is difficult to explain the O2 level drops if only considering Earth-bounded geochemical processes (Berner, 2005). An alternative possibility is that the O2 molecules are dissociated into oxygen atoms by solar radiation and subsequently ionized; these O+ ions are further energized in the ionosphere by the solar wind forcing and can thus overcome the containment by the magnetosphere and gravity, finally escape into the interplanetary space. Since the magnetospheric containment is expected to be severely weakened when the dipole collapses during geomagnetic reversals, we hypothesize that geomagnetic reversals cause O2 level drops, and the subsequent mass extinctions. To test this hypothesis, in this paper we will examine the newest databases and simulate O+ escape during reversals.
Fig. 1A shows the newest reversal rate data during the Phanerozoic era (solid line, Ogg et al., 2008). Because of several data gaps in this database, the relative reversal rate from an older database (dashed line, McElhinny, 1971) is also plotted to show the trend of reversal rate for a reference. The reversal rate, atmospheric O2 level (Fig. 1B) (Berner, 2009) and Marine diversity (Fig. 1C) (Alroy, 2010) show a strong correlation in support of our hypothesis. During the 2nd, 3rd and 4th mass extinctions identified by the diversity drops (red), the reversal rates increased and the O2 level decreased. On the contrary, when the reversal rate remained at zero or very low, namely during the superchrons (blue) (Merrill and Mcfadden, 1999), the diversity increased, and the O2 level also increased for 3 out of 4 superchrons. However, during 1st and 5th mass extinction, the reversal rate also increased despite that the O2 level just remained at low level or had no discernable change. These features suggest that some mass extinctions might be explained by our hypothesis: increasing geomagnetic reversals continually enhance oxygen escape, and this cumulative effect could cause a significant drop of O2 level over a few millions years. As a result, the global hypoxia might gradually kill numerous species.
Temporal evolution of reversal rate, O2 level and marine diversity over the Phanerozoic era. (A) Geomagnetic reversal rate. The solid line is total reversals within a 10-Myr bin from the new database (Ogg et al., 2008). The dashed line is the relative reversal rate to represent the trend of reversal rate from an older database (McElhinny, 1971). The blue blocks show the superchrons. KRS: Kiaman Reversed Superchron (267–313 Ma). CNS: Cretaceous Normal Superchron (83–125 Ma). MRS: Mayero Reversed Superchron (463–481 Ma). (B) Modeled percentage and amount of atmospheric O2 over time (Berner, 2009). (C) Number of marine genera (Alroy, 2010). The red blocks show the gradual pattern of 5 well-known mass extinctions, and the 6th mass extinction has not been confirmed. The blue blocks show the superchrons. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
2. Explanation and simulation
Planetary ion escape is universal in our solar system throughout its history (Lammer, 2008 and Lundin et al., 2007 and Moore and Horwitz, 2007). Many spacecraft observations and theories have revealed that ion escape from terrestrial planets is mainly driven by solar wind dynamic pressure, but the efficiency of this process highly depends on the intensity and morphology of the planetary intrinsic magnetic field (Lundin et al., 2007 and Moore and Horwitz, 2007). Earth’s present dipole field interacts with solar wind and forms an intrinsic magnetosphere, and its outer boundary, the magnetopause usually extends to 10 RE (Earth’s radius) at the subsolar point (Fig. 2B). An Earth-like intrinsic magnetic field can effectively prevent the planetary ionosphere from directly interacting with solar wind and thus constrains ion escape (Moore and Horwitz, 2007 and Seki et al., 2001). Mars and Venus have no global intrinsic field, thus their ionospheres directly interact with solar wind and form induced magnetospheres. However, an induced magnetosphere is much smaller in size; for example, Martian magnetopause only extends to 1.2 RM (Mars’ radius) at the subsolar point (Dubinin et al., 2006). Direct comparison of spacecraft observations has illustrated that a Mars-like ion escape is much more efficient in removing planetary O+ ions (Wei et al., 2012). No doubt, Earth will also have a Mars-like ion escape (Lundin et al., 2007) when Earth’s dynamo is substantially weakened (Fig. 2C). Collapse of the dipole during its polarity reversal (Merrill and Mcfadden, 1999) reduces the size of the magnetosphere, thus also causes a transition from the Earth-like ion escape to the Mars-like.
