That’s a very interesting read, thanks! I took my data and plotted it in the same way as represented here, this is actually even more telling.
Balanced diet and structured base:
Carb extravaganza 
:
I think it’s fair to say but even when I’m specifically training fat adaptation, I’m still very much carb reliant, but neglecting this training makes for a very extreme example.
This actually explains very neatly what happened to me this year during the Maratona dles Dolomites. I assume I have about 1500 kcal worth of muscle glycogen on board, I fueled with 80g carbs an hour (so that’s 320 kcal an hour) plus my pre-race breakfast was 120 g of carbs (480 kcal).
First 4 hours I felt great, I was doing a pace that felt really easy and I was planning to speed up in the last part: (notice the TEE at 2852 kj)
Around the 4 hour mark I started feeling gradually worse, I somehow battled through the climb, stopped for a few minutes and ate some more food, though the descent would help me recover, but I still crashed and burned at 5 hour mark (at 3496kj):
1500+480+(320x5)=3580kcal
Sorry for the long anecdotal plug in, but it kind of clicked for me now.
The testing was done in a training center in Amsterdam, Netherlands. The equipment they are using is Cosmed Quark with Omnia Metabolic module. The protocol is a graded test very similar to TR Ramp test actually, the full protocol name is GLI (Spirometry), Wasserman extended. Not sure if it helps but maybe good to know 
Your local training center will have a different soft/hardware, but that should not affect the results as long as it’s consistent. What’s important is that they are able to give you the raw data to play with.