Remarks on approximate methods in geophysical inverse problems