1. Introduction
Ocean stratification is a fundamental hydrographic feature represented by water density change with depth and is therefore mainly determined by the vertical distributions of temperature and salinity. Stratification can regulate the vertical mixing of heat, nutrients, carbon, and oxygen, and thus affect air–sea interaction, the ocean ecosystem, and biogeochemical processes (Vincent et al. 2014; Roxy et al. 2016; Bourgeois et al. 2022; Shackelford et al. 2022).
The Indian Ocean (IO) is the third largest ocean after the Pacific and Atlantic Oceans. Changes in its dynamic and environmental field have remarkable effects on the weather and climate and have attracted considerable attention (Hoerling et al. 2004; Luo et al. 2012; Tierney et al. 2013; Wang 2019; Zhang et al. 2019), wherein the IO stratification plays important roles in modulating the air–sea interaction. For instance, Li et al. (2017b) reported that the Indian summer monsoon’s significant freshwater input creates a strong salinity stratification, which alters the monsoon’s intraseasonal oscillations by impacting the sea surface temperature. Meanwhile, the strength of the IO stratification also affects the development of marine heat wave and tropical cyclones (Rathore et al. 2022; Jangir et al. 2016). Strong stratification can contribute to the marine heatwave development and the subsurface ocean warming, thereby increasing the heat potential of tropical cyclones (Rathore et al. 2022). Meanwhile, during the development of tropical cyclones, stable stratification can reduce the induced cooling effect of the cyclones, thus favoring an increase in the cyclone’s intensity (Vincent et al. 2014). Furthermore, several studies pointed out that the stratification can regulate the IO ecosystem. Roxy et al. (2016) and Venkataramana et al. (2021) suggested that the strong vertical thermal stratification caused by rapid warming leads to a decrease in surface phytoplankton and chlorophyll-a in the IO.
The ocean stratification has multiscale variations ranging from seasonal through interannual and interdecadal to multidecadal or even longer time scale (Vincent et al. 2014; Maes and O’Kane 2014; Cheng et al. 2016). However, studies on IO stratification have so far focused on seasonal to interdecadal variability (Li et al. 2017b; Kumari et al. 2018; Yadidya and Rao 2022), less on its long-term trend. Yamaguchi and Suga (2019) discovered in their study on global ocean stratification that the IO stratification has been increasing for more than 50 years, at a rate of 0.025 kg m−3 decade−1. But previous studies (Yamaguchi and Suga 2019; Li and Zheng 2022) usually used traditional statistical methods (e.g., straight-line fitting) to investigate the linear trend of stratification, which can only extract trends at a constant rate. Is the enhancement of the IO stratification over the last several decades temporally uniform? To address this question, this study will use the ensemble empirical mode decomposition (EEMD) to extract the trend of IO stratification. Unlike traditional straight-line fitting methods, the EEMD method has the ability to reflect the hidden nonlinear and nonstationary properties of the time series (Wu and Huang 2009). Currently, this method has been widely used to study nonlinear trends in meteorological elements (Franzke 2010; Ji et al. 2014; Cnossen and Franzke 2014; Franzke 2014).
Variations in temperature and salinity determine stratification variability (Joshi et al. 2017; Li and Zheng 2022). Since the industrial revolution, the greenhouse effect due to anthropogenic activities is expected to leave an imprint on ocean stratification by changing the temperature and salinity fields (Capotondi et al. 2012; Yamaguchi and Suga 2019). In the context of global warming, stratification strengthens mainly due to an increase in temperature. However, stratification changes are not perfectly synchronous with temperature variations (Somavilla et al. 2017). This is because the changes in salinity can further contribute to strengthening stratification in the upper ocean and can also offset the decrease in surface density due to the increased temperature (Li et al. 2020). Further, salinity can also influence sea surface temperature by contributing to oceanic stratification, but its variability is often complex (Gévaudan et al. 2021). A current study suggested that temperature and salinity variations in the IO are spatially heterogeneous (Ummenhofer et al. 2021), which raises another question: which regions of the IO are stratified by temperature and which by salinity? Here, we will employ the potential energy anomaly (PEA) to quantify the stratification and distinguish the respective contributions of temperature and salinity. The PEA proposed by Simpson and Hunter (1974) and Simpson (1981) can fully consider the vertical distribution of the density of ocean water (Yamaguchi and Suga 2019) and is an excellent measure of the ocean stratification, as suggested by de Boer et al. (2008).
The focus of this study is to reveal the nonlinear trends and spatial heterogeneity of IO stratification enhancement and the related mechanism. The rest of this paper is organized as follows. In section 2, we provide a brief description of the data and the methods employed to quantify the stratification and its trend. Section 3 presents the observed trends in stratification in the IO, while section 4 discusses the individual effects of temperature and salinity on these trends. Furthermore, the relevant dynamic and thermodynamic mechanisms are presented and implemented in section 5. Section 6 discusses some possible mechanisms and limitations of the results. Finally, conclusions are presented in section 7.
2. Data and method
a. Data
We will examine the IO stratification changes with oceanic observations and reanalysis data in this study. The observational datasets are from the Institute of Atmospheric Physics, Chinese Academy of Sciences (called IAP data hereafter; Cheng and Zhu 2016) and the products developed by Ishii (called Ishii data; Ishii et al. 2017). The reanalysis data are the widely used Estimating the Circulation and Climate of the Ocean (ECCO) version 4 release 3 product.
The IAP data applied the CH14 scheme recommended by the expendable bathythermograph (XBT) community for XBT bias correction and advanced and carefully evaluated mapping methods (Cheng et al. 2014, 2017), which used Ensemble Optimum Interpolation (EnOI) method combined with covariance from CMIP5 multimodel simulations. The models have the capability to simulate the general ocean circulation and could provide a better representation of the covariance. The use of the ensemble strategy effectively reduces the impact of model bias in each single model and, therefore, the IAP mapping technique provides a best guess on the covariance. The IAP publishes monthly resolutions from 1940 to the present, with a horizontal resolution of 1° × 1° and 41 vertical levels from 1 to 2000 m depth.
Ishii data were generated using temperature and salinity observations from the latest versions of the World Ocean Database, the Global Temperature-Salinity Profile Program, and the Argo project. They further added a set of XBT observations compiled by the Japan Oceanographic Data Center and eliminated for biases from XBT and mechanical bathythermograph (MBT) data. This dataset provides a gridded monthly mean data for the period 1955–2022, with a horizontal resolution of 1° × 1° and 28 vertical levels from the sea surface to 3000 m.
