Bedmass calculation along the boundary

[xkhmyx]

During the past a few days, I have been testing the conservation of sediment mass for the Northern Gulf of Mexico model.

The model domain is 64 (eta, longshore) x 128 (xi, crossshore), and northern boundary condition is NORTH_WALL. There are 12 rho points along northern boundary, and the Atchafalaya River delivers water and sediment through 12 v points (south of rho points).

In the initial conditions, all river bedmass is 0. Once the sediment is turned on, river sediment is deposited on sea bed. When I checked the RIVER bedmass on these northern-most rho points, they were calculated, probably based on flux divergence (?).

These RIVER bedmass on 12 rho points only represents sediment flux into the model domain, not actual sediment deposition on sea bed. If using bedmass (kg/m2)*area of grid cell(km2) (1/pm/pn) on pho points to test the mass conservation, one probably should EXCLUDE these 12 rho points along the boundary. Otherwise, sediment accumulation will be doubled.

Kevin Xu
Department of Marine Science
Coastal Carolina University
kxu@coastal.edu

[jcwarner]

Kevin-
I think the issue here is the boundary conditions imposed on the bed, bed_frac, bed_mass, and bottom arrays. Along the outer rows and columns of rho points, the values for these arrays need to be filled by boundary conditions. The model can not compute the flux divergence at those points to determine the sediment properties. So, .. what should go there. Well, the values in these arrays could stay at zero, or the initial value, or some other value. I think we chose a gradient condition for output purposes. The best way to deal with this is to just use the interior rho points to compute sed mass in the system.