Are those pics with no smoothing?
You can adjust the IR window to be shorter to see which reflection is causing the comb filtering. Look at the impulse response and see the peaks. The first one will be the fundamental wave front. The following peaks are reflections. If you shorten the time interval so that only the first peak and trail are included, you will get no comb filtering, but the bass measurements will suffer (since the longer wavelengths need longer windows for a good measurment).
Then you can gradually add in the peaks (make the window longer) and see when the comb filtering really becomes a problem. The peak associated with that can be translated to a distance (you know the time and speed of sound) and from there you can try to locate the path the sound took to cover that distance and that's your reflection point (or at least a likely reflection point).
I can tell you from experience that comb filtering is very hard to get rid of completely. Usually I just 1/3 octave smooth my measurements so I don't see it
