I am a physical scientist. When I look at a scatter plot like the one that you made for temperature anomaly versus El Nino, I see a poor correlation (perfect correlation, 100% likelihood, would be a line) but admittedly one with a positive slope if one exists. I was interested in your method of analysis that reveals a stronger correlation. How did you determine where to draw the lines you used to separate the data into high, medium and low?

## Question for the Author

I am a physical scientist. When I look at a scatter plot like the one that you made for temperature anomaly versus El Nino, I see a poor correlation (perfect correlation, 100% likelihood, would be a line) but admittedly one with a positive slope if one exists. I was interested in your method of analysis that reveals a stronger correlation. How did you determine where to draw the lines you used to separate the data into high, medium and low?