It makes an interesting preliminary study but it raises more questions than it answers:
1) What diagnostic criteria are being used? Did the research team themselves assess the level of autism in the children in the subject or are they simply using statistics available? If they are just using prevalence levels, then one might consider an alternative correlation - doctors who live in a high rainfall area may have a greater tendency to diagnose autism...
2) Is the study blinded? In other words, do the researchers know the rainfall amount/geographic locality of the children they are re-assessing for autism?
3) I know from my own observations of difficult child 3 and his best friend (both diagnosed with autism) that they watched A LOT of television preferentially to playing outdoors. It has actually been a learning strategy used by them in terms of developing social skills and language.
This is a correlation. Nothing more. We have to be very careful in understanding this before we can even begin to claim any connection at all, let alone any possible causality. Just because one research group has tentatively identified a possible link in terms of frequency doesn't mean that rain (or TV watching) causes autism. It may be that a parent having a greater opportunity of observing some of the more autistic-like behaviours of a child watching television, if it's raining outside.
We have an autism incidence in Australia of one in 100. I don't think we could be classified as a high rainfall country.
There are several directions now for this idea to be taken. Frankly, I would like to see it deliberately tested in an area known to be either of lower rainfall or high gradient rainfall. Sydney would be a good start - we get much more rain on the coastal strip than to the west; some of the outer areas to the west of Sydney are in rainfall shadow and yet looking at the contact list for the teens Pervasive Developmental Disorder (PDD) group, there are a lot of home addresses in the rainfall shadow area.
I feel that while ever there are differences in how autism is diagnosed; differences in how liberally of strictly the diagnostic criteria are interpreted; differences in understanding and distribution of medical experts capable of diagnosing autism; differences in the ability of the population in gaining access to a diagnosis; then there will continue to be interesting fluctuations like this in autism epidemiology. Crikey, it's hard enough in Australia to get a diagnosis, simply because of the difference in access to experts depending on where you live. And we have Medicare universally to help foot the bills. How much more inequitable is it going to be in a country without a universal health care system? And by that remark I do not want to start a debate - merely pointing out another possible variable to consider in this result.
Summary - an interesting observation, it needs more consideration and wider exploration before it's considered to be causal.
Marg