Suppression of marine heatwave activity by tropical cyclone–induced upper ocean cooling

Iam-Fei Pun1*, I-I Lin2, Chun-Chieh Wu2
Marine heatwaves (MHWs) have drawn substantial scientific attention due to their profound biological and eco- nomic impacts, especially in the context of climate change. This study highlights the crucial role of tropical cy- clones (TCs) in disrupting MHWs by altering ocean temperatures. As demonstrated by Tropical Cyclone Bavi, intense sea surface temperature (SST) cooling, reaching up to 7°C, immediately terminated a mature MHW and suppressed subsequent MHW development for up to 8 months. More generally, on the basis of post-TC MHW cases (reoccurring within 1 month) during 2014 to 2023, we found that MHW disruption duration is proportional to the magnitude of TC-induced SST cooling, ~5 to 7 days per degree of cooling. At a broader scale, a 20-year (2004 to 2023) analysis revealed 99% significant inverse relationships between seasonal MHW and TC activity. These findings suggest that TCs can act as natural regulators, effectively mitigating heat stress from MHWs and poten- tially serving as a lifeline for marine ecosystems.

INTRODUCTION
As the global average temperature record has been constantly reset in recent years due to the increasing concentration of greenhouse gases (1), extreme temperature events occur more often than at any time in history (2-6). These extreme events can exert adverse im- pacts on the environment, as well as the human society. Given the increasing number of high-impact events worldwide (7-13), marine heatwaves (MHWs; see the definition in Materials and Methods) have drawn high attention from researchers in recent years (3, 14). Accompanying a period of consecutive anomalously warm ocean temperatures, MHWs have prominent, sometimes devastating, con- sequences on marine ecosystems in many aspects. For example, the long-lasting warm water temperatures can lead to the bleaching of corals and the large-scale vanishing of seagrass fields (15-18); both serve as essential habitats and feeding grounds for marine organ- isms (19). The structure and community of species in the area struck by MHW or nearby areas are expected to shift and be modified markedly as different species have different tolerances to heat (20- 22). Moreover, these ecological impacts would immediately result in socioeconomic consequences. For example, fisheries, aquaculture, and even the tourism could be substantially affected by the changing ecosystem due to MHWs (19, 23, 24). The effects of MHWs can be disastrous and far-reaching; understanding the mechanisms for their formation, development and dissipation is, therefore, a pressing issue.
Over the past decade, the causes of MHWs have been extensively studied (25). It is clear that MHWs are driven by atmospheric and/ or oceanic processes through local or remote forcing. Increased solar insolation, reduced surface wind speed, and warmed air tempera- ture associated with quasistationary anticyclonic pressure systems are the common drivers from the atmosphere (26-28). From the ocean, air-sea heat flux, current advection, and stratification are considered as the main players (11, 27, 29-31). Note that these atmospheric and oceanic drivers more often work simultaneously to heat up the ocean and generate MHWs. Recently, Pun et al. (32) suggested that tropical cyclones (TCs) can contribute to the development of MHW events. The subsidence region at the TC periphery tends to favor the

presence of clear skies with lower surface wind speed. This situation substantially enhances the incoming solar radiation and reduces surface heat fluxes, favoring the formation of MHWs. Pun et al. (32) also found that such type of MHW, in turn, aided the subsequent TC’s intensification through the air-sea heat flux feedback mechanism. To our understanding, works on the interaction between these two kinds of extreme natural phenomena, i.e., MHWs and TCs, are still in its premature stage (32-36).
To date, most studies have primarily focused on identifying the mechanisms behind the formation and development of MHWs, with relatively less attention devoted to investigate the causes of their dis- sipation. Understanding the factors contributing to MHW dissipa- tion is equally important as these factors can influence the duration of MHWs, which is directly tied to the severity and harmfulness of such events. On the other hand, these dissipation factors can be time- ly relief from the heat stress for marine life. Recent studies have no- ticed that MHWs are very likely to dissipate following the passage of TCs (32, 34). This is due to the pronounced sea surface cooling as- sociated with the strong vertical mixing induced by TCs (37-49). However, detailed investigations of what the extent TCs can dissi- pate or disrupt MHWs are still lacking in the existing literature.
In this study, we will first present a case study of Tropical Cyclone Bavi, which completely eliminated an MHW in the East China Sea (ECS) in 2020 by generating substantial sea surface temperature (SST) cooling. Following this event, no more MHWs occurred again in this area for the rest ofthe year. Second, on the basis of a systematic statistical analysis with 85 well-defined TC-MHW cases identified in the western North Pacific during 2014 to 2023, we demonstrate that the duration of MHW interruption, if they reemerge in the post-TC period, depends on the magnitude of TC-induced SST cooling. Fur- thermore, analysis of the 20-year dataset (2004 to 2023) reveals that seasonal MHW activity is strongly modulated by TC activity. These findings provide clear evidence that TCs can act as effective regulators of MHWs, potentially mitigating their impacts on marine ecosystems.

