I would believe that CAESAR assumes a vacuum, for which air is a good representation.

While you can still use CAESAR to estimate natural frequency in a vacuum, you'd need to correct this value to meet that of water.

Here's a link that describes what you need.

https://www.comsol.com/blogs/natural-frequencies-immersed-beams/

I've also attached a pdf of the article, in case the link becomes broken.