So let’s talk about capacity expansion.
One of the big things that separates current thinking about freeway interchanges from 1950’s thinking about interchanges is the question of through capacity. Originally, the idea was that you could reduce capacity through the interchange because some traffic is getting on and off. So for instance you might have an eight-lane freeway which was six through the interchange area with one or two lanes cutting off to the crossroad and one or two rejoining. Texas does not have a lot of surviving examples of this setup (they tend to jam up, and we tend to fix that thing), but you can find it if you go elsewhere. Like Tacoma, Washington where I-5 is eight lanes throughout south Tacoma, all the way down to the Air Force base, but necks down to six lanes as the outer lanes leave for the 38th Street / SR 16 C-D roads. (WashDOT is actually widening this section as we speak so enjoy it while it lasts.)
The current thinking is that you maintain all the through lanes through the interchange area, and any additional lanes you pick up get spooled off one-by-one into downstream local exits. This is pretty clearly seen on the westbound Katy Freeway free lanes where you have four lanes going under Beltway 8, pick up another from the frontage/local feeder entrance, pick up another two from the BW8 interchange, then spool them off one by one into Kirkwood, Dairy Ashford, and then eventually SH 6.
The Downtown Houston freeway loop can really be thought of as one very large interchange rather than a bunch of separate freeways and their associated interchanges. For instance if I’m at the Market and I decide I want to get to the UH, I’m going to just jump on 45 and that’s that. But if I’m at say the TxDOT Houston District HQ and I want to go to the car wash on OST I’m lookin’ at 10 East to 45 South to 59 South to 288 South. I don’t see where one journey is appreciably more complex than the other; in both cases I got on one freeway, transited the Downtown Houston interchange complex, then ended up on another. But numerology suggests otherwise.
If you think of the Downtown Loop as one big ol’ interchange instead of a bunch of small ones it logically follows that the loop capacity should be slightly greater than the sum of the roadways entering it. So for instance if the Gulf/South/Southwest lane combination is 22 dropping to 18, the loop combination at that point should be… 20. This suggests adding capacity somewhere. But where?
45 and 59 are both adjacent to rapidly-developing areas in the Fourth Ward and EaDo. But the area around IH 10 is largely vacant. When IH 10 went through in the 70’s, there was an actual, functioning rail yard up there. Now there’s a heap of gravel and the ashes of a very grandiose yet low-capacity rail station plan. Conceivably, then, you could shift IH 10 to the north and open up more of the area north of downtown for other land uses besides bail bondsmen… something like the Buffalo Bayou Partnership’s North Canal proposal (PDF), for example.
And 45 and 59, while both in vital areas, have completely different contexts. The Pierce Elevated causes minimal interruption to the street grid beneath it, and in fact there’s a decent amount of pedestrian traffic there. The Downtown Transit Center is there. Any addition to that facility would have some serious negative impacts on the street life. But 59, on the other hand… well, just look.
Basically, the GRB kills any and all street life behind it. Even if you put on your best urbanist bona fides and depressed 59 with a park over the top of it, you wouldn’t change the fact that there’s a 2000-foot wall between EaDo and Downtown. What could it hurt, then, to add a second level? Especially if it was supported on wide, single-column bents which sprout through the median of the existing roadway. Then there wouldn’t even be any adverse lighting effect on Chartres.
And it’s in that context that maybe the Giant Traffic Circle really isn’t all that crazy an idea. For it to work, though, one has to drop the pretext of getting more use out of the existing capacity and instead use it as a framework for adding capacity.