Can we identify the sub-scale processes caused by tropical storms and their role in recovering temperatures from cold wakes? Can we explore their role in flooding and climate hazards in the summer season, when tropical monsoon winds hit the coasts of Iran due to blocking and atmospheric cold troughs?

Submesoscale processes (submesoscales) play a crucial role in vertical tracer transport, which are typically stronger in winter than summer. However, recent studies revealed enhanced submesoscale activities following tropical cyclones (TCs) during summer months, though their characteristics, mechanisms, and dynamical impacts remain obscure. Through combining high‐resolution satellite and simulation data, we demonstrate that energetic submesoscales are present within TC‐induced cold wakes across the northwestern Pacific, which can produce strong upward vertical heat flux (VHF) in the upper ocean. Further analysis reveals thatsubmesoscales are more active in stronger cold wakes induced by slower‐moving and stronger TCs and they may be generated through a combination of baroclinic instability and strain‐induced frontogenesis. Crucially, submesoscale VHF can accelerate the post‐TC temperature recovery of cold wakes which may influence the sequential TC intensification. Our results improve the understanding of TC‐ocean feedback and underscore the necessity of resolving or parameterizing submesoscales in TC prediction models. Plain Language Summary Oceanic submesoscale processes with horizontal scales of O(0.1–10) km are key drivers of vertical tracer transport such as heat and nutrients. While these processes are typically more energetic in winter, recent observations show that they can also intensify during summer following tropical cyclones (TCs). Using high‐resolution satellite data and simulations, we found that strong submesoscale processes occurred within cold wakes left behind by TCs in the northwestern Pacific. These submesoscale processes generate intense upward heat transport and are particularly pronounced in strong cold wakes caused by powerful and slow‐moving TCs. They are likely generated by a combination of baroclinic instability and frontogenesis in the cold wakes. Importantly, the upward heat transport caused by submesoscale processes helps warm the cold wakesfaster after a TC passes, potentially affecting how rapidly later TCsintensify over the same area. Our findings advance the understanding of how TCs interact with submesoscale processes and highlight the need to include submesoscale dynamics in TC prediction models to improve forecast accuracy.

Introduction Submesoscale processes (submesoscales) encompass dynamic features including fronts, filaments, and eddies, with typical horizontal scale of O(0.1–10) km (McWilliams, 2016; Thomas et al., 2008). These processes are characterized by order one Rossby number (Ro) and Richardson number and their generation and evolution are accompanied by both frontogenesis and hydrodynamic instabilities (Callies et al., 2015; McWilliams, 2021; Z. Zhang, 2024). Their capacity for bi‐directional energy cascade enables unique contributions to ocean energy budgets through both forward and inverse energy transfers (Balwada et al., 2022; Dong et al., 2024; Naveira Garabato et al., 2022; Qiu et al., 2022; Yu et al., 2024; Z. Zhang et al., 2023). Distinct from quasi‐geostrophic mesoscale eddies, submesoscales can induce much stronger vertical velocities (w) with magnitude reaching O (10–100) m/day (Cao et al., 2024; D’Asaro et al., 2018; Yu et al., 2019; Z. Zhang et al., 2021). Consequently, submesoscales play important roles in vertical tracer transport in the upper ocean (e.g., heat, nutrients, dissolved oxygen, etc.) and thus significantly modulate air‐sea interaction and biogeochemical processes (Guo et al., 2024; Mahadevan, 2016; Su et al., 2018; Taylor, 2018; H. Yang et al., 2024). Existing studies based on both observations and simulations consistently demonstrate pronounced wintertime intensification of submesoscales (Buckingham et al., 2016; Callies et al., 2015; Dong et al., 2020; Mensa et al., 2013; Sasaki et al., 2014; Su et al., 2018; J. Zhang et al., 2021). This seasonality stems from the deeper mixed layers in winter. Specifically, the deeper mixed layer can accumulate more available potential energy (APE) for the development of submesoscale baroclinic instability and frontogenesis, which are recognized as two key mechanisms driving submesoscale generation in open oceans (McWilliams, 2016; Taylor & Thompson, 2023). Nevertheless, this canonical seasonality doesn't mean the absence of active summer submesoscales. Notably, mooring observations by Z. Zhang et al. (2021) captured vigorous submesoscale events following a tropical cyclone (TC) passage in the northwestern Pacific in September 2018. Generation of these TC‐related submesoscales does not contradict the existing theories because TCs can deepen the mixed layer and cause strong fronts surrounding their cold wakes (D’Asaro et al., 2007, 2014; Jacob et al., 2000; Mei & Pasquero, 2013; Price, 1981; H. Zhang, 2023; H. Zhang et al., 2021), which provide favorable conditions for the development of submesoscales. Recent high‐resolution simulations and observations have also corroborated enhancements of submesoscales following TC passages (Tang et al., 2025; Wu et al., 2024; Yi et al., 2024). However, these case studies focus on individual TCs leaving critical questions unresolved: (a) the universality of TC‐submesoscale linkages, (b) mechanistic controls on submesoscale variability, and (c) their roles in influencing the TC‐ induced cold wakes. This study addresses the above knowledge gaps through synergistic analysis of high‐resolution satellite observations and numerical simulations, focusing on the northwestern Pacific—a global hotspot for strong TC activity (D’Asaro et al., 2011). The remaining contents of this paper are organized as follows. Section 2 describes the data and methods. Section 3 presents results from case studies, statistical analyses, and evaluations of submesoscale impacts on the TC‐induced cold wakes. Finally, summary and discussion are given in Section 4.Data and Methods 2.1. High‐Resolution Simulation Data The 1/48° global realistic simulation LLC4320 was analyzed to investigate TC‐related submesoscales. The simulation was conducted using the Massachusetts Institute of Technology general circulation model (Marshall et al., 1997) on a Latitude‐Longitude polar Cap (LLC) grid (Forget et al., 2015). The forcing fields include the 6‐ hourly atmospheric forcing from the 0.14° European Centre for Medium‐Range Weather Forecasts (ECMWF) operational model reanalysis and tidal forcing from the 16 most important tidal constituents. The simulation