ECCO data are a global ocean model product, simulated by the Massachusetts Institute of Technology General Circulation Model (MITgcm; Marshall et al. 1997a,b). The data have 50 vertical layers with a spatial variation resolution of 1° longitude and 1/3°–1° latitude, covering the period 1992–2015. This dataset combined the MITgcm model and most of the available satellite (including satellite altimetry, AVHRR sea surface temperature, aquarium surface salinity) and in situ data [including Argo, conductivity–temperature–depth (CTD), and XBT] to produce estimates that are physically consistent with the closing property budgets. The estimation uses the neighbor-joining method to iteratively minimize the influence of the sum of squares weighted model and control adjustments. The baseline ECCO satisfies the equations of motion and conservation laws without unidentified sources of heat and buoyancy (Forget et al. 2015). Thus, ECCO is a dynamically consistent state estimate. This feature enables term-by-term diagnostics of the heat and salinity budget to explore the dynamic mechanisms of temperature and salinity variations (Hu et al. 2019).
b. Method
1) PEA
2) EEMD method
The empirical mode decomposition (EMD) method was proposed by Huang et al. (1998), which can well analyze nonstationary and nonlinear time series. EMD has been widely used in the decomposition of various signals (Franzke 2010; Franzke and Woollings 2011; Ji et al. 2014; Cnossen and Franzke 2014; Franzke 2014; Chen and Wallace 2016). But a major disadvantage of EMD is the frequent occurrence of mode mixing, which is often a result of signal intermittency. To overcome this problem, Wu and Huang (2009) proposed a new noise-assisted data analysis method, the EEMD, which defines the true intrinsic mode function (IMF) component as the mean of an ensemble of trials, each consisting of the signal plus a finite amplitude composed of white noise. The principle of EEMD can be simply described as: when the additional white noise is uniformly distributed over the entire time–frequency space, the time–frequency space consists of different scale components split into filter banks. The overall averaging of the corresponding IMFs obtained from multiple EMDs is used to offset the added white noise, thus effectively suppressing the generation of modal mixing (Wu and Huang 2009).
Additional significance testing is required to validate the trend derived from EEMD analysis. In this study, we use a Monte Carlo–based method to test the statistical significance of long-term trends. Specifically, we first shuffle a time series and compute the trend by EEMD method. This process is repeated 500 times to generate 500 synthetic trends. If a trend is outside the two standard deviations bound from the 500 synthetic trends, then it is significant; otherwise, it is not significant.
3. Long-term trend of stratification in IO
We calculate the PEA using the three datasets to demonstrate the robustness of the results. As shown in Fig. 1, the climatology and standard deviation of the PEA for different datasets are in good agreement. The Bay of Bengal and the Indonesian–Australian basin in the tropical southeast Indian Ocean (SEIO) are the regions with the strongest stratification. The strong stratification of the Bay of Bengal is the result of precipitation exceeding evaporation combined with enormous river runoff (Da Silva et al. 2017). The Indonesian Throughflow water, which has low salinity, can lead to increased stratification and the formation of barrier layers in the upper ocean (Makarim et al. 2019). The southern Arabian Sea has significant stratification variability, and the greatest variability occurs in the central IO at 10°S. This region is often referred to as the Seychelles–Chagos thermocline ridge, which is characterized by the local Ekman-driven upwelling associated with wind stress, resulting in a relatively shallow thermocline (Resplandy et al. 2009). This specific feature renders the area highly responsive to both atmospheric and surface forcing, leading to significant changes in variables such as temperature within this region (Hermes and Reason 2008; Duvel et al. 2004), which determines the PEA change.
(a)–(c) Climatology and (d)–(f) standard deviation spatial distributions of the PEA for (left) ECCO, (center) IAP, and (right) Ishii datasets, respectively.
Citation: Journal of Climate 37, 7; 10.1175/JCLI-D-23-0255.1
To quantitatively indicate the variations of the stratification, the time series of area-averaged PEA anomaly for the whole IO after removing the seasonal cycle and performing a 24-month low-pass filter are plotted in Fig. 2 for three datasets. The time series obtained from different datasets are highly correlated with minimum correlation coefficient of 0.94, indicating the consistency of the results. We additionally calculate the time series of the strength of stratification using the buoyancy frequency N2 and the vertical average of temperature differences between adjacent depth levels ΔT. The correlation coefficient between PEA and N2 (or ΔT) reaches ∼0.8, verifying the robustness of the PEA results (see Fig. S1 in the online supplemental material). To obtain the trend of the stratification, we perform an EEMD analysis for each time series. The black line is a nonlinear trend averaged over the three datasets, showing an increasing trend in regional average PEA anomalies since 1956, which validates the increasing trend in IO stratification. In contrast to Yamaguchi and Suga (2019), in which an increasing trend with a constant rate is identified, we find that the PEA exhibits temporal inhomogeneity. Therefore, in this study, we divide the entire time series into several periods based on different characteristics of changes. For the nonlinear trends, we calculate the value increment rather than the rate of change, following the suggestion of Ji et al. (2014). In this situation, we need to divide the time periods into equal lengths to ensure the equity. Furthermore, the mid-1970s are widely recognized as a significant turning point in the global climate pattern, marked by substantial changes in atmospheric circulation across numerous regions worldwide (Wang et al. 2010; Castino et al. 2016). This transition is also evident in our time series, with relatively weak variability in stratification observed before 1975/76, and increased stratification variability after that year. Remarkably, the fastest enhancement of PEA has occurred around the last 20 years. Based on the above consideration and the time length of the dataset used in the study, we divide it equally into three parts with the interval of 20 years. To present a more accurate growth of stratification from 1996 to 2015, we perform an EEMD analysis of the stratification in this period, which shows an explosive enhancement of stratification over the past 20 years (red line).
Time series of the area-mean PEA anomalies for the whole IO from 1956 to 2015 after removing the PEA seasonal cycle and performing a 24-month low-pass filter for ECCO (green), IAP (orange), and Ishii (blue) datasets. Black (red) line denotes the PEA trend during 1956–2015 (1996–2015) obtained by the EEMD method. Note that the trend is the mean of three datasets.
Citation: Journal of Climate 37, 7; 10.1175/JCLI-D-23-0255.1
In addition to assessing the overall trend of PEA, its spatial characteristics across the three periods are also analyzed. Figure 3 displays that the spatial distributions vary significantly across different periods, with a pronounced spatial heterogeneity during 1956–75 and 1976–95. Particularly, for the former period, there are significant differences in the trend between the eastern and western IO regions: in the western IO, the stratification weakens, while it enhances in the eastern IO. The spatial distribution of the IO stratification during the latter period is manifested by a strengthening of the stratification that is confined to only ∼40% of the grids, and a significant weakening trend of the stratification in the SEIO. During the period of 1996–2015, the figure reveals a significant strengthening of the PEA trend across the entire IO area, with a higher intensity compared to the previous periods. Statistically, the ratio of the grids number in the enhanced stratification region to the total number of grids has increased from 43.87%–49.63% and 33.95%–41.65% in the former two periods to 84.75%–96.17% in the third period (Table 1).
