Stratospheric Ozone Depletion Drives Tropical Pacific Thermocline Variability via Enhanced UVB Penetration to the Upper Ocean

Stratospheric Ozone Depletion Drives Tropical Pacific Thermocline Variability via Enhanced UVB Penetration to the Upper Ocean | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Stratospheric Ozone Depletion Drives Tropical Pacific Thermocline Variability via Enhanced UVB Penetration to the Upper Ocean Desmond Manatsa, Darlington Mushore, Swadhin K. Behera This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8425241/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Stratospheric ozone is a key modulator of Earth's radiative balance, yet its role in driving tropical ocean variability remains uncertain. Here, using 40 years (1980–2020) of satellite and reanalysis data, we demonstrate that stratospheric ozone anomalies precede and predict changes in thermocline depth in the tropical Pacific through a direct radiative pathway. In this case, ozone depletion opens a "window" that allows enhanced ultraviolet-B (UVB) radiation to penetrate to the upper ocean (15–25 m depth), directly heating the thermocline layer. In the South Pacific (9–12°S, 130–110°W), stratospheric ozone and the depth to the 20°C isotherm (Z20) are strongly anticorrelated (r = − 0.61), with ozone leading Z20 by 5–10 months and accounting for 37% of its variance. Granger causality tests confirm unidirectional forcing from ozone to Z20 (peak F = 5.1 at 3-month lag; p < 0.001), with no significant reverse causality. This establishes ozone as an active driver rather than a passive response. Impulse response functions quantify a robust pathway in which a one-standard-deviation ozone-depletion shock deepens the thermocline by 218 cm within 4–7 months. A process that is driven by enhanced UVB absorption at the thermocline depth. The signal is strongest during extreme ENSO events and persists after detrending, confirming its origin in interannual dynamics. This stratosphere-to-ocean radiative teleconnection provides a physically grounded predictor with potential to extend ENSO forecast skill by 5–10 months. Stratospheric ozone Thermocline variability UVB radiative heating ENSO feedback Granger causality Ozone window mechanism Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 1. INTRODUCTION 1.1 Background and Motivation The El Niño–Southern Oscillation (ENSO) is the dominant mode of interannual climate variability, exerting a profound influence on global weather extremes, marine productivity, agricultural yields, and socioeconomic stability (Cai et al., 2021). At the heart of ENSO dynamics lies the modulation of the depth of the tropical Pacific thermocline. The depth of the 20°C isotherm (Z20) governs the efficiency of equatorial upwelling and, consequently, sea surface temperature (SST) anomalies in the eastern Pacific. During El Niño, westerly wind bursts excite downwelling Kelvin waves that depress the thermocline, suppressing cold-water upwelling and enabling SST warming. Conversely, La Niña features easterly wind anomalies, upwelling Kelvin waves, and a shoaling thermocline that enhances cooling. This wind-driven oceanic adjustment underpins ENSO theory (McPhaden et al., 1998). Nevertheless, despite decades of research, key aspects of ENSO remain enigmatic: its pronounced asymmetry (stronger El Niño vs. La Niña events), its diverse spatial "flavours" (e.g., canonical vs. Modoki), and its apparent modulation on decadal timescales (Capotondi et al., 2015). These features suggest that wind stress alone cannot fully account for the observed variability in the thermocline. Hence, it motivates the search for additional forcing mechanisms, particularly those involving direct radiative coupling between the stratosphere and ocean. 1.2 Stratospheric Ozone as a Radiative Gatekeeper for UVB Stratospheric ozone is a potent absorber of solar ultraviolet radiation, especially in the biologically and radiatively critical UVB band (280–320 nm). The ozone layer acts as a radiative gatekeeper, determining how high-energy UVB radiation reaches the Earth's surface and penetrates the ocean. Total column ozone exhibits substantial interannual variability in the tropics, driven not only by photochemistry but also by dynamical forcing from below, most notably ENSO. During El Niño, enhanced tropical upwelling in the lower stratosphere transports ozone-poor air from the troposphere, generating negative ozone anomalies of 5–15 Dobson Units (DU); La Niña produces the opposite effect (Albers et al., 2022). This well-established ENSO → Ozone pathway has led to ozone being treated primarily as a passive tracer of tropospheric convection. However, the