RESULTS
Disruption of MHW by Bavi
In 2020, Bavi intensified to its peak intensity (100 knots or 51.4 m s-1) in the ECS and became the strongest TC in the region in the past

decade. The rare intensification was found to be linked to an MHW preexisting on Bavi’s path (32). From the daily SST evolution in the ECS in late August 2020 (fig. S1), it is found that this MHW (indi- cated by “x” symbols) gradually developed around mid-August. A few days later (i.e., 19 to 24 August; fig. S1, E to J), it became more pro- nounced and covered a larger area. However, the MHW dissipated quickly after the passage of Bavi (i.e., 25 to 26 August; fig. S1, K and L).
For a quantitative comparison, we define 21 August and 26 Au- gust as the reference times before and after the TC, respectively. The SST map from 21 August reveals that the entire ECS was extremely warm before theTC, with SST values well above 30°C (Fig. 1A). Com- pared to the climatology, it is found that the SST anomalies, which are usually used to refer to the magnitude of MHWs when they hap- pen (50), reached up to 3°C, expanding from the southwest of the Korean Peninsula (Fig. 1B). A large area of SST well exceeded the

90th percentile of the long-term daily values, indicating the pres- ence ofan MHW (Fig. 1C). Note that not all of this area (>90th per- centile) is qualified as an MHW because it must also meet the minimum persistence requirement of 5 days. Nevertheless, the pre- TC ocean condition in the ECS was dominated by a prominent MHW, providing a favorable environment for TC intensification. By contrast, the SST map from 26 August illustrates that, right after the passage of Bavi, the MHW completely disappeared and replaced with a patch of cold SST (~22°C), particularly on the right-hand side of the TC track along the west coast of the Korean Peninsula (Fig. 1D). The SST dropped to 5°C below the climatology, or 3°C lower than the 10th percentile level of the long-term-averaged daily values, suggesting the occurrence of a potential cold spell (Fig. 1, E and F). This result clearly illustrates Bavi’s effective role in destroy- ing a mature MHW along its path.

It is well known that TC-induced vertical mixing with colder, deeper water primarily drives the decrease in SST. The SST cooling on 26 August 2020 with respect tothe SST on 21 August 2020 is shown in Fig. 2A. It can be seen that Bavi generated marked SST cooling in the ECS. The maximum magnitude of the SST cooling was up to 7°C, located southwest of Jeju Island. In addition to the high wind speeds, the strong SST cooling was also likely attributed to the presence of semipermanent cold subsurface water in the ECS (51, 52). This in- tense SST cooling erased the warm anomaly and brought the SST below the 10th percentile (Fig. 1F).
The impact of TC-induced SST cooling on subsequent MHW de- velopment and behavior is of particular importance and the focus of the present study. Notably, following the Bavi-induced cooling event, almost no MHWs reemerged in the affected area, as delin- eated by the 3°C SST cooling contour in Fig. 2B, for the rest of the year. However, a huge contrast is outside the cooling area, where up to 20 days of MHWs occurred in the northwest and southeast re- gions adjacent to the cooling zone. In the southern latitudes near Taiwan and the east of the Korean Peninsula, the cumulative MHW days can reach 30 for the rest of 2020. This indicates that large-scale environmental conditions were still favorable for MHW develop- ment. In addition, the area experiencing intense SST cooling took a substantially longer time, on average about 240 days (~8 months), to produce another MHW (Fig. 2C). In other words, MHW activity in this region was suppressed until April of the following year.
To further explore the impact of TC-induced SST cooling on MHW dissipation and subsequent development, the evolution of SST averaged over the main MHW area along the Bavi’s track was analyzed (fig. S2, A and B). It is observed that the MHW gradually developed before the TC but collapsed almost instantly when the TC arrived (fig. S2A). Again, this is due to the strong SST cooling (~7°C) induced by the TC. As a result, SSTs dipped below the 10th percentile, although this situation only lasted for 2 days. From the extended time series, it is found that, despite some fluctuations shortly after the TC, the post-TC temperatures remained low and lin- gered around the 10th percentile curve until early October (fig. S2B).

