This is to accompany my answer to How vulnerable are we to a catastrophic asteroid impact? Are there people whose job is to watch for them?


Asteroids as large as a hundred kilometers would be devastating to life. They would pretty much sterilize the entire planet, even boil the oceans away, but they can’t hit us at all. The larger craters on the Moon date from well over 3 billion years ago. Earth cleared the larger asteroids out of its orbit long ago, and the only ones left of that size are in the asteroid belt, as asteroids like Ceres, Vesta, Pallas etc. Perhaps it’s because we are protected from asteroids and comets as large as that from beyond Jupiter by Jupiter itself?

For details, see Debunked: Earth could be struck by a huge asteroid hundreds of kilometers across and my answer to If a comet is heading to earth, like the Armageddon scenario, what are the options for earth to survive?


Comets in the range 10 km to 100 km could cause mass extinctions, and the larger ones could be very hazardous, especially the long period ones that can hit Earth with higher velocities than asteroids or short period comets.

We know the orbits of the short period ones very well, and they are no problem for a long time into the future. Comet Halley is a good example. It is in a steeply inclined orbit, crossing the ecliptic close to Venus. Although it passes fairly close to Earth’s orbit, it can’t hit us in this orbit, and will most likely evaporate over the next 25,000 years. For more about this and other possible fates of Halley’s comet, see my answer to Can Halley's comet strike a planet in the solar system?

The most hazardous of these is Comet Swift Tuttle, diameter 27 km. If it hit Earth, it would be devastating, with 300 times the energy of impact of the asteroid that ended the dinosaur era. It misses us by a third of the distance to the Moon in 2126 and like Halley, though with much closer flybys of Earth, it is in an orbit which misses Earth every time. But it has a one in a million chance of getting perturbed into an orbit that can hit us in the year 4,479 CE. That means we are 99.9999% certain it will miss.

It’s not very likely and we have thousands of years to develop the technology to do something about it if needed. For more about this see my answer to What will be the effects if comet Hale-Bopp hits the earth?

Comet Swift Tuttle is the only large “dinosaur extinction” type object in the JPL Small-Body Database Search Engine to pass closer than one lunar distance before 2200. To search their database for yourself, go to this query page and click “Generate table” at the end. The MOID there is the minimum orbital distance from Earth’s orbit. It doesn’t mean that the object gets that close to Earth, as it depends on where Earth is in its orbit during the encounter. You’ll notice that Comet Swift Tuttle is the only object listed of 4 km diameter or larger that has an orbit that can take it within one lunar distance of Earth.

Techy note: The diameter of most asteroids and comets is estimated based on its brightness. If you don’t know whether an object is light or dark coloured, A darker object will seem smaller than it is and a brighter object will seem larger than it is. The objects could be several times larger than their listed diameter if they are very dark and if they haven’t had their albedo measured (a measure of how light or dark in colour they are).

The largest ones like comet Swift Tuttle have been carefully studied, and have their properties measured and you can be more confident about their listed diameters.

For details see the section on “uncertainties in size estimates” in the supplementary material..

The longer period comets are the big unknown here. Some come back only every few thousand years or even only after millions of years and they can only be seen when close to Earth, within a few years of the flyby even for the larger ones.

The chance of one of these hitting Earth is very remote, though the exact figure is hard to work out. According to one estimate, less than one in eight impact craters are from comets, and we also seem to be living in a time with few comet flybys. Only 1 in 155 of the NEOs are comets. I haven’t yet found a paper that attempts an estimate of the impact probability per century of a 10 km diameter long period comet. Do comment and say if you know of anything. However, it’s obviously far less than a one in a million chance per century as that was the estimated risk of a 10 km asteroid from crater counts, before they found them all


As an example, Comet Hale Bopp, a long period comet at 40 - 80 km in diameter is even more hazardous than Swift Tuttle, but it’s in an orbit that can’t hit Earth at present. A comet as large as this would sterilize the land and destroy just about all photosynthetic life in the sea too.

