additonal considerations for water routing changes

[amoodie]

#133 will fix a number of problems with the water routing implementation, but doesn't completely fix the root issue though, which I believe is that sometimes very few parcels reach the boundary and the flow field kind of "dries up". This seems to happen in very large domains with very low sea-level rise (i.e., low accommodation), but can happen any time a restrictive configuration is reached (just by chance). I think a better/additional fix to consider may be to 1) continue routing water parcels until Np_water parcels have actually reached the boundary (or ~XX% of them anyway), or 2) somehow renormalize the parcels that did make it to the boundary to represent as though all Np_water parcels did reach the boundary. Things to consider...