[jarvus lunt]

Sound like you need to model the river in question, the larger river it flows into, and then use a time series stage boundary at the downstream end the matches the tide.

As far as the tidal boundary itself, this is from a previous post:

"The key to using a tidal boundary with unsteady flow is to have some additional large cross sections at the downstream end. Essentially, just extend the reach out into the ocean (or bay) so that you have some large cross sections that have a greater area than the river cross sections. The tidal boundary is then applied to the downstream, large cross section. The reason for this is at the end of the river, the water surface does not match the tide level. The mouth of the river is influenced by the river flow. So the cross sections need to be extended out to where the water surface does match the tide."