From records of previous impacts over three billion years ago, it seems that the air reaches temperatures of 500 C for weeks, the oceans boil for a year, and around 100 meters depth of the sea is boiled away. The deep sea remains just fine and there’d be life there to recolonize the land. Perhaps the land would be recolonized by the descendants of octopuses, and crabs, which have the ability to survive on land for a short while already.

I have tried to find an article on the frequency of Hale Bopp or Swift Tuttle sized impacts but can’t find anything. Do say if you know of anything in the comments?

Anyway, as a “back of the envelope” type first estimate, perhaps they happen at intervals of at least three billion years since that’s how long since the last one? That’s probably an over estimate of how often they happen, as back then Earth was still getting hit by more large objects than it is today, in the very tail end of the so called “Late Heavy Bombardment”.

That would make the risk per century as less than one in thirty million, and probably far less.

So, how does that fit with the existence of Hale Bopp? Well it’s in an orbit that will let Jupiter deflect it from time to time. Over that time, it can hit Jupiter or the Sun, or, since it’s made of ice, then it can gradually evaporate away during close approaches as will probably happen to comet Halley. Also, it’s likely to be disrupted by Jupiter into multiple smaller comets during close flybys, as happened to comet Shoemaker Levy before it finally hit Jupiter. It could also just break up into two or more comets as it evaporates away some of the ice holding it together.


With the risk from larger asteroids retired or soon to be retired, and the chance of a comet impact tiny, the attention has turned to the most likely asteroids of all, the ones of tens to hundreds of meters in diameter. The smaller asteroids are much more numerous.

These hit once per century or so. Much of Earth is uninhabited still, even on land, so they are most likely to hit the sea or an uninhabited area with no wider effects, just a crater or flattened trees, as for the Tunguska impact. This still could kill a few people if it strikes without warning. The Tunguska impact actually killed two reindeer herders, and there have been quite a few other recorded deaths by meteorite, small scale, only one person or a few people at a time in most cases. Records of deaths by what we now recognize to be meteorites may go back as far as a record from the Sui dynasty in China in 616 AD. See my answer to Has anyone died from being hit by an asteroid?

If they hit just off shore of a populated area, say Rio de Janiro, they could cause a smaller tsunami several meters high, which could kill millions if there is no warning. They could devastate a city and much of a small country if they hit on the land.

Asteroids this size could be deflected if there is enough warning. You just need to shift its trajectory by two centimeters per second to change a bulls eye hit a decade later into a miss (about the speed of a garden snail), and if it does a flyby in between, then there may be a “gravitational keyhole” of only a few hundred meters diameter it has to fly through to hit next time, which then can make it possible to deflect it with the gentlest of nudges of microns per second. If we have less warning, we can at least evacuate the impact zone.

Though the astronomers are doing a lot, there’s much more that could be done with funding that’s minor compared with the sums of money spent on defense. We could retire the threat of even the smaller asteroids almost completely with an outlay of around $50 million to send eight cubesats fitted with synthetic tracking telescopes, which could find more than 70% of the asteroids of 45 meters or larger in less than six years. For more on this see my How did we miss the Chelyabinsk asteroid?

Asteroid impacts are rare. Though more than 16,000 Near Earth Objects are currently being tracked, Earth is such a tiny target in interplanetary space that none of the ones so far discovered have a significant chance of hitting Earth before 2200. However, it’s the one natural disaster we can not only predict, to the minute, but also prevent too, given enough warning.


Asteroids much smaller than one kilometer in diameter will not have global effects. They are not large enough for mass extinctions or global climate effects, but would be catastrophic locally.

At one kilometer or larger they may have global effects. The main risk is to agriculture and an impact winter. The impact could put enough dust into the atmosphere to turn day to night, and there could be a year without crops. This would be survivable for countries that can stock pile enough food for a year.