Distribution of trends in upper-ocean stratification calculated for different periods of the three datasets. Hatched areas indicate statistically significant trends above the 95% confidence level.
Citation: Journal of Climate 37, 7; 10.1175/JCLI-D-23-0255.1
Ratio of the grids number in the enhanced stratification region to the total number of grids. All three sets of data are interpolated onto a 1° × 1° grid resolution.
The widespread and rapid strengthening of the PEA trend during the third period has potential to yield significant impacts on weather, climate, and ecosystems in the IO region. Therefore, in the following, we mainly focus on the fast enhancement of the stratification in this period.
4. Temperature and salinity contributions
Trends of (a)–(c) PEAT and (d)–(f) PEAS during the period 1996–2015 for three datasets. Hatched areas indicate statistically significant trends above the 95% confidence level.
Citation: Journal of Climate 37, 7; 10.1175/JCLI-D-23-0255.1
Throughout the period from 1996 to 2015, all three datasets consistently show a notable increase in PEAT in 89.74%–96.54% grids (Table 2), indicating that the variability of temperature is a primary driver underlying the enhanced density stratification. In contrast, the contribution of salinity variation to the overall stratification trend appears to be relatively weak, with only a partial influence on the stratification enhancement in the SEIO. In particular, salinity changes only result in increased stratification in 58.05%–70.61% region (Table 2) and its spatial changes are more complex than those of temperature.
Percentage of grids with enhanced temperature and salinity stratification and percentage of grids with stratification trend dominated by temperature and salinity, respectively.
To distinguish the contributions of temperature and salinity to ocean stratification, we divided the IO into temperature or salinity dominated regions, as shown in Fig. 5. In the temperature (salinity) dominated region, |PEAT| > |PEAS| (|PEAS| > |PEAT|). The analysis of the three datasets reveals that in most of the IO, the stratification trends are primarily determined by temperature changes, with only a small portion of the region being dominated by salinity changes. Specifically, the three datasets indicate that the region where salinity dominates is mainly located in the SEIO, while the ECCO dataset shows that salinity also contributes to the stratification in the northwestern IO. More precisely, the temperature dominated region accounts for about 81.12%–93.82% of the total IO, while the remaining 6.18%–18.88% is dominated by salinity and is mainly distribute in the SEIO region (Table 2).
The blue area indicates where temperature dominates the trend of stratification, and the yellow area indicates where salinity dominates the trend of stratification.
Citation: Journal of Climate 37, 7; 10.1175/JCLI-D-23-0255.1
Overall, temperature changes are found to be the dominant driver of the increase in stratification, but salinity changes also played a crucial role in SEIO. These findings provide important insights into the drivers of oceanic changes in the IO stratification and highlight the necessity to better understand the underlying mechanisms.
5. Temperature and salinity budget analysis
In this section, we will perform the temperature and salinity budget analysis to reveal the mechanism of the long-term trend of stratification in the IO.
a. Heat budget analysis
As mentioned in section 4, temperature changes play important roles in the enhancement of IO stratification. The disparity in temperature trends between 0–400 and 400–800 m is a significant factor contributing to the overall increasing stratification in the IO. The contrasting warming patterns in these two layers play a crucial role in shaping the vertical structure of temperature trend. Besides, considering that the remarkable warming signal extends to 400 m around 10°S (see Fig. S2), we choose the depth of 400 m to include this signal. To explore the concrete effects of temperature, we indicate the trends of its vertical averages at different depth intervals in Fig. 6. All datasets reproduce the overall warming: the temperature increasing trend is remarkable in the upper 400 m (Figs. 6a–c), exceeding 1°C in the central IO at 10°S. The region appears to correspond to the Seychelles–Chagos Thermocline Ridge region. This phenomenon may result from a reduction in oceanic cooling associated with upwelling (Pfeiffer et al. 2017). Swapna et al. (2014) pointed out that in recent decades, a weakening trend in the cross-equatorial flow during the summer monsoon has been observed, which alters the zonal wind meridional gradient, consequently reducing the upwelling near the central IO region, especially around the Seychelles area. In contrast, its long-term trend in the lower layer (Figs. 6d–f) exhibits a weaker warming in northern and equatorial IO and a cooling south of 15°S. This vertical disparity in temperature trends directly causes the larger temperature difference between the upper and lower layer. Consequently, the upper seawater becomes lighter, which, in turn, leads to stronger stratification.
Trends of vertical average of temperature in the (a)–(c) upper 400 m and (d)–(f) lower layer (400–800 m) for different datasets. Hatched areas indicate statistically significant trends above 95% confidence level.
Citation: Journal of Climate 37, 7; 10.1175/JCLI-D-23-0255.1
Trend of time integration of the right-side terms of the upper 400 m heat budget [Eq. (5)]. Hatched areas indicate statistically significant trends above the 95% confidence level.
Citation: Journal of Climate 37, 7; 10.1175/JCLI-D-23-0255.1
We further examine the contribution of heat flux and advection to temperature change and the role of global warming. As such, we regress the heat flux (Y) of each grid on the global temperature change index (X): Y = kX + b. Figure 8a shows the regression coefficient k reflecting the impact of global warming on heat flux, while the constant term b representing the characteristics of heat flux in the absence of global warming is shown in Fig. 8b. It can be inferred that global warming can lead to a greatly decrease in heat flux loss in the south IO and a further increase of heat flux input in the north IO. Here, the heat flux significant loss in the south IO (Fig. 8b) may be attributed to the latent heat release caused by wind-induced evaporation (Jones et al. 1995). The thermal input brought by heat flux in the north IO is subsequently transported to the south IO through the meridional velocity (Fig. 8b), leading to warming north of 15°S (Fig. 7c). This phenomenon can be attributed to the wind-driven meridional overturning circulation, responsible for exporting the climatological net heat flux to the southern IO (Miyama et al. 2003; Wacongne and Pacanowski 1996). This circulation comprises a southern overturning cell and a cross-equatorial overturning cell, both transporting warm water southward and downwelling around the south of 15°S (Lee 2004). After conducting regional averaging of temperature change and vertical flow velocity in the area south of 15°S, we observe that the meridional advection can only lead to surface warming over the upper 80 m (Fig. 8c). Subsequently, the vertical advection transfers the heat from the surface to deeper layers through the downward velocity (Fig. 8d), resulting in the warming in the upper 400 m (Fig. 7d).
