Using your 1" base for convenience, a tube, seen from the side, would always look 1" wide. A square block, seen from the side, would vary from 1" wide (as seen looking straight into the side) to sqrt(2)" wide (as seen looking straight into a corner). So, it depends on the wind direction, but the cross-section presented to the wind will be, on average, in the neighborhood of 1.2 times as much for a 1" square as it would be for a 1" circle. That, of course, leaves out any turbulence effects of the airflow in the neighborhood of the corners, etc. I'm guessing there would be some negative effects from the discontinuity of the curve at those boundaries.