They have found 881 asteroids of one kilometer diameter or larger, and find a new one every one to two months, with eight discovered so far this year. A recent estimate suggests there is a total population of 920±10 which makes the survey so far 95% complete with 29 - 49 still to find. They expect to have found at least 99% of the one kilometer asteroids by some time in the 2020s.

Of these 881 asteroids found so far, only a handful have orbits that could take them closer to Earth’s orbit than the Moon, and none will hit us by 2200.

As you can see, it’s highly unlikely that any of the 50 or so asteroids of one kilometer or larger that remain to be discovered will hit us before 2200. As I said in the answer, the figure of five there is not to be treated as an exact number because of the issues with actually working out the size of the asteroid from the data. But the number is going to be of that order. A handful of asteroids that large may come closer to us than the Moon by 2200. Whatever the number is, none of them have a significant chance of hittng Earth.

There are 16,524 Near Earth Objects in total. None of those are likely to hit Earth before 2200, though there are a few small ones that have a tiny chance of impact. Note, these numbers change frequently, as they find new asteroids. They found 998 new asteroids in the 229 day period from 1st Jan and 17th August 2017, so right now they are averaging around four new discoveries a day.

All the 10 km NEOs are already found as I said in the answer. That risk is retired. The one kilometer diameter risk is pretty much retired also.

As of 7th August 2017, we know of 106 comets that count as Near Earth Objects out of 16,512 in total, so one in 155 is a comet.


Summary: The risk from long period comets of ten kilometers or larger is tiny, much less than one in a million for the next century, though hard to estimate exactly. There’s a small risk from smaller long period comets but they are rarer than asteroids and short period comets.


We get hit by a 10 km asteroid roughly every hundred million years (table 2 of this paper), so the chance per century is about 1 in a million. But comets are very rare at present. I’ve tried to find an estimate for 10 km or larger comets. The nearest I found was:

So it’s hard to say, I’ve tried to find accurate estimates but can’t yet. Do say if you know of anything. At any rate it’s a lot less than the 1 in a million per century for asteroids of that size for sure.

Also, this is something that can’t hit us unawares. We’d also spot a comet that large many years in advance. That’s because of its coma. Even if a comet is headed straight for us, we’d spot the evaporating ices of its coma at a great distance.

Discovery photograph of Comet Holmes from 1892 near to the Andromeda Galaxy. If a comet approaches us head on, then you can still detect it easily by its extended coma which makes it look circular. Photograph from 1892 by Edwin Holmes who discovered it during regular observations of the Andromeda Galaxy.

There are occasional fast moving asteroids from the Oort cloud. Amongst the trillions of comets there, there are an estimated eight billion asteroids, for instance, flung outwards from the formation of the Moon. They could approach Earth without warning. But these are very rare indeed. According to Sky and Telescope, Andrew Shannon’s team estimates that the Large Synoptic Survey Telescope, which will be able to spot them could find perhaps a dozen Oort Cloud asteroids over a decade. For more on this: Eight Billion Asteroids in the Oort Cloud? - Sky & Telescope

So I don’t think we need to consider long period asteroids in a comet like orbit. They are just too small a proportion of the population.


There aren’t that many 10 km objects that get close to Earth. One of them is very famous though, Halley’s comet.

Halley’s comet can’t hit us in the near future, as its orbit intersects the plane of the ecliptic somewhere close to the orbit of Venus. Longer term, there are many possibilities, but one likely future is that it just evaporates away completely in the next 25,000 years as a result of its repeated close approaches to the Sun. See my answer to Can Halley's comet strike a planet in the solar system?

Comet 1P/Halley as taken xMarch 8, 1986 by W. Liller, Easter Island - it may evaporate away completely within 25,000 years as a result of the evaporation of its ices during its regular flybys of the inner solar system.

The long period comet Hale Bopp meanwhile is in an orbit that takes it nowhere near Earth either. So, though it gets within 20 lunar distances, it can’t hit us for thousands of years at least, though Earth would be devastated if it did.

