I think most of it is reaching the river via the tributaries, many of which are mined right to the river banks or even destroyed. If I look at all of our evidence, it looks like the worst contamination occurs during the first few years after a watershed is exposed, and that's very common in watershed disturbance. Any chemical that's in the geological substrate increases dramatically once the surface biological layer is removed, and then the amount of contamination tails away with time.
On March 30th, 2010. See this statement in context.