Fig. 5 | Nature Communications

Fig. 5

From: A model for super El Niños

Fig. 5

Surface wind anomalies during IOD from an observational estimate and simulations. We estimated the impact of IOD on surface wind anomalies (left) by regressing observed wind anomalies from 1958 to 2015 onto the Dipole Mode Index (DMI) and Nino3.4 SST anomaly time series. The vectors show the partial regression coefficients of wind anomalies with DMI, controlling for Nino3.4, for a May—July, b August—October, and c November—January. The shaded contours on the left are for marine cloudiness, which were analyzed in the same way. The brown-colored contours in (ac) mean less cloudiness (and by inference, less rainfall), while green colors indicate more cloudiness (more rainfall). The right panels are for the same seasons as the left, but they show surface wind anomalies from a simulation. The synthetic thermal forcing (shaded contours in df) that was used to drive the atmospheric model mimics the patterns of marine cloudiness anomalies shown in (ac); anomalously low thermal heating (brown contours in df) was prescribed in regions where anomalously low rainfall is noted in the observations (a–c); enhanced thermal heating (green shading in d–f,) in regions with anomalously high rainfall. A partial regression measures the amount by which the dependent variable (e.g., surface wind anomaly) changes when one of the independent variables (here, DMI) is increased by one unit while keeping the other independent variable (here, Nino3.4) constant (or controlled). The maps in the figure were rendered with the NCAR Command Language software (https://doi.org/10.5065/D6WD3XH5) from the Global Self-consistent, Hierarchical, High-resolution Geography Database (GSHHG). The GSHHG is available online at https://www.ngdc.noaa.gov/mgg/shorelines/gshhs.html

Back to article page