Introduction

The Indian Ocean Basin (IOB) mode is the first mode of the interannual variability of sea surface temperature in the tropical Indian Ocean (TIO), which is characterized by basin-warming/cooling (Klein et al. 1999; Yang et al. 3b). The downwelling Rossby waves deepened the thermocline, leading to thermocline warming in the south TIO. Meanwhile, the mixed layer is deepened by the Ekman pum**. During the mature phase (SON) of IOD, the positive wind stress curl anomaly became stronger and covered a larger area over the south TIO, forcing and strengthening the westward-propagating downwelling Rossby waves (Fig. 3b). Thus, the thermocline became deeper, and the resultant warming of the thermocline became more significant. Subsequently, the downwelling Rossby wave propagated westward to the thermocline dome in the SWTIO, which deepened the thermocline and caused the SST warming in the SWTIO (Fig. 3c; Chowdary et al. 2009). In DJF, the downwelling Rossby waves and associated thermocline fluctuation reached their peak (Fig. 3d). The SST warming in the SWTIO was maintained by the slow-propagating downwelling Rossby waves and local wind forcing until the 2020 summer (Fig. 3e). This breaks through our previous understanding that SWTIO SST warming cannot persist until the following summer during the pure IOD events.

Fig. 3
figure 3

Hovmöller diagrams of (a) SSTa (shading, ℃) and precipitation anomalies (contours, mm/day), and b SSHa (shading, m) and wind stress curl anomalies (contours, 10–8 N/m3) averaged in 8–12°S. Green arrows denote the westward-propagating Rossby waves. c–e Vertical distribution of potential temperature anomalies (shading, ℃), MLD (black dashed lines), and D20 (solid black lines) averaged between 8 and 12°S in 2019 SON, 2019 D(2020)JF, 2020 MAM, respectively. Gray lines in c–e are climatological means for the period 2004–2020

Full size image

The SST warming in the SWTIO-induced heavy precipitation by strengthening local deep convection during spring to summer (Fig. 4). Therefore, equatorially antisymmetric SST and convection patterns appeared in the TIO (Fig. 4). Consistently, the surface wind anomalies feature an antisymmetric circulation structure over the TIO (Fig. 4a; Kawamura et al. 2001; Wu et al. 2008; Wu and Yeh 2010). A cross-equatorial “C-shaped” wind anomaly over the TIO, with northeasterlies north and northwesterlies south of the equator, was forced by the north–south temperature gradient and Coriolis effect. In turn, the cross-equatorial “C-shaped” wind anomaly also favored the north–south SST gradient, because surface wind caused less latent heat flux loss over the southern TIO. Thus, in spring, positive WES feedback works in the southern TIO, supporting the maintenance of the antisymmetric mode (Wu et al. 2008; Wu and Yeh 2010). In summer, the anomalous northeasterlies weakened the climatological monsoon southwesterlies over the northern TIO, leading to an increase in SST due to the positive WES feedback mechanism (Fig. 4; **e and Philander 1994; Du et al. 2009; Chen et al. 2019). Meanwhile, the enhanced precipitation extended to the northern TIO, with the development of SST warming and summer monsoon in the TIO (Fig. 4; Annamalai et al. 2005; Izumo et al. 2008).

Fig. 4
figure 4

a SSTa (shading, ℃) and wind anomalies (vectors, m/s) zonally averaged between 40° and 100°E. b latent heat flux anomalies (shading, W/m2) and precipitation anomalies (contours, mm/day) zonally averaged between 40° and 100°E

Full size image

In addition, the westward-propagating downwelling Rossby waves in the south TIO transformed into the equatorward-propagating coastal-trapped waves after reaching the western boundary and then reflected as the equatorial Kelvin waves that propagate eastward along the equator (Le Blanc and Boulanger 2001; McPhaden and Nagura 2014; Wang et al. 2016; Chen et al. 2019). Thus, the downwelling Rossby waves and reflected-equatorial Kelvin waves sustained the warming of the western TIO and west–east temperature gradient from 2019 winter to the following spring (Fig. 5a, b; Jury and Huang 2004). In early spring, the west–east temperature gradient forced the equatorial easterly wind anomaly (Figs. 4a, 5b; Du et al. 2020), which favored the generation of downwelling Rossby waves and westward heat advection, further maintaining the warming in the western TIO.

