Figure 4. Regressed thermocline (color shading) and surface wind stress (vector) on DMI for (a) observations and (b) GFDL-ESM4 model. For thermocline and wind stress only significant regions at the 95% confidence interval are displayed. Regressed upper 50m (c) Nitrate and (d) Phosphate on GFDL-ESM4 model. For nutrients, significant regions at the 95% confidence interval are displayed. The contour lines indicate the regressed nutricline (m) on DMI. Here, thermocline and nutricline are defined as the depth of the maximum vertical gradient of potential temperature and nutrients, respectively.
To understand how IOD modulates surface chlorophyll in the tropical IO, we next examine IOD-induced changes in atmosphere-ocean coupled systems (e.g., surface wind stress, thermocline, and nutrients such as nitrate and phosphate, Fig. 4). As shown in Fig. 4a, southeasterly wind stress is significantly enhanced in response to IOD in the eastern equatorial IO. Given the importance of strong air-sea coupling in this region (Bjerknes feedback; Bjerknes 1969), IOD-induced southeasterly winds are favorable for thermocline shoaling in the eastern equatorial IO (Fig. 4a). This results in cooler SSTs (Fig. 1a), indicating strong upwelling in the southeastern IO, west of Sumatra. On the other hand, the wind stress and thermocline response in the western equatorial IO remains weaker, suggesting that IOD-induced upwelling/downwelling in this region is weaker or even negligible during late summer to autumn. Interestingly, consistent with the chlorophyll response (Fig. 1c), the southwest region of Indian subcontinent has stronger southeasterly winds and a deeper thermocline, suggesting that the IOD-induced upwelling/downwelling is stronger in this region compared to the western equatorial IO. Thus, upwelling (downwelling) is enhanced in the WS (SWI) in response to IOD. The IOD-induced changes in surface wind stress and thermocline in GFDL-ESM4 are consistent with the observations (Fig. 4b).
In general, nutrients such as nitrate and phosphate are essential for phytoplankton blooms in the tropical IO (i.e., Roxy et al 2016). To further illustrate the variability in nutrient supply to the surface in response to IOD, we have examined the regression patterns of nitrate (Fig. 4c), phosphate and nutricline (Fig. 4d). The nutricline is defined as the depth of the maximum vertical gradient of each nutrient. Consistent with the changes in the thermocline, the results clearly show an increase in nutrients in the eastern equatorial IO and a decrease in the south-southwestern region of India. Thus, IOD-induced changes in the upwelling/downwelling process are a key factor in chlorophyll variability, leading to a biological dipole in response to IOD in the tropical IO.
Discussion
Marine primary productivity in the IO is highly sensitive to the ocean-atmosphere dynamics associated with the IOD (Currie et al2013, Wiggert et al 2009), and it is therefore essential to understand the profound effect of the IOD on oceanic chlorophyll. Previous studies (Shi and Wang 2021, 2022) have attempted to explain the biological response of IOD using BDMI. However, there are discrepancies between physical response and biological response in the index region that prevent an accurate representation of the IOD-induced chlorophyll variability. As can be seen from the composites of positive and negative IODs, the change in the western equatorial IO is negligible. Thus, despite the dipole pattern in the biological response to IOD, the regions used to explain the BDMI in previous studies have limitations to express the actual biological variability in the IO. On the other hand, even if there is a biological dipole response, the BDMI is not able to capture the chlorophyll variability well. For example, the BDMI does not well reflect the extreme events (i.e., 2016 negative IOD, Fig. 3b and Fig. S5). Therefore, it suggests the need for an accurate index that can deeply explain the IOD-induced chlorophyll variability in the tropical IO.
Furthermore, we found discrepancies between observations and CMIP6 models. Although GFDL-ESM4 is closer to the observations, the remaining 12 models do not well capture the climatology and interannual variability of chlorophyll in the tropical IO, suggesting biases in current biogeochemical models (Roxy et al 2016, Lim et al2018). In addition, we have not used observed nutrients to quantify IOD-induced changes, so it will be important to construct advanced observational systems to understand the associated biophysical interactions in detail. Nevertheless, the observations and the model used in our study provide sufficient evidence for a general understanding of the IOD-induced changes in marine phytoplankton in the tropical IO.
As marine phytoplankton play a key role in the marine food web, environment, and global climate, understanding their changes in response to tropical climate variability has important implications. Studies have shown that IOD-induced changes in upwelling/downwelling and surface chlorophyll concentrations (an indicator of biological productivity) are likely to affect the abundance of economically valuable tuna stocks, catch rates, fishing grounds and other pelagic fish species in the IO (Wiggert et al 2009, Lan et al 2013, Gaol et al2015, Setyohadi et al 2021, Wang et al 2023). Therefore, changes in phytoplankton blooms can modulate fisheries production and ecosystem functions, and also can cause major environmental problems (Dai et al 2023). Thus, understanding IOD-induced changes in marine phytoplankton has direct and important implications for fisheries and environmental management in the IO.
Conclusion
We have investigated the IOD-induced surface chlorophyll responses in the IO using both observations and a CMIP6 earth system model, and show the strong zonal contrast of chlorophyll blooms between the southeastern IO and the south-southwestern region of the Indian subcontinent. This strong zonal contrast indicates a biological dipole in the IO in response to IOD. Furthermore, we show that the IOD-induced changes in the nutrient upwelling/downwelling process are a key factor in the chlorophyll variability leading to the biological dipole in the tropical IO. Finally, since the previously defined BDMI has limitations in representing the IOD-induced chlorophyll variability, we propose a new biological dipole index (BDI) that can robustly explain surface chlorophyll changes in the tropical IO in response to IOD. In addition, we highlight the need for a better understanding of the associated biophysical interactions in the tropical IO.