On the flat disc model, east and west curve toward north. A train sufficiently long going on a track due east or west should therefore have a perceptible difference on one side (convex south side versus concave north side) if viewed from the end. Discers have to prove a different curve. In Canada there's a very long stretch of East to West highway that could be used for this. You could use a sufficiently large laser to see if the east to west highway deflects toward North, which would prove the curve required for the Azimuthal Equidistant Projection to represent an actual flat earth disc.