Ants can sense earthquakes a day in advance

There is no link to the source (here: http://www.mdpi.com/2076-2615/3/1/63/pdf), no supporting material presented, it makes some odd claims starting with "ants prefer to live on active faults" (this paper says nothing about that - it's talking about ants that happen to be found in a geologically interesting faulting area) and ends up with "ants can predict earthquakes". Which then goes up to the headline.

Stop stop stop stop stop.

This paper is a decent exploration of performing day+night video analysis to describe above-ground ant behavior over large periods. The rest of it seems reaching and improbable, but I don't have the background to comment on any of the math or geology presented.

http://www.radiolab.org/2007/aug/14/

tl;dl - Longtime ant researcher is resigned to the fact _individual_ ants are incredibly stupid and ineffective, sometimes just carrying a stick back and forth for days. But in colonies - somehow brilliance can emerge. They'll can predict major thunderstorms well enough in advance to build defensive walls around the entrance to their colony to direct floods around it.

No. Seriously. They've used 5 data points in the research. Five. Calling that a 'possible correlation' is just calling for trouble.

Here is a link to the original article. http://www.mdpi.com/2076-2615/3/1/63/pdf