Underpredicted tidal amp: improvements with 3d instead of 2d?

Message

stef · #1 Unread post by **stef** » Fri Aug 20, 2021 2:27 pm

I'm modelling tides with 2d configuration in the north-eastern Bay of Bengal, as a precursor to a storm-surge modelling exercise.

A figure of the northern part of the domain (Ganges-Brahmaputra-Meghna Delta):

The white contour is 10m. The labels are UHSLC tide gauge stations, I did harmonic analysis on some of them:

The problem is that the amplitudes at Chittagong and Charchanga are much too low.I think in the wide shallow shelf, there is too much dissipation.

I'm using

quadratic bottom friction
8km grid spacing with a minimum depth of 5m (extracted from GEBCO 2020)

I have tried lowering bottom friction in parts of the domain (towards the north), but to get even close to the observed amplitudes, I have to use extremely low values that will cause numerical instabilities.

So far I tried to reproduce the results of the following paper, which uses a similar configuration as I, except for the (potentially important) difference that they have seemed to mask out the very northward end of the bay between Charchanga and Chittagong. This is not explicitly mentioned in their paper, I'm just guessing from their figures:

Sindhu, B., Unnikrishnan, A.S., 2013: Characteristics of tides in the Bay of Bengal.
https://doi.org/10.1080/01490419.2013.781088

Do you think that going from 2d to 3d could solve this?

Do you have literature suggestions for local effects of 3d vs 2d in very shallow regions, and the effects of tuning the bottom boundary layer (or whatever appropriate) that are particularly relevant to my case?

Thanks a lot for your help!
Stefan