(a) Regression coefficient (k) of heat flux on global temperature change index. Hatched areas indicate statistically significant above the 90% confidence level. (b) Regression equation constant term (b), with arrows indicating the climatological meridional velocity averaged over a depth of 400 m. (c) The area-mean temperature trend caused by meridional advection south of 15°S. (d) The area-mean climatological vertical velocity (upward is positive) for south of 15°S.
Citation: Journal of Climate 37, 7; 10.1175/JCLI-D-23-0255.1
As in Fig. 7, but for (a)
Citation: Journal of Climate 37, 7; 10.1175/JCLI-D-23-0255.1
b. Salinity budget analysis
As mentioned in section 4, the enhancing trend of stratification in the SEIO is determined by salinity variability. As such, we further analyze the salinity change mechanism in this subsection. Similar to the temperature case, we split the 0–800 m into 0–400 and 400–800 m to perform the following analysis. To validate the rationality of this choice, we examine salinity stratification in different depths and find that the strong stratification is mainly located in the upper 400 m. The spatial pattern of the long-term trend of salinity at the upper and lower layers is shown in Fig. 10. From 1996 to 2015, the decreasing trend of the average salinity in the IO from 0 to 400 m is larger than that of the lower layer of 400–800 m. Specifically, the freshness trend of salinity in the upper 400 m in the SEIO is as high as 0.15 psu (at the 95% confidence level), but the change in the lower layer does not exceed 0.05 psu. The change of salinity is spatially uneven. In the western IO, especially in the Arabian Sea, the salinity increasing trend can be observed. In the SEIO, where salinity stratification has increased, the upper salinity has significantly decreased. This freshening trend contributes to a significant reduction in the density of the surface seawater, ultimately enhancing the stability of the stratification.
Trends of vertical averaged salinity in the (a)–(c) upper 400 m and (d)–(f) 400–800 m. Hatched areas indicate statistically significant trends above the 95% confidence level.
Citation: Journal of Climate 37, 7; 10.1175/JCLI-D-23-0255.1
Figure 11 shows each term in the salinity budget Eq. (7). In different regions, the mechanisms controlling the salinity trend may be different. In the northern IO, although advection has a huge positive contribution, it is offset by diffusion and net freshwater flux resulting in a weak salinity trend in the north. We further examine the horizontal and vertical diffusion processes and find that the vertical diffusion plays a dominant role in the northern IO (figure not shown), while in the SEIO, horizontal and vertical diffusion processes cancel out each other, leading to a weak diffusion (Fig. 11c). Moreover, the freshness of the seawater in SEIO is remarkable compared to other regions (Fig. 11e). This is mainly due to ocean advection processes (Fig. 11b), which desalinate the salinity in the region with large amounts of freshwater transported from other places. Previous studies show that the desalination in SEIO is the result of the large transport of freshwater by the Indonesian Throughflow, the intensification of the global hydrological cycle and the negative phase of the interdecadal Pacific oscillation (IPO) (Hu et al. 2019; Rathore et al. 2020, 2021; Durack et al. 2012; England et al. 2014). During the considered period of 1996–2015 (corresponding to IPO negative phase), intensified easterly winds in the western Pacific drive the current of warm and freshwater from the western Pacific into the SEIO through the Indonesian Throughflow (Li et al. 2017a), consequently leading to a reduction in salinity in this region (Hu and Sprintall 2017). Additionally, in the negative IPO phase, the cooling of sea surface temperatures in the eastern tropical Pacific results in alterations in the Walker circulation (Sun et al. 2021). Consequently, this leads to increased rainfall in the Maritime Continent and the SEIO, contributing to a reduction in salinity (Sun et al. 2021). We have also calculated the precipitation trend during this period, and the result indicates a distinct increase in rainfall in SEIO (figure not shown). However, it is important to acknowledge that contribution of freshwater flux to salinity reflects the relative magnitudes of evaporation and precipitation. Despite the increase in precipitation, there is still a positive contribution to salinity in the SEIO as evaporation exceeds precipitation (Fig. 11a).
Trend of time integration of the right-side terms of the upper 400 m salinity budget [Eq. (7)]. Hatched areas indicate statistically significant trends above the 95% confidence level.
Citation: Journal of Climate 37, 7; 10.1175/JCLI-D-23-0255.1
However, in the regions of Sumatra/Java Island, an increasing trend in ocean salinity that differs from the surroundings can be observed, which can be attributed to the upwelling. This region is influenced by the southeast monsoon, located within the seasonal upwelling area (Du et al. 2008). Upwelling brings high-salinity water from the lower layers to the upper layers, thus counteracting the effect of the Indonesian Throughflow (Susanto et al. 2001; Zhang et al. 2016; Horii et al. 2020). In the western Arabian Sea, the increasing salinity is mainly a result of evaporation over precipitation and the concurrent effects of diffusion processes (Fig. 11). These imply that surface freshwater fluxes and ocean dynamics are both critical to the formation of salinity trend patterns in the SEIO.
6. Discussion
In this study, we utilize the EEMD method in conjunction with three datasets to analyze the trend of upper 800 m stratification in IO and find a fast enhancement of the stratification over the past 20 years. As we know, the datasets may have some uncertainties. For example, they involve complex mapping approaches that may yield a large spread in the IO (Savita et al. 2022). To avoid the effects of the mapping method as much as possible, we directly use the in situ observed temperature and salinity profiles archived in the World Ocean Database 2013 (WOD13; Boyer et al. 2013) for 1956–2015 to calculate the PEA and its trend. It is found that the growth of stratification in WOD13 data also exhibits a nonlinear trend over time and are highly correlated with ECCO data (Figs. S3 and S4). Meanwhile, we use the Research Moored Array for African–Asian–Australian Monsoon Analysis and Prediction (RAMA) data in the IO to test the data quality of the ECCO: despite the spatial and temporal limitations in the RAMA data, which is not feasible to accurately calculate the PEA, the temperature and salinity data have good consistency with ECCO at the corresponding grids (figure not shown). These results validate the robustness of our results.
