Two vocabularies exist for the same problem, and specifications get written in both.
What shear stress is
Bed shear stress is the tangential force per unit area that flowing water exerts on the bed. It is what physically dislodges material.
For steady uniform flow it is commonly estimated from depth and energy slope — deeper and steeper means more. That makes it a function of the flow’s geometry and gradient rather than of velocity alone.
Two sites at the same velocity can produce different shear stress if their depth or slope differ, which is exactly why the two vocabularies are not interchangeable.
Why velocity is used anyway
Because everyone has it.
Velocity comes out of every hydraulic model, it is intuitive, it can be measured, and permissible-velocity methods are long established and embedded in agency guidance.
Shear stress is the more fundamental quantity for describing what moves bed material, and it is less commonly the thing on a drawing.
Which approach suits which problem
Permissible velocity methods suit situations where the flow is reasonably uniform and comparable to the conditions the method was calibrated on: channels, revetments, straightforward reaches.
Tractive force or shear stress methods suit channel design where depth and slope vary along the reach, vegetated and lined channel design, and cases where the geometry differs substantially from the calibration basis of a velocity method.
Neither works well at a structure. At a pier or an abutment, the mechanism is a three-dimensional vortex system rather than a uniform boundary layer, and both approaches are approximations. That is why countermeasure design at structures generally uses relationships expressed in terms of the approach condition with a location factor, rather than a first-principles boundary shear calculation. See depth-averaged vs local velocity.
When a specification is in one and your data is in the other
This is the practical question, and it comes up regularly.
Conversion between them is possible in principle, using the relationships that connect depth, slope, roughness, and velocity for the flow condition. It is not a lookup, and the conversion carries its own uncertainty.
The workable approach:
- Establish what the specification actually requires — a permissible velocity, a permissible shear stress, or performance under a design condition
- Establish what your hydraulic data gives you, with its attributes. See taking a design velocity from a model
- If they differ, do the conversion explicitly and document the basis, rather than converting silently
- State the assumption, because a reviewer will want to know
What does not work is treating a shear stress figure and a velocity figure as if one implies the other without the intervening geometry.
What the product data is expressed in
The rated conditions for each size are published on product specifications and applied by the rock bag size calculator, and this article does not restate them.
That is a deliberate rule across this cluster: the site keeps rated figures in one place with their conditions attached, and copying them into method articles is how a number ends up separated from what it depends on. If you need the per-size figures, take them from the specification pages.
If your specification is written in a quantity the published rating is not expressed in, raise it at specification stage rather than at submittal. See writing a spec that stays open and getting a non-standard product approved.
What neither approach covers
Both are about initiation of motion of the armor. Neither addresses:
- Winnowing of fines from beneath the armor, which is a filter problem
- Undermining at the edges, which is a termination problem
- The bed continuing to lower across the reach, which is a degradation problem
Those three end more installations than under-sized units do. See why scour protection fails.
An armor design that gets the unit sizing exactly right and the filter, toe, and terminations wrong will fail. The reverse is much less true.