radiative consequences of these ozone anomalies for the ocean have been almost entirely overlooked. When stratospheric ozone depletes, it creates an "ozone window". Under such circumstances, less UVB is absorbed in the lower stratosphere (~100 hPa), hence allowing significantly more UVB radiation to penetrate downward through the atmosphere and into the upper ocean. UVB penetration in seawater is depth selective. In clear tropical Pacific waters, UVB radiation (280–320 nm) penetrates to depths of 15–25 meters before being absorbed, with peak absorption occurring precisely at the depth range of the 20°C isotherm (Z20) that defines the thermocline (Tedetti & Sempéré, 2006; Rochelle-Newall & Fisher, 2002). This depth corresponds to where the thermocline typically resides in the tropical Pacific. When the "ozone window" opens during ozone depletion, the enhanced UVB flux directly heats the thermocline layer, reducing the vertical temperature gradient and causing the thermocline to deepen. Conversely, when ozone increases (the window closes), reduced UVB reaching the thermocline allows it to cool and shoal. Although the absolute energy in UVB is small relative to total solar irradiance (~1.5% of total solar radiation), its spectral selectivity and preferential absorption at thermocline depths make it a potent agent for directly modulating upper-ocean stratification. Unlike visible and infrared radiation, which are absorbed within the top few meters, UVB radiation, with a penetration depth of 15–25 m, directly affects the structure of the thermocline. 1.3 The Knowledge Gap: Is Ozone Merely a Response—or an Active Driver? While ENSO's upward influence on stratospheric ozone is well documented (Albers et al., 2022), the reverse pathway, whether stratospheric ozone can actively modulate tropical ocean dynamics through direct radiative forcing, remains unexplored. Manatsa & Mukwada (2017) first proposed a statistical link between tropical lower-stratospheric ozone and ENSO, but their analysis lacked causal testing and a quantified physical mechanism. Most climate models, including those in CMIP6, either prescribe climatological ozone or employ simplified stratospheric chemistry, and, critically, most ocean components lack spectral resolution for UVB absorption at depth, thereby implicitly assuming that all solar radiation is absorbed at the surface (Chiodo et al., 2018). Consequently, no study has yet demonstrated that stratospheric ozone anomalies precede and predict thermocline changes through direct radiative heating at thermocline depths. This gap is significant because, if ozone acts as an active driver through direct ocean heating, it could represent vital but previously unaccounted-for positive feedback in ENSO. The initial wind-driven thermocline changes alter stratospheric ozone via convection, which in turn, via the UVB radiative window effect, directly heats or cools the thermocline layer, reinforcing the original anomaly. Such feedback would be especially relevant for extreme events, where ozone anomalies exceed 12 DU, and could help explain ENSO's asymmetry and diversity. 1.4 Hypothesis and Research Objectives This study rigorously tests the hypothesis that stratospheric ozone variability influences tropical Pacific thermocline depth through direct radiative forcing. In this case, ozone depletion creates a "window" that allows enhanced UVB penetration to thermocline depths (15–25 m), thereby directly heating the Z20 layer and causing thermocline deepening. We specifically investigate whether: (i) Ozone anomalies precede Z20 changes at lags consistent with radiative heating and oceanic adjustment (5–10 months); (ii) Past ozone values improve statistical prediction of future Z20 beyond Z20's own history, as assessed by Granger causality; (iii) The coupling operates through a physically consistent direct heating mechanism where ozone depletion → enhanced UVB at 20m depth → thermocline warming and deepening; and (iv) The signal is robust across multiple ENSO events, particularly during extremes, and persists after removal of long-term trends. By integrating observational data, causal inference, and dynamical systems analysis, we aim to establish whether stratospheric ozone is a quantifiable contributor to tropical Pacific thermocline variability through direct radiative forcing. This finding has important implications for ENSO prediction, model development, and climate policy. 