Thanks. Yes, I should have clarified, I misspoke in my OP. The first sleeper was empty except for staff. The other two had a handful of empty berths. And what was notable was there was almost no wait list for meals. Ordinarily there is a substantial wait list to work through for every meal.
There is a plausible explanation that people canceled or didn't book due to the high number of train delays this year. On the other hand, many travelers I would assume make their summer plans for rail travel in winter and book in the spring (at least those planing on cross country sleeper berth travel).
For the last eight years, the #5 & 6 have been reliably fully booked in June and July when we travel. This was a substantial difference, on the order of 40% or 50% less full than normal - in both directions across a week. That will significantly impact ridership, revenues, and diner usage if it reflects a trend.
As I mentioned above, we are feeling a little paranoid re Amtrak. I have some familiarity with the type of shenanigans that can be used to create self-fulfilling data points about business operations. So it is plausible that Amtrak management's efforts (and non-efforts) are showing up in usage.
If I was trying to kill long distance rail service, I would want to drive down usage statistics to reverse long term increasing trends (an inconvenient data point that contradicts my story). I would also need to drive down the revenues from sleepers and diners, to justify reductions in those high-touch services. A precipitous drop in overall revenues and cash flow would be handy as well.
The season's floods, storms, heat and accidents is creating a perfect storm to help make the case that Amtrak's CEO wants to make. Taking his foot off the gas on marketing and service, plus a price mismatch w/r/t alternatives could really put the hurt on revenue and ridership.