This comet at 40 - 80 km diameter is large enough to sterilize our planet of all land based life and boil the oceans leaving only life in the deep sea to recolonize the land. However the chance of it hitting Earth is tiny as we haven’t had an impact this large for over three billion years, and nor has the Moon, Mercury, or Mars. So the chance is clearly pretty low of an impact like this in the long term future, and if there is a risk, it’s an issue for a future civilization many thousands of years from now. See my answer to What will be the effects if comet Hale-Bopp hits the earth?

Comet Hale-Bopp photographed on 4th April 1997 by E. Kolmhofer, H. Raab; Johannes-Kepler-Observatory

This could be devastating for Earth, large enough to sterilize it of all land life. But it’s not in an orbit that can hit us at present. We haven’t been hit by a comet this large for well over three billion years, back in a time when the Earth was hit by many more impacts than today. So, it’s a very unlikely scenario. Probably such large comets get broken up by Jupiter, or hit the sun or Jupiter before they can endanger Earth, as happened with Shoemaker Levy

The object that’s often mentioned as most dangerous for Earth in the very long term is Comet Swift Tuttle. We now know that it can’t get closer than 80,000 miles (130,000 km) for the next several centuries. But over thousands of years it can get closer. There’s a there is a one in a million chance that it will hit us in 4,479 AD. It’s not very likely and we have thousands of years to develop the technology to do something about it if needed. See my answer to What will happen if Comet Swift-Tuttle strikes the Earth in 2126?

Comet Swift Tuttle does close approaches to Earth - most recently in 1737, and 1862. The approaches below the shaded area are visible from Earth. The next such are in 2126, 2261 and 2392. However it can’t hit Earth on its current orbit as we now know, after working it out more exactly. There’s a chance its orbit as it slowly evolves can hit Earth 4479 years from now but it is only a one in a million chance. It would be devastating too, not as bad as Hale Bopp but far worse than the dinosaur era asteroid impact. But we have more than 2000 years to develop the technology to do something about it in the remote case that we need to.

Even large asteroids and comets can be deflected given a timeline of decades or more to do something about it as then, only a tiny nudge to change its velocity will be plenty to change an impact into a miss decades later.


The much smaller Siding Spring 400 - 700 meter comet is a bit like a dress rehearsal for a comet in a close flyby of Earth. This comet did a close flyby of Mars was spotted on 3 January 2013, a year and nine months before its flyby on 19 October 2014. It seemed to have a tiny chance of hitting Mars when first discovered, but as they refined its orbit, they found it would miss. Since Mars is a very small target, that’s what you’d expect.

It eventually did a very close flyby of Mars at a distance of about 36% of the distance to our Moon.

This is a very long period comet with an orbital period of millions of years.

It would pan out similarly for Earth. Though we’ve never found a long period comet with a possibility of impacting Earth, if we found a small comet headed our way, we’d know of this many years in advance, and to start with we might not know if it would hit, but as astronomers refined their calculations, the most likely thing we find is always that it misses because Earth is such a tiny target compared to interplanetary space.


For smaller asteroids, by far the most likely, then the main effects are right away.

For instance, a 50 meter diameter comet or asteroid exploding in an air burst or a 200 meter one leading to a crater would be most likely to have almost no effect. Land in the sea or a remote uninhabited area, at most a few people killed.

But - if a 50 meter asteroid were to cause an airburst over London, it could kill 2.82 million people. A 200 meter crater forming impact could kill 8.76 million people, according to this study, page 23

An impact of a 200 meter asteroid angle 45°, impact speed 20 km/s and asteroid density 3100 kg/m3 could cause between 7.6 million casualties right next to the city, 2.35 million at a distance of 10 km and scaling down to 11 thousand at a distance of 300 km according to the same study. The tsunami is only 4.3 meters high in their model, and it doesn’t matter much how far out the impact is so long as there’s a slowly sloping continental shelf as impacts in deeper water create higher waves, but then they get attenuated by distance as they approach the shore.