configuration comprised 90 vertical layers, with resolutions ranging from 1 m near the surface to approximately 500 m near the bottom, spanning a maximum depth of 6,760 m. Hourly outputs were generated for 14 months from September 2011 to November 2012. The LLC4320 simulation has been demonstrated to resolve submesoscales effectively at low and mid‐latitudes (Dong et al., 2020; Rocha et al., 2016; Su et al., 2018; Z. Zhang et al., 2023). Furthermore, the ECMWF atmospheric fields used in the simulation can to a certain degree reproduce the TCs realistically, making the outputs suitable for investigating TCs‐related submesoscales. In addition to LLC4320, outputs of its 1/24° sister version LLC2160 were also utilized. Except for the half horizontal resolution, LLC2160 shares identical configurations with LLC4320. Given that the LLC2160 can hardly resolve submesoscales whereas the LLC4320 can, comparing these simulations enables the isolation of submesoscale effects (Su et al., 2018; Z. Zhang et al., 2023), particularly their roles in the evolution of TC‐ induced cold wakes. Outputs from both simulations, including temperature, salinity, and three‐dimensional velocity fields in the subtropical northwestern Pacific (120°E–150°E, 15°N–40°N) were analyzed in this study. 2.2. Observational Data Sets To validate the LLC4320 simulation and provide observational evidence of post‐TC submesoscales, Level‐3 daily sea surface temperature (SST) data from the Moderate‐Resolution Imaging Spectroradiometer (MODIS) were utilized. This product, with a 4 km horizontal resolution, is capable of resolving submesoscales comparable to those simulated by LLC4320. However, due to frequent data gaps during TC passages caused by heavy cloud cover in MODIS product, daily Optimum Interpolation SST (OISST) from the National Oceanic and Atmospheric Administration were also employed to evaluate the intensities and recovery rates of TC‐induced cold wakes (Banzon et al., 2016; Huang et al., 2021; Reynolds et al., 2007). The OISST data integrate observations from multiple platforms onto a regular 0.25° global grid, which are gap‐filled and have been widely employed in TC‐related research (Murakami et al., 2015; Stansfield & Reed, 2023; S. Wang & Toumi, 2021; G. Wang et al., 2016; Wood & Ritchie, 2015). TC track information for the study period (September 2011 to November 2012) is obtained from the International Best Track Archive for Climate Stewardship data set version 4r01 (Knapp et al., 2010). The data set offers 6‐ hourly records of TC center location, minimum central pressure, 1‐min maximum wind speed, and TC size, which have been extensively used in TC research (e.g., Darby et al., 2016; Kossin et al., 2013; N. Lin & Emanuel, 2016). In thisstudy, TC intensity was classified according to the Saffir‐Simpson Hurricane Wind Scale, with tropical depressions excluded from the analysis. Tropical storms and Category 1 TCs were categorized as weak TCs, while Category 2–5 TCs were classified asstrong TCs. This classification was established to ensure there are enough cases of strong TCs to conduct a robust statistical analysis. For each TC track, the 6‐hourly translation speed was calculated based on time and TC center positions (i.e., translation distance divided by time interval). Then, we defined the TCs with translation speed larger and smaller than 4 m/s as fast‐ and slow‐moving TCs, respectively (S. Lin et al., 2017; Price, 1981; Stramma et al., 1986). 2.3. Calculation of Submesoscale Quantities In order to evaluate the occurrence and intensity of submesoscales, the Ro and submesoscales‐induced vertical heat flux (VHFsm) are calculated using the hourly simulation data. Their calculation formulas are shown in Equations 1 and 2.Here, u, v, and w denote zonal, meridional, and vertical velocities, respectively, T is the temperature, f is the planetary vorticity, ρ0 = 1,030 kg/m3 is the reference density, and Cp = 3,900 J kg− 1 °C− 1 is the specific heat capacity of seawater. The prime represents the submesoscale anomaly obtained by subtracting the 0.5° spatially running mean (Su et al., 2018; Y. Yang et al., 2021; J. Zhang et al., 2023; Z. Zhang et al., 2023). The overbar denotes the 48‐hr low‐pass filter, which aims to remove the effects of tides, internal gravity waves, and near‐ inertial waves (the mean inertial period is ∼1.5 days in the study region). To have a fair comparison with the satellite data, the results were averaged on a daily basis in the formal analysis. The VHFsm was also estimated using the submesoscale parameterization proposed by J. Zhang et al. (2023) (Zhang23 parameterization hereafter) to examine the generation mechanism of submesoscales. The parameterized VHFsm (VHFp) is calculated using VHFp = cpρ gαT [− 4 3 z H · exp(− 8 9 z2 H2 + 1 2 ) ·Ce · SR2 f 2 · H2 · |∇hbm| 2 | f | ], (3) where z is the depth, H is the mixed layer depth, SR is the strain rate of background currents (i.e., meso‐ to large‐ scale currents), bm is background buoyancy, Ce = 8 is the empirical constant (J. Zhang et al., 2023), g is the acceleration of gravity, and αT = 2 × 10− 4 K− 1 is the thermal expansion coefficient. Here, the mixed layer depth is defined as the depth with potential density difference of 0.03 kg m− 3 from the sea surface density (de Boyer Montégut et al., 2004; Toyoda et al., 2017), and the SR is calculated using SR = ̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ ( ∂um ∂x − ∂vm ∂y ) 2 + ( ∂um ∂y + ∂vm ∂x ) 2 √ , (4) where subscript “m” denotes 0.5° spatially running mean. 2.4. Composite Analysis To statistically quantify TC‐induced submesoscales and their role in cold wakes recovery, we employed composite analysis following established methodologies in oceanic TC response studies (Hart et al., 2007; Mei &.Figure 1. Spatial distributions of sea surface temperature (SST) following tropical cyclone (TC) Prapiroon on 20 October 2012 from (a) Moderate‐Resolution Imaging Spectroradiometer observation and (c) LLC4320 outputs. (b, d) Same as (a, c) but for the modulus of the corresponding SST gradients. Colored lines show the TC track with circles denoting the position of TC center at 12:00 UTC of each day from 8 to 17 October 2012. Green, yellow, orange, and red colors of the TC track represent TC intensities range from tropical storm to Category 3.Pasquero, 2013; G. Wang et al., 2022). Our data set comprised 29 TC tracks observed in the northwestern Pacific from September 2011 to November 2012, yielding 471 spatiotemporal sampling points at 6‐hr intervals along TC trajectories. For each point, we established a 2° × 2° analysis domain centered on the TC center and calculated the absolute values of Ro and VHFsm on the 6th day preceding and the 6th day following TC passage. We identified submesoscales based on two distinct criteria: structures with (a) |Ro| > 0.3 and (b) VHFsm exceeding 50 W/m2 (Z. Zhang et al., 2021). The relative intensity of submesoscales was characterized through area fraction calculations, obtained by dividing the submesoscale‐occupied area by the total 2° × 2° domain area. The composite results were subsequently classified based on the TC intensity and translation speed. Sensitivity tests confirmed the robustness of our findings across parameter variations, including temporal windows (5–8 days), domain size (1° × 1°–3° × 3°), and variations of Ro and VHFsm thresholds (±20%). Note that the VHFsm (Ro) at 20‐m depth (sea surface) was used in this study because on average it is largest therein (Figure S1 in Supporting Information S1). 