Asteroid impact prediction
Asteroid impact prediction is the prediction of the dates and times of asteroids impacting Earth, along with the locations and severities of the impacts.
The process of impact prediction follows three major steps:
- Discovery of an asteroid and initial assessment of its orbit which is generally based on a short observation arc of less than 2 weeks.
- Follow up observations to improve the orbit determination
- Calculating if, when and where the orbit may intersect with Earth at some point in the future.
Most asteroids are discovered by a camera on a telescope with a wide field of view. Image differencing software compares a recent photograph with earlier ones of the same part of the sky, detecting objects that have moved, brightened, or appeared. Those systems usually obtain a few observations per night which can be linked up into a very preliminary orbit determination. This predicts approximate positions over the next few nights and follow up can then be carried out by any telescope powerful enough to see the newly detected object. Orbit intersection calculations are then carried out by two independent systems, one run by NASA and the other by ESA.
Current systems only detect an arriving object when several factors are just right, mainly the direction of approach relative to the Sun, the weather, and phase of the Moon. The result is a low rate of success, especially for smaller asteroids. A few near misses by medium-size asteroids have been predicted, years in advance, with a tiny chance of actually striking Earth. A handful of actual impactors have successfully been detected hours in advance, but all of those were small, struck wilderness or ocean, and hurt nobody. Unfortunately the majority of impacts remain unpredicted and occasionally hit a populated area, resulting in widespread damage and injuries. Performance is improving as existing systems are upgraded and new ones come on line, but the blind spot issue which all current systems face around the Sun can only be overcome by a dedicated space based system.
History
In 1992 a report to NASA recommended a coordinated survey to discover, verify and provide follow-up observations for Earth-crossing asteroids. This survey was scaled to discover 90% of all objects larger than one kilometer within 25 years. Three years later, a further NASA report recommended search surveys that would discover 60–70% of the short-period, near-Earth objects larger than one kilometer within ten years and obtain 90% completeness within five more years.In 1998, NASA formally embraced the goal of finding and cataloging, by 2008, 90% of all near-Earth objects with diameters of 1 km or larger that could represent a collision risk to Earth. The 1 km diameter metric was chosen after considerable study indicated that an impact of an object smaller than 1 km could cause significant local or regional damage but is unlikely to cause a worldwide catastrophe. The impact of an object much larger than 1 km diameter could well result in worldwide damage up to, and potentially including, extinction of the human race. The NASA commitment has resulted in the funding of a number of NEO search efforts, which made considerable progress toward the 90% goal by the target date of 2008 and also produced the first ever successful prediction of an asteroid impact. However the 2009 discovery of several NEOs approximately 2 to 3 kilometers in diameter demonstrated there were still large objects to be detected.
Three years later, in 2012, the small asteroid 367943 Duende was discovered and successfully predicted to be on close but non-colliding approach to Earth again just 11 months later. This was a landmark prediction as the object was only, and it was closely monitored as a result. On the day of its closest approach and by coincidence, a smaller asteroid was also approaching Earth, unpredicted and undetected, from a direction close to the Sun. Unlike 367943 Duende it was on a collision course and it impacted Earth 16 hours before 367943 Duende passed, becoming the Chelyabinsk meteor. It injured 1,500 people and damaged over 7,000 buildings, raising the profile of the dangers of even small asteroid impacts if they occur over populated areas. The asteroid is estimated to have been 17 m across.
In April 2018, the B612 Foundation stated "It's 100 per cent certain we'll be hit , but we're not 100 per cent sure when." Also in 2018, physicist Stephen Hawking, in his final book Brief Answers to the Big Questions, considered an asteroid collision to be the biggest threat to the planet. In June 2018, the US National Science and Technology Council warned that America is unprepared for an asteroid impact event, and has developed and released the "" to better prepare.
