Placing a mylar sheet between, and in compressive contact with, the ground and a sleeping pad provides no additional insulating value.
Placing the mylar sheet between a pad and sleeping bag also does nothing. The only reason manufacturers can show a numerical difference in R values of diffuse-surface closed-cell foam pads and closed-cell foam pads with a reflective coating on the surface ... is just owing to a quirk of the testing setup. The standardized test involves a heated flat plate applied to the pad with x pounds of force - the flat plate creates meaningful voids when compressed against a corrugated closed cell foam pad, and reducing radiation heat transfer at these air voids does make a measurable difference in a laboratory environment. It's an air void. Not contact conduction. A real life application reduces the thermal radiation resistance to near zero.
Winter rated air-core mattresses work because there is a physical gap between the mylar surfaces. Ideally, this gap would be a vacuum. That's why they are called space blankets. With a pad, you have a convective air gap, and convection is a very finicky thing.
Turn the ASTM F3340-18 test upside down, literally, and the air-core mattresses will have a lower R-Value compared to the standard test. Put a box fan in the lab, and the air-core mattresses will have a lower R-Value compared to the standard test, whereas a closed-cell foam pad will show little to no change; depending on the design, self-inflating open-cell-foam air mattresses can be largely immune to outside air convection, like the closed-cell pads, or they can be degraded by this convection, just depends on the geometry of any weight-saving cutouts made in the foam.