Prediction of Northern Hemisphere regional surface temperatures using stratospheric ozone information

Here we quantify for the first time the potential use of ozone information for Northern Hemisphere seasonal forecasts, using observations and a nine-member chemistry climate model ensemble. The ensemble composite correlations between March total column ozone (TCO) and April surface temperatures display a similar structure to observations, but with slightly lower correlation magnitudes.

This is likely due to the larger number of cases smoothing out sampling error in the pattern, which is visible in the difference between correlations calculated from individual ensemble members. Using a linear regression model with March TCO as the predictor, predictions of the following April surface temperatures in regions that show large correlations are possible up to 4 years following the regression model end date in individual ensemble members, and up to 6 years in observations.