So there’s no doubt that a 200 meter asteroid can be devastating and even a 50 meter airburst, not unlike the Tunguska impact, could kill millions of people in the remote chance that it hits a populated region.

The 50 meter impactor would hit on average every 850 years and the 200 meter impactor on average only once in 40,000 years. Most of the deaths are from wind and from the thermal effects - burns etc.

For larger asteroids, the most recent paper I found to discuss global effects is rather old, from 2003 and a similar table online by a different author.

So anyway it will give a first idea. The table is here: Environmental Damage from Asteroid and Comet Impacts

In this image I’ve deleted most of the rows to focus on the essentials for agriculture:

This paper from 2004 puts the average interval between impacts of a 1 km object as 600,000 years. At any rate they are very rare.

As you see, a one kilometer asteroid impact has some global crop failures. At 2 kilometers there are global drops of temperature by 8 C for weeks and plant growth disrupted for years. At 5 kilometers, it’s so dark as a result of all the dust that photosynthesis stops for months. At 10 kilometers then day becomes night for months, and freezing conditions away from coastlines continue for years.

However the 10 kilometers are of no practical concern at present. The one kilometers upwards are very unlikely and will probably be ruled out by the 2020s. Luckily!

Longer term, it would be possible to survive any of these. Even with global darkness for months and freezing for years. I think the authors who are pessimistic about human survival haven’t really taken account of what we are capable of as a species.

The authors Seth Baum et al also say that species of plants could be preserved through seed banks, as for animals, then it’s not so easy but you would need only 100 individuals per species so you could preserve many species by using only a small fraction of the amount of food produced to feed billions of humans during the disaster.

This is of mainly theoretical interest however I think. We almost certainly have anything from centuries to hundreds of thousands of years or most likely millions of years to prepare for the more devastating 5 to 10 km asteroids.

That’s plenty of time to have a complete survey of our solar system and to deflect them centuries in advance. So long as we continue with space technology over those timescales then surely we will be able to prevent even the largest impacts, deflect them. A civilization over those timescales would discover long period comets with orbits of thousands of years long before impact, and maybe deflect them during an earlier fly through of the solar system. Or we may have fusion power, settlements in the Kuiper belt, one way or another if we keep up a technological presence in space, I can’t imagine it being impossible to prevent large asteroid and comet impacts between centuries and millions of years from now.


Lexell's Comet passed at only 6 lunar distances on July 1st 1770. It was a Jupiter family comet and has never been seen again. It had an orbital period of 5.58 years at the time, but only temporarily. It was put into that orbit only a few years earlier in 1776. Probably it had another encounter with Jupiter which either ejected it from the solar system or put it into such a distant orbit that it can’t be seen easily even with our modern telescopes.

So, it must be in a distant orbit now, or has left our solar system and is no threat to Earth.

Comet IRAS-Araki-Alcock (9.2 km in diameter) is a long period comet with an orbital period of around 970 years. It did a close flyby at only 12.14 lunar distances in 1983 and is due back some time around 2,953.

55P/Tempel–Tuttle (26 km diameter) is the parent body of the Leonids meteor shower with an orbital period of 33 years. It passed at a distance of 8.9 lunar distances on 26th October 1366.

P/2016 BA14 (PANSTARRS) a fragment of 252P/LINEAR only 60 - 200 meters in diameter, did a flyby at 9.2 LD in October 2016

252P/LINEAR is 100 - 400 meters in diameter did a flyby at a distance of 13.9 in October 2016

Apart from that, there are the Marsden group comets which are sun grazers studied by SOHO as they pass the sun. They could approach us undetected from the direction of the sun, but are small and harmless.

The tiny comet 1999 J6 found by SOHO was perhaps only 35 meters in diameter spotted as a Marsden family sun grazer. It probably passed just 1.3 million miles from Earth a month later. Even if it had collided with us, as a comet, it would probably have just burned up harmlessly in the upper atmosphere.

