Hard to tell, but I believe it shows that all the drivers in HIII and the sub (wow) are working.
I looked at the Rives site and it shows their 1/3 octave sources on their "II" version. The x-axis looks to be set out out 1/3rd points.
IIRC, 1/3 octave noise is distinctly hissy. 1/3 octave warble tones sound like a Star Trek phasor. BTW.
The use of 1/3rd octave sources is a way of averaging out peaks and valleys in the measurement. It is not really averaging out what is measured (the speaker and the room) but rather what is fed to the speaker -- which amounts to pretty much the same thing.
Your graph looks ragged at first. But much of it fits in a plus or minus 3 dB window. I'd say that is not too bad.
The smoothing is supposed to remove (average out) the reflections from the room walls. But I think you might still be measuring the effects of floor, ceiling,and wall bounces.
Why don't you tell us more about where the speaker and microphone were located.
One technique for testing is to move the speaker outdoors and set it on a hard surface, like a driveway, away from buildings. The speaker box is placed upside down so the tweeter is near the ground. Then put the microphone about 10 feet away and cant the speaker front to aim at the microphone. Maybe more than you want to do.
It is difficult to convince people just how much the room reflections affect measurements. One time I used a single tone to a speaker (say 200 Hz). At some places in the room, the tone could not be heard at all. People have reported something similar with higher freqs (maybe 1000 Hz) and walking around the room with an RS meter. Meter readings (more obvious if you have a moving needle type) vary widely with movements of one foot.