Pathline particle tracking time flips sign

[ncl88]

I have a particle simulation that tracks particles backwards from a well in a model that has grid refinement around the well (disv). The well is placed in layer 12 (out of 14 layers) and layer 2 - 6 are inactive (idomain -1). Both with and without Newton Rapshon formulation, the particles show a behavior that I haven't encountered before: after 3 steps, the sign of the time flips from positive to negative. Many of the particles seems also to travel on a straight line following the edge of the gridcells. Does anyone have an explanation for this behavior?

[aprovost-usgs]

@ncl88 There's definitely something wrong, as tracking times should never be negative. Would you be able to email me (aprovost@usgs.gov) your zipped MODFLOW and MODPATH input files, or a link to them if they're too large to send directly?