Visually, it looks as though low temperatures are better at predicting La Nina than warm temps are for El Nino. Is there a measurable change in variance (homoscedastisity)? Yes, I use that word whenever I can find the opportunity.