Add rock volume representation

[shiyuning]

@cmw29 :

First, bulk density is estimated based on soil texture by Saxton and Rawls pedotransfer. I assume that this estimation of bulk density should still apply to the particles in the <2mm size fraction. The next issue is how to include the rock volume in the overall density of the soil layer. For soil hydraulic and thermal properties, the rocks act like consolidated soil particles, such that you would want to include rock mass in the overall density of the layer. For the calculation of nutrient concentrations, however, rocks are inert material, so should be excluded from the overall density of the soil layer.

So it would seem simple enough that when initializing the soil variables two versions of bulk density are calculated, a version including rock mass and a version excluding rock mass. Then we just need to make sure each process algorithm is using the appropriate representation of bulk density.