Local Scour Depth at a Bridge Pier (HEC-18)

Scour is the leading cause of bridge failure in North America. Not overload, not fatigue, not collision — the river digging the bed out from under a foundation. The Schoharie Creek collapse in New York in 1987, which killed ten people, was a scour failure at a pier, and the national scour evaluation programme that followed it is why this equation is in every bridge engineer's hands today.

The equation is the Colorado State University pier scour formula, recommended by the Federal Highway Administration in Hydraulic Engineering Circular No. 18, Evaluating Scour at Bridges. The mechanism it describes is worth picturing. Flow approaching a pier is stopped at the nose, so the velocity head converts to pressure, and because the approach velocity is greater near the surface than near the bed, that pressure varies with depth. The imbalance drives a downflow at the pier face. That downflow strikes the bed, rolls upstream, and organises itself into a horseshoe vortex that wraps around the pier and trails downstream — and the vortex is a very efficient excavator. It digs a hole, and the hole deepens until the vortex, now further from the bed, can no longer lift material out of it.

The factors follow from that picture. Deeper approach flow and a wider pier both make a stronger vortex, hence the $y_1$ and $a$ terms; a higher Froude number means more energy to convert, hence $Fr^{0.43}$. $K_1$ rewards a nose shape that separates the flow less. $K_2$ penalises a pier skewed to the current. $K_3$ accounts for the bed condition, and $K_4$ allows a reduction where coarse, well-graded material will armour the bottom of the hole and stop it deepening.

Two things about this equation that a user needs to hold in mind at once. First, it is DESIGN-CONSERVATIVE by intention. It was fitted as an envelope over flume data so that it would rarely underpredict, and comparisons against field measurements show it overpredicting more often than not. For a structure whose failure drops a span into a river that is the right bias — and it is exactly the wrong bias if you are trying to explain a hole someone has surveyed, or to back-calculate the flood that made it. Second, this is LOCAL scour only. Total scour at a bridge is long-term degradation of the whole reach, plus contraction scour through the constricted opening, plus this local hole, and the three stack. A foundation designed against the local component alone has been designed against perhaps half the problem.

The factor that most often goes wrong is $K_2$. A pier aligned with the flow gets 1.0; a long pier at a substantial skew can carry 3 or more, which swamps every other term in the equation. Rivers move their thread of high velocity from flood to flood, and the alignment measured on a calm survey day is not necessarily the alignment at the design flood. The factor most often abused is $K_4$, because it is the only one that can reduce the answer: HEC-18 permits an armouring credit only for coarse, well-graded material with a median size of at least 2 mm and a coarse fraction big enough to armour, with a floor of 0.4. Claiming it on a sand bed is claiming a protection that will not appear. Worth knowing too is the practical ceiling: local scour at a round-nosed pier aligned with the flow saturates near 2.4 pier widths regardless of how deep or fast the approach flow is, and HEC-18 allows that limit to be applied.