Lactate production doesn't shut down but continues to rise until it peaks out when do fairly frequent intense sprints, then it (lacate) starts to drop as speed decreases (due to fatigue). Thus, it would appear that buffering is stronger at this point and production is not as high, but this is false because it is the speed that has declined and with it quantities of lactate. Lactate is directly related to carbohydrate burning and hence quantitatively linked to intensity. If speed drops, so does lactate.
I can give an example of this at the other end of the spectrum. I can tell a runner to go for a long run and I will measure your lactates every 3 miles. Over time lactate values will drop gradually. The assumption, which is incorrect, is the buffering of lactate is improving. Truth is, either speed slowed down as one travels many miles, which effects lactate production - i.e. quantity - or there was a shift in fuel (substrate) utilization. If you burn more fat to generate ATP, you use fewer carbs. If you use fewer carbs (glycogen or glucose) you produce less lactate.
A runner can never sprint-float sprint 2 miles on the track and achieve the same time as running at a steady, fast pace. Why? Sprinting and floating in alternation produces far more fatigue due to very high muscular and blood acidosis.
I think the over-riding benefit of surge training (sprint-float-sprints) is multi-faceted: neuromuscular, biochemical, and mental. Changes of paces improves neural control and firing. Changes of paces requires concentration. Surge training for 2 miles on the track is very, very hard to do and it toughens one's mentally. Additionally, physically, it recruits fast twitch fibers and thus demands anaerobic contribution at a high level of output per unit time. It is a very strong stimulus, but if not tempered with overdistance running, plenty of post-speed training cool down running, morning jogs, it can flatten you in a hurry.
One must really know how to coordinate training elements in order to do it well (like Lydiard). One must have plenty of aerobic conditioning in the "bank" to do it well and benefit from it. It also is a strong aerobic stimulus for fast oxidative fibers (fast twitch A fibers) and fast glycoltic (explosive)(fast twitch B) fibers. Through frequent use, fast twitch B fibers can alter their constitution such that they process oxygen to create ATP energy, not just process pyruvate to generate ATP anaerobically. If one doesn't have many of these fibers, their isn't going to be a big improvement with this type of workout as far as biochemical processing of energy is concerned. Thus, if one is a "slow twitch" runner (little natural speed), then an even shorter workout of surge training should be planned. A long surge training session is not advisable for a slow twitch runner, in my opinion. It will work magic for middle distance runners with great natural speed, I have noticed. Tinman