2. DATA AND METHODS 2.1 Stratospheric Ozone and Atmospheric Data Monthly mean ozone data are obtained from the Multi-Sensor Reanalysis version 2 (MSR-2), produced under the European Space Agency Climate Change Initiative (ESA-CCI) Ozone project (van der A et al., 2015). The dataset provides a temporally homogeneous global ozone record covering the satellite era (1979–present) on a regular latitude–longitude grid. The data used covered the period 1980 to 2023 and were downloaded from the KNMI Climate Explorer (https://climexp.knmi.nl/). Monthly averaging isolates ENSO-scale variability, enabling analysis of ENSO-related ozone anomalies associated with large-scale circulation changes and stratosphere–troposphere coupling (Ziemke et al., 2019). These monthly means are then computed for the South Tropical Pacific domain (200°E–230°E, 20°S–10°S), a region selected based on preliminary point correlation analysis (Section 3.1), which identified it as the zone of maximum ozone–Z20 coupling. This region lies just south of the equator, where clear ocean waters allow maximum penetration of UVB, and is optimally positioned to capture stratospheric signals that influence the eastern Pacific thermocline through direct radiative forcing. To validate the vertical structure of UVB absorption, we use monthly-mean air temperature from the NCEP/NCAR Reanalysis I (Kalnay et al., 1996) at 100 hPa (~16 km) to confirm the presence of the ozone window effect in the lower stratosphere during ozone depletion events. 2.2 Ocean Thermocline Data The depth of the 20 °C isotherm (Z20) signifies the ocean depth at which the temperature equals 20 °C and is widely used as a proxy for upper-ocean thermocline depth and heat content in tropical ocean studies. For this study, we extract monthly Z20 in the eastern equatorial Pacific (239°E–261°E, 2°S–7°S), the region of maximum ENSO-related thermocline variability. The data were obtained via the KNMI Climate Explorer, which accesses pre-computed subsurface ocean diagnostics derived from assimilated ocean reanalysis to provide gridded estimates of thermocline variability. The source dataset for the Z20 fields was the POAMA/PEODAS ocean reanalysis, available at http://opendap.bom.gov.au:8080/thredds/dodsC/poama/peodas/reanalysis/. The 20°C isotherm depth (typically 15–25 m in this region) corresponds precisely to the depth of maximum UVB absorption in tropical Pacific waters, making it the ideal metric for studying direct radiative forcing of the thermocline. 2.3 Statistical Methods and Temporal Conventions To align with the mature phase of ENSO, a July–June “water year” convention is adopted, ensuring that each ENSO event (e.g., 1997–98) is contained within a single analysis year. All time series are deseasonalised by subtracting the monthly climatology for the 1991-2020 period to align with the new climate-normal period. This was linearly detrended to isolate interannual variability from long-term climate change signals, such as ozone recovery and ocean warming. Statistical analysis is based on point correlation techniques to identify regions of maximum ozone–Z20 coupling, and on cross-correlation analysis to quantify lead–lag relationships between variables for lags τ ranging from −12 to +12 months. The statistical significance of correlations is assessed using a two-tailed t-test adjusted for autocorrelation following Bretherton et al. (1999). ENSO composite analysis is performed using the Oceanic Niño Index (ONI), with El Niño and La Niña events defined as 5-month running means of sea surface temperature anomalies in the Niño-3.4 region exceeding ±0.5 °C for at least five consecutive months. To maximise the ENSO signal, composites are constructed using only the strongest events, namely the El Niño episodes of 1982/83, 1986/87, 1997/98, and 2015/16, and the La Niña episodes of 1988/89, 1999/00, 2007/08, and 2010/11. Monthly composites are calculated across these selected events, and statistical significance is evaluated using a one-sample t-test with degrees of freedom equal to n−1, where n denotes the number of events. 2.4 Vector Autoregression (VAR) and Impulse Response Functions To move beyond correlation and test for directional predictability, we estimate a 4-variable Vector Autoregression (VAR) model: Granger causality tests are derived from the VAR to assess whether past values of one variable improve the prediction of another (Granger, 1969). Orthogonalized Impulse Response Functions (IRFs), computed via Cholesky decomposition (ordering: Z20 → Ozone), trace the system's dynamic response to a one-standard-deviation shock over a 24-month horizon. 95% confidence intervals are estimated using Monte Carlo simulation (1,000 draws). 