The dashed contours shown on each map indicate where the mean of the ensemble forecast was different to the mean of the model climatology at the 90% significance level according to a students T-test. Although not the ideal test statistic it indicates where the model is more 'confident' that the forecast mean is different to climatology. There are no guarantees that a significant deviation will be realised, though a simple filter (see here) shows that the model is more likely to get the sign of the anomaly right than wrong over certain regions when only high T-test scores are used.

Notes:

Comparison with last months forecast - click on link below
- Regional Seasonal Mean plots