Interesting modelling exercise but until we have physical evidence that this is occurring, and can calibrate the model to what degree, this is just an interesting hypothesis/story. Models are always subject to the GIGO and built-in-assumption confirmation problems and are always wrong, just sometimes useful.
Find the evidence to prove/disprove the model.