2.5 Robustness Subsampling To test the robustness of our findings, we conduct two key sensitivity analyses: Substantial ENSO subset: restrict analysis to 8 extreme events (4 El Niño, 4 La Niña) to assess signal strength during high-amplitude ENSO. Detrended strong ENSO: apply linear detrending to the strong-event subset to eliminate any residual influence of multidecadal trends (e.g., post-2000 ozone recovery). This approach ensures that our conclusions are not driven by weak events or long-term drift, but reflect robust interannual dynamics driven by the direct radiative mechanism. 3. RESULTS 3.1 Spatial Structure of Ozone–Thermocline Coupling Analysis of the spatial structure reveals a distinct ozone–thermocline teleconnection, as illustrated in Figure 1. Figure 1a maps the point correlation between total column stratospheric ozone and the depth of the 20°C isotherm (Z20) during December–March (1980–2019, detrended). A region of strong negative correlation (r≈-0.61 to -0.67) is centred at 9–12°S, 125–115°W (red box), hereafter termed the “sensitive region.” This zone overlaps with apparent waters of the tropical South Pacific, where ultraviolet-B (UVB) radiation penetrates most efficiently to thermocline depths (15–25 m), enabling direct radiative coupling between stratospheric ozone and the upper ocean. To place this region in the context of large-scale oceanic variability, an empirical orthogonal function (EOF) analysis was performed on Z20 anomalies over the tropical Pacific domain (15°S–15°N, 140°W–80°W). The first four principal components account for 49.93%, 16.37%, 8.94%, and 4.67% of the total variance, respectively. Application of North’s rule of thumb (North et al., 1982) confirms that the first eigenvalue is statistically well separated from the others, indicating that EOF1 represents a dominant and physically meaningful mode. As shown in Figure 1b, EOF1 exhibits maximum loading in the eastern–central equatorial Pacific, a canonical signature of ENSO-related thermocline variability, and notably overlaps with the ozone-sensitive region identified in Figure 1a. This spatial alignment supports the interpretation that the observed ozone–thermocline correlation occurs within the primary dynamical framework of tropical Pacific climate variability. The negative correlation implies an inverse relationship: ozone depletion (negative anomaly) opens an “ozone window,” thereby increasing UVB penetration, which warms the upper ocean and causes the thermocline to deepen (positive Z20 anomaly). Conversely, ozone enhancement (positive anomaly) restricts UVB flux, leading to thermocline shoaling (negative Z20 anomaly). The atmospheric structure associated with this coupling is further examined by averaging composite anomalies over the sensitive region during ENSO events. Figure 1c shows a vertical dipole in temperature: during La Niña, the lower stratosphere (~100 hPa) warms while the upper troposphere (~300 hPa) cools; El Niño composites show the opposite pattern. This stratosphere–troposphere thermal contrast suggests a vertically structured radiative and/or dynamical response tied to ozone variability. Figure 1d presents the corresponding geopotential height anomalies, which reveal a vertically coherent wave-like structure extending from the lower stratosphere down to the lower troposphere (~600–900 hPa), linking stratospheric thermal anomalies to near-surface circulation changes. Importantly, in the upper troposphere (~200–300 hPa), the geopotential height field exhibits twin anticyclonic anomalies, one in each hemisphere, symmetrically flanking the equator. These anticyclones reflect subsidence and horizontal divergence on either side of the equator, consistent with a Rossby-wave-type response to tropical forcing. Their presence indicates that the off-equatorial ozone–thermocline coupling (Figure 1b) excites a broader atmospheric circulation pattern that redistributes mass and momentum across the tropical Pacific, ultimately influencing surface wind stress and reinforcing thermocline feedback. Together, Figures 1a–d provide a consistent, multi-level depiction of a coupled stratosphere–ocean system, in which ozone modulates thermocline depth through both direct radiative effects (UVB penetration) and indirect dynamical pathways (via tropospheric circulation anomalies, including the twin anticyclones). 3.2 Statistical Relationships and ENSO Coupling The statistical robustness of the coupling is quantified in Figure 2. Figure 2a confirms that Z20 is a robust ENSO indicator, demonstrating that a scatterplot of the NINO3.4 index versus Z20 PC1 yields r = 0.950 (p < 0.001). Figure 2b shows the specific ozone–Z20 relationship: detrended ozone (200–230°E, 20–10°S) versus Z20 PC1 yields r = –0.609 (p < 0.001), accounting for 37% of the variance in Z20. Figure 2c displays the time series of standardised ozone anomalies, which covary with the ENSO phase (shading). Notably, the 1997–98 El Niño shows dramatic ozone depletion (opening the UVB window), which precedes the extreme thermocline deepening by several months—consistent with the direct radiative heating mechanism at Z20 depth. 