I wouldn't discount atmospheric effects that we haven't seen much yet. Or at least not with high enough frequency to understand their origins and behavior. I'm

I wouldn't discount atmospheric effects that we haven't seen much yet. Or at least not with high enough frequency to understand their origins and behavior. I'm thinking of some form of ball-lightning, a self-contained plasma. It's still an unexplained phenomenon with a history of observation (or multiple phenomena classified under one label). A self-contained plasma could very much be opaque/reflective for radar (depending on the plasma frequency) and also emit light. Maybe it'll even be absorptive to some of the active radars trying to paint it, making some of the involved airplane and ship radars actually stabilize and guide the phenomenon (e.g. it 'escaping' as soon as a painting radar gets closer). Lots of maybes but infrequent natural phenomena that cannot be easily observed with human senses and where humans usually don't go are still a valid hypothesis in my opinion.