Discovery of Near Earth Asteroids
The first step in predicting impacts is detecting asteroids and determining their orbits. Finding faint Near-Earth objects against the background stars is very much a search. It is achieved by sky surveys that are designed to discover near Earth asteroids. Unlike the majority of telescopes that have a narrow field of view and high magnification, survey telescopes have a wide field of view to scan the entire sky in a reasonable amount of time with enough sensitivity to pick up the faint Near-Earth objects they are searching for.NEO focused surveys revisit the same area of sky several times in succession. Movement can then be detected using image differencing techniques. Anything that moves from image to image against the background of stars is compared to a catalogue of all known objects, and if it is not already known is reported as a new discovery along with its precise position and the observation time. This then allows other observers to confirm and add to the data about the newly discovered object.
Cataloging vs Warning
The existing asteroid surveys have a fairly clear-cut division between 'cataloging surveys' which use larger telescopes to mostly identify larger asteroids well before they come very close to Earth, and 'warning surveys' which use smaller telescopes to mostly look for smaller asteroids on their final approach. Cataloging systems focus on finding larger asteroids years in advance and they scan the sky slowly, but deeply. Warning systems focus on scanning the sky relatively quickly. They typically cannot detect objects that are as faint as cataloging systems but will not miss an asteroid that brightens for just a few days when it passes very close to Earth. Some systems compromise and scan the sky say once per week.Cataloging Systems
For larger asteroids, prediction is based on cataloging the asteroid, years to centuries before it could impact. This technique is possible as they can be seen from a long distance due to their large size. Their orbits therefore can be measured and any future impacts predicted long before they are on their final approach to Earth. This long period of warning is important as an impact from a 1 km object would cause worldwide damage and a long lead time would be needed to deflect it away from Earth. As of 2018, the inventory is nearly complete for the kilometer-size objects which would cause global damage, and approximately one third complete for 140 meter objects which would cause major regional damage. The effectiveness of the catalogue is somewhat limited by the fact that some proportion of the objects have been lost since their discovery. Their orbits were not determined sufficiently well to allow their future position to be predicted with accuracy, so we no longer know what part of their orbit that they are on.Warning Systems
Smaller near-Earth objects are far more numerous. They therefore impact Earth much more often, though with much less damage, and the vast majority remain undiscovered. They seldom pass close enough to Earth on a previous approach that they become bright enough to observe, and so most can only be observed on final approach. They therefore cannot usually be cataloged well in advance and can only be warned about, a few weeks to days in advance. This is much too late to deflect them away from Earth, but is enough time to mitigate the consequences of the impact by evacuating and otherwise preparing the affected area. Warning systems can also detect asteroids which have been successfully catalogued as existing, but whose orbit is insufficiently well determined to allow a prediction of where they are now.Current mechanisms for detecting asteroids on final approach rely on ground based telescopes with wide fields of view. Those currently can monitor the sky at most every second night, and therefore miss most of the smaller asteroids which are bright enough to detect for less than two days. Such very small asteroids much more commonly impact Earth than larger ones, but they make little damage. Missing them therefore has limited consequences. Much more importantly, ground-based telescopes are blind to most of the asteroids which impact the day side of the planet and would miss even large ones. These and other problems mean very few impacts are successfully predicted.
Surveys
The main NEO focussed surveys are listed below, along with future telescopes that are already funded. They have a fairly clear-cut division between 'cataloging surveys' and 'warning surveys'. The existing warning surveys have enough capacity between them to scan the northern sky once per clear night. However, they are concentrated in a relatively small part of the planet and therefore miss some asteroids that come close to Earth while the Sun is up on that part of Earth. Two surveys are in Hawaii, which means they see the same parts of the sky at the same time of day, and are affected by similar weather. Two others are located in the southwestern United States and so suffer from similar overlap. These surveys do complement each other to an extent in that some are cataloging surveys and some are warning surveys. However, the resulting coverage across the globe is imperfect. In particular, there are currently no major surveys in the Southern Hemisphere. This coverage limitation is most relevant for warning surveys, since cataloging surveys also have opportunities to detect the same asteroids when their orbit brings them to the Northern sky.This clustering together of the sky surveys in the Northern hemisphere means that around 15% of the sky at extreme Southern declination is never monitored, and that the rest of the Southern sky is observed over a shorter season than the Northern sky. Moreover, as the hours of darkness are fewer in summertime, the lack of a balance of surveys between North and South means that the sky is scanned less often in the Northern summer. Once it is completed, the Large Synoptic Survey Telescope will cover the southern sky, but being at a similar longitude to the other surveys there will still be times every day when it will be in daylight along with all the others. The 3.5 m Space Surveillance Telescope, which was originally also in the southwest United States, was dismantled and moved to Western Australia in 2017. When completed, this would make a significant difference to the global coverage. However, due to the new site being in a cyclone region, construction has not been completed. Unless this issue is resolved, the two new ATLAS telescopes planned for construction in the Southern hemisphere by 2020, are currently the only ones which will cover this gap in monitoring the skies.
