I figured for the particles you should use...
<emissive>false</emissive>
<lighting>true</lighting>
to get the lighting right.
Yep, that's what I use, which is why the water becomes grey rather than white in the morning. I guess part of the problem is that all the other light computations in the scene are done by the terrain shaders, whereas the particle system doesn't know what light the terrain shaders assign to the scene. So maybe making them emissive tied to a property is actually a better idea...
I was asking why would you do the dot(faceNormal, upVector), since that is most likely to be equal to faceNormal.z (if my math is still right)?
Depends on what the coordinate system is at the point you evaluate this - if the up vector is indeed (0,0,1) then you're right, if we are still in some global coordinate system where (0,0,1) points to the North Pole then it doesn't work.
Thorsten - you can probably suggest an algorithm off the top of your head
I define the dot of normal and up as 'steepness' and then do a smoothstep in the steepness to determine the bias factor for vegetation in the shaders. I would simply try a few locations to see what looks good - or maybe even give us some property control over the limits, because it actually depends on the vegetation belt you're in - tropical vegetation is quite a bit more clinging to steep slopes than arctic vegetation...
IMHO, I don't think Yosemite falls should die halfway down
On most pictures it appears to do pretty much that. But there's also a technical issue: Particles don't 'flow' over terrain when they hit it, they just pass through. So even if the real waterfall is not free-fall but in contact with the rockface, we can't really simulate that unless the rock face is actually vertical, we can only do free-fall. But again, in Flightgear the rock faces are usually a bit less steep than in reality. So to make this work, the waterfall has to actually start with a high horizontal velocity to be thrown out far enough such that it doesn't disappear immediately beneath the terrain. Which is a bit of a limit determining how far down we can make it fall. Adding a second particle source underneath for a multiple overlapping cascade effects seems to be tricky (I got some texturing artefacts), so... without burning excessive resources for the waterfall, which in the whole scheme is just some addition to the scene, this is probably as good as it gets (if I were willing to burn excessive resources, there's Nasal and shader magic, but I'm not sure I want to develop a hi-fidelity waterfall simulator with the option to see this from an airplane )