3. Results 3.1. Case Study: TC Prapiroon (2012) TC Prapiroon originated as a tropical depression at 12:00 UTC on 7 October 2012 near 18°N, 136°E, tracking northwestward with progressive intensification. It attained its peak intensity at 18:00 UTC on 11 October and then gradually weakened and eventually dissipated on 19 October. The offshore trajectory of Prapiroon, coupled with minimal MODIS data gaps, rendered it an ideal candidate for investigating TC‐related submesoscales. Figure 1a displaysthe spatial distribution of MODIS SST on 20 October 2012 (post‐TC), which reveals an evident cold wake on the right side of Prapiroon's track, with maximum SST cooling reaching approximately 7°C. The most remarkable SST cooling occurred at the TC track‐turning region rather than at the peak TC intensity region. This phenomenon is attributed to the prolonged ocean response time during the track‐turning stage (Guan.Figure 2. Spatial distributions of (a) LLC4320‐derived surface Ro, (b) LLC4320‐derived “true” VHFsm at 20 m depth, and (c) parameterized VHFsm following tropical cyclone (TC) Prapiroon on 20 October 2012. (d–i) are same as (a–c) but for the results following TCs Guchol and Sanba on 19 June and 18 September 2012, respectively. Purple rectangle indicates the domain analyzed in Figure S1 in Supporting Information S1. Black lines denote TC tracks and circles mark 6‐hourly TC locations.et al., 2025). Notably, abundant finescale structures are observed on the periphery of the cold wake, with SST gradient magnitudes ranging from 0.1 to 0.3°C km− 1 (Figure 1b). These structures manifest as elongated fronts and vortices with widths of O(10) km, falling within the submesoscale category. By comparing the post‐TC SST and its gradient derived from the LLC4320 simulation with the concurrent MODIS observations, we find that LLC4320 performs well in simulating the oceanic response to TC Prapiroon (Figures 1a and 1c). Specifically, the model successfully reproduces the spatial distribution of the cold wake, capturing SST cooling similar in magnitude to the observations. Furthermore, the simulated SST gradients exhibit submesoscale features with intensity comparable to observations (Figures 1b and 1d). These results confirm the model's capacity to resolve transient submesoscale dynamics and provide evidence that TC Prapiroon can trigger submesoscales. To investigate dynamical characteristics of submesoscales associated with TC Prapiroon, we calculated the Ro and VHFsm using the LLC4320 outputs and present their post‐TC distributions in Figures 2a and 2b, respectively. Corresponding to the active submesoscales as seen from SST gradients in Figure 1, both Ro and VHFsm exhibit large magnitudes on the periphery of the cold wake. Ro and VHFsm are dominated by positive values, with localized maxima exceeding 1.0 and 150 W/m2 , respectively. This positive skewness of Ro likely occurs because anticyclonic vorticity is prone to causing centrifugal instability, resulting in the dissipation of submesoscales (Buckingham et al., 2016; Z. Zhang et al., 2021). The strongest intensity of submesoscales occurred several days after TC passage (Figure S2 in Supporting Information S1), suggesting that direct TC forcing via wind stress and buoyancy flux cannot fully explain their generation and may even destroy the original submesoscale structures (Liu et al., 2023). The upward VHFsm

corresponds to the release of APE via isopycnal flattening, indicating that baroclinic processes such as frontogenesis and baroclinic instability play key roles in energizing these submesoscales. In open oceans, strain‐induced frontogenesis and submesoscale baroclinic instability typically co‐occur (McWilliams, 2016; Taylor & Thompson, 2023), and their combined effects determine the spatiotemporal variations of submesoscales (J. Zhang et al., 2021; Z. Zhang et al., 2020). To further clarify this relationship, we calculated the VHFp using the Zhang23 parameterization, which integrates these generation mechanisms (J. Zhang et al., 2023). The VHFp (Figure 2c) shows spatial patterns and magnitudes closely matching those of the “true” VHFsm directly derived from LLC4320 outputs (Figure 2b). This result confirms that strain‐induced frontogenesis and submesoscale baroclinic instability jointly drive submesoscale generation within the cold wake of TC Prapiroon. 3.2. Statistical Analysis In the middle and bottom panels of Figure 2, we examined the submesoscales associated with two additional TC cases: TC Guchol in mid‐June 2012 (Figures 2d–2f) and TC Sanba in mid‐September 2012 (Figures 2g–2i). These two cases also demonstrate systematic generation of submesoscales following their passages. However, the submesoscale strength in these cases, quantified by Ro and VHFsm, is weaker than that observed in the TC Prapiroon case. This discrepancy arises because the faster translation speeds of TCs Guchol and Sanba resulted in weaker cold wakes, which provided less APE (corresponding to weaker fronts and reduced mixed layer deepening) to fuel submesoscales. Similar to the first case, the VHFp also agrees well with the “true” ones in the latter two cases. This demonstrates that the combination of strain‐induced frontogenesis and submesoscale baroclinic instability may be a unified mechanism for submesoscale generation in TC‐induced cold wakes. To assess the generalizability of the above findings, we extended our analysis to all 29 TCs (471 TC domains) in the northwestern Pacific during our study period. In Figures 3a and 3b, we show the mean area fractions of submesoscales in TC domains defined using |Ro| > 0.3 and VHFsm > 50 W/m2 , respectively (see Section 2.4). TC domains were classified into the following four types based on TC intensity and translation speed: (a) weak and slow‐moving, (b) weak and fast‐moving, (c) strong and slow‐moving, and (d) strong and fast‐moving TCs. We can see from Figure 3a that the area fractions of submesoscales are generally smaller than 5% (for the