Bathymetry smoothing in ROMS: A new approach

Mathieu Dutour Sikiric, Ivica Janekovic and Milivoj Kuzmic
Rudjer Boskovic Insitute

Terrain-following (sigma) coordinates are often the system of choice for both large scale basins and regional coastal applications. This coordinate system exhibits an intrinsic error in the discretization of the horizontal pressure gradient (HPG) term in areas with tall and steep bathymetry.

We present a new linear-programming (LP) procedure to minimize HPG errors. This new procedure has been applied to an Adriatic Sea test case using the Rutgers implementation of ROMS. Its level of success in alleviating the HPG error is judged by comparing against two implementations of the Shapiro filter, the Mellor-Ezer-Oey (1994) algorithm, and the recently proposed scheme of Martinho and Batteen (2006). The comparisons reveal that even an improved Shapiro filter creates very large modifications to the bathymetry and should be avoided. The Martinho-Batteen procedure’s performance is comparable to that of LP, when the slope factor (r-factor) is larger than about 0.2, and inferior to it at smaller values. The difference between the studied solutions are more pronounced for coarser grids.