“a Marsden comet that reached perihelion in May of 1999, C/1999 J6, likely passed just 1.3 million miles from Earth a month later. It is believed to be the same object that returned in November 2004 as C/2004 V9, which allowed Brian Marsden to calculate its orbit with unusual accuracy for a SOHO object. Even had the comet collided with us, it probably would have burned up harmlessly in the upper atmosphere. There is no evidence of any Marsden comets large enough to do us any damage. The daytime Arietid meteor shower, which peaks around June 10, are believed to be related to Marsden comets, Comet 96P/Machholz, and the Kracht group. “

A SOHO and Sungrazing Comet FAQ

There’s a NEO table of comet close approaches before 1900


These larger objects are very rare, indeed we’ve already talked about most of them.

Here is the table of objects from 4 km upwards that pass within 50 times the distance to the moon (LD) from the JPL small body search engine. There often are uncertainties in the diameter - more on that later. But the larger ones tend to be more thoroughly studied. It’s sorted by the minimum distance from Earth (MOID).

To generate this table go here, and then click Generate Table at the bottom of the page.

Note, the MOID is the minimum distance from the orbit of the object to the orbit of Earth. It doesn’t mean that there are any approaches that close before 2200.

Of the objects that have minimum orbital distance less than 10 lunar diameters, only comet Swift Tuttle and comet Araki-Alcock are close to 10 km in size. Then there are the comets Halley and Hale Bopp that we’ve already looked at, which pass at greater distances.

The ones left that are either larger than 10 km or approaching it in size are:

NB to convert AUs to Lunar Distances, just multiply by 389.17794


This is a technical detail. So long as it misses us, it doesn’t really matter if it is a ten kilometer asteroid or a 1 km asteroid. But in those figures I gave, for instance of five one kilometer asteroids that come closer to Earth than the Moon by 2200, then the estimate of the diameter is often very approximate.

It all depends on whether we have a measured albedo. This is a measure of how bright or dark its surface is. The Moon is very dark, albedo 0.12, similar to worn asphalt. It looks so bright because we see it in the night sky lit up by the very bright sun.

For many asteroids, especially with a short observation arc, all we have to go on is the observed brightness of the asteroid. This also is often only known to within perhaps half a magnitude or so because the amateur astronomers who make the observations may not have had the ability to estimate it better than that with their equipment.

So, we have to work out its size just from its brightness.


Brightness is measured using the scale of magnitudes. This is a bit unintuitive if you aren’t used to it. The lower the magnitude, the brighter it is. Sirius, Venus at its brightest, the Moon and the Sun have negative magnitudes, as do Canopus, Alpha Centauri and Arcturus. See Brightest Stars: Luminosity & Magnitude

Asteroids vary in brightness in our skies depending on how far they are from the sun and how far they are from us. So we use the scale of absolute magnitudes. For stars, this is how bright it would appear at a distance of 32.6 light years, or 10 parsecs. But that is not very practical for asteroids.

For asteroids the absolute magnitude is the brightness it would have if seen from Earth at a distance of one AU, with the asteroid also at a distance of 1 AU from the sun. See The H and G magnitude system for asteroids for details.


If an asteroid is darker than expected, then we will estimate it to be smaller than it really is, If brighter, it will seem larger than it really is.

Unfortunately for easy size estimates, the albedos of asteroids vary hugely. The albedo ranges from 0, reflects no light, to 1 meaning it is 100% reflective -all the light scattered, brighter even than white snow.

Examples of albedos:

So anyway asteroids and comets as it turns out have a very wide range of albedos. Some have albedoes as low as 0.01 or lower, so darker than charcoal. Others have albedos all the way up almost to 1, brighter than fresh snow.

The ones at about the samE distance from the sun as Earth tend not to be quite as bright as that because ice is not stable on their surface. But ones with eccentric orbits that come as close as Earth occasionally could easily have higher albedos.