| Survey | Telescope diameter | Number of telescopes | Time to scan entire visible Sky | Limiting magnitude | Hemisphere | Activity | Peak yearly observations | Survey category |
| ATLAS | 0.5 | 2 | 2 nights | 19 | Northern | 2016–present | 1,908,828 | Warning survey |
| ATLAS | 0.5 | 2 | 1 night | 19 | Southern | 2021 | NA | Warning survey |
| Catalina Sky Survey | 1.5 | 1 | 30 nights | 21.5 | Northern | 1998–present | see Mount Lemmon Survey | Cataloging survey |
| Catalina Sky Survey | 0.7 | 1 | 7 nights | 19.5 | Northern | 1998–present | 1,934,824 | Cataloging survey |
| Catalina Sky Survey | 0.5 | 1 | ? | ? | Southern | 2004–2013 | 264,634 | Warning survey |
| Large Synoptic Survey Telescope | 8.4 | 1 | 3–4 nights | 27 | Southern | 2022 | NA | Both |
| Lincoln Near-Earth Asteroid Research | 1.0 | 2 | ? | ? | Northern | 1998–2012 | 3,346,181 | Cataloging survey |
| Lowell Observatory Near-Earth-Object Search | 0.6 | 1 | 41 nights | 19.5 | Northern | 1998–2008 | 836,844 | Cataloging survey |
| Mount Lemmon Survey | 1.52 | 1 | ? | ~21 | Northern | 2005–present | 2,920,211 | Cataloging survey |
| Near-Earth Asteroid Tracking | ? | 2 | ? | ? | Northern | 1995–2007 | 1,214,008 | Cataloging survey |
| NEOSM | 0.5 | 1 | ? | ? | SEL1 | 2025 | NA | Cataloging survey |
| NEO Survey Telescope | 1 | 1 | 1 night | 21 | Northern | 2022 | NA | Warning survey |
| NEOWISE | 0.4 | 1 | ~6 months | ~22 | Earth Orbit | 2009–present | 2,279,598 | Cataloging survey |
| Pan-STARRS | 1.8 | 2 | 30 nights | 23 | Northern | 2010–present | 5,254,605 | Cataloging survey |
| Space Surveillance Telescope | 3.5 | 1 | 6 nights | 20.5 | Northern | 2014–2017 | 6,973,249 | Warning survey |
| Space Surveillance Telescope | 3.5 | 1 | 6 nights | 20.5 | Southern | 2022 | NA | Warning survey |
| Spacewatch | 1.8 | 1 | ? | ? | Northern | 1980–1998 | 1,532,613 | Cataloging survey |
| Spacewatch | 0.9 | 1 | ? | 22 | Northern | 1980–1998 | 1,532,613 | Cataloging survey |
| Zwicky Transient Facility | 1.2 | 1 | 3 nights | 20.5 | Northern | 2018–present | 483,822 | Warning survey |
ATLAS
ATLAS, the "Asteroid Terrestrial-impact Last Alert System" uses two 0.5-metre telescopes located at Haleakala and Mauna Loa on two of the Hawaiian Islands. With a field of view of 30 square degrees each, the telescopes survey the observable sky down to apparent magnitude 19 with 4 exposures every two clear nights. The survey has been fully operational with these two telescopes since 2017, and in 2018 obtained NASA funding for two additional telescopes. Both will be sited in the Southern hemisphere, with one at the South African Astronomical Observatory, and one in Chile. They are expected to take 18 months to build. Their southern locations will provide coverage of the 15% of the sky that cannot be observed from Hawaii, and combined with the Northern hemisphere telescopes will give non-stop coverage of the night sky.Catalina Sky Survey (including Mount Lemmon Survey)
In 1998, the Catalina Sky Survey took over from Spacewatch in surveying the sky for the University of Arizona. It uses two telescopes, a 1.5 m Cassegrain reflector telescope on the peak of Mount Lemmon, and a 0.7 m Schmidt telescope near Mount Bigelow. Both sites use identical cameras which provide a field of view of 5 square degrees on the 1.5-m telescope and 19 square degrees on the Catalina Schmidt. The Cassegrain reflector telescope takes three to four weeks to survey the entire sky, detecting objects fainter than apparent magnitude 21.5. The 0.7 m telescope takes a week to complete a survey of the sky, detecting objects fainter than apparent magnitude 19. This combination of telescopes, one slow and one medium, has so far detected more near Earth Objects than any other single survey. This shows the need for a combination of different types of telescopes.CSS used to include a telescope in the Southern Hemisphere, the Siding Spring Survey. However operations ended in 2013 after funding was discontinued.
