Fig. 6: Atmosphere-ocean interaction for PNA in the coupled and uncoupled simulation and its difference.

Anomalies for DJF-mean SLP (a; contours at 2hPa intervals), THF (a; shading, downward positive, Wm⁻²), T850 (b; contours at 0.5°C intervals), SST (b; shading, °C), 10 m wind (b; vectors, ms⁻¹), oceanic Ekman transport (c; vectors, m²s⁻¹) and associated heat transport expressed as equivalent surface energy flux (c; shading, SST warming direction is positive, Wm⁻²) from the C-run, regressed onto the EC1_C. Contours are dashed if negative. Coloured only where the statistical confidence exceeds 99% level based on two-sided t test. The 10 m wind anomalies in (b) are shown in red if they increase scalar wind speed and in blue if they reduce it. Contoured in (c) are SST climatology (3°C intervals). d, e, Same as (a, b), but for anomalies from the A-run, regressed onto the EC1_A. f Difference in the variance of SLP (contours at 1hPa² intervals, dashed if negative) and of THF anomalies (shading; coloured only where the statistical confidence exceeds 95% level based on two-sided F-test) associated with the EOT1 (C-run minus A-run). The black triangle indicates the centre of SLP anomalies in the North Pacific.