But if they are on LEO the atmospheric drag shouldn't be too negligible no?

It isn't negligible on the timescales we care about. Skimming through the graphs in [0], at 900km we're talking 100+ years of orbital lifetime. And even if most fragments were to disappear sooner[1] - giving them a generous 20 years - that's still 20 years of dealing with a swarm of untrackable hypervelocity bullets coming at you from random directions. While space is big, it's not that big in LEO, and any such bullet will, upon impact, spawn more bullets.

I don't think we'll have a catastrophic cascade leading to Kessler syndrom even if these two satellites collide today, but few more such events might be enough.

EDIT:

Also, the collision will be off-center and, eyeballing the orbital diagrams, some 15 degrees shy of being head-on. That means you can expect most of the both satellites not being directly involved in the collision, but instead gaining spin and breaking apart under stress. Those surviving pieces will then keep most of their orbital velocity and heading, but scatter like buckshot, each piece on slightly different orbit, migrating further over time.

--

[0] - https://www.spaceacademy.net.au/watch/debris/orblife.htm

[1] - Orbital decay due to atmospheric drag is a function of mass to area ratio. I'm not sure whether the debris from a hypervelocity collision tends to be more like irregular, high-drag shatter, or like low-drag spheres of superheated metal that cooled off radiatively.

> Also, the collision will be off-center and, eyeballing the orbital diagrams, some 15 degrees shy of being head-on.

How can we tell it will be off-centre if there's enough uncertainty in the orbital prediction that they are only "1% to 20%" to collide at all?

Because if there's 1% to 20% chance they'll collide at all, then the chance they'll collide centrally is even smaller.

It depends on LEO altitude (exponentially) and aerodynamic properties of the object.

~ 150 km making single orbit becomes impossible.

~ 300 km orbital decay is ~km/month, so it already takes years to fall down to earth. ISS is at 400 km and decays 2 km/month. It has unfavourable mass to area ratio due to all those solar panels.

~ 500 km it takes few decades to fall down. (Hubble space telescope is here)

990 km is the altitude where these two satellites might collide. Debris can stay up thousands of years .

991km is pretty high for LEO - ISS is at ~400km, Hubble is at ~500km.

Those fragments will scatter in all directions. Some will go down, some will go up or in eccentric orbits, some wil go sideways, some will stay in the same or similar orbits. There will be A LOT of those fragments, all going at orbital speeds in various directions.

From up stream in the comments, one was launched 31 years ago... so yeah, this stuff could stick around a while.

For every 100-200 km in height the lifetime seems to increase by a factor of 10. At 900 km the lifetime of big objects is 1000 years. Smaller objects can still orbit 10s to 100s of years.