switch coordinate input
At the moment z denotes the depth in model and seismogram(header) input/output and the input files. In the code "y" is used as depth. y should be used therefore as depth again to be consistent. So the z<->y transformation is no longer needed.