I have recently changed my hydro simulations to the Refluxing strategy implemented by Erik Schnetter. For this I have changed the standard configuration to a cell centred refinement, staggered boundaries and low order spatial prolongation for the hydro variables. Since then I experience problems with BSSN or, more specifically, the QuasilocalMeasures thorn (a=0.9 black hole). After an initially fine behaviour, no Eigenvector with value one is found and the spacetime gradually crashes.

I am still puzzled where this is coming from. Does anybody have some experience with the combination spacetime evolution + hydro refluxing, i.e. the somewhat different grid setup? I would be glad for any input to see where I start to fix my work.