Firstly, I wish Zach (that I never met) to run very fast marathon, because I respect every athlete who train hard and is motivated with facts, not only words, for reaching his goal.
So said, I want to do a technical comment.
1) Every distance inferior 3/4 of the total distance NEVER CAN BE CONSIDERED SPECIFIC TRAINING.
2) 16 miles are 25,750 km. and at the speed of 4'47" per mile means a final time of 1 h 16'30" about. We are still 50' far from the final marathon time, and only at the beginning of the total consumption of glycogen.
3) If the training is like this, can't be a prediction for the marathon, and is because these kind of training are considered "specific workouts" that still we face the "wall" after 30 km.
4) If an athlete want to try his best in a marathon, during the preparation (lasting not less than 3 months) needs to run at least 3 times 40 km, one every month, with the last (about 4 weeks before the race) at 96-97 % of marathon pace.
5) I give some example of the last long run we used in Kenya with European athletes because their PB : for the group running Berlin (Amanal Petros 2:05:58, Tadesse Abraham 2:05:10, Boniface Kimutai 2:05:05 coming from 2:09:50, plus other Kenyans running under 2:06) the last 40 km run in Kaptuli (2200m altitude, hilly but good tarmac) happened 32 days before the race, and was between 2:06:30 and 2:07:10. I think the difference per km between Kaptuli and sea level is about 3-4 secs per km, this means the same workout at sea level could finist between 2:04 and 2:05, practically the time of 40 km is like the final time of the race:
6) 40 km is 94,78% of the total distance, this means that the level of the effort of the last long run is about 94-95% of the total marathon.
7) With Sondre Moen before the European Record in Fukuoka in 2017 (normal shoes) we had a long run in Nangili (20 km going + 20 km coming back, first part uphill the return downhill) in 2:06:20. At that time, I used to calculate a difference of 6 secs per km comparing with sea level, not because the altitude only (like now), but because 3 further secs depended on the surface (rough road very bad, with bad grip). This means like 40 km at sea level in 2:02:20 (3'03"5 / km) that is the 97.4 % of the speed of the race (in Fukuoka ran at 2'58"8 / km).
8) Recently we saw the training of Kiptum, around 270 - 280 km per week during the last 3 months, including one session per week (starting with 30 km) at 95-96 % of Marathon Pace. The evolution of this training, after running 30 km 3 month before the competition, is the EXTENSION of the distance, maintaining the same speed. So, don't be surprised if, instead finding the "wall", Kiptum after 30 km always has the ability to increase the speed.
9) In every marathon of today the winners use negative splits, because still have a lot of fuel after 1 and half hour of competition. This because with the continuity in the preparation with a lot of workouts longer than 30 km they become able to use, for the same speed, a mixture of fuel increasing every time the percentage of fatty acids, that everybody has in big quantity, while the tank of glycogen is very much smaller.
10) At the end, remeber my mantra : "If you want to run in competition long (like the marathon) and fast, you can't prepare with long and slow combined with short and fast, but sometimes need to run, in training too, LONG and FAST".