Ro criteria) prior to the TC passage, which aligns with the known seasonal suppression of submesoscales in summer (Buckingham et al., 2016; Callies et al., 2015; Dong et al., 2020). After TC passage, however, submesoscale area fractions increased across all types. The most pronounced enhancement occurred for strong and slow‐moving TCs, followed by strong and fast‐moving and weak and slow‐moving TCs, with the least pronounced response observed for weak and fast‐moving TCs. In regions affected by strong and slow‐moving TCs, the post‐TC submesoscale area fraction reached 11% (16%) for the Ro‐based (VHFsm‐based) definition, which increases by 247% (493%) compared with the pre‐TC value. This pattern aligns with the understanding that stronger and slower‐moving TCs generate more intense cold wakes (Lloyd & Vecchi, 2011; Mei et al., 2012, 2015), thereby supplying greater APE to drive submesoscale activity via the combination of frontogenesis and submesoscale baroclinic instability. Collectively, these findings demonstrate that strong and slow‐moving TCs are highly effective in triggering submesoscales and that TC‐triggered submesoscales are not isolated occurrences but are widespread across the northwestern Pacific. 3.3. Influence of Submesoscales on Cold Wakes The results in Section 3.1 and 3.2 demonstrate that active submesoscales occur in the strong cold wakes and they can induce large upward VHFsm in the upper ocean. During the 14‐month study period, seven TCs in the study region generated prominent cold wakes with maximum surface cooling greater than 1°C. TC Bolaven induced two such cold wakes in different areas. These eight prominent cold wakes were selected for further analysis (Figure S3 in Supporting Information S1). In Figure 3c, we compared the post‐TC weekly mean VHFsm with the air‐sea net heat flux (Qnet) within each cold wake (averaged over the 2° × 2° box in Figure S3 in Supporting Information S1). For each cold wake, the mean VHFsm and Qnet are always upward and downward, which means that VHFsm and Qnet together heat the upper mixed layer and tend to reduce the intensity of cold wakes. Notably, the VHFsm accounts for 15%–53% of Qnet, highlighting its important role in mixed‐layer heat budget and thus post‐TC temperature recovery. To further investigate the influence of VHFsm, the SST recovery rates of cold wakes were calculated using the LLC4320 and LLC2160 data. The SST recovery rate here is defined as the linear SST trend during the week when.

Figure 3. (a) Composite‐mean submesoscale area fractions in the tropical cyclone (TC) domains. Submesoscale points are defined using |Ro| > 0.3. Yellow and orange columns denote the results on the 6th day preceding and post‐TC passage, and green columns denote their differences (i.e., later minus former). Vertical bars indicate the 95% confidence interval calculated using bootstrap method. The number of TC domains are marked at the bottom of panel (b). (b) Same as (a) except that submesoscale points are defined using VHFsm > 50 W/m2 . (c) Comparison between the post‐TC weekly mean VHFsm (red) and Qnet (blue) spatially averaged within a 2° × 2° box in each prominent cold wake (see Figure S3 in Supporting Information S1). The positive VHFsm and Qnet values mean upward and downward heat fluxes, respectively. The ratio between VHFsm and Qnet is marked above the red column. (d) Area‐mean sea surface temperature recovery rate of each cold wake calculated using satellite (blue), LLC4320 (red), and LLC2160 (yellow) data. The ratios between results from LLC4320 (LLC2160) and satellite data are marked above the red (yellow) columns.

cold wake began to warm up (averaged over the 2° × 2° box in Figure S3 in Supporting Information S1). For comparison, the observational estimates of recovery rates were derived from daily OISST data. Figure 3d shows that the SST recovery rate in the LLC4320 (submesoscale permitting) is always larger than that in the LLC2160 (submesoscale unresolved), and the former is much closer to the observed result. Specifically, the SST recovery rates in LLC4320 and LLC2160 account for 89%–130% (mean percentage of 105%) and 54%–83% (mean percentage of 70%) of the observed results, respectively. For the observation, LLC4320, and LLC2160 data, the averaged SST recovery rates of the 8 cases are 0.19, 0.19, and 0.13°C/day, respectively. Given that both LLC4320 and LLC2160 employ the same bulk formula for Qnet (Large & Yeager, 2009) and yield closely aligned Qnet values post‐TC passage, the accelerated recovery in LLC4320 primarily results from submesoscale‐driven VHFsm. 4. Summary and Discussion Through case studies and composite analyses, we have demonstrated for the first time that submesoscales are widespread within the TCs‐induced cold wakes across the northwestern Pacific. We find that the generation of these submesoscales is jointly regulated by baroclinic instability and strain‐induced frontogenesis, which also explain the greater activity of submesoscales in the stronger cold wakes left by strong and slow‐moving TCs. Furthermore, through comparative analysis, we have revealed that these submesoscales can significantly expedite the SST recovery of TC‐induced cold wakes via the substantial upward VHFsm they produced. The key aspects of our study are summarized in Figure 4. The above findings of this study improve the understanding of TCs‐ocean feedback dynamics. While cold wakes traditionally suppress the development of the overlying and subsequent TC intensification via reduced enthalpy fluxes (Balaguru et al., 2014; I.‐I. Lin et al., 2013; Wu & Li, 2018), submesoscale‐mediated SST recovery may weaken this negative feedback by accelerating the post‐TC temperature recovery. Consequently, the unresolved submesoscales in current coarse‐degree climate models could introduce biases in predicting the intensity and trajectory of TCs. In order to reduce this bias, the suitable parameterizations of VHFsm are called for in the climate models. We acknowledge that the high‐resolution ocean model used in this study is not coupled to the atmosphere and covers only a 14‐month period, which may lead to possible uncertainties in VHFsm calculation and limited cold wakes used for statistical analysis. To further advance this research and quantitatively evaluate submesoscales' contributions to the upper ocean heat budget, future efforts will focus on: strengthening submesoscale‐resolving field observations to quantify VHFsm under varying TC regimes (e.g., H. Zhang et al., 2024), conducting high‐ resolution air‐sea coupled simulations to capture TC‐submesoscales interactions, and testing submesoscale parameterization schemes in air‐sea coupled models. All of these interesting topics deserve to be explored in future studies.

