Let me start by stating the obvious: we’re all talking about achieving warmth from a comfort perspective, not for any survival purposes. With this in mind, let’s think about what it takes for us to be comfortable and how our bodies provide feedback to us. I’m only going to focus on the heat element of comfort and ignore anything else like humidity, air movement, bad dreams etc.
From my own experiences, I’ve found that it’s not as simple as looking at heat transfer from my body as a system. According to the EN standard, the Lower Limit is the temperature down to which a standard male is in thermal equilibrium, curled up, but not shivering. My body as a system can be in thermal equilibrium but if a certain body part falls below the “coldness threshold”, I’m going to be uncomfortable. This is exacerbated at night when my metabolism drops and my circulation lessens compared to when I’m awake and active. I’ve had nights where I feel the heat getting sucked out of me by only the ground (very low R-value pad), and nights where I feel the heat getting sucked out of me by only the air above — in both situations, I felt cold. What this tells me is that there’s a certain RATE of heat transfer from different parts of my body (my back, my chest, my arms and legs) where I’m comfortable, and outside of that range I’m not. My body disagrees with over-supplementing an area that’s already warm to offset another area that’s cold. Your mileage may vary.
With ground insulation, a simplified way of thinking about heat loss is that the greater the difference between the ground temperature and your body, the greater the rate of heat transfer away from your body. So the colder the ground, the more R-value you’ll need from your sleeping pad to decrease that rate of heat transfer to a “comfortable level.” Anecdotally, I’ve used my NeoAir XTherm (R-value 5.7) in conditions from consecutive -5F winter nights to mid-40F summer nights. It’s never been too warm for me but sometimes it’s just barely enough for a frozen ground.
A paper by Kalev Kuklane shows the correlation between sleeping pad R-value and the measured warmth of the sleeping bag. If the EN standard tested sleeping bags using a pad with an R-value of 3.0 (what most sleeping pads tend to be at currently), then those temperature ratings would be higher than if they used a pad with an R-value of 4.8 (like they use for calibration).
Aside: I find it hard to be precise with language when talking about this notion of comfort and how we achieve it through two independent pieces (sleeping bag + pad). Not only that, but we have different units of measurement for bag insulation and pad insulation. Pads are measured in R-value, while sleeping bags are represented with an air temperature rating. Even if we started measuring sleeping bags in terms of clo (a similar idea as R-value with the same units) or created a new unit of measurement for the entire system, it’s a challenge because us ground-sleepers interface with two different mediums (the air and the ground). McCullough 2009’s paper suggests that “the insulation of the system components is not additive,” which further complicates the analysis. It’s easy enough to measure air temperature but who’s sticking a probe into the ground to determine soil temperature, density, and composition? Despite this lack of detail, user experience has resulted in acceptable pad recommendations for winter (R-value around 6.0 or greater) vs. summer conditions.