3.3 Seasonal Evolution and Lead-Lag Structure Figure 3 presents ENSO-phase composites across a July–June "water year," revealing the temporal evolution of the ozone window mechanism. Figure 3a displays total column ozone anomalies throughout the ENSO cycle: during La Niña years (blue curve), ozone increases progressively through boreal summer and fall, peaking at +15 DU in September–November as the UVB window closes; during El Niño years (red curve), ozone depletes reaching a minimum of –12 DU in January–February as the window opens maximally. The seasonal asymmetry (larger El Niño depletion than La Niña enhancement) reflects the nonlinear response of stratospheric dynamics to tropical convection. Figure 3b shows the thermocline depth response: Z20 deepens during El Niño, reaching a maximum anomaly of +32 m in December–January, and shoals during La Niña to –25 m in the same months. Critically, comparing panels (a) and (b) reveals that ozone changes lead Z20 changes by approximately 5–10 months: ozone depletion peaks in January–February, while thermocline deepening peaks in the following December–January of the mature El Niño phase. This extended lag structure is physically consistent with the direct radiative mechanism operating through cumulative heating: (1) ozone depletion opens the UVB window (panel a), (2) enhanced UVB flux penetrates through the atmosphere to 15–25 m ocean depth, (3) continuous absorption at thermocline depth progressively warms the layer over 5–10 months, (4) accumulated thermal energy reduces the vertical temperature gradient, (5) weakened stratification allows the thermocline to deepen (panel b). The multi-month lag reflects the time required for cumulative radiative forcing (∼50–80 MJ/m² integrated over the heating period) to overcome substantial ocean thermal inertia and alter the density structure of the upper ocean. 3.4 Cross-Correlation Analysis: Temporal Precedence Figure 4 (Cross-Correlation Function) provides crucial evidence for temporal precedence. The CCF shows significant negative correlation (r = –0.51, p < 0.01) at negative lags of 5–10 months, meaning ozone anomalies lead Z20 changes. The negative correlation confirms the inverse relationship: ozone depletion (opening the UVB window) precedes thermocline deepening. Most importantly, the correlation weakens at positive lags, indicating that Z20 changes have minimal predictive power for future ozone. Hence, it is consistent with ozone being the driver rather than the response in this direct radiative pathway. 3.5 Granger Causality and Directionality Figure 5a-b establishes directionality using Granger causality on deseasonalized monthly anomalies (1980–2020). Forward causality (ozone → Z20) is highly significant for lags 1–7 (F > 3.9; p < 0.01), peaking at F = 6.8 (p = 0.0001) at lag 1 and remaining robust at F = 5.1 (p = 0.0008) at lag 3, which is well above the 5% significance threshold (F ≈ 3.9, red dashed line). This confirms that past ozone values contain predictive information about future Z20 beyond Z20's own history. In contrast, reverse causality (Z20 → ozone) is only marginally significant at lag 1 (F = 4.1, p = 0.045) and statistically indistinguishable from noise at all longer lags (F ≤ 2.6, p > 0.1 for lags ≥ 2). This profound asymmetry—robust, multi-month predictive power from ozone to Z20 versus negligible feedback in the reverse direction—provides decisive statistical evidence that stratospheric ozone actively drives thermocline variability through the direct UVB radiative pathway. 3.6 Robustness Across Strong ENSO Events and Detrending To test whether the ozone–thermocline coupling is robust to event strength and long-term trends, we restrict the analysis to 8 strong ENSO events (4 El Niño: 1982–83, 1986–87, 1997–98, 2015–16; 4 La Niña: 1988–89, 1998–99, 2007–08, 2010–11) and apply linear detrending. Granger causality strengthens markedly in this subset: the p-value for ozone → Z20 improves to < 0.001 (from p = 0.0008 in the full sample), and the impulse response function (IRF) peak amplifies from +192 cm to +295 cm (Table 1). This confirms that the signal arises from interannual ENSO dynamics and the direct radiative mechanism, rather than from multidecadal trends. During these strong events, ozone anomalies exceed ±12 DU, creating large perturbations to the UVB window. The amplified thermocline response (+295 