Figure 4. Schematic diagram of tropical cyclone (TC)‐triggered submesoscales that lead to accelerated recovery of the cold wake. Large red swirl, red line, and red arrow denote TC, TC track, TC moving direction, respectively. Background colored shading represents the three‐dimensional temperature distribution with blue denoting colder water. Green swirls and green ellipses with arrows represent submesoscales and submesoscales‐induced secondary circulations, respectively. Yellow arrows represent equivalent vertical heat fluxes (i.e., VHFsm) associated with submesoscale secondary circulations which flatten the tilted isopycnals and restratify the upper layer. Purple arrows denote the negative feedback of the cold wake to the TC, which could be suppressed by rapid sea surface temperature recovery due to VHFsm. The dashed black arrow line indicates increasing days after the TC passage. The location of the mixed layer base is marked using a blue dashed line.

Data Availability Statement The LLC4320 and LLC2160 simulation outputs are available via NASA Open Data Portal (NASA, 2023a, 2023b). The MODIS SST data are obtained from NASA/JPL (2020). The OISSTv2.1 data is provided by the National Oceanic and Atmospheric Administration (Huang et al., 2020). The TC best track data are obtained from the International Best Track Archive for Climate Stewardship (Gahtan et al., 2024). References

Balaguru, K., Taraphdar, S., Leung, L. R., Foltz, G. R., & Knaff, J. A. (2014). Cyclone‐cyclone interactions through the ocean pathway. Geophysical Research Letters, 41(19), 6855–6862. https://doi.org/10.1002/2014GL061489 Balwada, D., Xie, J.‐H., Marino, R., & Feraco, F. (2022). Direct observational evidence of an oceanic dual kinetic energy cascade and its seasonality. Science Advances, 8(41), eabq2566. https://doi.org/10.1126/sciadv.abq2566 Banzon, V., Smith, T. M., Chin, T. M., Liu, C., & Hankins, W. (2016). A long‐term record of blended satellite and in situ sea‐surface temperature for climate monitoring, modeling and environmental studies. Earth System Science Data, 8(1), 165–176. https://doi.org/10.5194/essd‐8‐165‐ 2016 Buckingham, C. E., Naveira Garabato, A. C., Thompson, A. F., Brannigan, L., Lazar, A., Marshall, D. P., et al. (2016). Seasonality of submesoscale flows in the ocean surface boundary layer. Geophysical Research Letters, 43(5), 2118–2126. https://doi.org/10.1002/2016gl068009 Callies, J., Ferrari, R., Klymak, J. M., & Gula, J. (2015). Seasonality in submesoscale turbulence. Nature Communications, 6(1), 6862. https://doi. org/10.1038/ncomms7862 Cao, H., Jing, Z., & Fox‐Kemper, B. (2024). Scale‐dependent vertical heat transport inferred from quasi‐synoptic submesoscale‐resolving observations. Geophysical Research Letters, 51(12), e2024GL110190. https://doi.org/10.1029/2024gl110190 Darby, S. E., Hackney, C. R., Leyland, J., Kummu, M., Lauri, H., Parsons, D. R., et al. (2016). Fluvial sediment supply to a mega‐delta reduced by shifting tropical‐cyclone activity. Nature, 539(7628), 276–279. https://doi.org/10.1038/nature19809 D’Asaro, E. A., Black, P., Centurioni, L., Harr, P., Jayne, S., Lin, I., et al. (2011). Typhoon‐ocean interaction in the western North Pacific: Part 1. Oceanography, 24(4), 24–31. https://doi.org/10.5670/oceanog.2011.91 D’Asaro, E. A., Black, P. G., Centurioni, L. R., Chang, Y.‐T., Chen, S. S., Foster, R. C., et al. (2014). Impact of typhoons on the ocean in the Pacific: ITOP. Bulletin of the American Meteorological Society, 95(9), 1405–1418. https://doi.org/10.1175/BAMS‐D‐12‐00104.1 D’Asaro, E. A., Sanford, T. B., Niiler, P. P., & Terrill, E. J. (2007). Cold wake of hurricane Frances. Geophysical Research Letters, 34(15), 2–7. https://doi.org/10.1029/2007GL030160 D'Asaro, E. A., Shcherbina, A. Y., Klymak, J. M., Molemaker, J., Novelli, G., Guigand, C. M., et al. (2018). Ocean convergence and the dispersion of flotsam. Proceedings of the National Academy of Sciences of the United States of America, 115(6), 1162–1167. https://doi.org/10.1073/pnas. 1718453115 de Boyer Montégut, C., Madec, G., Fischer, A. S., Lazar, A., & Iudicone, D. (2004). Mixed layer depth over the global ocean: An examination of profile data and a profile‐based climatology. Journal of Geophysical Research, 109(C12), C12003. https://doi.org/10.1029/2004JC002378 Dong, J., Fox‐Kemper, B., Wenegrat, J. O., Bodner, A. S., Yu, X., Belcher, S., & Dong, C. (2024). Submesoscales are a significant turbulence source in global ocean surface boundary layer. Nature Communications, 15(1), 9566. https://doi.org/10.1038/s41467‐024‐53959‐y Dong, J., Fox‐Kemper, B., Zhang, H., & Dong, C. (2020). The scale of submesoscale baroclinic instability globally. Journal of Physical Oceanography, 50(9), 2649–2667. https://doi.org/10.1175/jpo‐d‐20‐0043.1 Forget, G., Campin, J. M., Heimbach, P., Hill, C. N., Ponte, R. M., & Wunsch, C. (2015). ECCO version 4: Ainverse modeling and global ocean state estimation. Geoscientific Model Development, 8(10), 3071–3104. https://doi.org/10.5194/gmd‐8‐ 3071‐2015 Gahtan, J., Knapp, K. R., Schreck, C. J. I., Diamond, H. J., Kossin, J. P., & Kruk, M. C. (2024). International best track archive for climate stewardship (IBTrACS) project, version 4.01 [Dataset]. NOAA National Centers for Environmental Information. https://doi.org/10.25921/ 82ty‐9e16 Guan, S., Liu, P., Zhang, Y., Lin, I.