Rupturing the creeping segment came up at the Sept 2012 Southern California Earthquake Center meeting. I think the context was the Indian Ocean M8.6 strike-slip quake. Yes, it could be a good idea to trench the creeping section. But I think the odds of it rupturing through are really tiny (these are just opinions; I don't work on this). It seems more likely that a southern quake that stops near Parkfield would trigger a northern quake if both sides were ready to go, than rupturing through.

The reason is, the creeping section is not accumulating any measurable strain. A talk by Chuck de Mets (sp?) in November and by someone else after his showed GPS data that show the flexing of the crust across locked sections of the fault, and a step function in velocities across the creeping section. The fault is just sliding with no flexing.

New Madrid zone also is not accumulating strain and it is the middle of a big bulls-eye for earthquake hazard on the national maps. Liquefaction features like sand-blows show repeated quakes somewhere in the region. But this is different: the blocks on either side of the zone are not moving with respect to each other (within errors of the GPS). The faults could be locked. The hypothesis that justifies the hazard maps is that only part of the differential/deviatoric stress is released with each quake (I don't understand how it is possible that you get some stress drop with a quake and then can get another quake without rebuilding the stress).

Chris