Bug tracer with periodic BC Gerris

[Jean-Pierre Issartel]

Hi, I am using Gerris and should probably move to Basilisk. I am using a 32x32x32 cube with three periodic BC. I have computed a purely turbulent flow in this domain, corresponding to a torus without boundaries. This flow has same behaviour statitically in all meshes, with zero mean. However, when I want to scatter a scalar tracer in this domain using this flow off-line, there is clearly an anomalous behaviour at the faces with locally created tracer. How can I fix that. Is it difficult to transfer my computations to Basilisk using the same flow ?