chongo's Journal: Hazardous NEO Q&A session #0

Shanes asked the following excellent questions:

Implied Q #0: I'd say this should be worth a Q&A session on Slashdot's front page.

A #0: Feel free to suggest it to the /. folks. I'd be happy to do a more general Hazardous NEO Q&A.

Q #1: I notice that many potential impacts have 2 dates that are virtually the same. 1997 XR2's 2 impacts, for instance, differ with just 0.01 days. Why aren't they considered the same potential impact (which in some cases could give it a higher Torino scale number)?

A #1: [[UPDATED]] The collection of asteroid observations may not be well defined / may have significant uncertainties require the model to split into parallel paths. If more than one of these paths intersect or closely approach the Earth, then multiple close encounters and/or potential impacts will be listed.

Over a span of 100 years, the Yarkovsky effect has a significant enough affect on 1997-XR2 to warrant its consideration. Our 1997-XR2 model has two major paths. One path assumes a clockwise rotation and the other a counter-clockwise rotation of 1997-XR2. Like JPL, we show two potential encounters whose close approach differs by about 1.4 minutes on 1 June 2101.

I cannot state for certain that the Varkovsky effect is why JPL lists 2 encounters for 1997-XR2, but it does for our model.

There are other reasons why a model may show multiple close encounters over a short span of time. Sometimes a model's will be split into several unique paths that represent a range of possible outcomes after a close encounter.

Sometimes 2 or more potential encounters within a brief period of time represent different paths (previously split as noted above) with different numbers of revolutions around the sun.

It is also possible that an object is really multiple objects (such as an asteroid with a moon or a fragmented comet) each of which has its own potential impact / close encounter. However this is rare.

Q #2: [[UPDATED]] What's the deal with the Hazardous NEO Technical Reviews Does it mean that a potentially hazardous impact with a Torino value > 1 would be kept off the JPL and NEODyS lists (and even in your journal!) until the Committee find it in its heart to tell the rest of us?

I find it hard to believe that such things could be kept secret since the observations are public and everyone can (in theory) make their own calculations based on them. Right?

A number of years ago, a hazardous NEO was discovered (I forget which one) that was picked up by the press. It was a slow news day. Hollywood had or was about to release several asteroid/comet doomsday movies. The press went wild and blew the report way out of proportion.

Then as usual, a number of new observations came along. Better models were developed. As I also recall, the initial model may even have had an error in it. The result of this was that the object was removed from the Hazardous NEO list.

Some of the press reported that Astronomers had made a blunder. Other press groups continued to publish doomsday reports, finding some person who discounted the revised model (claiming that the new data was part of a cover-up).

Nobody was happy with the result.

At the urging of notable Astronomers (I recall Marsden and Morrison), a new protocol was developed.

One aspect of this protocol is when an objects appears to have a >= 0 Palermo Technical Scale is discovered, the orbit model undergo a review by a technical committee before their results are released. Typically this review takes less than 72 hours.

After the review, the results are published. Additional material is also released help the press more accurately report the finding. The object is also posted on the various NEO lists after the technical review.

A summary of the 20 Dec 2002 Hazardous NEO Technical Review process is available.

Permit me to add a few remarks of my own:

• The ~72 hour review is to improve the accuracy of significant NEO potential impactors.
• A review is not unusual in Science. Many technical publications go through a referee process prior to journal publication. In some cases, the publication delay can be weeks to months. The ~72 hour technical review is a rather rapid review process when compared to most.
• A ~72 hour review / sanity check should not pose a risk to you personally. If we have less then 72 hours before something significant impacts the Earth, then there may be little that anyone can do about it anyway.
• The review process does not include any government approval. There is no censorship. The review team and 2nd level advisers are an International team. So single institution or nation has a lock on the process.
• The review team has promised to keep the information confidential during the ~72 hour review process. While I believe they will keep the information confidential during the ~72 hour period, the risk in the process is that someone will leak data instead of hiding/burying data.
• At the end of the review process, something is always released.
  • It may be that the object shows up as a new entry in the hazard list .
  • It may be that the object is determined to be a re-discovery on a previously known NEO object and only an IAU circular of the observation is published.
  • It may be that the object is a previously lost object that is not a NEO object and an only IAU circular of the observation is published.
  • It may be that the object is a newly discovered object that is a non-NEO object and a IAU circular of the observation and orbit parameters is published, along with perhaps a new entry in the minor planet database.

Q #3: How computational expensive are the the models?

A #3: Models can be somewhat computationally expensive. Detailed models can be computationally expensive.

Sometimes an observer / observatory's data has a systematic error.Discovering and correcting or ignoring bad data can be a combinatorial process that requires multiple models to be run.

Q #4: The JPL model takes the gravitational effects of the 3 biggest asteroids into consideration and I've noticed that 2002NT7 is listed with a close approach to the second largest asteroid, Pallas, in 2020. So I guess even the small masses of these asteroids are worth having in the model. But why do they stop there? Why isn't Juno and other large asteroids accounted for? Is this a trade off of accuracy vs computational time?

A #4: The more bodies that you add to your model, the more complex your model becomes. One must not only take steps to ensure that the new body data is accurate, one must also add the uncertainty of the new body data to the overall model error.

Most of the time, minor planets do not have a significant or even measurable impact on a NEO orbit model over the next 100 years. Most of the time even details about the Earth (such the Earth's non-spherical shape, non-uniform mass distribution, and time of day (how the non-uniform mass is rotated)), and the Moon (nutation, etc.) usually to not matter over a 100 year period.

It is only when an object has a close encounter with an another significant object that one has to resort to more detailed and complex models.

It is somewhat a trade off of accuracy vs computational time. It also an issue of model complexity and accuracy of additional orbits.

Q #5: Watching the hazard lists it seems that the only thing influencing the numbers are more observations, not more time to calculate. How long does it take to calculate a trajectory 100 years ahead given a set of observations?

A #5: The hazard lists produced by the JPL use their standard ephemeris model. Tables can be produced in short amount of time.

When we find that our model that suggests a close encounter with another significant object, we adjust our model. For example, a close encounter with Pallas might warrant adding estimated Pallas positions and mass over a brief period of time. More complex encounters such as the J002E3's 6 orbit eventrequires even more detailed model improvement.

Usually it takes more time to adjust and check a new model than it does to run the new model.

[[recent addition]]

Q #6: Has there ever been a potential impact listed with a Torino value >1?

A: #6: No.

Well that is all.

Corrections and additions to the above are welcome.