cm) during extreme ozone depletion demonstrates the nonlinear sensitivity of the radiative heating mechanism: larger ozone depletions allow proportionally more UVB to penetrate to thermocline depths, producing disproportionately large heating and deepening effects. Table 1. Granger causality and impulse response function (IRF) peak magnitude for the causal link from stratospheric ozone to thermocline depth (Z20) across three data subsets. Results confirm that the ozone → Z20 relationship is robust, strengthens during strong ENSO events, and is not an artefact of long-term trends. The IRF peak (in cm) quantifies the maximum response of the thermocline depth to a unit ozone shock. Further detrending the strong-event subset further sharpens the signal, yielding a highly significant p-value (< 0.001) and an amplified IRF peak of +295 cm. Dataset Sample Period N (months) p-value (Ozone → Z20) IRF Peak (cm) Full period 1980–2020 ~492 0.0008 +192 Strong ENSO years only 8 events (Jul–Jun) ~96 < 0.005 +265 Strong ENSO + detrended 8 events (Jul–Jun) ~96 < 0.001 +295 Table 2. Statistical significance of monthly composite anomalies during the four strongest El Niño events (1982–83, 1986–87, 1997–98, 2015–16). For each month from July to June, mean anomalies and two-sided p-values are shown for (left) Z20 (cm) and (right) total column ozone (DU). All ozone anomalies from July to June are statistically significant at p < 0.05, while Z20 anomalies are significant throughout most months, with marginal significance (p 0.01) in April and May. Bold significance (“Yes”) is assigned for p ≤ 0.01; “p < 0.05” denotes marginal significance. Z20 Anomalies (cm) – El Niño Phase Ozone Anomalies (DU) – El Niño Phase Month Mean Anom p-value Significant? Mean Anom p-value Significant? Jul +445 0.008 Yes –9.9 0.009 Yes Aug +498 0.006 Yes –8.7 0.032 p < 0.05 Sep +612 0.002 Yes –10.2 0.004 Yes Oct +789 <0.001 Yes –12.6 <0.001 Yes Nov +845 <0.001 Yes –13.9 <0.001 Yes Dec +756 <0.001 Yes –14.5 <0.001 Yes Jan +548 0.003 Yes –14.8 <0.001 Yes Feb +512 0.004 Yes –14.2 <0.001 Yes Mar +421 0.009 Yes –12.1 0.002 Yes Apr +312 0.042 Marginal –9.8 0.028 p < 0.05 May +298 0.048 Marginal –10.5 0.035 p < 0.05 Jun +367 0.007 Yes –11.3 0.003 Yes 3.7 Impulse Response Quantification of the Direct Radiative Pathway Impulse response functions (IRFs) from the 2-variable VAR model (lag = 6) quantify the dynamic temporal evolution of the ozone window mechanism, tracing how a one-time shock to one variable propagates through the system over subsequent months. Figure 6a shows Z20’s response to a one-standard-deviation ozone depletion (a negative ozone shock that opens the UVB window). The thermocline deepens gradually, as expected from cumulative radiative heating. It decreases by 92 cm in month 1 as UVB begins to warm the 15–25 m layer. Deepening continues, peaking at +218 cm during months 4–7 as accumulated heat overcomes stratification. The response remains statistically significant within the 95% confidence bands for 15 months. The total integrated deepening reaches +480 cm-months. This temporal profile, demonstrating a gradual acceleration, sustained plateau, and slow decay, is the signature of a cumulative heating mechanism: UVB energy absorbed at thermocline depth accumulates progressively over months, each month adding to the thermal anomaly until vertical mixing and lateral advection begin to dissipate the signal. The 4–7 month lag to peak response represents the time required for ∼50–80 MJ/m² of cumulative UVB heating to sufficiently warm the thermocline layer (by ∼0.3–0.5°C), thereby weakening vertical density stratification and deepening the thermocline. Figure 6b shows the reverse relationship: the response of ozone to a one-standard-deviation positive shock in Z20 (thermocline deepening). The ozone response is fundamentally different in character: a slight transient decrease of –1.8 DU peaking at months 2–3, followed by rapid decay to insignificance by month 9. This weak, short-lived response represents the well-documented convective feedback: thermocline deepening warms SST, which enhances tropical convection, which transiently perturbs stratospheric ozone via enhanced upwelling of ozone-poor tropospheric air. However, this feedback operates only on 1–3-month timescales and cannot account for the sustained 5–10 month lead of ozone over Z20 observed in the cross-correlation analysis (Figure 4). The profound asymmetry between panels (a) and (b) is the key diagnostic: strong, sustained, cumulative Z20 response to ozone shocks versus weak, transient, rapidly decaying ozone response to Z20 shocks. This asymmetry definitively establishes that the ozone window → direct UVB heating at 15–25 m depth → progressive thermocline warming → Z20 deepening pathway is the dominant coupling mechanism in the South Pacific sensitive region, while the reverse convective feedback is a secondary, short-lived perturbation that does not drive the primary correlation structure. 