None of you answered the question.
To the person who is quibbling over the significant figures - I say that it does not compromise accuracy to carry these digits throughout your calculations - as long as when you present you findings in report format, you have rounded the values to the reasonable amount, otherwise he's right, you do look like you're fooling someone.
I looked at the model. A few things stand out.
When viewing the geometric data, your cross sections cross one another, making it appear impossible to have a positive reach length between one section and the one downstream of it. Also, this makes your bridge deck look like it is getting skinnier as you go across.
That being said, HEC-RAS doesn't "see" this picture in the calculations and as long as you do have positive reach lengths reflecting the distance downstream to the next section on the left overbank as well as the right, then you'd be fine. But it looks like you didn't reflect this properly:
If the stream curves as it appears to in your picture, then your LOB and ROB reach lengths should not be equal to the channel. What you have modeled is a very straight stream with very simple and common n-values. This is not usually the case and therefore will give you the wrong answers. When stream curves to the right, the LOB reach length should be the greatest, then the channel, then the right.
Next: Boundary conditions. For subcritical flow, you need at minimum one boundary condition downstream. You have used critical depth as both conditions. This is usually because the user doesn't know what else to do and if you use critical depth, you need nothing else. Problem is that the calculations are extremely dependent on the downstream conditions. Only if you have thousands of feet of stream between the downstream boundary conditions and your point of interest, can you put any confidence in your output.
You need a FEMA FIS study or something from you local environmental agency that tells you the elevation of the 100-year storm at your downstream-most point. Without this, you can estimate the elevation using normal depth. You would use the channel slope, which equals the energy slope for uniform flow. Put this in as the downstream boundary conditions for normal depth. Then HEC-RAS will compute the elevation and go from there. Take the upstream condition out unless you think you will find supercritical flow.
If you change those few things, I think your model may give you more reasonable-looking answers.
One more thing: the EGL was not coinciding with the WSE. When you zoom in, they separate.
