My problem is with the assignment of numbers to the higher levels, all the way up to a "red alert" level 10. I think this degree of detail is unnecessary and confusing to the general public. After all, we really don't know the diameters of the objects (and can only make guesses from often-questionable observations of the brightness and an assumption of the albedo, or fraction of incident sunlight reflected by the object's surface). Also, the impact probability is not always as well defined as some might think. The way the "red alert" level 8 abuts the "no-consequence" level 0 for an infinitesimal change in size is quite illogical, and the presence of a point on the diagram that is ambiguously level 2, 4, 5, 6 or 7 must have been remarked upon by others.
The scale would surely be much more satisfactory if the numbers were scrapped completely, and it utilized only the colors. For the larger impacting bodies, there are levels corresponding to an "orange alert" between the yellow and red. To have a sequence white-green-yellow-orange-red is fine, except that I would adjust the probability-size relations such that this sequence is consistently used throughout the diagram, thereby eliminating the crazy white-red boundary, as well as the yellow-red and white-yellow boundaries of the current version.
Use of colors, rather than numbers, could remove any tendency to compare the scale to the Richter Scale. Although Binzel and others have stressed the similarity, it really makes no sense to compare a scale involving tentative predictions of impacts far into the future with that giving the well-defined Richter numbers of past earthquakes. The impact probabilities change continually as new observations are obtained, and by using colors, rather than numbers, I think one can also better convey these changes, avoiding the precision that the use of numbers implies.
Although our experience to date is with four objects having demonstrated impact possibilities over the course of the next half-century, two never emerged beyond a white alert, one moved up from white to green, and one (1997 XF11 in 2040) just made it to a yellow alert. With the availability of further observations, three of the four dropped down to the extreme white region (i.e., essentially zero-impact probability), while one of the white alerts remains a non-zero white alert because of the absence of further data. It is perhaps not generally appreciated that, if the 1997 XF11 orbit relative to the earth had been shifted by something like 20,000 miles, there would have been an orange alert for this object in 2028. In Binzel's notation this would have been at level 6, verging on level 7. At such levels he considers the object to have a "significant threat of a collision capable of causing a global catastrophe."
Would that therefore have been any cause for alarm? Of course not! The very next day, the recognition of the 1990 observations would still have sent the impact probability straight down to zero (for the next thousand years or more), just as actually happened. The actual result of reaching an orange alert would therefore have been no different whatsoever from reaching yellow, green or staying in the white! The public needs to be reassured that this kind of thing will happen--and not frightened by explanations of the scale that speak of a "significant threat".
In almost every case, at some point the acquisition of further data will eliminate all the danger. There might be some slight cause for worry until further data are acquired (either from observations yet to be made or from the examination of old records), but it is really not worth losing sleep over this. Nevertheless, observations are of supreme importance until the threat has completely gone. In this sense, one might in fact say that small objects are more troublesome than the larger objects, because the existence of past records of a smaller object would be less likely, and there would tend to be fewer opportunities for further observations before the possible impact is to occur. Of course, if there is really going to be a hit, the impact probability will eventually reach 100 percent.
That brings up my final point, namely, that the scale needs to have built into it another parameter, and this is the amount of time to elapse between the present and the possible impact event. It makes a great deal of difference whether we are talking about an impact event that might occur in three decades, three days--or somewhere in-between. As I have suggested, the public at large really does not need to get worked up if there is a lead time of several decades. And that's the way it will be -- most of the time. These are situations for the astronomers to monitor.
So how can we accommodate this in the scale? The obvious way would be to indicate the shade or intensity of the color. If the lead time was less than one year, say, the color would be dark. Above one year, the color would be light. Very dark colors would apply under one-tenth of a year (37 days), very light colors beyond 10 years. If necessary, this scheme could be extended to extremely dark -- below 0.01 year (4 days) and extremely light -- beyond a century.
Thus, 1997 XF11 made it to a "very light yellow" alert, 1999 AN10 to a "very light green," before going to white -- the shade of which does not apply. There is in fact still some slight possibility that 1999 AN10 will be a danger a century or so from now, but at this stage it really does not matter whether it is "extremely light green" or "extremely light yellow". Of course, the worry in all of this is that we should find an imminent high-probability impact event that would qualify as "very dark red" or "extremely dark orange". This has not yet happened, and we hope it remains that way...
Brian G. Marsden
Harvard-Smithsonian Center for Astrophysics
November 7, 1999
advertisement
