Will a Nearby Supernova Endanger Life on Earth? Michael Richmond [last revised Dec 5, 2009] --------------------------------------------------- Since this topic seems to come up every year or so, I decided to try to work out some of the dangers quantitatively. Let me list the various sources of danger I've considered. 1. optical and near-optical light 2. X-rays from the explosion itself 3. X-rays from the supernova remnant 4. gamma rays* 5. neutrinos 6. energetic particles* I haven't been able to find much information on those items marked with "*", but I'll tell you what I know. My current best guess is that items 2 and/or 4, energetic photons, are the most dangerous to those nearby. I've tried to estimate the amount of X-ray or gamma-ray radiation which poses a threat to unshielded humans in Appendix A. What dose of high-energy radiation is lethal? I use several units of distance in this text which may be unfamiliar to some readers. They may find a very terse explanation in Appendix B. Units of distance I've listed references to other sources of information on the effects of supernovae on the biosphere near the end of the text. --------------------------------------------------------- 1. Optical and near-optical light After a very brief (but significant? I don't think we know enough to say for sure) "flash", the optical output of SNe rises over a period of several weeks, peaks for a few days (types Ia, Ib) to a few months (type IIP), then fades relatively slowly. The absolute magnitudes of of supernovae at peak, like everything else, vary according to the astronomer who answers the question :-), but rough values are distance at which object would Object peak M(V) have same apparent V mag as Sun --------------------------------------------------------------------- Sun +4.8 1 AU type Ia -19.0 58,000 AU = 0.3 pc type II >= -18.5 <=46,000 AU = 0.2 pc Clearly, SNe must be very close to affect the Earth via optical photons. 2. X-rays from the explosion I could find data for X-rays from type II SNe only, but I'll make estimates for Ia as well. Satellites detected <= 80x10^(-12) erg/(cm^2-s) in the 6-28 keV range from SN 1987A (which was probably less luminous than most). Let me guess that, over the entire X-ray range, at peak, about 800x10^(-12) erg/(cm^2-s) would be observed from SN 1987A. The Sun, on the other hand, during a large flare, emits around 0.35 erg/(cm^2-s) in the same X-ray band. In order to produce the same X-ray flux as a large solar flare, then, SN 1987A would have to be closer by the sqrt(0.35/[800x10^(-12)]), or at a distance of ~2 pc. Another type II SN which was detected in X-rays was SN 1993J, in M81. The ASCA satellite measured a flux of about 1x10^(-11) ergs/(cm^2-s) in the range 1-10 keV. Using the same value as above for the X-ray flux of the Sun during a large flare, we find that the SN would have to be moved closer by a factor of 190,000 to equal the solar flare; that would correspond to a distance of about 20 pc (from its actual distance of 3.6 Mpc). The Type IIP supernova 1979C has remained bright at X-ray (and optical) wavelengths for several decades. Observations by the XMM satellite in 2001 at energies 0.3 - 2 keV reveal an X-ray luminosity of roughly 0.8 x 10^(39) erg/s, which is about the same as the X-ray luminosity measured six years earlier by ROSAT. We believe that the prolonged X-ray emission is due to shock waves running through rich circumstellar material. While this X-ray luminosity is lower than that of some other Type II SNe, it suggests that in some cases, X-ray emission could continue for many years. The Type Ic supernova 1998bw was observed by the BeppoSAX satellite. It had a peak X-ray flux of about 4.6 counts per thousand seconds in the BeppoSAX MECS instrument; using a distance of 43 Mpc to the supernova, this corresponds to an X-ray luminosity of about 5x10^(40) erg/s in the range 2 to 10 keV. The X-ray flux decreased by a factor of 3 over a six-month period. The Type IIc supernova 2003bg was observed with the Chandra X-ray Observatory; the low-energy (0.5 to 10.0 keV) flux indicates a luminosity of about 4x10^(39) erg/s. This event would provide roughly the X-ray flux of a big solar flare if it were placed 10 pc away. SN 2001em was an example of a type Ib/Ic supernova. We believe these are very massive stars which gently and gradually shed their outer layers