This is weak evidence. I would be more convinced by a controlled experiment. Do a study where two demographically equivalent groups are given equal sun exposures one with sunscreen the other without, and then compare their rates of melanoma. The author doesn't have data on sun exposure or evidence that the sunscreen sold is actually used, etc. etc.