simple question
 

I wrote a simple Boltzmann method solver for 2D flow around a circular cylinder. My question concerns the density distribution on the wall of the cylinder because I think my boundary conditions are messing up. The density builds up in front of the cylinder and decreases behind it like I would expect, but right on the cylinder wall it sharply decreases ( at the front, or increases at the back) to a value close to the free field value. This sharp shift takes place between the lattice point adjacent to the wall and the point on the wall. Is this correct? I would expect the density directly on the wall to be higher than the free field value.