‐I., Zhou, L., Yang, Q., et al. (2025). Enhanced sea surface cooling and suppressed storm intensification during slow‐moving track‐turning stage of tropical cyclones. Journal of Geophysical Research: Oceans, 130(2), e2024JC022234. https://doi.org/10. 1029/2024JC022234 Guo, M., Xing, X., Xiu, P., Dall'Olmo, G., Chen, W., & Chai, F. (2024). Efficient biological carbon export to the mesopelagic ocean induced by submesoscale fronts. Nature Communications, 15(1), 580. https://doi.org/10.1038/s41467‐024‐44846‐7 Hart, R. E., Maue, R. N., & Watson, M. C. (2007). Estimating local memory of tropical cyclones through MPI anomaly evolution. Monthly Weather Review, 135(12), 3990–4005. https://doi.org/10.1175/2007MWR2038.1 Huang, B., Liu, C., Banzon, V., Freeman, E., Graham, G., Hankins, B., et al. (2021). Improvements of the daily optimum interpolation sea surface temperature (DOISST) version 2.1. Journal of Climate, 34(8), 2923–2939. https://doi.org/10.1175/JCLI‐D‐20‐0166.1 Huang, B., Liu, C., Banzon, V. F., Freeman, E., Graham, G., Hankins, W., et al. (2020). NOAA 0.25‐degree daily optimum interpolation sea surface temperature (OISST), version 2.1 [Dataset]. NOAA National Centers for Environmental Information. https://doi.org/10.25921/RE9P‐ PT57 Jacob, S. D., Shay, L. K., Mariano, A. J., & Black, P. G. (2000). The 3D oceanic mixed layer response to Hurricane Gilbert. Journal of Physical Oceanography, 30(6), 1407–1429. https://doi.org/10.1175/1520-0485(2000)0302.0.CO;2 Knapp, K. R., Kruk, M. C., Levinson, D. H., Diamond, H. J., & Neumann, C. J. (2010). The international best track archive for climate stewardship (IBTrACS): Unifying tropical cyclone data. Bulletin of the American Meteorological Society, 91(3), 363–376. https://doi.org/10.1175/ 2009BAMS2755.1 Kossin, J. P., Olander, T. L., & Knapp, K. R. (2013). Trend analysis with a new global record of tropical cyclone intensity. Journal of Climate, 26(24), 9960–9976. https://doi.org/10.1175/JCLI‐D‐13‐00262.1 Large, W. G., & Yeager, A. S. G. (2009). The global climatology of an interannually varying air‐sea flux data set. Climate Dynamics, 33(2–3), 341–364. https://doi.org/10.1007/s00382‐008‐0441‐3 Lin, I.‐I., Black, P., Price, J. F., Yang, C.‐Y., Chen, S. S., Lien, C.‐C., et al. (2013). An ocean coupling potential intensity index for tropical cyclones. Geophysical Research Letters, 40(9), 1878–1882. https://doi.org/10.1002/grl.50091 Lin, N., & Emanuel, K. (2016). Grey swan tropical cyclones. Nature Climate Change, 6(1), 106–111. https://doi.org/10.1038/nclimate2777

Lin, S., Zhang, W.‐Z., Shang, S.‐P., & Hong, H.‐S. (2017). Ocean response to typhoons in the western North Pacific: Composite resultsfrom Argo data. Deep Sea Research Part I: Oceanographic Research Papers, 123, 62–74. https://doi.org/10.1016/j.dsr.2017.03.007 Liu, J., Han, B., & wu, R. (2023). Numerical simulation of upper ocean responses to Typhoon Maria (2018) in the Northwest Pacific: Behavior of sub‐mesoscale processes. Deep Sea Research Part I: Oceanographic Research Papers, 201, 104164. https://doi.org/10.1016/j.dsr.2023. 104164 Lloyd, I. D., & Vecchi, G. A. (2011). Observational evidence for oceanic control on Hurricane Intensity. Journal of Climate, 24(4), 1138–1153. https://doi.org/10.1175/2010jcli3763.1 Mahadevan, A. (2016). The impact of submesoscale physics on primary productivity of plankton. Annual Review of Marine Science, 8(1), 161– 184. https://doi.org/10.1146/annurev‐marine‐010814‐015912 Marshall, J., Adcroft, A., Hill, C., Perelman, L., & Heisey, C. (1997). A finite‐volume, incompressible Navier‐Stokes model for studies of the ocean on parallel computers. Journal of Geophysical Research, 102(C3), 5753–5766. https://doi.org/10.1029/96JC02775 McWilliams, J. C. (2016). Submesoscale currents in the ocean. Proceedings of the Royal Society A. Mathematical, Physical and Engineering Sciences, 472(2189), 20160117. https://doi.org/10.1098/rspa.2016.0117 McWilliams, J. C. (2021). Oceanic frontogenesis. Annual Review of Marine Science, 13(1), 227–253. https://doi.org/10.1146/annurev‐marine‐ 032320‐120725 Mei, W., Lien, C. C., Lin, I. I., & Xie, S. P. (2015). Tropical