I had intimate knowledge of a couple of local running events where the courses were impeccably certified. One event was a 5mi/10mi race. The other event was a 10K/half marathon/marathon combo. In both cases...all the races started from same exact place...usually at the same time...and all the races used the same finish line. The problem was the certified finish line was a bit different for each distance. You could see this when you viewed the certified maps.
In both events, the race organizers did not use the multiple certified finish lines...and used one finish line that either split the difference...or made the most important race accurate... while letting the others be off.
Year after year, the 5mi/10mi race had a 5 miler that was about 50-60 meters short while the 10 miler was 80-100 meters long. The other race made the marathon exactly right and let the half marathon by 80 meters long with the 10K being very long (~250-300m?).
This went on for quite a while. Maybe that's what happened here? A course can be certified but you need to actually run the certified course for it to be accurate.
Still seems a bit unfair to deny the trials qualifiers entry as they did nothing wrong. But extrapolating the results could be one way to correct things...assuming they can figure out how much to adjust.