Large Synoptic Survey Telescope
The Large Synoptic Survey Telescope is a wide-field survey reflecting telescope with an 8.4-meter primary mirror, currently under construction on Cerro Pachón in Chile. It will survey the entire available sky around every three nights. Science operations are due to begin in 2022. Scanning the sky relatively fast but also being able to detect objects down to apparent magnitude 27, it should be good at detecting nearby fast moving objects as well as excellent for larger slower objects that are currently further away.Near-Earth Object Surveillance Mission
A planned space-based 0.5m infrared telescope designed to survey the Solar System for potentially hazardous asteroids.NEO Survey Telescope
The Near Earth Object Survery TELescope is an ESA funded project, starting with an initial prototype currently under construction. The telescope is of a new "fly-eye" design that combines a single reflector with multiple sets of optics and CCDs, giving a very wide field of view. When complete it will have the widest field of view of any telescope and will be able to survey the majority of the visible sky in a single night. If the initial prototype is successful, three more telescopes are planned for installation around the globe. Because of the novel design, the size of the primary mirror is not directly comparable to more conventional telescopes, but is equivalent to a conventional 1-metre telescope.The telescope itself should be complete by end of 2019, and installation on Mount Mufara, Sicily should be complete in 2020 but was pushed back to 2022.
NEOWISE
The Wide-field Infrared Survey Explorer is a 0.4 m infrared-wavelength space telescope launched in December 2009, and placed in hibernation in February 2011. It was re-activated in 2013 specifically to search for near-Earth objects under the NEOWISE mission. By this stage, the spacecraft's cryogenic coolant had been depleted and so only two of the spacecraft's four sensors could be used. Whilst this has still led to new discoveries of asteroids not previously seen from ground-based telescopes, the productivity has dropped significantly. In its peak year when all four sensors were operational, WISE made 2.28 million asteroid observations. In recent years, with no cryogen, NEOWISE typically makes approximately 0.15 million asteroid observations annually. The next generation of infrared space telescopes has been designed so that they do not need cryogenic cooling.Pan-STARRS
, the "Panoramic Survey Telescope And Rapid Response System", currently consists of two 1.8 m Ritchey–Chrétien telescopes located at Haleakala in Hawaii. It has discovered a large number of new asteroids, comets, variable stars, supernovae and other celestial objects. Its primary mission is now to detect near-Earth objects that threaten impact events, and it is expected to create a database of all objects visible from Hawaii down to apparent magnitude 24. The Pan-STARRS NEO survey searches all the sky north of declination −47.5. It takes three to four weeks to survey the entire sky.Space Surveillance Telescope
The Space Surveillance Telescope is a 3.5 m telescope that detects, tracks, and can discern small, obscure objects, in deep space with a wide field of view system. The SST mount uses an advanced servo-control technology, that makes it one of the quickest and most agile telescopes of its size. It has a field of view of 6 square degrees and can scan the visible sky in 6 clear nights down to apparent magnitude 20.5. Its primary mission is tracking orbital debris. This task is similar to that of spotting near-Earth asteroids and so it is capable of both.The SST was initially deployed for testing and evaluation at the White Sands Missile Range in New Mexico. On 6 December 2013, it was announced that the telescope system would be moved to the Naval Communication Station Harold E. Holt in Exmouth, Western Australia. The SST was moved to Australia in 2017, but due to the new site being in a cyclone region, construction has been delayed, pending a redesign that can withstand cyclone force winds.