cyclone‐induced ocean response: A comparative study of the South China sea and tropical northwest Pacific. Journal of Climate, 28(15), 5952–5968. https://doi.org/10.1175/jcli‐d‐14‐00651.1 Mei, W., & Pasquero, C. (2013). Spatial and temporal characterization of sea surface temperature response to tropical cyclones. Journal of Climate, 26(11), 3745–3765. https://doi.org/10.1175/jcli‐d‐12‐00125.1 Mei, W., Pasquero, C., & Primeau, F. (2012). The effect of translation speed upon the intensity of tropical cyclones over the tropical ocean. Geophysical Research Letters, 39(7), L07801. https://doi.org/10.1029/2011gl050765 Mensa, J. A., Garraffo, Z., Griffa, A., Özgökmen, T. M., Haza, A., & Veneziani, M. (2013). Seasonality of the submesoscale dynamics in the Gulf stream region. Ocean Dynamics, 63(8), 923–941. https://doi.org/10.1007/s10236‐013‐0633‐1 Murakami, H., Vecchi, G. A., Underwood, S., Delworth, T. L., Wittenberg, A. T., Anderson, W. G., et al. (2015). Simulation and prediction of category 4 and 5 hurricanes in the high‐resolution GFDL HiFLOR coupled climate model. Journal of Climate, 28(23), 9058–9079. https://doi. org/10.1175/JCLI‐D‐15‐0216.1 NASA. (2023a). Global MITgcm simulation ‐ llc2160 [Dataset]. Ecco Data Portal. Retrieved from https://data.nas.nasa.gov/ecco/eccodata/llc_ 2160/ NASA. (2023b). Global MITgcm simulation ‐ llc4320 [Dataset]. Ecco Data Portal. Retrieved from https://data.nas.nasa.gov/ecco/eccodata/llc_ 4320/ NASA/JPL. (2020). MODIS aqua level 3 SST thermal IR daily 4km daytime V2019.0 [Dataset]. NASA Physical Oceanography Distributed Active Archive Center. Retrieved from https://search.earthdata.nasa.gov/search/ Naveira Garabato, A. C., Yu, X., Callies, J., Barkan, R., Polzin, K. L., Frajka‐Williams, E., et al. (2022). Kinetic energy transfers between mesoscale and submesoscale motions in the open ocean's upper layers. Journal of Physical Oceanography, 52(1), 75–97. https://doi.org/10. 1175/jpo‐d‐21‐0099.1 Price, J. F. (1981). Upper ocean response to a hurricane. Journal of Physical Oceanography, 11(2), 153–175. https://doi.org/10.1175/1520‐0485 (1981)0112.0.co;2 Qiu, B., Nakano, T., Chen, S., & Klein, P. (2022). Bi‐directional energy cascades in the Pacific Ocean from equator to subarctic gyre. Geophysical Research Letters, 49(8), e2022GL097713. https://doi.org/10.1029/2022GL097713 Reynolds, R. W., Smith, T. M., Liu, C., Chelton, D. B., Casey, K. S., & Schlax, M. G. (2007). Daily high‐resolution‐blended analyses for sea surface temperature. Journal of Climate, 20(22), 5473–5496. https://doi.org/10.1175/2007jcli1824.1 Rocha, C. B., Gille, S. T., Chereskin, T. K., & Menemenlis, D. (2016). Seasonality of submesoscale dynamics in the Kuroshi

Geophysical Research Letters, 43(21), 11304–11311. https://doi.org/10.1002/2016gl071349 Sasaki, H., Klein, P., Qiu, B., & Sasai, Y. (2014). Impact of oceanic‐scale interactions on the seasonal modulation of ocean dynamics by the atmosphere. Nature Communications, 5(1), 5636. https://doi.org/10.1038/ncomms6636 Stansfield, A. M., & Reed, K. A. (2023). Global tropical cyclone precipitation scaling with sea surface temperature. Npj Climate and Atmospheric Science, 6(1), 60. https://doi.org/10.1038/s41612‐023‐00391‐6 Stramma, L., Cornillon, P., & Price, J. F. (1986). Satellite observations of sea surface cooling by hurricanes. Journal of Geophysical Research, 91(C4), 5031–5035. https://doi.org/10.1029/JC091iC04p05031 Su, Z., Wang, J., Klein, P., Thompson, A. F., & Menemenlis, D. (2018). Ocean submesoscales as a key component of the global heat budget. Nature Communications, 9(1), 775. https://doi.org/10.1038/s41467‐018‐02983‐w Tang, H., Wang, D., Shu, Y., Yu, X., Shang, X., Qiu, C., et al. (2025). Vigorous forced submesoscale instability within an anticyclonic eddy during tropical cyclone “Haitang” from glider array observations. Journal of Geophysical Research: Oceans, 130(1), e2024JC021396. https://doi.org/ 10.1029/2024JC021396 Taylor, J. R. (2018). Accumulation and subduction of buoyant material at submesoscale fronts. Journal of Physical Oceanography, 48(6), 1233– 1241. https://doi.org/10.1175/JPO‐D‐17‐0269.1 Taylor, J. R., & Thompson, A. F. (2023). Submesoscale dynamics in the upper ocean. Annual Review of Fluid Mechanics, 55(1), 103–127. https:// doi.org/10.1146/annurev‐fluid‐031422‐095147 Thomas, L. N., Tandon, A., & Mahadevan, A. (2008). Submesoscale processes and dynamics. Ocean Msodeling in an Eddying Regime, 177, 17– 38. https://doi.org/10.1029/177gm04 Toyoda, T., Fujii, Y., Kuragano, T., Kamachi, M., Ishikawa, Y., Masuda, S., et al. (2017). Intercomparison and validation of the mixed layer depth fields of global ocean syntheses. Climate Dynamics, 49(3), 753–773. https://doi.org/10.1007/s00382‐015‐2637‐7 