Based off the exped dimensions and fill weight, they're using somewhere between 600 and 700 fp down. It depends on whether they went with fully lofted down or overstuffed a little, I think fully lofted would probably be best in a down mat since it will be compressed anyway by body weight, so most likely 600 fp was use. This would also make sense why the fill power is not advertised. 700+ is considered "quality" and useful for marketing, 650 or less is considered "cheap" (which it is but still more than functional).
So using 900fp would potentially save you about 1.5oz in weight but the cost/oz does go up too.
1 clo ~ 0.88R (in customary/english units, not SI). Does anyone know which R value exped is using? SI R-values are considerably different between the systems, but it does appear to be English Units, despite most of their other specs are also offered in SI units.
At this R value of 5.9 I get ~2.4 clo/inch for the mat. Some of this clo is the shell material and most will be the down/air. I've seen some reports that state 600fp down has a clo/inch of 2.31 which is in line with the design parameters of the exped mat.
So based off that study 800fp has 3.26 clo/inch and 900 is 3.57. But these numbers are in no way experimentally verified unfortunately, it's based off a linear regression on a very limited data set, although exped mats seem to approximately verify the numbers for the 600-650 down. So maybe the better fill power numbers are close to accurate.
Like I said there's a lot of variables that go into this. there is no way to make an oz to R value statement.
What you need to do is find out the R value you want. Then based off your down choice you can calculate thickness. From that thickness and the overall profile dimensions of the mat you are making you can calculate internal volume (how accurately is up to you, do you factor in the cylindrical nature of the tubes and all that). Finally you now have a volume that you can divide by the fill power and that will give you the ounces you need to properly fill it and get a rough estimate of the R value.
Alternatively you can work backwards and let the down choice decide the end R-value but this then may give you a much warmer/colder mat than you need for the expected conditions.
Regardless I would err on the side of making it warmer than necessary unless you have the time and resources to prototype a few and have them tested or you aren't overly particular if the mat turns out too cold and will just adjust what trips it is taken on.