Same return trip today, similar conditions, full lower grille block and snorkel back on. Again averaging about 52 mpg or maybe even slightly better, the display shows litre per 100 km: 4.5 so it is 12 times as crude as a mpg readout
It is safe to say that under these conditions a snorkel delete is just about as effective as a full grille block.
I think they should not be combined (an air intake temp regulator would then be needed) so I stay to the latter option, which seems safer.
You say it is better to leave the grille partly unblocked; do you mean like just the part above the bumper or that and a bit of the lower grille?
Have to add that testing was not ideal due to the effect of sidewind.
What I noted about sidewind and vehicle stability;
Most of the time the sidewind does not affect handling at all, but with a strong wind coming from about 45 degrees behind (!) that produces a stream of erratic jerks to the body. Not enough to make the car change course, but clearly distinguishable nonetheless. FC drops somewhat when that happens; the aerodynamic properties take a blow too.
On the whole sidewind seems bad for economy; when the motorway swung round an made me face the wind head-on the FC got better!
This may be due to the design of the rear bumper. Its sides and rear join at a rounded edge.
When the airflow along its sides is laminar it will curve around that edge due to the Coanda effect; that will reduce the wake but also create a vacuum at the edges, pulling hard on the bumper.
It may even produce a vortex roll like a horizontal tornado behind your car which low pressure center will again pull the car back (this often happens behind trucks hit by sidewind).
But the side airflow will not be laminar at all times. Straight against the wind it will be disturbed by the wheel wells and the gaps between the body panels; the induced turbulence will carry on and disrupt the Coanda effect at the back of the bumper.
Only a strong sidewind will iron out those turbulences to produce laminar flow at the back.
Problems arise mainly when the sidewind is just about strong enough to do that. The airflow keeps interchanging between laminar flow and turbulence so the Coanda effect will bump in and out.
I expect breaking the Coanda effect on the bumper to bring improvement on handling and MPG.
Deleting the bumper is one way to make that happen (I know someone who did
Another way is reshaping the bumper. Some cars, like the Prius III, have sharp edges at the back. To do that on the existing bumper seems a bit crude, and might even hurt the safety of other road users bumping into you.
Attaching a tiny 1 millimeter (1/24 inch) thick strip just at the point where the curve starts to induce turbulence would also do the trick. At the Nagano 1998 winter Olympics the Dutch speed skating team used those on their suits on top of their head and running down their legs and brought home some gold medals.
Nowadays the strips are gone at the ice rink; the suits use a mixture of very slippery and very rough patches, like shark skin. So maybe just sanding the bumper with the coarsest sanding paper you can find is best. Wont get that plan past my wife, I fear
(You might need to move parts of this to a new thread, if you think it is worthwhile)