Ray P. Norris,

CSIRO Australia Telescope National Facility,

PO Box 76, Epping, NSW1710, Australia



This paper considers the factors that determine the probable age of a civilisation that might be detected in a SETI search. Simple stellar evolution considerations suggest an age of a few Gyr. Supernovae and Gamma-ray-bursters could in principle shorten the lifetime of a civilisation, but the fact that life on Earth has survived for at least 4 Gyr places a severe constraint on such factors. If a civilisation is detected as a result of a SETI search, it is likely to be of order 1 Gyr more advanced than us.


When we conduct searches for extra-terrestrial intelligence, we often make implicit assumptions about the age of the civilisation that we are trying to find. For example, our strategy for searching for a life-form of a similar age to us is likely to be different from that for a civilisation billions of years more advanced than us. Similarly, in the event of a confirmed detection, the way in which we plan our response will also depend on how advanced that civilisation may be. In this paper, I estimate the likely age of the civilisation that we are most likely to detect, should we be successful in our searches.

The two key factors that determine how old a detected civilisation is likely to be are (a) the length of time since intelligent life first appeared in our Galaxy and (b) the median lifetime of a civilisation. The second of these is more problematic, since the development of a civilisation can be cut short by a wide range of events, including disease, war, global mismanagement, asteroids, supernovae, and gamma-ray bursters. We should also acknowledge the possible existence of other hazards, of which we are not yet aware. For example, the devastating effect of gamma-ray busters has only been appreciated in the last 2-3 years, and there are probably other phenomena yet to be discovered. Events such as disease, war, and global mismanagement are almost impossible to quantify, and so in this paper I concentrate on those events that we can quantify: asteroids, supernovae, and gamma-ray bursters. But in the first section of this paper, I consider what the maximum lifetime of a planetary-bound civilisation might be.

Throughout this paper, I make a very conservative assumption that an extraterrestrial civilisation (ET) resembles us in most significant respects (other than age and evolution). In other words, ET lives on a planet orbiting a solar-type star, and has taken as long after the formation of their star to evolve to "civilisation" as we have, which is ~5 Gyr (Gigayears, or billion years). I therefore estimate the longevity of ET by looking at the hazards that confront the Earth.



I assume that stars like our Sun have been forming since the formation of the Galaxy some 10 Gyr ago. Observed changes in metallicity since then are not sufficient to alter this simple assumption significantly. Our Sun is now about 5 Gyr old, and has an expected total lifetime of 10 Gyr.

For the first 5 Gyr of the life of the Galaxy, there would not have been enough time for a civilisation to develop, and so ET did not exist. Between 5 and 10 Gyr, assuming a constant rate of star formation, the number of civilisations would increase linearly until the present day. At around the present time, some of those first solar-type stars will be dying at the same rate as others are forming, and so, assuming their civilisations die at the same rate as they do, the number of civilisations is then level from now on.

The median age of a civilisation is therefore the median age of those civilisations that started between 5 and 0 Gyr ago, which is 1.7 Gyr. Therefore, in the absence of other factors, any civilisation that we detect via SETI is likely to be 1.7 Gyr more advanced than we are.


A supernova results from the explosion of a high-mass star after its hydrogen and helium fuels are used up, at the end of its lifetime. A supernova exploding within 50 ly of the Earth will have a catastrophic effect. The 1040 J of energy produced in the first few days bathes the earth in a total amount of ionisation some 300 times greater than the annual amount of ionisation from cosmic rays. Surprisingly, little of this radiation reaches Earth. Instead, Most of it ionises atmospheric nitrogen, which reacts with oxygen to form nitrous oxide, which in turn reacts with ozone3. The effect will be to reduce the amount of ozone in the Earth's atmosphere by about 95%, resulting in a level of UV on the Earth's surface some four orders of magnitude greater than normal, which continues for a period of 2 years. This will certainly result in almost 100% mortality of small organisms and most plants. The effect on mammals is not clear, and some might survive. However this 2-year period is followed by a longer (80 years) period of bombardment by the cosmic rays from the supernova, which have similar, although slightly reduced, effects. It is difficult to see how anything other than an advanced civilisation could survive such an extended holocaust.

A supernova such as this goes off in our galaxy roughly every 5 years, and we expect one within 50 ly (light-years) of the earth roughly once every 5 Myr. We expect one even closer (within 10 ly) every 200 Myr. Therefore all life would be expected to be destroyed at this interval. Clearly this has not happened, since we are still here, and I will return to possible reasons in a later section.