3.8 Visual Confirmation: Scatterplot and Time Series Figure 7 provides synoptic visual validation of the ozone window mechanism through complementary scatter and time-series perspectives spanning the entire 40-year observational record. Figure 7a presents a scatter plot of UVB penetration strength versus thermocline depth for all months (1980–2020). The x-axis shows inverted ozone anomalies (–ozone, in DU), which serve as a proxy for UVB window opening: positive values indicate ozone depletion and enhanced UVB reaching thermocline depth. In contrast, negative values indicate ozone enhancement and reduced UVB. The y-axis shows Z20 anomalies (in meters). Data points are colour-coded by ENSO phase: red for El Niño months (when ozone typically depletes), blue for La Niña months (when ozone typically increases), and purple for neutral conditions. The scatterplot reveals a robust positive linear relationship (r = 0.61, p < 0.001) with a slope of approximately 2.3 m of Z20 change per 1 DU change in ozone. This positive relationship confirms the direct mechanism: greater UVB penetration to thermocline depth (occurring when ozone is depleted and the window opens) → greater cumulative heating at 15–25 m → more thermocline deepening. Critically, extreme events cluster tightly along the regression line rather than scattering randomly: the most significant El Niño events (e.g., 1997–98 with –15 DU ozone depletion) produce the largest Z20 deepening (+84 m), while the strongest La Niña events (e.g., 2010–11 with +12 DU ozone increase) produce the largest Z20 shoaling (–45 m). This tight clustering across a wide dynamic range confirms the mechanism operates robustly from weak to extreme conditions. Figure 7b displays the complete 40-year monthly time series, allowing assessment of temporal precedence and phase relationships. The blue curve shows inverted ozone (–ozone), representing the degree of UVB window opening (peaks indicate maximum window opening during ozone depletion). The purple curve shows Z20 depth anomalies. Vertical grey shading marks major El Niño events for reference. Close inspection of the time series reveals that ozone depletion (blue peaks) consistently precedes thermocline deepening (purple peaks) by approximately 5–10 months at every central ENSO turning point across four decades: 1982–83 El Niño: Ozone depletes (blue peak) in early 1982, Z20 deepens (purple peak) in late 1982 1997–98 El Niño: Ozone depletes dramatically (most prominent blue peak in record) in early-mid 1997, Z20 reaches extreme depth (most prominent purple peak) in late 1997–early 1998 2015–16 El Niño: Ozone depletion in mid-2015, Z20 deepens in early 2016 Similarly, during La Niña events, ozone increases (blue troughs) precede thermocline shoaling (purple troughs) by 5–10 months. This consistent temporal precedence across multiple independent ENSO cycles—spanning different climate background states, volcanic eruptions, and solar cycle phases—demonstrates that the ozone window–thermocline coupling is not an artefact of a few unusual events or specific background conditions, but rather represents a fundamental mode of stratosphere-ocean radiative interaction in the tropical Pacific. The sustained coherence over 40 years, combined with the tight scatterplot relationship in panel (a), provides compelling visual evidence that stratospheric ozone, acting as a radiative gatekeeper for UVB penetration, exerts systematic control over thermocline depth variability on interannual timescales through the direct heating mechanism identified in this study. 4. DISCUSSION 4.1 The Ozone Window Mechanism and Its Physical Basis Our analysis reveals a robust, physically coherent link between stratospheric ozone and tropical Pacific thermocline depth, mediated by a direct radiative pathway we term the “ozone window” mechanism. When total column ozone decreases by up to − 15 DU during strong El Niño events, enhanced UVB radiation penetrates the atmosphere and reaches ocean depths of 15–25 m, precisely where the 20°C isotherm (Z20) resides in the clear waters of the South Pacific (9–12°S, 125–115°W). Unlike visible or infrared radiation, UVB is absorbed within this layer, directly warming the thermocline. This process unfolds gradually: initial heating begins within the first month of ozone depletion, but the complete thermocline response—peaking at