Ouch.
Your mileage seems to be pretty good - close to 70 fairly often, right?
For me, there were three big differences in my training when changing the emphasis from 5k/10k to the marathon.
First, of course, my overall volume went up. I went from a 50 mpw runner, to maxing out anywhere between 80 and 100 mpw. And my long runs go from 12 to 16 miles to 18 to 20 miles.
Second, the emphasis on my quality days shifted from the shorter, more intense intervals that you do now, to longer slower tempo efforts. I still do typical vo2max intervals from 18 weeks out until 12 weeks out, but other than races, in the last 12 weeks of my training cycle for a marathon, I will never have anything on the books faster than tempo pace (although, as a practical matter, that tends to trend toward 10k pace fairly often). Often, those tempo efforts are integrated into long runs.
Finally, I try to focus on glycogen management. That means that I try to get in as many runs over 10 miles as possible. My reasoning for that is that we store about half of our glycogen in our blood stream and liver, and about half in our muscles. During the marathon, you transition from using one source of glycogen to the other Since the marathon is as much about efficient use of energy as it is about aerobic fitness, I really want to become efficient at making that switch from one source of glycogen to the other. So running for long enough to cause my body to switch sources becomes an emphasis as presumably that will lead to some adaptation on that front. From what I understand, that switch occurs somewhere after the 1 hour mark of running, thus the emphasis on lots of 10 milers that you see in my training log each week.
In fact, if you look at each of the three points I emphasize, they all have to do with burning glycogen - (I) my long runs get up over the 2 hour mark (at which point I either burn through or nearly burn through all of my stores); (ii) my integrated long run/tempo efforts also result in a relatively glycogen burn (and allow me to get used to running hard during the time when my glycogen stores are about to be depleted, and certainly after the point at which the source switch occurs); and (iii) I avoid doubling so as to maximize the number of times I make that switch over (such that it should, in theory, happen on as many as 25 or more of my runs each month).
I guess what I am getting at is that it may not so much be a lack of overall volume (although your mileage is lower and your long runs are shorter than one would expect of a runner your ability level at the marathon distance), or going out too fast, but rather a lack of emphasis on glycogen management.
The other interesting thing is that it sounds like you bonked at about the 1:42-1:45 mark of your first marathon. Most folks who bonk tend to bonk a bit later - at about the two hour mark (I bonked at about 21 or 22 in my first marathon). Did you do a good job of carbo loading for that first race? Did you have and implement a nutrition plan during the race? I really emphasized that for my second marathon and avoided bonking as a result (although I did go out too hard on a warm day and had really bad calf cramping late in the race- which now has me thinking through using salt tabs for my third marathon in a couple of months).
So, if you decide to race this weekend, I would really emphasize carbo loading and in-race nutrition (and stick to the plan, even if you don't think you need that last gel - once you actually feel like you need it, it is too late). The other stuff is just food for thought should you decide to emphasize the marathon in a future training cycle.