k-1 says: Lessee, 1200 lbs over 500 miles wide = 12 lbs per mile in a strip that goes all the way around the world.

A little sloppy with the arithmetic, but:

The potential impact zone is 500 miles wide x the circumference of the earth (~24,000 miles), or 12,000,000 square miles. If a 1200 pound object is evenly dispersed over that area, the "impact density" would be 1200/(12*10⁶) = 0.0001 lb/square mile. At this density, the object would be so dispersed that it would be essentially an aerosole that would merely float in a cloud.

A single 1200 lb object (remaining size at impact) would produce a shock wave with a damaging overpressure perhaps 50 miles in radius, or covering 7,854 square miles, equally likely to strike anywhere in the 12*10⁶ square mile zone. This gives a probability that any given point will be affected of around 0.00065, or "odds" of 1/1538. Recent estimates are that there are 6.8*10⁹ persons in the world, with 60%(?) in the impact zone. If 1 of every 1538 persons in the zone is killed or seriously injured, that gives a casualty count of about 0.6*6.8*10⁹/1538 = 2,652 persons.

Of course the object might land where the population density is very much lower than the "average" of 0.05 persons per square mile, say perhaps in NY City or Paris, where it's nearer a million persons per square mile.(??). Impact of a single 1200 lb object at one of those places could easily wipe out a million or more people, and would possibly completely obliterate many cities only a little smaller.

The news reports are unclear whether the NASA statements mean that a single 1200 lb object might impact the surface, or that a total of 1200 pounds of objects might make impact, and also whether the 1200 pounds is an estimate of mass(es) entering the atmosphere with the potential to have some reduced (fragmented and ablated) mass portion reach the surface, or the total of masses that might actually make impact. (Reports that appear to be clear disagree with others that appear equally clear.)

Fragmentation into many pieces would significantly reduce the effects of each individual impact, but might significantly raise the odds that more populated areas could be affected.

Of course all the analyses that can be done are "probabilistic" and very much like "statistics" so we can get any estimate(s) we want, with all the guesses being equally unsupportable.

Anybody want to shoot some CRAPS while we wait?

John