Flow around channel bends

Henderson, p. 256, says that subcritical flow in a channel can become critical along the inner wall of a channel bend.  A lab observation of this can be seen here. He doesn't mention if subcritical flow can become supercritical however.
Centrifigual force causes the flow to thin on the inner wall and thicken on the outer wall.  By mass conservation the flow accelerates on the inner wall.

Can subcritical flow become supercritical around a bend?

Chow p. 447 has a formula to find the depth and speed of fluid passing through the bend of a channel. The speed and depth across the channel depend on the radius of curvature of the bend and the channel width, as well as the specific energy E of the flow before it reaches the bend.

I wanted to find out if subcritical flow could become supercritical going around the 32 degree bend of Cape Mendocino. For subcritical upstream flow and a wide channel, I chose 10 km as the radius of curvature of the Cape Mendocino bend and used Chow's formula.
A plot of the variation in speed (top) and depth (bottom) across the channel is shown here.  Near the bend (onshore) is to the right, with offshore to the left. The flow is supercritical (green on lower plot) until 15 km offshore. The speed and depth before the flow reached the bend is shown by the purple lines. For higher Fr upstream or sharper bends, the depth of the flow near the wall can become negative.  The formula doesn't consider friction (which also makes subcritical flow more critical.)

Supercritical flow also thins and accelerates around a bend, with jumps possible. Chow, p. 439 says "subcritical flow shows smooth water suface and slight superelevation, whereas supercritical flow exhibits characteristic cross wave disturbance patterns on the surface and thus exaggerates the superelevation." (Superelelevation is the change in depth as the fluid thins on the inner wall and deepens on the outer wall of a bend.)  Jumps develop off of the outer wall which turns inward on the flow. The inner wall which turns away from the flow sends out an expansion fan.

I'd like to reconcile the fact that subcritical flow slows in a channel which widens (if you average across the channel), but accelerates around a channel bend at the wall due to centrifugal forces.   Is the difference whether the coastline redirects the flow (then you get centrifugal forces) or opens away from the flow, so that the flow must "fill in" the wider area?  Does the Earth's rotation confine the flow to follow the coastline, causing the flow to accelerate and thin?

Some model results

The COAMPS model (Burke et al, Mesoscale Simulation of Supercritical, Subcritical, and Transcritical Flow along Coastal Topgraphy, submitted) was used to find the momentum balance upstream and downstream from a coastline bend and found that:
  1. Centripetal acceleration as the flow rounds the bend is accompanied by thinning for sub and supercritical flow
  2. Subcritical flow can become supercritical for a big enough bend or for flow that starts close to critical
Audrey Rogerson's model (Rogerson, Transcritical Flow in the Coastal Marine Atmospheric Boundary Layer, submitted) showed
  1. Near-critical flow becoming supercritical around a bend.
  2. The supercritical region extends far offshore and downshore, while the width of a thickened/slowed region  of subcritical  flow at an inward bend was the Rossby radius.

Some questions

  1. Do we get secondary flow at a bend (flow away from the bend at the top of the MABL, and towards the bend at the bottom of the MABL) like in a river channel?  This flow arises because centripetal acceleration is constant through the layer, but friction near the bottom reduces the Coriolis force, creating a toward-bend flow at the bottom and a compensating flow away from the bend above.
  2. What controls the width of the region adjusting to the coastline change?  Why does subcritical adjustment seem to be confined by the Rossby radius, while the supercritical region is much wider?  Does a new lengthscale come out of the major terms in the momentum balance of flow rounding a bend?  ...of supercritical flow?
  3. Can I use Cherniawsk and LeBlond's analytical solution  (Cherniawsky and LeBlond, Rotation Flows along Indented Coastlines, 1986) to reproduce a supercritical region near a bend?  Why is this region's width ~ Rossby radius in his paper?