However, the SSTs were still below the climatological values until mid-November. Even if the SSTs appeared to return to normal at the end of 2020, no qualified MHWs were established. For comparison, the SST time series during a similar period (i.e., from mid-July to the end of August) in 2016 was analyzed (fig. S2C). The year of 2016 was chosen because the ECS experienced one of its warmest Augusts (30) and the TC activity in the ECS was quiet during that period. Therefore, it provides an opportunity to differentiate the influence of TCs on MHW dissipation. During this period, two MHW events were identified, i.e., 23 July to 1 August 2016 (lasting for 10 days) and 10 to 26 August 2016 (lasting for 17 days). These two MHWs were close together, with only 5 days whose SSTs were just slightly below the 90th percentile threshold in between (fig. S2C). Compared to the short-lived MHW (6 days) encountered by Bavi in 2020 (fig. S2A), it is suggested that MHWs can last substantially longer in the ECS without the interference of TCs. According to this com- parison, the duration of the MHW was reduced by about 65%, i.e., ( 1 − × 100%, due to the influence ofTCs, highlighting the cru- cial role of TCs in modulating MHW activity.
As noted above, MHW development within the cold wake re- gion following Bavi’s passage appeared to be suppressed for up to 8 months. Because atmospheric conditions also plays a crucial role in MHW formation, this prolonged MHW suppression may have been partially influenced by unfavorable atmospheric factors. To clarify the atmospheric contribution, we examined atmospheric conditions using the European Centre for Medium-Range Weather Forecasts (ECMWF) fifth generation of atmospheric reanalysis (ERA5) data- set (see Materials and Methods). In particular, surface net shortwave radiation (SWR) and 10-mwind speed (U10) were analyzed as these parameters are commonly recognized as key atmospheric drivers of MHWs (53, 54). During the 8-month period from 27 August 2020 (immediately following Bavi’s passage) to 30 April 2021 (when an MHW event reappeared in the cold wake area), four distinct MHW episodes were identified in the ECS. These occurred in November 2020, February 2021, March 2021, and April 2021, respectively

(Fig. 3). Notably, during the first three episodes, MHWs were only observed in regions surrounding the cold wake, whereas SSTs with- in the cold wake still remained relatively low (Fig. 3, A to C). Ac- cording to the reanalysis data, the cold wake region exhibited large-scale positive anomalies in SWR during the first three MHW episodes, suggesting generally favorable conditions for MHW de- velopment (fig. S3, A to C). However, the surface winds were less supportive, with negative values appearing only during the March 2021 event (fig. S4, A to C). Note that negative wind anomalies are considered favorable conditions as they limit heat loss to the atmo- sphere through reduced enthalpy fluxes. These findings indicate that atmospheric conditions during the first three MHW episodes were generally conductive to MHW development. Therefore, the pro- longed absence of MHWs within the cold wake region is unlikely attributed to unfavorable atmospheric forcing.
Another notable evidence appears in the final MHW episode in April 2021, when the cold wake region eventually experienced an

MHW event (Fig. 3, D to F). During this period, atmospheric condi- tions were generally favorable, with positive SWR anomalies (fig. S3, D to F) and negative wind speed anomalies prevailing across the ECS (fig. S4, D to F). However, the onset of MHWs within the cold wake region markedly lagged behind that of the surrounding waters. This delay further highlights the suppressive effect of Bavi-induced SST cooling on subsequent MHW formation, even under otherwise favorable atmospheric conditions.
Moreover, to better characterize favorable atmospheric condi- tions, we proposed a simple index, referred to as the atmospheric condition index (ACI), based on the parameters of SWR and U10 (see Materials and Methods). As shown in fig. S5, the areal-averaged ACI over the ECS reveals that consecutive favorable atmospheric conditions, as reflected by positive ACI values, were intermittently present throughout the post-Bavi period. This finding further sup- ports the notion that atmospheric forcing was unlikely the limiting factor inhibiting MHW formation in Bavi’s cold wake. Therefore,

the prolonged MHW suppression is more likely attributable to the SST cooling induced by Bavi.