3.1. Dipole pattern in surface chlorophyll
First, we examined the IOD-induced changes in surface temperature, circulation, and chlorophyll in the tropical IO during the boreal summer to autumn, and the results are shown in Fig. 1. It can be clearly seen that the southeastern IO is cooler and western IO is warmer, and the southeasterly winds are intensified in response to IOD in the observations (Fig. 1a). Consistently, GFDL-ESM4 also shows similar responses, suggesting its ability to reproduce the air-sea interactions in the IO (Fig. 1b). To investigate the influence of IOD on surface chlorophyll changes, we examined the regression patterns of chlorophyll in the tropical IO. The chlorophyll anomalies in the southeastern IO, west of Sumatra, are positively enhanced in response to positive IOD, and vice versa for negative IOD (Fig. 1c). Thus, consistent with previous studies (Currie et al 2013, Shi and Wang 2021), high chlorophyll blooms are evident in the southeastern IO during positive IOD events. However, the chlorophyll response in the western equatorial IO remains weaker, raising questions about the accuracy of using chlorophyll in the western equatorial IO to define the BDMI in previous studies. On the other hand, the changes in chlorophyll in the south-southwest coastal region of Indian subcontinent are strong and significant, implying that this region has a more dominant response to IOD than the western equatorial IO. Thus, the strong zonal contrasts of chlorophyll responses exist between the southeastern IO and south-southwest coastal region of Indian subcontinent, indicating a biological dipole in the IO in response to IOD. Despite some changes in the Bay of Bengal and Arabian Sea regions, GFDL-ESM4 also shows consistent results with observations (Fig. 1d). Thus, both composites (Fig. S1) and regression patterns (Fig. 1c) show the emergence of a biological dipole in the tropical IO in response to IOD during the boreal summer to autumn.
To further elucidate the characteristics of the biological dipole, we next examine the correlations between DMI and surface chlorophyll anomalies. Since previous studies showed that the effect of IOD on chlorophyll is strongest in late boreal summer and autumn (Thushara and Vinayachandran 2020, Shi and Wang 2021), we examined the relationship from August to November. Consistent with the regression patterns (Fig. 1c), the correlation shows more distinctive dipole pattern between the regions of southeastern IO and south-southwest of Indian subcontinent (Fig. 2a). In addition, the correlation remains weaker in the western equatorial IO region, but it is dominant for the chlorophyll response in the south-southwest of Indian subcontinent. Therefore, although there is a dipole pattern in the biological response to IOD, the regions used to explain the BDMI in previous studies might to be less straightforward to represent the chlorophyll responses to the IOD.
Hence, to clearly represent the IOD-induced biological dipole, we select two regions based on the regression (Fig. 1c) and correlation (Fig. 2a) patterns. As shown in Fig. 2b, the strongest chlorophyll response is observed in the regions of Southwest of Indian subcontinent (SWI, 4° N-16° N, 70° E-79° E) and West of Sumatra (WS, 10° S-2° S, 95° E-105° E). Further, to show the dipole-like response of chlorophyll, we examine the correlation between the SWI (WS) chlorophyll index and chlorophyll anomalies in the tropical IO, and the results are shown in Fig. 2b (Fig. 2c). The results clearly show that when chlorophyll anomalies are positively increasing in the SWI (WS) region, chlorophyll anomalies are decreasing simultaneously in the WS (SWI) region. Thus, our results clearly show a dipole pattern in surface chlorophyll in the tropical IO in response to IOD during late summer to autumn. This spatial pattern of correlation has been well captured in GFDL-ESM4 consistent with the observations (Fig. S3). Therefore, we suggest an improved biological dipole index (BDI) as the difference between the average chlorophyll anomaly in the SWI and the WS regions. To show further evidence for the biological dipole, we have also examined the correlation between chlorophyll anomalies of the SWI and WS regions and found a significant negative relationship (Fig. S4). It is worth noting that the correlation during June to November is –0.43, and the correlation becomes stronger (–0.48 for the observations, and –0.53 for the GFDL-ESM4) during August to October, implying that the dipole pattern in chlorophyll is more dominant during late summer to autumn.