of (mostly) hydrogen via strong stellar winds, leaving behind a layer of (mostly) helium, for type Ib, or (mostly) heavier elements, for type Ic. When the star runs out of fuel at its center, the core collapses and then explodes, just like a Type II supernova. Back to 2001em: in 2004, some 2.5 years after the explosion, the Chandra X-ray satellite measured X-rays in the range from 0.5-to-8.0 keV which indicate a luminosity of 10^(41) erg/s for an object at 80 Mpc. That makes 2001em one of the most luminous X-ray emitters among supernovae. This supernova could produce as much X-ray flux as a strong solar flare at a large distance, something like 50 pc away from us. SN 2008D was another type Ib/c supernova. It was caught in the act of exploding by the SWIFT satellite; that is, the radiation emitted as the shockwave of the explosion reached the photosphere of the star was detected. Soderberg et al. (Nature 453, 469, 2008) estimate that the peak X-ray luminosity during the brief (400 seconds) breakout phase was about 6 x 10^(43) erg/s, much more luminous than the late-time X-ray emission. Note that it doesn't last very long. This supernova could produce as much X-ray flux as a strong solar flare at a distance of around 1200 pc from the Earth. The authors did not detect any gamma rays from this breakout, however. Models of type Ia SNe (Shigeyama et al., A&AS 97, 223 [1993]) predict X-ray luminosities of around 10^42 erg/sec at peak. This is about 10-100 times as luminous as SN 1993J, and so a type Ia SN could be 3-10 times farther away (60-200 pc) and still equal a solar flare. I suspect that X-rays (and gamma-rays, see below) are the most deadly of a nearby SN's effects. See Appendix A for a discussion of the amount of X-ray radiation needed to kill a human being. 3. X-rays from the supernova remnant As material in the ejecta slams into the surrounding ISM, it produces shock waves that can heat material up to millions of degrees and produce X-rays. For example, the SNR Cas A, at a distance of about 3kpc, has a flux of something like 5x10^(-15) erg/(cm^2-s) according to X-ray satellites (I had to assume a LOT about the size and efficiency of the instruments here, so I'm probably way off, but it won't matter). Comparing again with the flux from a solar flare, we find that Cas A would have to be located less than 0.001 pc from the Earth to produce the same flux. Now, this flux would be very long-lasting, but even so, is clearly less important than the X-rays produced in the explosion. 4. Gamma-rays from the explosion (much based on words of wisdom from David Palmer - thanks, Dave!) One way to estimate the effect of SNe in gamma-rays is to compare the amount of power they produce in gamma-rays ALONE with that from the Sun at ALL wavelengths: distance at which power power is equal to Sun's total Sun 10^33 erg/s (all wavelengths) 1 AU SN II 10^39 erg/s (gamma + X-rays) 1200 AU ~ 0.006 pc SN Ia 2x10^41 erg/s (gamma + X-rays) 17000 AU ~ 0.08 pc In somewhat more detail: both X-rays and Gamma-rays from SN 1987A were due to the decay of radionucleides, primarily Co56 from the Ni56->Co56->Fe56 decay chain. Gamma rays, primarily at 0.847 and 1.238 MeV, were downgraded by Compton scattering in the envelope (keeping the envelope hot and luminous) and then emerged at lower energies in the X-ray and gamma-ray range. The unscattered photons at 0.847 and 1.238 were also seen. In greater detail: observations of the flux in X-rays and gamma rays from the Sun reveal that most of the energy is in the X-rays, with relatively small fractions in the gamma-ray regime. Using data from Colhane et al. (Solar Physics 153, 307 [1994]), Baoz et al. (Solar Physics 153, 33 [1994]) and McConnell et al. (Adv. Space Res., v 13, n 9, 245 [1993]), I find Energy from Sun during flares Satellite energy range duration total power (erg/cm^2) ----------------------------------------------------------------------- Yohkoh 20-? keV 10 sec 350 COMPTEL (GRO) 1-10 MeV 900 sec 0.01 GAMMA-1 > 30 MeV 600 sec 0.0002 By comparison, the flux in the 847 and 1238 keV lines due to the decay of Ni-56 in a type Ia SN is estimated to be (data from Ruiz-Lapuente et al., ApJ 417, 547 [1993]) at a distance of 1000 pc SN Ia ~1 MeV ~60 days ~40,000 This is significant -- a type Ia SN, at the distance of 1000 pc, dumps as much gamma-ray radiation onto the earth as 1,000 solar