Fig. 5
figure 5

Hovmöller diagrams of (a) monthly SSTa (shading, ℃) and precipitation anomalies (contours, mm/day), (b) daily SSHa (shading, m) and zonal wind velocity anomalies (contours, m/s), and (c) monthly net air–sea heat flux anomalies (shading, W/m2) and latent heat flux anomalies (contours, W/m2) averaged in 1°S–1°N. Green arrows in b denote the eastward-propagating Kelvin waves

Full size image

In the eastern TIO, the SST warming was affected by the air–sea interface exchanges associated with cloud-radiation-SST feedback and WES feedback, resulting from the weakening of anomalous subsidence of Walker Circulation over the east TIO. During the boreal winter of 2019, a reduction of cloud cover, induced by the cooling pole of IOD off the Sumatran southwest, favored an increase in the shortwave radiation (Figure not shown; Cai and Qiu 2013; Liu et al. 2014). In addition, the strong southeasterly wind anomaly weakened during the decay phase of IOD, resulting in the SST warming due to a decrease in the heat latent flux loss of the ocean (Figs. 4b, 5c; Tokinaga and Tanimoto 2004). In later spring, the reflected-equatorial Kelvin reached the Sumatra-Java coasts, sustaining SST warming in the eastern TIO (Fig. 5a, b). An increase in precipitation followed the SST warming in the eastern TIO (Figs. 2c, 5a).

Summary and discussion

The TIO experienced a basin-wide warming in 2020, following an extreme and prolonged positive IOD event instead of an El Niño event. This is the first record since the 1960s. Persistent warming occurred in the SWTIO from late 2019 to early 2020, sustained by oceanic downwelling Rossby waves associated with the extreme 2019 IOD via thermocline-SST feedback. During 2020 spring to early summer, the SWTIO warming triggered the equatorially antisymmetric SST, precipitation and surface wind patterns over the TIO. The cross-equatorial “C-shaped” wind anomaly, with northeasterly–northwesterly wind anomaly north–south of the equator, weakened the climatological surface winds then led to basin-wide warming via reducing surface evaporation. Moreover, the westward-propagating Rossby waves reflected as the eastward-propagating equatorial Kelvin waves, which favored the persistence of warming in the western TIO in 2020 spring. In the eastern TIO, air–sea interface exchanges play an important role in SST warming. The importance of air–sea coupling processes associated with the independent and extreme IOD in the TIO basin-warming mode has been clarified in this study, which allows us to rethink the relationship between the Indo-Pacific climate modes.

In addition to the local forcing of the TIO, the remote forcing from the Pacific might also impact the development of IOB warming in 2020. A weak El Niño Modoki developed in the tropical Pacific during 2019. Zhou et al. (2021) suggested that the warming in the central-western tropical Pacific during the weak El Niño Modoki contributed to the sustained anticyclone wind curls over the south TIO in early 2020 according to the results of the atmospheric model experiments. Zhang L et al. (2021) further indicated that the warming of the tropical Pacific contributes to the tropical Indian Ocean wind anomalies, regardless of whether the warming is in the western or eastern tropical Pacific. The tropical Pacific warming causes the Pacific convection center to shift eastward, and then triggers atmospheric Rossby waves over the Indian Ocean, resulting in a pair of low-level anomalous anticyclones occurring on both sides of the equator. Nevertheless, our results in this study challenge the perception that only ENSO triggers the IOB. The relationship between ENSO and IOB would change with ocean mean state and ENSO activity on long timescales (**e et al. 2010; Zheng et al. 2011, 2016; Hu et al. 2013; Tao et al. 2015; Liu et al. 2021). In a warming climate, both extreme positive IOD events and El Niño events are projected to become more frequent (Cai et al. 2014; Freund et al. 2020). Thus, it is worth paying attention to changes in the relationship between the Indo-Pacific climate modes, and their relative contributions to climate variability in the Indo-Pacific.