How big an impact could humanity survive?

Factual Questions
1

Inspired by this thread, how big would a meteor impact - or series thereof a la Shoemaker-Levy - have to be to wipe out humanity? I don’t mean necessarily immediately kill, but sufficient to ensure our extinction.

How big an impact would kill half? One in 10? One in 100? Obviously the location of the impact is very important given the concentrations of population in India and China.

2

What are you asking, precisely? How many would be killed within a few seconds of impact, directly from the heat, impact itself, earthquakes, tsunamis, and massive boulders thrown into the air? How many would die within a few weeks from the massive collapse of civilization? How many would die before we got everything back together and working again?

3

As I said,

4

Descriptions of the Chicxulub impact that I’ve read (but I don’t know how accurate they are, of course) indicate that when the column of vaporized rock that was initially knocked out into space began falling back to earth, it raised surface temperatures to around 160 or 170 degrees over the entire face of the earth for several hours, so only animals that were hiding deep in burrows or who were deep underwater survived. That was an impactor about 10km in diameter, IIRC.

I guess it depends on how prepared humanity was. I suspect that a group of people far enough away from any major impact sites who were in a bomb shelter or underground lair, who had sufficient air and food and water, would probably be able to survive long enough to wait out most of the immediate effects and, if they had enough food (or enough energy to power growing lights to grow more food underground), they might be able to survive long enough for the climate effects to dissipate enough for them to be able to start growing crops on the surface again, allowing them to repopulate the earth.

Of course, if the impacting object is the size of Pluto, then the whole crust is going to turn molten from the impact and you probably won’t see anything at all survive. So there’s certainly an upper limit, though I’m not sure where it is precisely.