A 2008 study from Cornell University suggests a statistical link between the occurrence of autism and the rainfall in some US states. The authors suggest it's because children tend to spend more time indoors watching TV in those states. The idea fits conveniently with the conclusion from an earlier paper by the same authors that autism rates are positively related to the percentage of households with cable TV.
But the authors are economists with no medical or psychological expertise. Rainfall varies widely from year to year, but there is no corresponding variation in autism rates.