Spacewatch
Spacewatch was an early sky survey focussed on finding near Earth asteroids, originally founded in 1980. It was the first to use CCD image sensors to search for them, and the first to develop software to detect moving objects automatically in real-time. This led to a huge increase in productivity. Before 1990 a few hundred observations were made each year. After automation, annual productivity jumped by a factor of 100 leading to tens of thousands of observations per year. This paved the way for the surveys we have today.Although the survey is still in operation, in 1998 it was superseded by Catalina Sky Survey. Since then it has focused on following up on discoveries by other surveys, rather than making new discoveries itself. In particular it aims to prevent high priority PHOs from being lost after their discovery. The survey telescopes are 1.8 m and 0.9 m. The two follow up telescopes are 2.3 m and 4 m.
Zwicky Transient Facility
The Zwicky Transient Facility was commissioned in 2018, superseding the Intermediate Palomar Transient Factory. It is designed to detect transient objects that rapidly change in brightness as well as moving objects, for example supernovae, gamma ray bursts, collisions between two neutron stars, comets and asteroids. The ZTF is a 1.2 m telescope that has a field of view of 47 square degrees, designed to image the entire northern sky in three nights and scan the plane of the Milky Way twice each night to a limiting magnitude of 20.5. The amount of data produced by ZTF is expected to be 10 times larger than its predecessor.Follow up observations
Once a new asteroid has been discovered and reported, other observers can confirm the finding and help define the orbit of the newly discovered object. The International Astronomical Union Minor Planet Center acts as the global clearing house for information on asteroid orbits. It publishes lists of new discoveries that need verifying and still have uncertain orbits, and it accepts the resulting follow up observations from around the world. Unlike the initial discovery, which typically requires unusual and expensive wide-field telescopes, ordinary telescopes can be used to confirm the object as its position is now approximately known. There are far more of these around the globe, and even a well equipped amateur astronomer can contribute valuable follow-up observations of moderately bright asteroids. For example, the Great Shefford Observatory in the back garden of amateur Peter Birtwhistle typically submits thousands of observations to the Minor Planet Center every year. Nonetheless, some surveys have their own dedicated follow up telescopes.Follow up observations are important because once a sky survey has reported a discovery it may not return to observe the object again for days or weeks. By this time it may be too faint for it to detect, and in danger of becoming a lost asteroid. The more observations and the longer the observation arc, the greater the accuracy of the orbit model. This is important for two reasons:
- for imminent impacts it helps to make a better prediction of where the impact will occur and whether there is any danger of hitting a populated area.
- for asteroids that will miss Earth this time round, the more accurate the orbit model is, the further into the future its position can be predicted. This allows recovery of the asteroid on its subsequent approaches, and impacts to be predicted years in advance.
Estimating size and impact severity
One example of such a follow up observation was for 3671 Dionysus by UKIRT, the world's largest infrared telescope at the time. A second example was the 2013 ESA Herschel Space Observatory follow up observations of 99942 Apophis, which showed it was 20% larger and 75% more massive than previously estimated. However such follow ups are rare. The size estimates of most near-Earth asteroids are based on visible light only.
If the object was discovered by an infrared survey telescope initially, then an accurate size estimate will already be available, and infrared follow up will not be needed. However none of the ground-based survey telescopes listed above operate at thermal infrared wavelengths. The NEOWISE satellite had two thermal infrared sensors but they stopped working when the cryogen ran out. There are therefore currently no active thermal infrared sky surveys which are focused on discovering near-Earth objects. There are plans for a new space based thermal infrared survey telescope, Near-Earth Object Surveillance Mission, due to launch in 2025.
