Soon it will be time to remove the turbulator bar.
I think I figured out the step from 36 to 37 mpg: around mid-February I adjusted tire pressure, and overshot a bit, going from about 38 psi to 45 psi. I just checked, it's still at 45 (front 44, rear 45 psi).
I think one can use that to get an idea of the fraction of the total energy spent due to rolling resistance:
Tire pressure went up by 18% (45/38~1.18). That shortens the contact patch by the same amount, and the radial tire deformation (assuming constant contact patch width and vehicle weight) by 40% (1.18^2~1.40). Assuming (to first order) that the energy spent on deforming the tire goes linear with tire deformation, that part of the total energy loss per distance went down 40%. This change corresponds to 1 mpg in 36 mpg, or 2.8%; so the fraction of the total energy loss spent on tire deformation would be 0.028/0.40, or about 7% at 38 psi. At 45 psi, it is still about 4%. That's only part of the rolling resistance, but it's the part one can influence.
The last tank was over 40 mpg; for a while it looked like it might end up over 600 miles, but in the end the MFD range prediction leveled out at about 575 miles.