Ion escape scenarios for the reversing field, present field and no field of Earth. (A) Mars-like ion escape in a quadrupole-dominated magnetosphere for reversing field. This is the view from duskside with two neutral points on the dusk-dawn line. The red region illustrates the O+ dominated ionosphere, and the four color belts on the Earth denote possible auroral zones. BS: bow shock. AW: auroral wind. MP: magnetopause. MT: momentum transfer. (B) Earth-like ion escape for present field (Moore and Horwitz, 2007): Present Earth’s magnetosphere has a Dungey cycle up to 120 RE tail ward, which can return most of outflowing O+. DNL: distant neutral line. (C) Fully Mars-like ion escape when the geodynamo totally stops. The magnetosphere is created by the interaction of solar wind with the upper ionosphere, where the crustal field complicates the interaction. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
To quantitatively estimate the O+ escape rate during reversals, we first make an estimate of the evolution of solar wind dynamic pressure (PSW) over Phanerozoic era. Studies of solar analogs suggest that the stellar wind dynamic pressure can be represented as a function of the star’s age or X-ray flux (Wood et al., 2002). By using this method the solar wind pressure 600 million years ago (Ma) was 1.3–1.5 times larger than its present value PSW0. Details are given in supplementary material, see Appendix A. If we use X-ray flux, we consider a Sun-like star, 18 Scopii (HD 146233), which is identical to the Sun but ∼300 million years (Myr) or more younger than the Sun (Wright et al., 2004). The stellar wind dynamic pressure of 18 Scopii may represent the upper limit of the Phanerozoic PSW because its age overlaps the early Phanerozoic era. This method gives an upper limit of PSW as 3–3.7 times larger than PSW0. Details are given in supplementary material, see Appendix A. The averaged PSW0 over one solar cycle is observed as 2–8 nPa, without consideration of space weather effects (Richardson and Wang, 1999). Therefore, the PSW over Phanerozoic era was about 3–30 nPa.
The intensity and morphology of a reversing field are still in debate due to lack of global-scale ancient field records, and they are also probably different from one reversal to another (Merrill and Mcfadden, 1999). Simulations suggest that the total intensity severely decreases; meanwhile, the quadrupole component dominates over the dipole but also decreases by almost one order of magnitude in intensity (Glatzmaier and Roberts, 1995). Following the method proposed by Clement (2004), we construct a reversing field through modifying the present field by reducing the dipole term to zero, multiplying the quadrupole term by 10% and keeping the other terms unchanged. Considering the balance between the Phanerozoic PSW and Earth’s magnetic pressure, the averaged position of the magnetopause varies between 1.3 and 1.8 RE, here we take 1.5 RE to simplify the calculations. Details are given in supplementary material, see Appendix A. Such a small magnetosphere is morphologically similar to the induced field surrounding Mars, especially concerning the altitude of the magnetopause. Therefore, the ion escape during reversals should be highly Mars-like.
Fig. 2A illustrates the ion escape in the conceptual quadrupole magnetosphere during reversals. The Dungey cycle (Fig. 2B) that returns a substantial fraction of O+ outflow back to Earth (Nilsson et al., 2013; Moore and Horwitz, 2007 and Seki et al., 2001) does not exist anymore, thus we expect that a major fraction of the O+ outflow finally escapes. In analogy to the fully Mars-like scenario (Fig. 2C), the ionosphere directly supplies O+ ions to the near-magnetopause region, where they gain momentum from solar wind and thus are accelerated beyond gravitational escape speed (Lundin et al., 2007 and Dubinin et al., 2006). Moreover, the auroral wind (Moore and Horwitz, 2007) can also deliver O+ ions to higher altitude by indirectly absorbing solar wind kinetic energy via magnetic reconnection between the magnetic field frozen-in solar wind and the magnetospheric field. The auroral wind might be highly important during space weather events when denser and faster solar wind impacts Earth. The escaping O+ ions are usually observed over a wide energy range, but most of them are cold and low-energy ions, as observed by many spacecraft for both Earth-like (André and Cully, 2012) and Mars-like escape (Lundin et al., 2008). Here we simply assume that all O+ ions which have a speed above 10 km/s (the gravitational escape speed near the magnetopause) do escape.
To test the hypothesis, we consider the Triassic–Jurassic event (4th mass extinction) for example. There were 88 reversals within 30 Myr (Ogg et al., 2008), and the O2 level decreased from 23% to 14% with a total O2 loss of 15×10/18 mol (Berner, 2009). Taking an averaged reversing duration of 104 yr (Clement, 2004) for each reversal …
In other words, if geomagnetic field reversals could explain the drop of the O2 level during the Triassic–Jurassic event, the O+ escape rate should increase to 3 orders of magnitude greater than the present outflow rate of up to 6×1026 #/s (Yau and Andre, 1997).