Impact calculation
Minimum orbit intersection distance
The minimum orbit intersection distance between an asteroid and the Earth is the distance between the closest points of their orbits. This first check is a coarse measure that does not allow an impact prediction to be made, but is based solely on the orbit parameters and gives an initial measure of how close to Earth the asteroid could come. If the MOID is large then the two objects never come near each other. In this case, unless the orbit of the asteroid is perturbed so that the MOID is reduced at some point in the future, it will never impact Earth and can be ignored. However, if the MOID is small then it is necessary to carry out more detailed calculations to determine if an impact will happen in the future. Asteroids with a MOID of less than 0.05 AU and an absolute magnitude brighter than 22 are categorized as a potentially hazardous asteroid.Projecting into the future
Once the initial orbit is known, the potential positions can be forecast years into the future and compared to the future position of Earth. If the distance between the asteroid and the centre of the Earth is less than Earth radius then a potential impact is predicted. To take account of the uncertainties in the orbit of the asteroid, several future projections are made. Each simulation has slightly different parameters within the range of the uncertainty. This allows a percentage chance of impact to be estimated. For example, if 1,000 simulations are carried out and 73 result in an impact, then the prediction would be a 7.3% chance of impact.NEODyS
NEODyS is a European Space Agency service that provides information on near Earth objects. It is based on a continually and automatically maintained database of near earth asteroid orbits. The site provides a number of services to the NEO community. The main service is an impact monitoring system of all near-Earth asteroids covering a period until the year 2100.The NEODyS website includes a Risk Page where all NEOs with probabilities of hitting the Earth greater than 10−11 from now until 2100 are shown in a risk list. In the table of the risk list the NEOs are divided into:
- "special", as it is the case of Apophis
- "observable", objects which are presently observable and which critically need a follow up in order to improve their orbit
- "possible recovery", objects which are not visible at present, but which are possible to recover in the near future
- "lost", objects which have an absolute magnitude brighter than 25 but which are virtually lost, their orbit being too uncertain; and
- "small", objects with an absolute magnitude fainter than 25; even when those are "lost", they are considered too small to result in heavy damage on the ground.
Sentry prediction system
's Sentry System continually scans the MPC catalog of known asteroids, analyzing their orbits for any possible future impacts. Like ESA's NEODyS, it gives a MOID for each near-Earth object, and a list of possible future impacts, along with the probability of each. It uses a slightly different algorithm to NEODyS, and so provides a useful cross-check and corroboration.Currently, no impacts are predicted.
Impact probability calculation pattern
The ellipses in the diagram on the right show the predicted position of an example asteroid at closest Earth approach. At first, with only a few asteroid observations, the error ellipse is very large and includes the Earth. Further observations shrink the error ellipse, but it still includes the Earth. This raises the predicted impact probability, since the Earth now covers a larger fraction of the error region. Finally, yet more observations shrink the ellipse revealing that the Earth is outside the error region, and the impact probability is near zero.For asteroids that are actually on track to hit Earth the predicted probability of impact continues to increase as more observations are made. This initially very similar pattern makes it difficult to quickly differentiate between asteroids which will be millions of kilometres from Earth and those which will actually hit it. This in turn makes it difficult to decide when to raise an alarm as gaining more certainty takes time, which reduces the time available to react to a predicted impact. However raising the alarm too soon has the danger of causing a false alarm and creating a Boy Who Cried Wolf effect if the asteroid in fact misses Earth.
In December 2004 when Apophis was estimated to have a 2.7% chance of impacting Earth on 13 April 2029, the uncertainty region for this asteroid had shrunk to 83000 km.
Response to predicted impact
Once an impact has been predicted the potential severity needs to be assessed, and a response plan formed. Depending on the time to impact and the predicted severity this may be as simple as giving a warning to citizens. For example, although unpredicted, the 2013impact at Chelyabinsk was spotted through the window by teacher Yulia Karbysheva. She thought it prudent to take precautionary measures by ordering her students to stay away from the room's windows and to perform a duck and cover maneuver. The teacher, who remained standing, was seriously lacerated when the blast arrived and window glass severed a tendon in one of her arms and left thigh, but none of her students, whom she ordered to hide under their desks, suffered cuts. If the impact had been predicted and a warning had been given to the entire population, similar simple precautionary actions could have vastly reduced the number of injuries. Children who were not in her class were injured.
