Subseasonal variability and predictability of the Arctic Oscillation/North Atlantic Oscillation in BCC_AGCM2.2

- The AO/NAO index is skillfully predicted from daily to monthly timescales.
- Underestimation of synoptic eddy forcing may lead to rapid damping of the index.
- Stratospheric initial conditions have an impact on the prediction skill of the index.

The subseasonal variability and predictability of the Arctic Oscillation/North Atlantic Oscillation (AO/NAO) is evaluated using a full set of hindcasts generated from the Beijing Climate Center Atmospheric General Circulation Model version 2.2 (BCC_AGCM2.2). It is shown that the predictability of the monthly mean AO/NAO index varies seasonally, with the highest predictability during winter (December-March) and the lowest during autumn (August-November), with respect to both observations and BCC_AGCM2.2 results. As compared with the persistence prediction skill of observations, the model skillfully predicts the monthly mean AO/NAO index with a one-pentad lead time during all winter months, and with a lead time of up to two pentads in December and January. During winter, BCC_AGCM2.2 exhibits an acceptable skill in predicting the daily AO/NAO index of ∼9 days, which is higher than the persistence prediction skill of observations of ∼4 days. Further analysis suggests that improvements in the simulation of storm track activity, synoptic eddy feedback, and troposphere-stratosphere coupling in the Northern Hemisphere could help to improve the prediction skill of subseasonal AO/NAO variability by BCC_AGCM2.2 during winter. In particular, BCC_AGCM2.2 underestimates storm track activity intensity but overestimates troposphere-stratosphere coupling, as compared with observations, thus providing a clue to further improvements in model performance.