flares. Even when the Sun is at the peak of its activity cycle, I don't think it flares ten times a day, so, even at a kiloparsec, a type Ia SN would outshine the Sun in gamma rays. However, while I _do_ know that we easily survive even the greatest solar flares, I don't know how a large increase in the gamma-ray flux over a period of several months would affect the earth's atmosphere. Steve Thorsett, in a preprint "Terrestial Implications of Cosmological Gamma-Ray Burst Models," quotes sources which suggest that considerably more than 100,000 erg/(cm^2) in gamma-rays are needed to destroy the ozone layer, so it seems that a type Ia would have to be closer than 1 kpc to cause significant damage. 5. Neutrinos The neutrino flux from SN 1987A was about 5x10^10 cm^(-2) in a burst a few seconds, which is similar to that from the Sun (6.5x10^10/(cm^2-s))! Calculation due to Robert W. Spiker, U. of Virginia: > This was part of a question on my PhD qual exams two years ago. The > answer I got was about an AU. Here's how I did it: > Total energy released in a SN in neutrinos is E_\nu \sim 10^53 ergs. > Cross section for interaction is \sigma = 10^{-44} cm^2. > Minimum lethal dose is 1000 rads = 10^5 ergs of energy absorbed per > gram of body weight = 8 x 10^9 ergs absorbed for a body weight of 80 > kg in order to die a horrible death. (This number I was given.) > Energy absorbed = Energy passing thru * cross section * path length * > number density of absorbing body > I figure the typical body presents 1 square meter of area and has a > path length of 30 cm (so we can look at the pretty star as it blows). > Number density of the body I chose to be 1 g cm^{-3} / 6 m_H; that is, > density equal to water (we float) and mean molecular weight of about 6 > (mostly H but lots of C, N, and O I figured). > The energy passing thru = "flux" * area = 10^53 ergs / (4 \pi d^2) * 1 m^2 > so the distance d needed to absorb a lethal dose is > d^2 = (E_\nu A \sigma n l) / (4 \pi E_{lethal}) > which if you plug in as I did comes to 1 AU. I think it's safe to say that the neutrino dangers are small compared to others. However, perhaps the simple calculations above are missing some important points. A paper appeared on the Astrophysics Preprint Server (May, 1995) that claims neutrinos can indeed be dangerous at larger distances. The paper is called "Biological Effects of Stellar Collapse Neutrinos", is written by J. I. Collar (University of South Carolina), and was submitted to Phys. Rev. Lett. Here's the abstract: Massive stars in their final stages of collapse radiate most of their binding energy in the form of MeV neutrinos. The recoil atoms that they produce in elastic scattering off nuclei in organic tissue create a radiation damage which is highly effective in the production of irreparable DNA harm, leading to cellular mutation, neoplasia and oncogenesis. Using a conventional model of the galaxy and of the collapse mechanism, the periodicity of nearby stellar collapses and the radiation dose are calculated. The possible contribution of this process to the paleontological record of mass extinctions is examined. You can find more information by looking up astro-ph/9505028 at http://xxx.lanl.gov/archive/astro-ph. 6. Cosmic-ray particles This possiblilty is one that may be important, but I just don't know enough to calculate HOW important. Here's what I could find: The solar wind, at the distance of the Earth, has a density of about 9 protons/cm^3, velocity ~470 km/s (and the particles have a temperature around 10^5K). Let me take as a measure of the impact on the Earth's atmosphere the product of density and (velocity*velocity): 2x10^16 in cgs units (protons/cm-s^2). Now, let me consider the material in the expanding shell of ejecta from a type II SN; assume a total mass of 5M(solar), an expansion velocity of 5,000 km/s and a shell thickness of 0.01 times its radius (I'll bet that the real thickness is greater, but this increases the impact). Then, assuming that the shell expands uniformly and ignoring all the material swept up in its path (which really IS significant over scales of >~ 1pc), I find time since shell radius explosion proton density density*velocity*velocity -------------------------------------------------------------------- 1 pc 200 yr 30 cm^3 7x10^18 10 pc 2000 yr 0.03 cm^3 7x10^15 So, this vastly over-simplified model