Summary - there are two peaks in the population at albedo 0.03 and 0.186. The rare outliers in the population extend all the way down at least to 0.01 up to at least 0.65. The average albedo is 0.14.


This table is of observations of albedos of asteroids by the WISE Infrared space telescope, sadly it only shows the semi-major axis (half the diameter of the elliptical orbit, with the sun at one focus) so you can’t deduce from this whether the asteroid or comet can cross Earth’s orbit.

Asteroid albedos: graphs of data - albedo versus semimajor axis

An asteroid or comet in an elliptical orbit could have semimajor axis significantly larger than 1 au and still hit Earth, for instance a Jupiter family comet. It could then be very bright or very dark.

Earth crossers tend to have low albedos not surprisingly because any comets that get close to the sun will lose their icy coverings and darken. The extremes there reading from the graph for 1 au are about 0.01 and 0.65. Those are rather isolated dots, especially the one at 0.65, so it seems that 0.65 is very bright for a NEA and 0.01 is very dark.

That’s just a first impression because many asteroids of course are Mars and Jupiter crossers and will have semimajor axis much more than 1 au (a Jupiter crosser, even if a sun grazer, has to have a semimajor axis of at least 2.5 au or so). However any bright comet that comes as close to the sun as 1 au is likely to lose its ices pretty quickly and darken

So anyway what about detailed studies of the population of NEAs? The average albedo is 0.14 according to this paper.

Then an analysis of the WISE data published in 2016 found evidence of two distinct populations there, after analysing 428 asteroids in the Near Earth Asteroids population.

About a quarter have an albedo peaking at 0.03 while the other three quarters (more precisely, 74.7%) peak at 0.168. See “The Albedo Distribution of Near Earth Asteroids”.

This is their best fit curve:

Albedo shown horizontally, figure 1 from this paper, this is not the data itself but a best fit curve to it, with two peaks at 0.03 and 0.186 with three quarters of the population (74.7% to be precise) in the population that peaks at 0.186

It’s a very good fit too. This graph plots the best fit double peak curve in black along with the raw data in red. It’s plotting it in a somewhat different way, so the double peak isn’t obvious but notice how closely the red and black curves match.

It’s a cummulative graph, for instance, you see from the blue line that 80% of the population they measured had an albedo less than 0.27 (just reading from the graph). Notice how well the black curve fits the red data.

Another study from 2011, using data from 447 NEOs studied using the Spitzer space telescope in the ExploreNEOs program goes more detail exploring the NEAs by taxonomic classes. They find evidence of many populations with albedos from the

For details see this paper

Perhaps a bit surprisingly, asteroids of just about all spectral types are found in the NEO population. There are quite a few of those very dark D type asteroids for instance, though not nearly as many as of the other types. So though those asteroids are rare, we can’t rule them out.

Table from this paper. For details of what the letters mean, see Asteroid spectral types


The Center for Near Earth Object Studies has a nifty tool for calculating diameters given an albedo and magnitude which is useful for seeing the influence of albedo on predicted diameter.

Asteroid Size Estimator

So, let’s try out the albedos from those two peaks in “The Albedo Distribution of Near Earth Asteroids”, and use 0.14 for the average albedo. Plugging in their albedos of 0.03 and 0.186, we get

Their graph tails off at 0.6, but the WISE data shows one dot at 0.65 with semimajor axis 1, so definitely an Earth crosser or close to it at albedo of 0.65. If that’s not a mistake, perhaps we should take that as the upper limit instead of 0.6, while keeping in mind that such high numbers are very rare.

So, plugging in those extremes of 0.01 and 0.65 into the Asteroid Size Estimator again:

Until we have good albedo measurements then wecan’t rule out those extremes.

Many asteroids of course do have good albedo measurements already - after all WISE and NEOExplore each found the albedos for over 400 asteroids But there are over 16,500 of them in total and for many of them we won’t know the albedo yet.