Wang, G., Wu, L., Johnson, N. C., & Ling, Z. (2016). Observed three‐dimensional structure of ocean cooling induced by Pacific tropical cyclones. Geophysical Research Letters, 43(14), 7632–7638. https://doi.org/10.1002/2016GL069605 Wang, G., Wu, L., Mei, W., & Xie, S. (2022). Ocean currents show global intensification of weak tropical cyclones. Nature, 611(7936), 496–500. https://doi.org/10.1038/s41586‐022‐05326‐4 Wang, S., & Toumi, R. (2021). Recent tropical cyclone changes inferred from ocean surface temperature cold wakes. Scientific Reports, 11(1), 22269. https://doi.org/10.1038/s41598‐021‐01612‐9 Wood, K. M., & Ritchie, E. A. (2015). A definition for rapid weakening of North Atlantic and eastern North Pacific tropical cyclones. Geophysical Research Letters, 42(22), 10091–10097. https://doi.org/10.1002/2015gl066697 Wu, R., & Li, C. (2018). Upper ocean response to the passage of two sequential typhoons. Deep Sea Research Part I: Oceanographic Research Papers, 132, 68–79. https://doi.org/10.1016/j.dsr.2017.12.006

Wu, R., Tong, C., & Liu, Y. (2024). A study of the vortex filament pool left by a super typhoon. Journal of Geophysical Research: Oceans, 129(8), e2024JC021198. https://doi.org/10.1029/2024JC021198 Yang, H., Chen, Z., Sun, S., Li, M., Cai, W., Wu, L., et al. (2024). Observations reveal intense air‐sea exchanges over submesoscale ocean front. Geophysical Research Letters, 51(2), e2023GL106840. https://doi.org/10.1029/2023gl106840 Yang, Y., McWilliams, J. C., San Liang, X., Zhang, H., Weisberg, R. H., Liu, Y., & Menemenlis, D. (2021). Spatial and temporal characteristics of the submesoscale energetics in the Gulf of Mexico. Journal of Physical Oceanography, 51(2), 475–489. https://doi.org/10.1175/JPO‐D‐20‐ 0247.1 Yi, Z., Qiu, C., Wang, D., Cai, Z., Yu, J., & Shi, J. (2024). Submesoscale kinetic energy induced by vertical buoyancy fluxes during the tropical cyclone Haitang. Journal of Geophysical Research: Oceans, 129(7), e2023JC020494. https://doi.org/10.1029/2023JC020494 Yu, X., Barkan, R., & Naveira Garabato, A. C. (2024). Intensification of submesoscale frontogenesis and forward energy cascade driven by upper‐ ocean convergent flows. Nature Communications, 15(1), 9214. https://doi.org/10.1038/s41467‐024‐53551‐4 Yu, X., Garabato, A. C. N., Martin, A. P., Buckingham, C. E., Brannigan, L., & Su, Z. (2019). An annual cycle of submesoscale vertical flow and restratification in the upper ocean. Journal of Physical Oceanography, 49(6), 1439–1461. https://doi.org/10.1175/jpo‐d‐18‐0253.1 Zhang, H. (2023). Modulation of upper ocean vertical temperature structure and heat content by a fast‐moving tropical cyclone. Journal of Physical Oceanography, 53(2), 493–508. https://doi.org/10.1175/jpo‐d‐22‐0132.1 Zhang, H., He, H. L., Zhang, W. Z., & Tian, D. (2021). Upper ocean response to tropical cyclones: A review. Geoscience Letters, 8(1), 1–12. https://doi.org/10.1186/s40562‐020‐00170‐8 Zhang, H., Tian, D., Sun, Y., Yang, M., Yang, S., Zhou, Y., et al. (2024). Unmanned vehicles probed inner‐core air‐sea conditions during Super Typhoon Koinu (2023). Science Bulletin, 69(24), 3789–3792. https://doi.org/10.1016/j.scib.2024.10.018 Zhang, J., Zhang, Z., & Qiu, B. (2023). Parameterizing submesoscale vertical buoyancy flux by simultaneously considering baroclinic instability and strain‐induced frontogenesis. Geophysical Research Letters, 50(8), e2022GL102292. https://doi.org/10.1029/2022gl102292 Zhang, J., Zhang, Z., Qiu, B., Zhang, X., Sasaki, H., Zhao, W., & Tian, J. (2021). Seasonal modulation of submesoscale kinetic energy in the upper ocean of the northeastern South China Sea. Journal of Geophysical Research: Oceans, 126(11), e2021JC017695. https://doi.org/10.1029/ 2021JC017695 Zhang, Z. (2024). Submesoscale dynamic processes in the South China Sea. Ocean‐Land‐Atmosphere Research, 3, 0045. https://doi.org/10. 34133/olar.0045 Zhang, Z., Liu, Y., Qiu, B., Luo, Y., Cai, W., Yuan, Q., et al. (2023). Submesoscale inverse energy cascade enhances southern ocean eddy heat transport. Nature Communications, 14(1), 1335. https://doi.org/10.1038/s41467‐023‐36991‐2 Zhang, Z., Zhang, X., Qiu, B., Zhao, W., Zhou, C., Huang, X., & Tian, J. (2021). Submesoscale currents in the subtropical upper ocean observed by long‐term high‐resolution mooring Arrays. Journal of Physical Oceanography, 51(1), 187–206. https://doi.org/10.1175/jpo‐d‐20‐0100.1 Zhang, Z., Zhang, Y., Qiu, B., Sasaki, H., Sun, Z., Zhang, X., et al. (2020). Spatiotemporal characteristics and generation mechanisms of submesoscale currents in the northeastern South China Sea revealed by numerical simulations. Journal of Geophysical Research: Oceans, 125(2), e2019JC015404. https://doi.org/10.1029/2019JC015404.