And less than 0. Also, what exactly is a "normalized reflectance"? Normalized to what?
A reflectance greater than one is caused by a conductive object traveling through a magnetic field. Due to its great speed, it is a massive generator, and the object takes on a very high voltage charge. When that voltage leaks away in a coronal discharge, it generates light. Kinda like a weak arc lamp.
This causes it to be brighter than simple reflectance.
I don't know. I can speculate: Reflectance is always measured relative to some reference; either a calibrated reflector or a calibrated source. They may be able to use the sun as a constant, known source. Hence, normalized to solar irradiance at some specific distance from Sol. However:
1) The thing is small and distant. Total light coming from it is very little, requiring long integration times to get any usable signal.
2) The thing is moving very fast, thus illumination on it is changing rapidly relative to integration time.
3) The thing seems to be spinning, varying both the presented surface and presented area rapidly relative to integration time.
4) The thing is small relative to pixel size. Are we seeing multiple pixels on it, or just a point spread function? (I don't know the answer to that.) This calls into question whether all measurements are of the same material. Material variations over the surface of the object are unknowable with available data.
All of this adds up to the huge error bars shown on the graph. Trying to glean meaning from its shape other than the slope is probably not supportable.