Table time selection

[abkfenris]

Currently the table displaces the intersection of the forecast times. That's already been shown to not be great when the forecasts don't have times that line up. It also isn't great when a forecast is lagging, as only a small subset of the possible forecast will be shown.

Currently BIO is lagging, so there are only a few hours that both it and FVCOM have.

I'm thinking we may just want to take all (unique) times, and filter by those that are greater than the current time, maybe minus an hour.

Do we also want some sort of greater than hourly stride, so we can get more days in the view at the same time? @youngmorse

[abkfenris]

Here is how it looks when a forecast doesn't have up to date data with all times greater than an hour ago.