In the paper “Mode shift in the Indian Ocean climate under global warming stress” by Nobuko Nakamura et al. (Geophysical Research Letters, 36, L23708, doi:10.1029/2009GL040590, 2009), coral IOD index was evaluated by determining its correlation coefficient with the observational data with “r2”; however, we made a simple mistake, and “r2” should be “r” in the text and figures. We corrected this mistake in page 2 line 3 (right side), page 4 line 15 (right side) and Figure S6 in the auxiliary material of Nakamura et al. . We also corrected the same mistake in page 4 lines 16, 18 and 20 (left side) of Kayanne et al. , the first report of this coral IOD series. Nakamura et al. reconstructed 115-year IOD variability related to East African Short Rain variability based on the coral IOD index derived from the Kenyan coral monthlyδ18O record. Our coral IOD index is defined as the coral δ18O values assigned to January of the following year, which reflects the Short Rain anomalies for the previous September to November.
 At this time, we would like to re-evaluate and reconfirm the validity of the coral IOD index by determining the confidence interval and its determining factors. The correlation coefficient (r) between the Mombasa rainfall (SON) anomaly and coral IOD index for 1986–1999 is 0.77 (Table 1), and the standard error for this estimation is 0.10‰. Figure 1 shows the scatter plots correlating the Mombasa rainfall (SON) anomaly with the coral IOD index for the last 40 years. The coefficient of determination between the Mombasa rainfall (SON) anomaly and coral IOD index for 1959–1999 is low (solid line in Figure 1: r2 = 0.33); however, it increases to r2 = 0.69 in the IOD years (red crosses and dashed line in Figure 1). This observation indicates that the coral index accurately separates East African Short Rain anomalies related to IOD and validates the reconstruction of past IOD variability using this coral index.
|Coral IOD Index \ DMI||SST-DMI (SON)||EARI-DMI (SON)||Mombasa Rain (SON)|
 The factor affecting the moderate correlation between the coral IOD index and Mombasa rainfall (SON) may be derived from the low correlation of non-IOD years (Figure 1). The coral IOD index constructed by raw data on the coral δ18O values also contains an SST anomaly in January. The relationship between Mombasa rainfall (SON) and SST (Jan) shows that the SST (Jan) anomaly ranges from −1 to 1°C (1σ = 0.34°C), and that low precipitation followed by a high SST in January was observed in several years (Figure 2). This trend has a contradictory effect on the coral δ18O. The scatter plots for non-IOD years show a moderate correlation in the overall plots. However, we show that the coral IOD index clearly separates anomalousδ18O values related to heavy rainfall and dry conditions for the Short Rainy period each year. These years correspond to positive and negative IOD years, and we identified the coral IOD years for the last 115 years based on approximately one sigma of this variance.
 We conclude that the coral index must be used within the confidence intervals; the correlation between the coral index and Mombasa rainfall (SON) is r = 0.77 for 1986–1999. Consequently, this coral δ18O index can be used to effectively analyze the history of the East African Short Rain variability (IOD variability) in longer coral cores.