Subsequently, we divide the time series into three periods, 1956–75, 1976–95, and 1996–2015, and mainly focus on the third period with a rapid strengthening of the stratification. Interestingly, some phenomena are also observed in the prior two periods (Fig. 3), such as the dipole pattern of stratification trend during the first period. Such a pattern is associated with the dipole distribution of temperature (Fig. S5), which may be related to the effect of the Indian Ocean dipole (IOD). The period from 1956 to 1975 coincides with a negative phase of the IPO, which creates favorable conditions for the development of negative IOD events (Ummenhofer et al. 2021). More negative IOD events occurred during this period, as evidenced by the number of IOD events counted among the three periods (Fig. S6). During negative IOD events, westerly wind anomalies in the equatorial region cause upper-ocean warm water to accumulate in the eastern IO (Zhang et al. 2018), resulting in an intensification of stratification. The situation is opposite in the western IO. Besides, for the second period (1976–95), the basin-scale temperature trend is relatively weak, except in the western IO around 10°S, where a remarkable warming can be observed (Fig. S5). This warming can be explained by the wind–evaporation–SST mechanism (Xu et al. 2021). In particular, it can be observed from Fig. S5b that the wind trend (red arrows) favors wind weakening, which results in a decrease of the latent heat release and an increase in temperature. Furthermore, Fig. S5 also presents the negative–positive–negative trends of the salinity during the three periods in SEIO, which may be related to the negative, positive, and negative phases of the IPO. During negative IPO, Indonesian Throughflow transports fresh seawater to the SEIO (Ummenhofer et al. 2021), leading to a decrease in salinity in the first and third periods and a strengthening of stratification there, as opposed to the second period. Of course, to explore the mechanisms in detail, we may need to perform numerical simulations using a climate model in the future.
We conduct a further analysis of the mechanism behind the trend of stratification changes from 1996 to 2015, mainly explaining significant changes in temperature and salinity at the upper 400 m. Through heat budget analysis, it can be evident that heat flux and advection are important processes leading to warming of the upper 400 m of the IO. And the advection processes are mainly provided by the climatological mean field for warming. Notably, the nonlinear processes of meridional advection further enhance warming in the equatorial region, possibly associated with Rossby wave dynamics (Nagura et al. 2014). To validate the effect of Rossby waves, we calculate the nonlinear meridional advection caused by Rossby waves following the method of Belonenko et al. (2018). The results show that the meridional nonlinear advection associated with Rossby waves (see Fig. S7) has similar spatial distribution characteristics to the total meridional nonlinear advection (Fig. 9d). Moreover, west of 80°E around 10°S, the warming trend appears to be mostly influenced by the residual term, as seen in Fig. 7e. However, unfortunately, the residual term involves some ocean processes that are difficult to quantify through data analysis using the ECCO dataset. Therefore, we cannot investigate the specifical ocean processes associated with the residual term. To further explore the underlying causes behind the nonlinear term and residual term, a numerical simulation needs to be performed, which will be our next work. In the salinity case, the analysis of the upper 400 m of salinity balance also helps us to understand the trend of salinity in the SEIO.
7. Conclusions
The ongoing global climate change is anticipated to induce alterations in the stratification of the IO. The growth of stratification exhibits a nonlinear trend over time, with a significantly faster rate and wider increase from 1996 to 2015 compared to the period from 1956 to 1995. This study focuses on the mechanisms responsible for the rapid strengthening observed in the past 20 years.
The density of seawater is influenced by changes in temperature and salinity, which in turn affects the strength of stratification. The spatial variability of temperature and salinity changes in the IO also leads to spatial heterogeneity in stratification. Specifically, temperature and salinity changes contribute to a strengthening trend in the stratification of the upper 800 m of the IO by 89.74%–96.54% and 58.05%–70.61%, respectively. Regions where temperature is the dominant factor account for 81.12%–93.82% of the total, while regions where salinity is the dominant factor only account for 6.18%–18.88% and are mainly located in the SEIO.
The significant changes in temperature and salinity in the upper 400 m of the ocean are the main cause of the variations in temperature and salinity stratification. These changes can be further decomposed into different physical processes by using the heat budget analysis method and salinity balance equation. It was found that vertical and meridional advection play a dominant role in the warming trend in the upper 400 m of the southern IO. The warming in the northern IO is partially regulated by the ocean–atmosphere heat flux and meridional advection. Specifically, in the context of global warming, heat is input into the north IO and transported to the south IO through meridional advection. Simultaneously, it contributes to warming above 400 m in the region south of 15°S through vertical advection. These advection processes are mainly provided by the climatological mean field for warming, but the nonlinear processes of meridional advection also enhance warming in the equatorial region.
Similarly, the salinity budget equation was used to analyze the salinity changes in the upper 400 m of the SEIO. The results indicated that the salinity decrease in this region is mainly due to the freshening effect of advection. This provides an explanation for the rapid decline in stratification observed in the SEIO over the past 20 years.
Acknowledgments.
This study was supported by the Strategic Priority Research Program of Chinese Academy of Sciences (XDA20060502), the National Natural Science Foundation of China (42076017), and the Fundamental Research Funds for the Central Universities (B200201011). The authors declare no conflicts of interest relevant to this study.
Data availability statement.
The ECCO, IAP, and Ishii datasets can be obtained from https://web.corral.tacc.utexas.edu/OceanProjects/ECCO/ECCOv4/Release3/, http://www.ocean.iap.ac.cn/pages/dataService/dataService.html?languageType=en, and https://climate.mri-jma.go.jp/pub/ocean/. The WOD13 data are downloaded from https://www.ncei.noaa.gov/products/world-ocean-database. The RAMA mooring data are obtained from https://www.pmel.noaa.gov/gtmba/pmel-theme/indian-ocean-rama. The global temperature change index can be accessed at https://www.ncei.noaa.gov/access/monitoring/climate-at-a-glance/global/time-series. The EEMD code can be downloaded from http://khsci.com/docs/index.php/2020/04/09/1/.
REFERENCES
Alory, G., and G. Meyers, 2009: Warming of the upper equatorial Indian Ocean and changes in the heat budget (1960–99). J. Climate, 22, 93–113, https://doi.org/10.1175/2008JCLI2330.1.
Alory, G., S. Wijffels, and G. Meyers, 2007: Observed temperature trends in the Indian Ocean over 1960–1999 and associated mechanisms. Geophys. Res. Lett., 34, L02606, https://doi.org/10.1029/2006GL028044.
Belonenko, T. V., I. L. Bashmachnikov, and A. A. Kubryakov, 2018: Horizontal advection of temperature and salinity by Rossby waves in the North Pacific. Int. J. Remote Sens., 39, 2177–2188, https://doi.org/10.1080/01431161.2017.1420932.
Bourgeois, T., N. Goris, J. Schwinger, and J. F. Tjiputra, 2022: Stratification constrains future heat and carbon uptake in the Southern Ocean between 30°S and 55°S. Nat. Commun., 13, 340, https://doi.org/10.1038/s41467-022-27979-5.
Boyer, T. P., and Coauthors, 2013: World Ocean Database 2013. NOAA Atlas NESDIS 72, 209 pp.
Capotondi, A., M. A. Alexander, N. A. Bond, E. N. Curchitser, and J. D. Scott, 2012: Enhanced upper ocean stratification with climate change in the CMIP3 models. J. Geophys. Res., 117, C04031, https://doi.org/10.1029/2011JC007409.