Since we have estimated the PSW and the O+ escaping speed, the escape rate can be evaluated with a model (Lundin et al., 2007 and Dubinin et al., 2006). The model considers the transfer of momentum from solar wind to ionospheric O+ ions, and has been successfully applied for the Mars-like ion escape on both Mars and Venus (Lundin et al., 2007 and Dubinin et al., 2006). Therefore, the efficiency of momentum transfer is a critical parameter. The efficiency has been determined as a function of the O+ density in the momentum transfer region by spacecraft observations on Mars (Perez-de-Tejada, 1998). … According to the relation between the efficiency of momentum transfer and the O+ density determined from observations at Mars, the efficiencies at different O2 level can be estimated. With those parameters derived above, the O+ escape rate for the reversals during the Triassic–Jurassic event can be simulated. More details are given in supplementary material, see Appendix A.
3. Results and discussion
The results of our simulations (Fig. 3) suggest that O+ escape driven by a somewhat enhanced solar wind could explain at least 50% of the O2 drop during the 4th mass extinction. Thus our hypothesis is a possible explanation. The predicted escape rate is much higher than that on present Earth (Moore and Horwitz, 2007), Mars (Lundin et al., 2008) and Venus (Nordström et al., 2013), but comparable with the water loss rate on comet Halley at Earth orbit (Krankowsky et al., 1986), and far less than that on comet Hale–Bopp at Earth orbit (Russo, 2000). The upper limit of O+ escape rate should be the global production rate of O+, which is estimated at about 5×10/30 #/s during solar maximum based on the International Reference Ionosphere (IRI) model (Bilitza et al., 2012). Our escape rate is also below this upper limit. In addition, from Fig. 3 one may note that Earth’s present outflow rate is comparable to that of Mars and Venus, and this recently led to a debate on whether a planetary magnetic field can protect a planetary atmosphere (e.g., Barabash, 2010). Wei et al. (2012) have presented observational evidence to support the protective role of the planetary magnetic field. Here our simulations also give an important clue: the Earth’s outflow rate should be much higher than Venus and Mars if the Earth’s magnetic field is significantly weakened, and the difference is mainly due to the fact that Earth has a higher O+ density in the interaction region between solar wind and ionosphere.
Simulated O+ outflow rate under averaged solar wind for O2 level at 21% and 14%, respectively. The red regions represent Phanerozoic PSW, while the blue segment represents present PSW. The grey bars show the O+ escape rates of two big comets at the Earth orbit and those of Mars and Venus. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
The uncertainties in this estimate are mainly introduced by poor knowledge of the solar wind dynamic pressure and the intensity of the reversing field in ancient times. However, two other factors, which have not been considered in our simulations, may significantly increase the O+ escape rate. One is the occurrence of space weather events. The “corotating interaction region” only impacts Mars for 15% of the total time, but can drive at least 33% of the total O+ escape (Dubinin et al., 2006). The “coronal mass ejection” can even cause 3–7 times higher O+ loss rates on Mars during high solar activity years than during low solar activity years (Luhmann et al., 2007). Some Sun-like stars younger than our Sun have been observed to be more active (Telleschi et al., 2005), and thus the Sun during Phanerozoic time should produce more frequent space weather events. The other factor is excursion, also called “aborted reversal” (Merrill and Mcfadden, 1999). It occurs more frequently but with shorter duration (several thousand years) than reversals, and could also cause a transition from Earth-like O+ escape to Mars-like.
There are at least two methods to test our hypothesis. Because ion escape favors the lighter isotopes (Lammer, 2008), severe O+ escape could have caused an increase of the O18/O16 ratio. The other is the stepwise pattern (Jin et al., 2000), of which each step should correspond to one or a series of geomagnetic reversals.
Recently, it has been hotly debated whether we are in the mid of the 6th mass extinction (Barnosky et al., 2011). As shown in Fig. 1, over the last 40 Myr, the reversal rate was increasing and the diversity was decreasing. This implicate that present Earth might be experiencing a new gradual pattern of mass extinction. However, the slight changes of O2 level cannot be regarded as a strong link between them. More data and work are required to understand these variations.