I'm working on a game where the player gets to (among other things) hurl meteors at a landscape and see the effects up close in 1st person. At present my meteor simulator is rather simple, making a somewhat tiered approach to the problem. The result is passable but lacking. I'm not simulating the actual impact. Rather I just let the user select a radius and a center and I use that information to deform the height map. I'd like to improve the meteor impact code to be more "realistic". I'm at a loss however as to what sort of mathematics to use. I would assume that the closest subject to cross reference might be simulating waves in a fluid simulation, but I'm not even that sure about it. Impact craters have, to my knowledge: 1. A crater, rather bowl shaped. 2. A rim, which is rock raised above the "flat" land around by the force of the impact. 3. Smaller secondary craters from ejected material falling back to earth. 4. Radial streaks if the impact occurred on an airless world. 5. Wiki also says "Complex craters are larger, and have uplifted centers that are surrounded by a trough, plus broken rims. The uplifted center is due to the "rebound" of the earth after the impact. It is something like the ripple pattern created by a drop of water into a pool, frozen into the Earth when the melted rock cooled and solidified. I don't necessarily need to simulate the latter two features, but it would be a nice touch. Has anyone played around with anything like this at all?

Rather than simulating a crater forming, which IMHO seems to be much more work than it is worth, have someone create a crater model. Simply scale it depending on the energy of the impact and rotate it to provide some variety. Then use the modified crater model to deform your terrain. Create secondary craters the same way.

Off topic, but interesting.

Get a picture of some craters on the moon, they look like craters.
Then turn it upside down, they look like bumps!

It's the same reason that inside out masks seem to rotate the opposite way to the way in which you turn them though, boringly. Normals and lighting.

I agree that simulating the crater is overkill. I would like to have the crater "template", however, be defined procedurally instead of "modelling" the crater. Other issues aside this just seems to be a kind of problem that can benefit from proceduralism.

Ideally, I see the function returning height changes given a certain distance away from the center of the crater. Basically a crater profile generator. I suppose what I really need is some general "rule of thumb"s for the ratio of the radius of a crater to its depth, to its rim size, as well as some sort of comparison of crater shape given size (larger craters seem to become complex rather than simple, which changes a number of ratios I'm sure).

I'm not even sure what sort of function a crater's profile would be. Presumably a fresh crater is fairly elliptical (perhaps even spherical, if it's small enough) inside the bowl with the walls "slumping" as the crater matures. I'm not sure what the rim would start out as. It would suffer the most from "slumping" for sure.

I don't necessarily need to simulate craters slumping though. It might not be too hard, though, if it can be modelled using a simple erosion algorithm.

After much searching, I think what I really need is Impact Cratering : A Geologic Process (Oxford Monographs on Geology and Geophysics) by H. J. Melosh. Well, it's probably overkill, but having researched impact craters a little I wouldn't mind a difinitive work on it ;) Unfortunately I seriously doubt if I'm ever going to find a copy. :/

check out:

this
and this

(note: there's a pdf at the bottom of the second link)

Thnaks, the links you provided will be useful when I decide to determine meteor size from crater size (perhaps to do a short animation or some such).

Doesn't really help me on forming the crater shape though.