On a clear day, San Francisco's runways 28L and 28R look like a textbook parallel pair. On a foggy day — and SFO gets roughly 200 of those annually — they're something closer to a single runway wearing a disguise.
The FAA has banned simultaneous parallel approaches at SFO, and delays have quadrupled as a result. But this isn't a safety story. It's a geometry story that was always going to end this way.
Standard FAA rules require 4,300 feet of separation between runway centerlines for independent IFR parallel operations. SFO's 28L/28R are separated by approximately 750 feet. That gap — less than a fifth of the required distance — meant the airport was never legally entitled to run simultaneous arrivals in instrument conditions under normal procedures.
What it had instead was a workaround. The Simultaneous Offset Instrument Approach, or SOIA, was a precision choreography: one aircraft flying the ILS straight in, the other flying an offset course and breaking off at a specific point. It required compatible avionics, specialized crew training, and a specific weather ceiling to execute safely. Remove any one variable, and the procedure collapses.
The FAA's ban removes the procedure entirely. Now, in low visibility, SFO operates its two arrival runways sequentially — effectively halving its arrival throughput.
The arithmetic is brutal. SFO handles 450 to 500 flights daily. When one arrival stream disappears, approach controllers must stretch miles-in-trail spacing, absorption queues back up over the Pacific, and the delay wave rolls east into every connecting bank from LAX to Denver.
The structural irony is almost elegant. SFO is a fog airport — built on a bay, named after a city defined by marine layer — that constructed its entire peak-capacity model around a fog-dependent procedure. The workaround worked, until regulators decided it didn't.
SFO didn't lose capacity when the ban landed. It lost the illusion that it ever had it.