open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi

I am trying to replicate a longshore current by imposing a gradient in the longshore water-level (LIHBOR = 5, LIUBOR = 4, LIVBOR = 4). Storm-tide water-levels and currents are available from a previously run (and validated) coastal 2D model hindcast. The resulting t2d modelled currents -forced by the gradient in the coastal model water-levels- looks good compared to the coastal model currents (attached), but I would like to improve the comparison. I would like to adjust/nudge the input water-level gradient at each time step within bord.f so that the difference between the T2D computed U and the coastal model U is reduced.
The problem I have is that at point p, U%R(p) in bord.f is empty #E-14 because the velocities have not been computed yet (?). How could I code-in the velocities at a point p from the previous time step?

Attached is a time-series plot, the black line is the imposed difference in longshore water-level (between two points separated by ~5km), red line is the resulting t2d U (m/s), green line is the coastal model U and blue dots are the observed U.



Thanks.

Julian

P.S I did try imposing the currents at the boundary (5,6,6 in the .cli file) but this over constrained the simulations, also tried the Thompson B.C (with 5,6,6) but it did not replicate a longshore current better than the just the longshore gradient in water levels (5,4,4).

Hello,

The velocity U contains the initial conditions or the values of the last time step when you are in bord.f, so I am not sure to understand your statement:

U%R(p) in bord.f is empty #E-14

However if you need the velocity of last time step, the components are in UN and VN, and you can "see" them in bord.f if you add UN and VN in the ONLY list of declarations_telemac2d.

With best regards,

Jean-Michel Hervouet

Sorry, user error, I defined point p at my first read statement, but didn't save p for future reads.
Thanks, UN and VN is what I was after.