As mentioned earlier, the reason you get afternoon backups in both directions on the Pierce Elevated is that 45, 59, and 288 comprise 18 lanes entering the Downtown Loop while the capacity of the Pierce Elevated and 59 combined is only 14 lanes.
Could the “traffic circle from hell” address that? Maaaaybe. By taking out the median on the Pierce Elevated, for example, you could get seven lanes instead of six. But how would the exits work? A decent chunk of downtown-bound traffic leaves the radial freeways before they interface with each other. For instance, if you’re coming in from Sharpstown, you get off at the Spur. But a number of exits exist within the downtown loop. How would those work?
Basically, with one-way operation you need contraflow lanes to handle the following four movements:
—Traffic from 10 EB and 45 SB desiring to get off at McKee
—Traffic from St Joseph/Pease desiring to get on 10 WB or 45 NB
—Traffic from San Jacinto desiring to enter 10 EB
—Traffic from Walker Street desiring to enter 45 NB or 10 WB.
Realistically, you could probably delete Walker and send that traffic up Louisiana or Travis. But then if you already have contraflow to handle the Pease/St Joe traffic it’s trivial to add Walker to that. Likewise, if you have contraflow to allow 45S/10E traffic to exit McKee, and more contraflow to allow San Jac to enter 10E/59N, there’s basically no reason to divert the IH 10 EB mainline flow down to the Pierce Elevated. Moreover, the length of the diversion – a full 3 miles – suggests that if you *did* construct it like this, eastbound traffic would exit McKee, double back on Sterrett or Providence, then get back on at San Jac.
So even if you start with the premise of “downtown traffic circle,” realistically the only one-way flow you get is 45 from Brazos to 59, and 59 from 45 to Ruiz. Which is to say, 2.2 miles out of a total loop length of 5.7, or less than 40% of the total mileage.