why are these events always forecasted so late? some hours earlier i could have witnessed it
Two related reasons. 1m is quite small, so it needs to be close for us to detect it. Since we don’t have many observations, the error in estimates of its orbital path is high. In general, it also is extra difficult to calculate the exact last little bit of a trajectory that will impact earth because the force models start getting really wild as something gets very close to earth - approximations stop working.
What kind of forces? I assume you're talking about the space part, before atmosphere, so I would assume it's the same gravity effects that you also get in space, just that gravity is a bit larger value. I wrote a gravity simulator at lgms.nl/p/badgravity so I have a bit of an idea but am definitely no expert (the thing doesn't do any sort of collisions and assumes 2d point masses), so I'm curious what you mean. Especially with the earth dominating the gravity and not needing to use N-body physics anymore, it should get more accurate if anything.
Could you elaborate or do you have a link? Or do you simply mean the compounding effect of having used N-body approximations for the interplanetary space part of the trajectory leading to different starting points for the final approach?