I think it is as follows:
In real coil packs, the conductors have a finite thickness (down into the plane of Fig 5, and horizontally, in Fig 6)
In Biot and Ampere's laws, wire can be assumed to be
infinitely thin so that the B at a distant point from the contributions of wire lengths dl along the wire length can be solved analytically as a simple first approximation.
So Fig 6 is being solved analytically for (Bx,y) as if the coil packs are infinitely thin, lying in the vertical plane shown in Fig 6 ( along the blue vertical axis) Fig 6 has rotated Fig 5 by 90 degrees.
They call the vertical blue line "equivalent plane" of the coil packs.
Then they get (9) and (10) contributions analytically using 8
In a 2D or more accurate 3D FEM, the real thickness of the packs would be drawn, and the numerical solvers would consider the J in elements inside the copper conductors across the thickness.
Back in early 20th C they were solving jobs like Fig 5 and including the finite thickness, analytically with slide rules, log tables and adders.
Eg Dwight H B "Coils and Conductors" 1945