specify whether shapefile boundaries go into pop_size estimations

[mannfred]

A attendee at the SCO workshop discovered that in our OSFL example (BCRs 60 and 81), the population size estimate for the entire AOI is different (either less or more, I can't remember) than the sum of the population sizes of the two BCRs estimated separately. We suspected this has to do with the boundaries splitting the BCRs double counting or not counting in the pop_size estimates.

[ecknight]

Probably has to do with the way the boundary cells are included in the crop argument. Manny to reproduce.