With 22.8 tonnes per Falcon 9 launch, and 50 of those per year? That's 1140 tonnes /year of new debris in the upper atmosphere, a 21% increase. I'd say that is quite a significant change.
So in ~5 years they've launched 5113 satellites[1]: 2 * 400 kg + 60 * 227 kg + 1665 * 260 kg + 2987 * ~300 kg + 399 * 800 kg = 1663 metric tons
So if they all burned up today it would be around a 6% yearly increase, globally? (1663 / 5200 * 5). And that's probably the most forgiving estimate I could come up with, I think your ~20% number is more accurate for future projection purposes given most of the mass was launched in a much shorter timespan.
Then the real question is what effect does space dust have on the earth's atmosphere (if any).
There are two more things to consider. 1) The whole satellite doesn’t burn up. Most of it crashes into the ocean. 2) much of the burning up happens in the thicker lower atmosphere where we’re already putting lots of material from jet engines and factory chimneys.
"I'm commuting every day to work with my private jet. 5200 tonnes of micrometeorites hit the earth each year, so my action isn’t going to make a difference."
"It's simple. I started a charity with $10 million that convinces 1 million middle/low class people to keep their heating off during the winter, meaning my jet is now carbon negative. Sure a few people die each winter, but that's another win for the earth."
This is basically the reality of government leaders jetting in to global climate change conferences.
We’re talking about the risks of satellites burning up in the upper atmosphere. The satellites deposit lots of material when they burn up. People are worried that this might be bad.
Lots of people incorrectly think that the upper atmosphere is pristine and you would never normally get metal up there. In fact thousands of tons of micrometeorites hit the earth each year, depositing lots of rock and metal up there.
This means that the burning satellites cause an increase in metals deposited in the upper atmosphere. We’re not going from nothing to many tones. We’re going from many tones to slightly more tones.
There is an open question if the increase is a problem, but it’s less likely to be a problem than if we were going from nothing to something.
Given a choice between bigger fuel containers burning up empty and more copies of all the other parts burning up I know what I would pick. The current very short lived design is a compromise to get to operational coverage fast and for quick rollout of improved designs. Future generations with a larger mass fraction for fuel won't get multiple orders of magnitude more runtime (in orbits that low this would require some bussard drive equivalent..), but a generous multiplication.
Unfortunately the physics of this don't work very well. And at least for now the satellites are getting generational upgrades frequently, so the old satellites are obsolete. Maybe in the far future it will make sense to try this, but not anytime soon.
Pie in the sky: in the future when they have the ability to put large starship fuel tankers into orbit, they could match the orbits of starlink and have some kind of drone move between satellites and tanker.
https://www.scientificamerican.com/article/antarctic-study-s...