If a more severe impact is predicted, the response may require evacuation of the area, or with sufficient lead time available, an avoidance mission to repel the asteroid. According to expert testimony in the United States Congress in 2013, NASA would require at least five years of preparation before a mission to intercept an asteroid could be launched.
Effectiveness of the current system
The effectiveness of the current system can be assessed a number of ways. The diagram below illustrates the number of successfully predicted impacts each year compared to the number of unpredicted asteroid impacts recorded by infrasound sensors designed to detect detonation of nuclear devices. It shows that the vast majority are still missed.One problem with assessing effectiveness this way is that the asteroids tend to be small. Missing small asteroids is unimportant as they generally do very little damage. However, missing a large day-side impacting asteroid is problematic. In order to assess the effectiveness for detecting larger asteroids, a different approach is needed.
Another way to assess the effectiveness is to look at warning times for asteroids which did not impact Earth, but came reasonably close. Looking at asteroids which came closer than the Moon, the below diagram shows how far in advance of closest approach the asteroids were first detected. Unlike actual asteroid impacts where, by using infrasound sensors, it is possible to assess how many were undetected, there is no ground truth for close approaches. The below chart therefore does not include any statistics for asteroids which went completely undetected. It can be seen however that about half of the asteroids that were detected, were not detected until after they had passed Earth. That is to say, if they had been on an impact trajectory, they would have been undetected before impact. This includes larger asteroids such as 2018 AH, which wasn't detected until 2 days after it had passed, and is estimated to be around 100 times more massive than the Chelyabinsk meteor.
It is worth noting that the number of detections is increasing as more survey sites come on line, and that approximately half of the detections are made after the asteroid passes the Earth.
One final statistic which casts some light on the effectiveness of the current system is the average warning time for an asteroid impact. Based on the few successfully predicted asteroid impacts, the average time between initial detection and impact is currently around 14 hours. Note however that there is some delay between the initial observation of the asteroid, data submission, and the follow up observations and calculations which lead to an impact prediction being made.
Improving impact prediction
In addition to the already funded telescopes mentioned above, two separate approaches have been suggested by NASA to improve impact prediction. Both approaches focus on the first step in impact prediction as this is the largest weakness in the current system. The first approach uses more powerful ground-based telescopes similar to the LSST. Being ground-based, such telescopes will still only observe part of the sky around Earth. In particular, all ground-based telescopes have a large blind spot for any asteroids coming from the direction of the Sun. In addition, they are affected by weather conditions, airglow and the phase of the Moon.To get around all of these issues, the second approach suggested is the use of space-based telescopes which can observe a much larger region of the sky around Earth. Although they still cannot point directly towards the Sun, they do not have the problem of blue sky to overcome and so can detect asteroids much closer in the sky to the Sun than ground-based telescopes. Unaffected by weather or airglow they can also operate 24 hours per day all year round. Finally, telescopes in space have the advantage of being able to use infrared sensors without the interference of the Earth's atmosphere. These sensors are better for detecting asteroids than optical sensors, and although there are some ground based infrared telescopes such as UKIRT, they are not designed for detecting asteroids. Space-based telescopes are more expensive, however, and tend to have a shorter lifespan. Therefore, Earth-based and space-based technologies complement each other to an extent. Although the majority of the IR spectrum is blocked by Earth's atmosphere, the very useful thermal frequency band is not blocked. This allows for the possibility of ground based thermal imaging surveys designed for detecting near earth asteroids, though none are currently planned.
Opposition effect
There is a further issue that even telescopes in Earth orbit do not overcome. This is the issue of illumination. Asteroids go through phases similar to the lunar phases. Even though a telescope in orbit may have an unobstructed view of an object that is close in the sky to the Sun, it will still be looking at the dark side of the object. This is because the Sun is shining primarily on the side facing away from the Earth, as is the case with the Moon when it is in a crescent phase. Because of the opposition effect, objects are far less bright in these phases than when fully illuminated, which makes them difficult to detect.This problem can be solved by the use of thermal infrared surveys. Ordinary telescopes depend on observing light reflected from the Sun, which is why the opposition effect occurs. Telescopes which detect thermal infrared light depend only on the temperature of the object. Its thermal glow can be detected from any angle, and is particularly useful for differentiating asteroids from the background stars, which have a different thermal signature.