Castino, F., B. Bookhagen, and M. R. Strecker, 2016: River‐discharge dynamics in the southern Central Andes and the 1976–77 global climate shift. Geophys. Res. Lett., 43, 11 679–11 687, https://doi.org/10.1002/2016GL070868.
Chen, X., and J. M. Wallace, 2016: Orthogonal PDO and ENSO indices. J. Climate, 29, 3883–3892, https://doi.org/10.1175/JCLI-D-15-0684.1.
Cheng, J., Z. Liu, S. Zhang, W. Liu, L. Dong, P. Liu, and H. Li, 2016: Reduced interdecadal variability of Atlantic meridional overturning circulation under global warming. Proc. Natl. Acad. Sci. USA, 113, 3175–3178, https://doi.org/10.1073/pnas.1519827113.
Cheng, L., and J. Zhu, 2016: Benefits of CMIP5 multimodel ensemble in reconstructing historical ocean subsurface temperature variations. J. Climate, 29, 5393–5416, https://doi.org/10.1175/JCLI-D-15-0730.1.
Cheng, L., J. Zhu, R. Cowley, T. Boyer, and S. Wijffels, 2014: Time, probe type, and temperature variable bias corrections to historical expendable bathythermograph observations. J. Atmos. Oceanic Technol., 31, 1793–1825, https://doi.org/10.1175/JTECH-D-13-00197.1.
Cheng, L., K. E. Trenberth, J. Fasullo, T. Boyer, J. Abraham, and J. Zhu, 2017: Improved estimates of ocean heat content from 1960 to 2015. Sci. Adv., 3, e1601545, https://doi.org/10.1126/sciadv.1601545.
Cnossen, I., and C. Franzke, 2014: The role of the sun in long-term change in the F2 peak ionosphere: New insights from EEMD and numerical modeling. J. Geophys. Res. Space Phys., 119, 8610–8623, https://doi.org/10.1002/2014JA020048.
Da Silva, R., and Coauthors, 2017: Salinity stratification controlled productivity variation over 300 ky in the Bay of Bengal. Sci. Rep., 7, 14439, https://doi.org/10.1038/s41598-017-14781-3.
de Boer, G. J., J. D. Pietrzak, and J. C. Winterwerp, 2008: Using the potential energy anomaly equation to investigate tidal straining and advection of stratification in a region of freshwater influence. Ocean Modell., 22 (1–2), 1–11, https://doi.org/10.1016/j.ocemod.2007.12.003.
Du, Y., T. Qu, and G. Meyers, 2008: Interannual variability of sea surface temperature off Java and Sumatra in a global GCM. J. Climate, 21, 2451–2465, https://doi.org/10.1175/2007JCLI1753.1.
Durack, P. J., S. E. Wijffels, and R. J. Matear, 2012: Ocean salinities reveal strong global water cycle intensification during 1950 to 2000. Science, 336, 455–458, https://doi.org/10.1126/science.1212222.
Duvel, J. P., R. Roca, and J. Vialard, 2004: Ocean mixed layer temperature variations induced by intraseasonal convective perturbations over the Indian Ocean. J. Atmos. Sci., 61, 1004–1023, https://doi.org/10.1175/1520-0469(2004)061<1004:OMLTVI>2.0.CO;2.
England, M. H., and Coauthors, 2014: Recent intensification of wind-driven circulation in the Pacific and the ongoing warming hiatus. Nat. Climate Change, 4, 222–227, https://doi.org/10.1038/nclimate2106.
Forget, G., J.-M. Campin, P. Heimbach, C. N. Hill, R. M. Ponte, and C. Wunsch, 2015: ECCO version 4: An integrated framework for non-linear inverse modeling and global ocean state estimation. Geosci. Model Dev., 8, 3071–3104, https://doi.org/10.5194/gmd-8-3071-2015.
Franzke, C., 2010: Long-range dependence and climate noise characteristics of Antarctic temperature data. J. Climate, 23, 6074–6081, https://doi.org/10.1175/2010JCLI3654.1.
Franzke, C., 2014: Nonlinear climate change. Nat. Climate Change, 4, 423–424, https://doi.org/10.1038/nclimate2245.
Franzke, C., and T. Woollings, 2011: On the persistence and predictability properties of North Atlantic climate variability. J. Climate, 24, 466–472, https://doi.org/10.1175/2010JCLI3739.1.
Gévaudan, M., J. Jouanno, F. Durand, G. Morvan, L. Renault, and G. Samson, 2021: Influence of ocean salinity stratification on the tropical Atlantic Ocean surface. Climate Dyn., 57, 321–340, https://doi.org/10.1007/s00382-021-05713-z.
Halkides, D., T. Lee, and S. Kida, 2011: Mechanisms controlling the seasonal mixed-layer temperature and salinity of the Indonesian Seas. Ocean Dyn., 61, 481–495, https://doi.org/10.1007/s10236-010-0374-3.
Hermes, J. C., and C. J. Reason, 2008: Annual cycle of the South Indian Ocean (Seychelles-Chagos) thermocline ridge in a regional ocean model. J. Geophys. Res., 113, C04035, https://doi.org/10.1029/2007JC004363.
Hoerling, M. P., J. W. Hurrell, T. Xu, G. T. Bates, and A. S. Phillips, 2004: Twentieth Century North Atlantic climate change. Part II: Understanding the effect of Indian Ocean warming. Climate Dyn., 23, 391–405, https://doi.org/10.1007/s00382-004-0433-x.
Horii, T., I. Ueki, and K. Ando, 2020: Coastal upwelling events, salinity stratification, and barrier layer observed along the southwestern coast of Sumatra. J. Geophys. Res. Oceans, 125, e2020JC016287, https://doi.org/10.1029/2020JC016287.
Hu, S., and J. Sprintall, 2017: Observed strengthening of interbasin exchange via the Indonesian seas due to rainfall intensification. Geophys. Res. Lett., 44, 1448–1456, https://doi.org/10.1002/2016GL072494.
Hu, S., and Coauthors, 2019: Interannual to decadal variability of upper-ocean salinity in the southern Indian Ocean and the role of the Indonesian throughflow. J. Climate, 32, 6403–6421, https://doi.org/10.1175/JCLI-D-19-0056.1.
Huang, N. E., and Coauthors, 1998: The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis. Proc. Roy. Soc., 454A, 903–995, https://doi.org/10.1098/rspa.1998.0193.
Ishii, M., Y. Fukuda, S. Hirahara, S. Yasui, T. Suzuki, and K. Sato, 2017: Accuracy of global upper ocean heat content estimation expected from present observational data sets. SOLA, 13, 163–167, https://doi.org/10.2151/sola.2017-030.
