I wonder if, without the rear fan, the side and fronts create turbulence and creates a "dead spot" near your northbridge in your setup. the rear fan would create a different pattern of turbulance.
I was a tad surprised by it, too, with about 1/3rd of that difference showing up in the case temp. I think the big reason is that it only has a passive heatsink and without the rear fan to draw away the hot air off the CPU cooler, it's being directed right onto the NB's HS
Im not surprised the front fans dropped a bit, the rear fans help lower the pressure so the fronts get an easier job of pushing air into the case.
The surprise for me was how much of an effect it had. It went from 1710 to 1670 - a difference of 2.4%. The fan will still spin, even if you bolt it straight onto a solid panel. So even at it's max pressure without it being able to move any air, it will still rotate. I might have to try doing that to the fan to see the maximum effect it can have on the rpm. This means you can't make a direct assumption of the affect to airflow for the change in pressure. I just wish I had a way that I could measure the airflow through the fan, and see the real impact of it.