This problem can also be solved without using thermal infrared, by positioning a space telescope away from Earth, closer to the Sun. The telescope can then look back towards Earth from the same direction as the Sun, and any asteroids closer to Earth than the telescope will then be in, and much better illuminated. There is a point between the Earth and Sun where the gravities of the two bodies are perfectly in balance, called the Sun-Earth L1 Lagrange point. It is approximately from Earth, about four times as far away as the Moon, and is ideally suited for placing such a space telescope. One problem with this position is Earth glare. Looking outward from SEL1, Earth itself is at full brightness, which prevents a telescope situated there from seeing that area of sky. Fortunately, this is the same area of sky that ground-based telescopes are best at spotting asteroids in, so the two complement each other.
Another possible position for a space telescope would be even closer to the Sun, for example in a Venus-like orbit. This would give a wider view of Earth orbit, but at a greater distance. Unlike a telescope at the SEL1 Lagrange point, it would not stay in sync with Earth but would orbit the Sun at a similar rate to Venus. Because of this, it would not often be in a position to provide any warning of asteroids shortly before impact, but it would be in a good position to catalog objects before they are on final approach, especially those which primarily orbit closer to the Sun. One issue with being as close to the Sun as Venus is that the craft may be too warm to use infrared wavelengths. A second issue would be communications. As the telescope will be a long way from Earth for most of the year communication would often be slow and at times impossible, without expensive improvements to the Deep Space Network.
Solutions to problems: summary table
This table summarises which of the various problems encountered by current telescopes are solved by the various different solutions.| Proposed solution | Global coverage | Clouds | Blue sky | Full moon | Opposition Effect | Thermal Infrared | Airglow |
| Geographically separated ground based survey telescopes | |||||||
| More powerful ground based survey telescopes | |||||||
| Infrared ground based NEO survey telescopes | |||||||
| Telescope in Earth orbit | |||||||
| Infrared Telescope in Earth orbit | |||||||
| Telescope at SEL1 | |||||||
| Infrared Telescope at SEL1 | |||||||
| Telescope in Venus-like orbit |
NEOCAM
In 2017 NASA proposed a number of alternative solutions to detect 90% of near-Earth objects of size 140 m or larger over the next few decades, which will also improve detection rates for the smaller objects which impact Earth more often. Several of the proposals use a combination of an improved ground based telescope and a space based telescope positioned at the SEL1 Lagrange point such as NEOCAM. However none of these proposals have yet been funded. As this is a global issue, and noting that to date NASA-sponsored surveys have contributed over 95% of all near earth object discoveries, in 2018 the Trump administration asked NASA to find international partners to help fund the improvements.List of successfully predicted asteroid impacts
Below is the list of all near-Earth objects which may have impacted the Earth and which were successfully predicted beforehand. This list would also include any objects identified as having greater than 50% chance of impacting in the future, but no such future impacts is predicted at this time. As asteroid detection ability increases it is expected that prediction will become more successful in the future.| Date of impact | Date discovered | Object | Observation arc | Warning period | Cataloged | Size | Velocity | Explosion Altitude | Impact Energy | |
| 2008-10-07 | 2008-10-06 | 2008 TC3 | 1,145 | 0.7 | No | 4.1 | 30.4 | 12.8 | 37 | 0.98 |
| 2014-01-02 | 2014-01-01 | 2014 AA | 69 | 0.8 | No | 2–4 | 30.9 | 35.0 | unknown | unknown |
| 2018-01-22 | 2018-01-22 | A106fgF | 39 | 0.4 | No | 1–4 | 31.1 | unknown | N/A | N/A |
| 2018-06-02 | 2018-06-02 | 2018 LA | 227 | 0.3 | No | 2.6–3.8 | 30.6 | 17 | 28.7 | 1 |
| 2019-06-22 | 2019-06-22 | 2019 MO | 138 | 0.5 | No | 3–10 | 29.3 | 14.9 | 25 | 6 |