Jangir, B., D. Swain, and T. V. Udaya Bhaskar, 2016: Relation between tropical cyclone heat potential and cyclone intensity in the North Indian Ocean. Proc. SPIE, 9882, 988228, https://doi.org/10.1117/12.2228033.
Ji, F., Z. Wu, J. Huang, and E. P. Chassignet, 2014: Evolution of land surface air temperature trend. Nat. Climate Change, 4, 462–466, https://doi.org/10.1038/nclimate2223.
Jones, C. S., D. M. Legler, and J. J. O’Brien, 1995: Variability of surface fluxes over the Indian Ocean: 1960–1989. Global Atmos. Ocean Syst., 3, 249–272.
Joshi, M., R. Glasow, R. S. Smith, C. G. M. Paxton, A. C. Maycock, D. J. Lunt, C. Loptson, and P. Markwick, 2017: Global warming and ocean stratification: A potential result of large extraterrestrial impacts. Geophys. Res. Lett., 44, 3841–3848, https://doi.org/10.1002/2017GL073330.
Kumari, A., S. P. Kumar, and A. Chakraborty, 2018: Seasonal and interannual variability in the barrier layer of the Bay of Bengal. J. Geophys. Res. Oceans, 123, 1001–1015, https://doi.org/10.1002/2017JC013213.
Lee, T., 2004: Decadal weakening of the shallow overturning circulation in the South Indian Ocean. Geophys. Res. Lett., 31, L18305, https://doi.org/10.1029/2004GL020884.
Li, G., L. Cheng, J. Zhu, K. E. Trenberth, M. E. Mann, and J. P. Abraham, 2020: Increasing ocean stratification over the past half-century. Nat. Climate Change, 10, 1116–1123, https://doi.org/10.1038/s41558-020-00918-2.
Li, K., and F. Zheng, 2022: Effects of a freshening trend on upper-ocean stratification over the central tropical Pacific and their representation by CMIP6 models. Deep-Sea Res. II, 195, 104999, https://doi.org/10.1016/j.dsr2.2021.104999.
Li, Y., W. Han, and L. Zhang, 2017a: Enhanced decadal warming of the Southeast Indian Ocean during the recent global surface warming slowdown. Geophys. Res. Lett., 44, 9876–9884, https://doi.org/10.1002/2017GL075050.
Li, Y., W. Han, W. Wang, M. Ravichandran, T. Lee, and T. Shinoda, 2017b: Bay of Bengal salinity stratification and Indian summer monsoon intraseasonal oscillation: 2. Impact on SST and convection. J. Geophys. Res. Oceans, 122, 4312–4328, https://doi.org/10.1002/2017JC012692.
Luo, J.-J., W. Sasaki, and Y. Masumoto, 2012: Indian Ocean warming modulates Pacific climate change. Proc. Natl. Acad. Sci. USA, 109, 18 701–18 706, https://doi.org/10.1073/pnas.1210239109.
Maes, C., and T. J. O’Kane, 2014: Seasonal variations of the upper ocean salinity stratification in the tropics. J. Geophys. Res. Oceans, 119, 1706–1722, https://doi.org/10.1002/2013JC009366.
Makarim, S., J. Sprintall, Z. Liu, W. Yu, A. Santoso, X.-H. Yan, and R. D. Susanto, 2019: Previously unidentified Indonesian throughflow pathways and freshening in the Indian Ocean during recent decades. Sci. Rep., 9, 7364, https://doi.org/10.1038/s41598-019-43841-z.
Marshall, J., A. Adcroft, C. Hill, L. Perelman, and C. Heisey, 1997a: A finite-volume, incompressible Navier Stokes model for studies of the ocean on parallel computers. J. Geophys. Res., 102, 5753–5766, https://doi.org/10.1029/96JC02775.
Marshall, J., C. Hill, L. Perelman, and A. Adcroft, 1997b: Hydrostatic, quasi-hydrostatic, and nonhydrostatic ocean modeling. J. Geophys. Res., 102, 5733–5752, https://doi.org/10.1029/96JC02776.
Miyama, T., J. P. McCreary, T. G. Jensen, J. Loschnigg, S. Godfrey, and A. Ishida, 2003: Structure and dynamics of the Indian-ocean cross-equatorial cell. Deep-Sea Res. II, 50, 2023–2047, https://doi.org/10.1016/S0967-0645(03)00044-4.
Nagura, M., Y. Masumoto, and T. Horii, 2014: Meridional heat advection due to mixed Rossby gravity waves in the equatorial Indian Ocean. J. Phys. Oceanogr., 44, 343–358, https://doi.org/10.1175/JPO-D-13-0141.1.
Pfeiffer, M., J. Zinke, W.-C. Dullo, D. Garbe-Schönberg, M. Latif, and M. E. Weber, 2017: Indian Ocean corals reveal crucial role of World War II bias for twentieth century warming estimates. Sci. Rep., 7, 14434, https://doi.org/10.1038/s41598-017-14352-6.
Ponte, R. M., and N. T. Vinogradova, 2016: An assessment of basic processes controlling mean surface salinity over the global ocean. Geophys. Res. Lett., 43, 7052–7058, https://doi.org/10.1002/2016GL069857.
Rathore, S., N. L. Bindoff, C. C. Ummenhofer, H. E. Phillips, and M. Feng, 2020: Near-surface salinity reveals the oceanic sources of moisture for Australian precipitation through atmospheric moisture transport. J. Climate, 33, 6707–6730, https://doi.org/10.1175/JCLI-D-19-0579.1.
Rathore, S., N. L. Bindoff, C. C. Ummenhofer, H. E. Phillips, M. Feng, and M. Mishra, 2021: Improving Australian rainfall prediction using sea surface salinity. J. Climate, 34, 2473–2490, https://doi.org/10.1175/JCLI-D-20-0625.1.
Rathore, S., R. Goyal, B. Jangir, C. C. Ummenhofer, M. Feng, and M. Mishra, 2022: Interactions between a marine heatwave and Tropical Cyclone Amphan in the Bay of Bengal in 2020. Front. Climate, 4, 861477, https://doi.org/10.3389/fclim.2022.861477.
Reichl, B. G., A. Adcroft, S. M. Griffies, and R. Hallberg, 2022: A potential energy analysis of ocean surface mixed layers. J. Geophys. Res. Oceans, 127, e2021JC018140, https://doi.org/10.1029/2021JC018140.
Resplandy, L., J. Vialard, M. Lévy, O. Aumont, and Y. Dandonneau, 2009: Seasonal and intraseasonal biogeochemical variability in the thermocline ridge of the southern tropical Indian Ocean. J. Geophys. Res., 114, C07024, https://doi.org/10.1029/2008JC005246.