Gamma-ray bursters (GRB) are a recently discovered phenomenon, in which some 1045 J of energy are released in a few seconds. The ones that have been observed on earth appear to be distributed uniformly across the observable Universe. Their power is such that we are able to detect GRB right up to the edge of the observable universe. The mechanism is still not known, but is likely to involve the merging of two neutron stars, possibly resulting in the formation of a black hole.

A GRB is some 5 orders of magnitude more energetic than a supernova, and could occur even at the Galactic centre,
25 000 ly away from us, and have a similar effect as a supernova within 50 ly. However, in this case there is an even more deadly effect, in that, should a GRB go off in the Galactic centre, the immediate blast of ionising radiation is followed by an intense blast of cosmic rays lasting perhaps a few weeks4. These cosmic rays will initiate a shower of relativistic muons in the Earth's atmosphere, causing a radiation level on the surface of the earth some 100 times greater than the lethal dose for a human being. The muons are so energetic that they would even penetrate nuclear air-raid shelters to a depth of perhaps hundreds of metres2.

We expect such a GRB roughly once every 200 Myr, and it would almost certainly result in the extinction of all life on earth other than that deep in the ocean. Again, clearly this has not happened, since we are here.


The geological and biological record shows a series of mass extinctions of life on Earth. The most famous is that at the Cretaceous-Tertiary (KT) boundary, which was almost certainly caused by an asteroid hitting the earth about 65 Myr ago. The KT mass extinction wiped out the dinosaurs, and paved the way for the emergence of mammals as the dominant species on Earth.

Less well known are a series of similar, and in some cases even more extreme, mass extinctions every few tens of Myr, and many smaller extinctions, the last of which was only 11000 yr ago. The cause of most of these is unknown. It is likely that a range of causes including asteroids, distant supernovae, and climatic changes are responsible for them.

All these mass extinctions are on a much smaller scale than the catastrophic events we expect from a nearby supernova or a gamma-ray burst in the Galactic centre. In each of these cases, a number of species (sometimes as many as 50%) were extinguished, but a sufficient range of diversity remained for the biota to recover in a relatively short time.


I have identified two causes that should wipe out essentially all life on Earth roughly every 200 Myr, and yet we are here. Two possible explanations are:

  • The calculation of either the timescales or the severity of the effects is erroneous, or
  • We have been very lucky!

In the first case, simply multiplying the timescale by a factor of a few is insufficient. We have been evolving for at least 4 Gyr, and so the interval between catastrophes must be at least 4 Gyr for us to survive so far. Presumably the precise interval will vary randomly around this figure, and so any surviving civilisation can look forward to a lifetime of between zero and a few Gyr. In this case, if we detect ET, then ET will have a median age of perhaps 1 or 2 Gyr, which is similar to the 1.7 Gyr derived from simple stellar evolution arguments. Thus, in this case, the supernovae and GRBs have not significantly changed the median age of ET.

In the second case, we have already survived for some 20 times the mean interval between catastrophes, which is very lucky indeed. Whilst it is not possible to quantify this without more detailed knowledge of the frequency distribution of supernovae and GRB, it is likely that the probability is so low that we are alone in the Galaxy. Apart from providing a solution to the Fermi paradox1, this implies that the median lifetime of ET is meaningless, as we will never detect ET!


Conventional models imply that supernovae and gamma-ray-bursters will extinguish life on planets at intervals of about 200 Myr. Since this has not happened on Earth, either these conventional models are wrong, or else life on Earth is probably unique in the Galaxy. The first case predicts a median age of ET as being of the order of 1 billion years. The second case predicts that we will never detect ET. Thus, if we do detect ET, the median age is of order 1 billion years. Note that, in this case, the probability of ET being less than one million years older than us is less than 1 part in 1000.

Therefore, any successful SETI detection will have detected a civilisation almost certainly at least a million years older than ours, and more probably of order a billion years older.


1. Annis, J., 1999, JBIS, 52, 19.

2. Leonard, P.J.T., & Bonnell, J.T, 1998, Sky & Telescope, 95, 28.

3. Rudermann, M.A., 1974, Science, 184, 1079.

4.Thorsett, S.E., ApJ, 444, L53.

Staff space