Fix all of Flint's pipes

The mayor of Flint, Michigan has announced a plan to replace all of the town's water pipes for $55 million. Considering that Michigan's governor, Rick Snyder, has estimated the total cost of ameliorating the Flint water catastrophe to be $712 million (per Rachel Maddow), $55 million seems like a relatively modest sum. If it won't make the town whole, it will certainly get it a long way toward that point, I would think.

Anyway, Flint's mayor wants to prioritize the pipe replacement:

Retired Brig. Gen. Michael McDaniel, who was tapped by [Flint mayor Karen] Weaver to oversee pipe removal, said the project will begin with high-risk households with children and pregnant women in neighborhoods where the highest levels of lead were found in kids’ blood.

I must be missing something, because I don't understand this.

To fix an entire water system, you have to start at its source and move downstream. That way, each successive fix keeps clean water, clean. You also must ensure that water in the system never moves "upstream" from dirty old pipes to clean new ones.

If all you had to do was to fix pipes in a "downstream" direction, then I could see a clear path: you would first fix the big distribution pipes, the ones that carry water to hundreds and thousands of houses, then you would fix the "last mile" pipes into individual houses on the priority basis the mayor described. But water only flows when somebody opens a tap. The rest of the time, it sits around in the pipes. To my knowledge all that water in the pipes is part of one big notional pool. The water doesn't move backwards in any great volume, but neither do contaminants remain isolated in those "downstream" old pipes. Those contaminants at minimum could spread out to other homes, including ones with new pipes; they might also damage the new pipes.

If I were a Flint resident, and this plan went into effect, I'd refuse to use the water coming out of those pipes until I had unequivocal proof that every inch of pipe in the town had been replaced and tests had showed the water running clean in multiple samples, over time, throughout the town.

An incremental-replacement plan that makes more sense to me would replace pipes neighborhood by neighborhood. Since contaminants can only spread slowly, by osmosis, you could be confident of decent-quality water once enough pipes had been replaced beyond your neighborhood to provide a "buffer", beyond which contaminants wouldn't have time to spread before the dirty water was drained by ordinary use (e.g., for landscaping). The downside of this plan is, it requires dirty water to be flushed out of home pipes through ordinary use, and what sane person in Flint would keep using this water for anything? Even landscape watering would create further problems by allowing lead to enter the soil (and eventually, ground water). Even this incremental-replacement plan, then, is problematic at best.

"Priority" replacement sounds good but it doesn't make sense if you think about it. Unless Flint takes extraordinary steps to isolate new pipes from old ones (and I'm not sure that's even possible), nobody should use the system's water until every pipe has been replaced.