Extended Data Fig. 3: Heat budget analysis based on observations.
From: Opposite response of strong and moderate positive Indian Ocean Dipole to global warming

a, Composite of heat budget terms over the equatorial Indian Ocean (10°S–10°N, 40°E–100°E) in August-September-October (ASO). The uncertainty bar on each composite represents the range over the three strong or three moderate pIOD events. b, c, Composite of ASO temperature tendency during strong and moderate pIOD events. The green box marks the area where the main warm temperature tendency is located over the southwestern Indian Ocean (10°S–0°, 50°E–80°E). d, Relationship between normalized temperature tendency and the Ekman pumping term in ASO averaged over the main warming area in the southwestern Indian ocean (10°S–0°, 50°E–80°E, green box in c), with statistical properties shown. e, Map of ASO correlation coefficients between the Ekman pumping term averaged over 10°S–0°, 50°E–80°E and zonal wind stress (TauX) anomalies. Correlation coefficients of ±0.27, ±0.32, ±0.41 are statistically significant at the 90%, 95%, and 99% confidence levels, respectively. The area indicated by the red box is used to calculate the average zonal wind stress anomalies. f, Time series of normalized ASO Ekman pumping term averaged over the western Indian Ocean (10°S–0°, 50°E–80°E) (red curve) and zonal wind stress (TauX) anomalies averaged over the equatorial Indian Ocean (5°S–5°N, 60°E–100°E, red box in e) (green curve). The Ekman pumping term has been multiplied by −1 for plotting. The data used in this figure are from ORA-S5 for the period of 1979–2018.