Long time no see, slack; welcome back. As Ken pointed out, no, the impedance doesn't change(appreciably), and yes, it's because of the action of the crossover. An "8ohm" speaker with two sections being connected in parallel through the crossover doesn't consist of a 16 ohm high frequency section and a 16 ohm low frequency section, resulting in an 8 ohm speaker because of the parallel operation. Each section is 8 ohms(nominally)and when one of the sections is presented with material out of its operating range(e.g., a 100Hz signal to the tweeter)the resistance in the high frequency section of the crossover feeding the tweeter is greatly increased so that, for example, the overall resistance to the tweeter at that frequency might be 160 ohms rather than its inherent 8 ohms. So, the combined woofer/tweeter impedance at 100 Hz would be the reciprocal of the sum of the reciprocals of the two sections, i.e., 1/8 + 1/160=21/160; 160/21 would be the effective net impedance of about 7.6 ohms. The fact that the resistance in the crossover causes the effective resistance of the driver in question to be much higher than its nominal value when the signal is out of its designed operating range results in the impedance changing relatively little, instead of being cut in half(e.g., 1/8+1/8=2/8; 8/2=4ohms)by the parallel operation.