predicts that the ejecta material will be comparable to the solar wind at a distance of a few parsecs. Again, I have no idea how much stronger the "SN wind" must be than the solar wind for it to pose a danger. However, I've left out the issue of the energy of the particles. It has been hypothesized that SN remnants are sites of cosmic-ray acceleration, which could produce a smaller population of MUCH more energetic particles than in the typical ejecta shell. Those very-high-energy particles could have a significant impact despite their small numbers. Since I know zero about acceleration mechanisms, or the effect of the energy of cosmic-ray particles on their interaction with the Earth's atmosphere, I'll just stop here. Conclusion: I suspect that a type II explosion must be within a few parsecs of the Earth, certainly less than 10 pc (33 light years), to pose a danger to life on Earth. I suspect that a type Ia explosion, due to the larger amount of high-energy radiation, could be several times farther away. My guess is that the X-ray and gamma-ray radiation are the most important at large distances. Additional reading: Larry Marschall provides a general overview of supernovae and their properties in "The Supernova Story." "Biological Effects of Stellar Collapse Neutrinos", by J. I. Collar (University of South Carolina), submitted to Phys. Rev. Lett. in 1995; You can find more information by looking up astro-ph/9505028 at http://xxx.lanl.gov/archive/astro-ph Brian Fields and John Ellis comment on a recent (1999) claim of the detection of radioactive material which _might_ have come from a nearby supernova only a few million years ago in "On Deep-Ocean Fe-60 as a Fossil of a Near-Earth Supernova". You can read the paper by looking up astro-ph/9811457 at http://xxx.lanl.gov/archive/astro-ph, or find less technical discussion at http://www.astro.uiuc.edu/~bdfields/NearbySN.html Steve Thorsett considers the possibilities that a local gamma-ray burst source (at the center of our own Milky Way galaxy) might affect adversely life on earth ... _IF_ the model of the burst is correct, See his paper in ApJ Letters, vol. 444, page L53 (1995). Neil Gehrels et al. calculate the effect of gamma rays from a nearby supernova on the ozone layer of the Earth. The preprint is http://xxx.lanl.gov/abs/astro-ph/0211361, and the paper is published in vol 585 of Astrophysical Journal. They conclude that a Type Ia SN would have to be within about 8 parcsecs of the Earth to destroy half the ozone layer, and therefore have a significant effect on the biosphere. You can read a press release at http://www.sciencedaily.com//releases/2003/01/030122072843.htm There have been in 2005 a burst of papers considering the effects of Gamma-Ray Bursts (GRBs) on the Earth. + Cosmic Rays from Gamma Ray Bursts in the Galaxy, Dermer and Holmes http://xxx.lanl.gov/abs/astro-ph/0504158 + Terrestrial Ozone Depletion Due to a Milky Way Gamma-Ray Burst, Thomas et al., ApJ 622, 153 (2005) http://arxiv.org/abs/astro-ph/0411284 ---------------------------------------------------------- Appendix A. What dose of high-energy radiation is lethal? Just how much high-energy radiation does it take to kill a human being? I'm not sure, but let me show one might take the appropriate factors into account. First, we measure the exposure and absorption of high-energy radiation by living creatures in units called "rems", "rads," and "grays." For simplicity, we consider only the total amount of energy to which a creature is exposed, ignoring the spectral shape of the radiation (that is, we lump together X-rays and gamma-rays of all energies). The units are: 1 gray = 1 joule of energy per kilogram of body mass = 10,000 ergs per gram 1 rad = 0.01 gray = 0.01 joule of energy per kilogram of body mass = 100 ergs per gram 1 rem = 1 rad * Q, Q = "quality factor" Here, the "qualify factor" is somewhat vague; I believe the intent is to account for a number of factors which can affect the amount of energy absorbed by the creature, and the biological effects thereof. For example, it might include factors for - protection by clothing - orientation of body to direction of the radiation - selective absorption by various biological organs or structures I'll assume that the "quality factor" has a value of order unity, although I suspect that in real life situations, Q may be closer to 0.1 