Previous studies consistently show that Sri Lanka receives its second-highest rainfall and standard deviation during spring, indicating that inter-annual rainfall variability is most pronounced during spring (Karunapala and Yoo, 2020). Our analysis of the monthly climatology further confirms that Sri Lanka receives its second-highest rainfall during the spring season (Fig. 1a), and among these, the southwestern region (wet zone) of Sri Lanka receives the highest rainfall, exceeding 8 mm/day (Fig. 1b). Next, to investigate the effects of pure CP and EP El Niño events on El Niño-induced rainfall during the El Niño decaying spring (hereafter spring) in Sri Lanka, first, we performed a multiple linear regression analysis based on Niño4, Niño3, and DMI indices to examine the sensitivity of spring rainfall of Sri Lanka (Fig. S3). It is found that the rainfall responses to CP and EP El Niño events are different, indicating that CP El Niño is likely to favor a wet spring, while EP El Niño is likely to favor a dry spring in Sri Lanka. Therefore, given the importance of these different effects, we secondly performed a composite analysis (Fig. 1 and S4). Since our interest in the present study is in pure El Niño events, we specifically selected two pure EP El Niño events (1986/87 and 1991/92) and two pure CP events (1987/88 and 2009/10), and the selected events are consistent with Zhang et al. (2022). Consistent with regression analysis (Fig. S3) it is found that during pure CP El Niño events, precipitation in the El Niño decaying spring is enhanced over Sri Lanka, leading to wetter conditions, especially in the country's wet zone (Fig. 1c). However, during pure EP events, spring rainfall is anomalously weakened, leading to drier conditions, especially over the wet zone of the country (Fig. 1d). Thus, pure CP (EP) El Niño events appear to be associated with a wet (dry) spring in Sri Lanka. This contrast in spring rainfall responses to pure El Niño events highlights two distinct effects. We also examined the differences in spring rainfall using ERA5 data and found consistent results with CHIRPS (Fig. S4). In terms of spatial variability, the study also shows substantial changes in rainfall patterns within the wet zone during the spring of both pure EP and CP El Niño events. These variations in spring rainfall across Sri Lanka in response to different types of pure El Niño events underline differences in the associated ocean-atmosphere dynamics. Therefore, we further investigate these differences in ocean-atmosphere dynamics to identify the reasons for the different responses of spring rainfall to CP and EP El Niño events.
We next examined the rainfall over the tropical Indo-Pacific region during the peak winter (here after winter) and spring of pure CP and EP El Niño events (Figs. S5 and S6), and the composite difference (pure CP minus pure EP El Niño) is shown in Fig. 2. Comparing the composite difference in two types of El Niño, it is clear that the positive rainfall anomaly over India and Sri Lanka is stronger during pure CP El Niño than during EP El Niño in winter (Fig. 2a). It is also clear that there is a positive precipitation anomaly over India and Sri Lanka in spring, which is stronger than in winter (Fig. 2b). Thus, it can be seen that pure CP El Niño events appear to be accompanied by strong wet conditions in spring than in winter over Sri Lanka and India. Previous studies have shown that El Niño events have a significant impact on precipitation through SST changes in the tropical Indo-Pacific region (Feng et al. 2011; Rifai et al. 2019). Therefore, we investigated the importance of El Niño-associated SST and lower-tropospheric circulation in the tropical Indo-Pacific during pure CP and EP El Niño events. We find that during winter, the CP El Niño-induced SST in the tropical IO is weakly positive and associated with weak cyclonic circulation (Fig. 2c). However, during spring, both the SST and the cyclonic circulation become stronger than in winter (Fig. 2d), favoring a strong increase in precipitation over the Sri Lanka, leading to an enhanced wetter condition. To investigate the response of precipitation and SST to El Niño during spring over the IO, we also performed a multiple linear regression analysis based on Niño4, Niño3, and DMI indices (Fig. S7). We found that the responses of precipitation and SST to CP and EP El Niño events are similar to those shown in Figs. S5 and S6. Thus, it is clear that CP and EP El Niño have a pronounced effect on the rainfall of Sri Lanka during the El Niño decaying spring.
Given the importance of moisture supply for rainfall of Sri Lanka (Karunapala and Yoo 2023; Kajakokulan et al. 2023b), and to understand the marked changes in spring rainfall, we examined the vertically integrated moisture flux convergence (VIMFC) and 850hPa wind anomalies associated with pure CP and EP El Niño over the IO (Fig. 3). In Fig. 3, positive values of VIMFC indicate convergence and negative values divergence. It can be seen that strong cyclonic circulation over the Arabian Sea (AS) during the pure CP El Niño causes anomalous moisture flux convergence, leading to increased precipitation over the Sri Lanka in spring (Fig. 3a). Conversely, moisture flux convergence anomalies become negative (divergent) due to the presence of anti-cyclonic circulation over the AS during pure EP El Niño events, suppressing precipitation over the Sri Lanka (Fig. 3b). Furthermore, it is found that during spring, there is a notable difference in the lower-level circulation between the pure CP and EP El Niño events, that is in agreement with previous studies by Chen et al. (2023), which is due to the presence of strong convergence over South Asia. Furthermore, the composite difference map also suggests that moisture convergence is higher during pure CP El Niño compared to pure EP El Niño in spring (Fig. 3c). The composite analysis of outgoing longwave radiation also reinforces the findings from the precipitation response in CP and EP El Niño events (Fig. S8). Thus, the present results confirm that pure CP El Niño events are typically associated with enhanced precipitation over Sri Lanka during spring, whereas pure EP El Niño events are associated with suppressed precipitation, and that these differences are associated with changes in SST, low-level circulation and moisture availability.
The SST contrast between the CP and EP El Niño can lead to different responses of the Walker circulation, affecting the large-scale descending and ascending motions over the tropical region (Xu et al. 2013). Therefore, to investigate the potential link between Sri Lanka rainfall and changes in the Walker circulation, we analyzed composites of vertical velocity anomalies during the pure CP and EP El Niño during El Niño decaying spring (Fig. 4). In Fig. 4, positive (negative) vertical velocity values indicate downward (upward) motion. It can be seen that the upward motion over the IO in spring is stronger during the pure CP El Niño, and therefore, the upward motion over Sri Lanka (79–82°E) is stronger during CP El Niño (Fig. 4a). However, during pure EP El Niño events, the upward motion in the IO becomes weaker and the subsidence over the Sri Lanka becomes stronger (Fig. 4b). Such vertical motion responses and the moisture availability are likely to lead to an enhancement (suppression) of spring rainfall during pure CP (EP) El Niño events. The composite difference map of vertical velocity also shows that the upward motion is positive during the pure CP El Niño compared to the pure EP El Niño in spring (Fig. 4c). As shown in Fig. 2, the presence of basin-wide warmer SST during pure CP El Niño events is consistent with a single Walker cell over IO, whereas the cooler SST during pure EP El Niño events is consistent with a double Walker cell. Yu et al. (2021) also investigated the Walker circulation cell in the IO and found that a single (double) cell is associated with the CP (EP) El Niño. Thus, our results show that there is an important relationship between SST changes and the Walker circulation in the tropical IO, which may influence the rainfall of Sri Lanka during pure El Niño decaying spring.