Weird long-period content in Mineos' displacement seismograms

Hi all,

I am running Mineos to generate synthetic seismograms for the Kaikoura earthquake. This works fine for acceleration and velocity seismograms, but introduces weird long-period signals in the displacement seismograms. The configuration of my run file are the exact same for both seismograms, which is why I cannot figure out what is going wrong. Has anyone else observed this behavior?

I can invert the curvature of the seismogram if I use different values for wmax (for example 166 instead of 200) in minos_bran, but there is no stable region where the seismogram is actually flat…

Best, Boris