than to 1.0. There have been many studies of the effects of high-energy radiation on humans and other animals. I grabbed the following table of effects of short-term exposure to high-energy radiation from http://pooh.chem.wm.edu/chemWWW/courses/chem105/projects/group5/page12.html REM Effect -------------------------------------------- 0-25 No detectable effects 25-100 Temporary decrease in white blood cell count 100-200 Nausea, vomiting, longer-term decrease in white blood cells 200-300 Vomiting, diarrhea, loss of appetite, listlessness 300-600 Vomiting, diarrhea, hemorrhaging, eventual death in some cases Above 600 Eventual death in almost all cases As for long-term exposure to radiation, the US Nuclear Regulatory Commision declared in January, 1994, that 5 rem/year is an acceptable dose for people. Now, let's calculate the amount of radiation absorbed by a human being. I'll approximate the cross-sectional area of a human by a rectangle 1.5 meters high and 0.4 meters wide; thus human cross-section area A = 0.6 m^2 = 6,000 cm^2 I'll also estimate a typical human body mass of human body mass M = 50 kg = 50,000 g We are now ready to calculate the effect of some source of high-energy radiation on an unshielded human being. Since we are concerned with astronomical sources, the ionizing radiation from which is blocked by the Earth's atmosphere, we will imagine a human being in space -- perhaps some astronaut in a spaceship on a voyage to Mars. Let's neglect the protection provided by the walls of his spaceship. (In the calculations below, I use "parcecs" as a measure of distance 1 parsec = 3.1 x 10^18 centimeters = 3.3 light years) Suppose a supernova explodes at some distance D pc from the solar system, and has an intrinsic total luminosity in X-rays and gamma-rays of L ergs/sec. Then the astronaut suffers a dose area L A dose = flux * ------- * Q = -------- * --- * Q mass 4*pi*D^2 M L (erg) 6000 (cm^2) = ------------- ---------- * ------ ------ * Q 1.2x10^38 D^2 (cm^2-sec) 50,000 (g) L = 1 x 10^(-39) ----- * Q (rem/sec) D^2 As discussed in section 2 above, a supernova might produce L = 10^42 erg/sec in X-rays at its peak (and perhaps a similar amount in gamma rays, as shown in section 4). Moreover, let's assume that Q = 1, to a very rough degree. Then the dose suffered by the astronaut every second is 1,000 dose = ----- (rem/sec) D^2 So, we can make a table showing the time it takes for her to accumulate a lethal dose of radiation, assuming that 300 rems will kill her: distance to SN time to lethal dose ----------------- ------------------------- 1 pc < 1 sec 10 pc 30 sec 100 pc 3,000 sec = 1 hour 1,000 pc = 1 kpc 300,000 sec = 3.5 day 10,000* pc = 10 kpc 30,000,000 sec = 1 year* So, any supernova within a kiloparsec presents an immediate threat to unshielded humans. However, more distant events aren't very dangerous, because their high-energy luminosity starts to drop after a few days, well before they impart a lethal dose to astronauts. For that reason, the final row of the table above has an asterisk (*). Following the measurements of the type Ib/c SN 2008, we can estimate the danger from the prompt X-ray emission due to the shock breakout. The peak luminosity in X-rays was about 6 x 10^(43) erg/sec, and the breakout lasted about 400 seconds, so that's a total of about 2.6 x 10^(46) ergs. Plugging into the formulae above, we find that this prompt X-ray dose would reach the lethal level (300 rems) if the SN was about 300 pc or closer. ---------------------------------------------------------- Appendix B. Units of distance The astronomical community often uses two base units, one designed to measure distances within our own solar system, the other designed to measure distances to other stars and to other galaxies. The "Astronomical Unit" (or AU) is approximately the distance from the Earth to the Sun. It's very useful for discussing situations within our own Solar System, or in other stellar systems. 1 AU = 150 million kilometers = 1.5 x 10^8 km = 1.5 x 10^11 m The "parsec" (or pc for short) is the distance at which a ruler 1 AU long would appear to subtend one second of arc -- that is, it would subtend 1/3600 degrees. 1 pc = 3.1 x 10^16 m = 206,265 AU = 3.26 light years We often use metric prefixes with parsecs to describe longer distances: 1 kpc = 1,000 pc = 3.1 x 10^19 m 1 Mpc = 1,000,000 pc = 3.1 x 10^22 m