[UPDATE: As pointed out in the comments, riders who access Link by bus will generally not be "new to transit", but will not count among the 1.2m "lost" from buses. For reasons described before, I don't think this is a particularly big group. Nevertheless, we should add this to the 7/36/39/106 riders described below and consider 5.8m rides to be a ceiling on the new riders figure.]
Towards the end of this Erica Barnett takedown of the Washington Policy Center was this tidbit:
Metro attributes [reduced ridership] to several factors, including the loss of 87,000 jobs in King County during the recession, four fare increases in four years, lower gas prices, and the opening of light rail, which lowered ridership by 1.2 million rides a year.
That’s out of a total of 6.96m rides on Link in 2010, implying over 80% of Link trips are new to transit. I asked Metro Planner Jack Lattemann, the source of the figure, how this quick-and-dirty estimate was computed, below the jump:
About 1 million annual riders were on former Route 194 between Sea-Tac Airport and downtown Seattle that has been replaced by Link. (Another 250,000 annual riders who used to ride Route 194 in Federal Way are now counted as part of ridership on Sound Transit ST Express routes 577 and 578.)
The remaining 200,000 annual riders were those who rode portions of former routes 42, 42X, and 48 along the M.L. King Jr. Way South corridor between South Henderson Street and Rainier Avenue South. About 60% of the riders on these former routes were within 1/4 mile walking distance of one of the Link stations along M.L. King. As Route 8 continues to gain more riders along M.L. King over time, this figure will shrink.
(I have not had time to analyze the fall 2010 data at the stop level so I can’t say yet to what extent Route 8 has made up some of the ridership loss.)
This estimation problem is hard because there are buses like the 36, 39, and 106 that prior to Link opening crossed the tracks and headed downtown. Link also pulls some riders off the 7. I’d guess the true number was a bit higher than Lattemann’s estimate, but it’s a good start.