Roxy, M. K., and Coauthors, 2016: A reduction in marine primary productivity driven by rapid warming over the tropical Indian Ocean. Geophys. Res. Lett., 43, 826–833, https://doi.org/10.1002/2015GL066979.
Santoso, A., A. Sen Gupta, and M. H. England, 2010: Genesis of Indian Ocean mixed layer temperature anomalies: A heat budget analysis. J. Climate, 23, 5375–5403, https://doi.org/10.1175/2010JCLI3072.1.
Savita, A., and Coauthors, 2022: Quantifying spread in spatiotemporal changes of upper-ocean heat content estimates: An internationally coordinated comparison. J. Climate, 35, 851–875, https://doi.org/10.1175/JCLI-D-20-0603.1.
Shackelford, K., C. A. DeMott, P. J. van Leeuwen, E. Thompson, and S. Hagos, 2022: Rain‐induced stratification of the equatorial Indian Ocean and its potential feedback to the atmosphere. J. Geophys. Res. Oceans, 127, e2021JC018025, https://doi.org/10.1029/2021JC018025.
Simpson, J. H., 1981: The shelf-sea fronts: Implications of their existence and behaviour. Philos. Trans. Roy. Soc., A302, 531–546, https://doi.org/10.1098/rsta.1981.0181.
Simpson, J. H., and J. R. Hunter, 1974: Fronts in the Irish Sea. Nature, 250, 404–406, https://doi.org/10.1038/250404a0.
Somavilla, R., C. González‐Pola, and J. Fernández‐Diaz, 2017: The warmer the ocean surface, the shallower the mixed layer. How much of this is true? J. Geophys. Res. Oceans, 122, 7698–7716, https://doi.org/10.1002/2017JC013125.
Sun, Q., Y. Du, S.-P. Xie, Y. Zhang, M. Wang, and Y. Kosaka, 2021: Sea surface salinity change since 1950: Internal variability versus anthropogenic forcing. J. Climate, 34, 1305–1319, https://doi.org/10.1175/JCLI-D-20-0331.1.
Susanto, R. D., A. L. Gordon, and Q. Zheng, 2001: Upwelling along the coasts of Java and Sumatra and its relation to ENSO. Geophys. Res. Lett., 28, 1599–1602, https://doi.org/10.1029/2000GL011844.
Swapna, P., R. Krishnan, and J. M. Wallace, 2014: Indian Ocean and monsoon coupled interactions in a warming environment. Climate Dyn., 42, 2439–2454, https://doi.org/10.1007/s00382-013-1787-8.
Tierney, J. E., J. E. Smerdon, K. J. Anchukaitis, and R. Seager, 2013: Multidecadal variability in East African hydroclimate controlled by the Indian Ocean. Nature, 493, 389–392, https://doi.org/10.1038/nature11785.
Ummenhofer, C. C., S. A. Murty, J. Sprintall, T. Lee, and N. J. Abram, 2021: Heat and freshwater changes in the Indian Ocean region. Nat. Rev. Earth Environ., 2, 525–541, https://doi.org/10.1038/s43017-021-00192-6.
Venkataramana, V., R. K. Mishra, P. Sabu, N. Anilkumar, A. Sarkar, R. K. Naik, M. A. Soares, and L. Gawade, 2021: Stratification governs the plankton community structure and trophic interaction in the southwestern tropical Indian Ocean during boreal summer. Reg. Stud. Mar. Sci., 48, 101987, https://doi.org/10.1016/j.rsma.2021.101987.
Vincent, E. M., K. A. Emanuel, M. Lengaigne, J. Vialard, and G. Madec, 2014: Influence of upper ocean stratification interannual variability on tropical cyclones. J. Adv. Model. Earth Syst., 6, 680–699, https://doi.org/10.1002/2014MS000327.
Wacongne, S., and R. Pacanowski, 1996: Seasonal heat transport in a primitive equations model of the tropical Indian Ocean. J. Phys. Oceanogr., 26, 2666–2699, https://doi.org/10.1175/1520-0485(1996)026<2666:SHTIAP>2.0.CO;2.
Wang, C., 2019: Three-ocean interactions and climate variability: A review and perspective. Climate Dyn., 53, 5119–5136, https://doi.org/10.1007/s00382-019-04930-x.
Wang, L., W. Chen, W. Zhou, J. C. L. Chan, D. Barriopedro, and R. H. Huang, 2010: Effect of the climate shift around mid-1970s on the relationship between wintertime Ural blocking circulation and East Asian climate. Int. J. Climatol., 30, 153–158, https://doi.org/10.1002/joc.1876.
Wu, Y., X.-T. Zheng, Q.-W. Sun, Y. Zhang, Y. Du, and L. Liu, 2021: Decadal variability of the upper-ocean salinity in the southeast Indian Ocean: Role of local ocean–atmosphere dynamics. J. Climate, 34, 7927–7942, https://doi.org/10.1175/JCLI-D-21-0122.1.
Wu, Z., and N. E. Huang, 2009: Ensemble empirical mode decomposition: A noise-assisted data analysis method. Adv. Adapt. Data Anal., 1 (1), 1–41, https://doi.org/10.1142/S1793536909000047.
Xu, X., L. Wang, and W. Yu, 2021: The unique mean seasonal cycle in the Indian Ocean anchors its various air-sea coupled modes across the basin. Sci. Rep., 11, 5632, https://doi.org/10.1038/s41598-021-84936-w.
Yadidya, B., and A. D. Rao, 2022: Interannual variability of internal tides in the Andaman Sea: An effect of Indian Ocean Dipole. Sci. Rep., 12, 11104, https://doi.org/10.1038/s41598-022-15301-8.
Yamaguchi, R., and T. Suga, 2019: Trend and variability in global upper‐ocean stratification since the 1960s. J. Geophys. Res. Oceans, 124, 8933–8948, https://doi.org/10.1029/2019JC015439.
Zhang, L., Y. Du, and W. Cai, 2018: A spurious positive Indian Ocean Dipole in 2017. Sci. Bull., 63, 1170–1172, https://doi.org/10.1016/j.scib.2018.08.001.
Zhang, L., W. Han, K. B. Karnauskas, G. A. Meehl, A. Hu, N. Rosenbloom, and T. Shinoda, 2019: Indian Ocean warming trend reduces Pacific warming response to anthropogenic greenhouse gases: An interbasin thermostat mechanism. Geophys. Res. Lett., 46, 10 882–10 890, https://doi.org/10.1029/2019GL084088.
Zhang, N., M. Feng, Y. Du, J. Lan, and S. E. Wijffels, 2016: Seasonal and interannual variations of mixed layer salinity in the southeast tropical Indian Ocean. J. Geophys. Res. Oceans, 121, 4716–4731, https://doi.org/10.1002/2016JC011854.