First there’s observer error -they may not fully understand how to calibrate and make good magnitude measurements, with an uncertainty of 0.25 to 0.5 magnitudes, especially for a short observation arcs. It’s not such a problem once you have a lot of measurements.

What’s the effect of a 0.5 magnitude error?

A 10 magnitudes change in the absolute magnitude leads to a multiplication of the diameter by 10.

So a 0.5 error would multiply or divide the result by 10^(0.5/10) leading to an error either way of about +12% or -11%.

So this is a relatively minor effect, but still could change the classification of an asteroid, for instance from 1 km to 890 meters, so that it no longer counts as kilometer scale.

Then as well the estimate of magnitude depends on your orbital calculation as you need the distance to convert the measured magnitudes into absolute magnitudes. So a slight change in orbit and so of its distance from Earth during the observations could affect the magnitude - again more so with a short observation arc.

The Sentry: Earth Impact Monitoring database is based on calculations by the Minor Planet center.

The discovery statistics pages uses data from the MPC (Minor Planet Center)’s tables of Atens, Apollos, and Amors:

The JPL Small-Body Database Search Engine uses their own internally generated orbits. Also it uses their own estimates of the absolute magnitude of the asteroid which uses a sightly different method based also on slightly different orbits. They also differ in their classification of Aten / Apollo asteroids.x

So for instance, if you search their database for NEOs, with diameter at least 1 km, you get 326 objects using their own internal database (click Generate Table at the bottom of this page to see the list).

(clipped to the first few rows)

If you go to their discovery statistics by size table.

then there are 881 of those 1 km objects. This page uses the Minor Planet Center database.

I didn’t know what to make of this until I contacted them to ask about it, and then got this explanation of how the tables work.

It’s not too surprising though, once you realize how much the diameter estimate depends on assumptions of albedo, observer error, and that they use different values for the absolute magnitude in both databases.

Most of those 1 km or larger asteroids will be close to 1 kilometer in size because there are far more of the smaller asteroids than larger asteroids at all size scales. So most of them will be close to the border line between >= 1 km and < 1 km so, combined with the uncertainties in the measurement of size, it’s no surprise that with a small change in the way they treat magnitudes, that over 550 of those 881 asteroids are shifted over the boundary to sub 1 km in the CNEOS database.

The CNEOS database has 186 of those 326 asteroids between 1 and 2 km. As before click Generate Table at the bottom of this page. If the proportions are the same for the minor planet center database, that would be 503 of those 881 asteroids between 1 and 2 km. The figures of the two databases differ by 555 asteroids.

I think this shows rather vividly how hard it is to estimate the diameter of an asteroid to enough precision to decide even whether it is 1 or 2 km in diameter. But it makes no difference to the calculations of whether they miss Earth, and since they are all predicted to miss us at least until 2200, this is mainly of technical interest.

For asteroids that do have their albedo measured directly or indirectly, then they can work out an accurate diameter, and even more so for the ones they are able to image via radar as they pass Earth.

These details from Alan Chamberlin (private communication) with much thanks!


Which of these two results is our best guess for the number of one kilometer NEOs? 326 or 881?

Alan Chamberlin says that they trust the Minor Planet Centers estimates of magnitude more than their own. These use the MPC’s unpublished algorithm. So his advice is that we should go with the Sentry table rather than the JPL search engine for the statistics here.

So our best guess at present is that there are 881 objects of 1 km or larger that get closer to Earth than about 20 lunar diameters (0.05 au) with maybe 30 - 50 or so left to find if we go by the recent estimate which suggests there is a total population of 920±10

However both approaches are using approximations.

I think it follows from this that we shouldn’t be surprised if that number 881 changes when eventually we know the diameters of them through accurate albedo measurements and radar measurements of their diameters during close flybys, and any other methods (e.g. close up observation by spacecraft).

This number 881 could go up or down, just through better estimates of the diameters of asteroids and comets, even after we complete the survey of all one kilometer asteroids and short period comets.