Is there any evidence of weather (outside of catastrophic weather) actually having an impact on elections?
In Austria, there has actually been a study on this:
Rain-weather increases turnout here, while sunny weather leads to lower turnout (mostly because people have better stuff like hiking or cycling to do).
Did you ever notice that shopping malls have like 3-4 times the number of customers when it rains rather than when it's sunny outside ? If it's sunny, shopping malls (like precincts) are virtually empty (well, not really, but there's still a difference).