But first, does the water actually run-off the gravel, or is it all collected in the stone? And what are you trying to model? The surface runoff, or the water that is collected in the stone?
Peter, thanks for the response. The floor of the pit is tightly packed gravel material, which is acting more like impervious area than pervious stone. I'm attempting to model runoff from the surface. I've started using a CN of 76 for "gravel roads, A soils". I've also considered using a CN 98 but I think this is a bit conservative.
Can you model it in place or is it too large or complicated to do that ? Just wait until it rains, record the rainfall at say, 15 minute intervals, and record the runoff hydrograph(s).
Use your Group-A soil values. Why penalize yourself. The curve nos are so radically different (gravel road to A-grass/woods) that you probably had no pre-development runoff and your actual final condition will be essentially the same, just with different slopes. Plus, because of your swing in CNs, the new composite CN would be somewhat higher and the results will say you've now got tons of runoff to deal with post versus none in the pre (not reality).