Mahmoud, thanks for clarifying the issue. I hope this one is easier to explain. It looks like you plot X projection of surface current along the line. The dipole arms has some width. Plot the same surface current across the dipole strip at several places and you see that the current is not uniform but increased at the edges. The distribution across the port can depend on the shape of the port (width vs. length), the width of the dipole and the mesh quality. This distribution can be slightly different from distribution across the dipole width - port has different formulation than PerfE BC in terms of fields. Plot the mesh and surface current on dipole and port. I prefer to have at least 10 mesh elements across the dipole width and alnog the port.
I want to repeat the question: why current distribution is so important for you? What is the simulation goals? The model itself involves a lot of "ideal" features. Based on simulation goals, you may need to include more realistic features like finite thickness and material properties.