May somebody else caculate from that (it can be done) how ...
Seems I lost You. The mike is positioned about 1m away from the speaker box, about 5" above the floor. By that the sound travels along two distinct pathes One goes direct from the driver to the mike, the other goes from the driver to the floor, where it is reflected towards the mike. Both pathes have different lengths. The individual contributions of both interferre in a characteristic manner, called comb filter effect.
That is what You see. Very deep 'nulls' in the frequency response, where the interferrence is destructive, and 'poles' up to a theoretical maximum of +6dB, where interferrence is constructive.
The pattern replicates itself over the frequency range, starting from some frequency, which is determined by the path length difference and the speed of sound. That simple.
Logically, if reflection is suppressed, namely by absorption, the second path loses amplitude, hence contributing less, and comb filter effect is mitigated. Taken the other way round, less comb filter effect means less reflection, means more absorption. I use this, to show how good the rug works. Again simple, right?
Here, with the rug in place the constructive interferrence vanishes nearly completly, the suck-outs are greatly reduced in depth. One might, as a first guess assume, that this relatively thin carpet, by whatever magic ;-) eats up all the sound, before it can reflect.
Two peculiarities, though. First the poles and nulls around 1,2--1,5kHz appear to be shifted! With the rug the null is lowered in frequency. How come? Look, the null is only there, because the rug loses appetite. Its effectivness declines towards lower freqeuncies. But still the heavy wool actually lowers the speed of sound in the medium. Lower speed of sound means bigger effective path length difference, placing the null deeper. Sidenote: this gives an indication on the frequency range the rug is effective in. It may start around 1kHz or even a bit below.
The other one is the irregularity around 4..5kHz. That is because the phase of the direct radiation towards the mike differs from the phase towards the floor. I missed that, sorry. The driver is a CD horn with a sharp edge in its dispersion, which (neccessarily) changes the phase as rapid over even a very small angle (way off the main axis, don't worry).
Sorry, it was just a very hasty check, because it was told, that my magic carpets don't work. I didn't attempt to go down. (I works by true faith, You know?)