2 Cameras, 1 Rendertexture, still have depth?

So I'm trying to make these denser forests on mobile. As you can imagine, even if all trees are billboarded, there's so much alpha clipping full screen that it just drains the fps.

I had this brilliant idea that what if I had a second camera that only rendered the trees, into a render texture that's 1/2 res lets say. And then using the UI, display that texture.

I did this, setting the second camera to clear solid color white, and then on the UI I used a custom shader to discard anything all white. That leaves me with a UI texture on screen that has all the terrain trees cutout just fine. However, a 3D object in between the trees is simply behind the entire texture, which makes sense.

I need to figure out if it's possible to still somehow get depth working here. This method really helps with my FPS, just doesn't do depth correctly.

I'd appreciate any insight!

In this first image you can see how the object is in between billboard trees, like it should with depth:



And here is what it looks like in the game view, the tree rendertexture is just overlayed on top:

 

Excuse my ignorance of how things worked, but I thought of another possible way?

What if I can get the depth information for both cameras, so two rendertextures. Then basically for each pixel compare the depth for that pixel and then choose which rendertexture too apply to that pixel?

I think something kind of like this guy is asking: http://answers.unity3d.com/questions/989314/compositing-two-rendertextures-based-on-depth.html