Fig. 1: Observed country-heterogenous economic impact.

a Global pattern of country-heterogenous teleconnection of the IOD, calculated as the sum of country-averaged, population-weighted climate teleconnection from the IOD in surface air temperature and rainfall (see “Climate teleconnections” in “Methods”). The bar chart below refers to the distribution of teleconnection, in which the red and black bar indicate the number of teleconnected and non-teleconnected countries, respectively. b Nonlinear relationships between the IOD index and IOD-induced impact on GDP growth for Indonesia (red), Kenya (yellow), and Australia (blue), with shadings indicating the 95% confidence interval based on a Bootstrap method (see “Bootstrap tests” in “Methods”). Dots on each line represents values of the 2019 strong pIOD event, with estimated losses in GDP value shown in the legend. c Cumulative effect of two strong pIOD events in 1997 (yellow) and 2019 (red), and two strong nIOD events in 1996 (green) and 2010 (blue) on global GDP cumulating from year 0 (IOD occurrence year) to year 2. Shadings indicate the 95% confidence interval based on the Bootstrap method. d Same as (b) but for IOD-induced impact on global GDP growth. Impact of the pIOD and nIOD on global economic is asymmetric, far greater during pIOD events and increases nonlinearly with IOD amplitude, and disproportionally large on developing economies.