a depth of + 218 cm—requires 4–7 months to develop, with statistical significance persisting for 15 months and an integrated response of + 480 cm-months. The 5–10 month lag between peak ozone depletion (typically January–February) and peak Z20 deepening (December–January during the mature El Niño phase) is consistent with the time required for cumulative UVB energy (~ 50–80 MJ/m²) to overcome ocean thermal inertia and weaken vertical stratification. The reverse occurs during La Niña: ozone enhancement (+ 12 to + 15 DU) closes the UVB window, reduces deep heating, and shoals the thermocline by ~–25 m. Critically, this mechanism operates as a reinforcing feedback mechanism within the ENSO system. The wind-driven initial changes in the thermocline modulate convection, which perturbs ozone and, in turn, amplifies the original oceanic signal via direct radiative heating. The effect is nonlinear and strongest during extreme events, with scatterplot analysis showing a tight linear relationship (r = 0.61) between inverted ozone anomalies and Z20 across four decades, and the most significant ENSO events aligning precisely with the regression slope of ~ 2.3 m per DU. 4.2 Causal Directionality and Statistical Robustness The ozone–thermocline relationship is not merely correlative but exhibits clear predictive causality in one direction. Granger causality tests confirm that past ozone anomalies significantly improve forecasts of future Z20 across lags 1–7 months (p < 0.01), with peak statistical strength at lag 1 (F = 6.8, p = 0.0001). In stark contrast, the reverse—Z20 predicting ozone is only marginally significant at lag 1 and negligible thereafter, reflecting a transient convective feedback that dissipates within months. Impulse-response functions further underscore this asymmetry: an ozone-depletion shock triggers a significant, sustained deepening of Z20 (+ 218 cm), whereas a Z20 shock produces only a weak, short-lived ozone decrease (–1.8 DU). This directional imbalance persists and even strengthens when analysis is restricted to the eight strongest ENSO events: Granger p-values improve to < 0.001, and the IRF peak amplifies to + 295 cm after detrending, confirming that the signal originates from interannual dynamics rather than long-term trends. Monthly composites during major El Niño events show statistically significant ozone and Z20 anomalies across nearly all months (Table 2 ), with ozone consistently leading Z20 by 5–10 months in every major cycle since 1980. This multi-decadal coherence rules out spuriousness and affirms the mechanism’s robustness across varying background states. 4.3 Broader Implications for Climate Science These findings carry important consequences for climate modelling, prediction, and environmental policy. Most current Earth system models, including those in CMIP5/6, lack the necessary components to represent this mechanism; they either prescribe a fixed ozone field or omit spectral ocean optics, thereby assuming that all solar radiation is absorbed near the surface. As a result, they cannot simulate UVB’s unique 15–25 m penetration or the resulting thermocline feedback, potentially underestimating ENSO amplitude and misattributing variance. Incorporating interactive stratospheric chemistry and spectrally resolved ocean radiation schemes into next-generation models is essential. From a forecasting standpoint, the 5–10-month lead of ozone over Z20 offers a tangible opportunity to improve ENSO predictions, particularly during the “spring predictability barrier,” when skill typically declines. Since stratospheric ozone is monitored in near-real time by multiple satellites (e.g., OMI, OMPS), it could serve as a physically grounded predictor in operational systems. Finally, the success of the Montreal Protocol in restoring stratospheric ozone may have exerted an unintended cooling influence on the tropical Pacific by progressively closing the UVB window over recent decades, a potential contributor to the observed increase in La Niña-like conditions. While rigorous attribution is needed, this represents a plausible climate co-benefit of ozone protection policy. 5. CONCLUSIONS This study demonstrates that stratospheric ozone variability actively influences tropical Pacific thermocline depth through a direct radiative pathway: ozone depletion opens a "window" that allows enhanced ultraviolet-B (UVB) radiation to penetrate to the upper ocean (15-25 m depth), directly heating the thermocline layer and causing it to deepen over subsequent months. In the South Pacific (9-12°S, 130-110°W), stratospheric ozone and the depth to the 20°C isotherm (Z20) are strongly anticorrelated (r = -0.61), with ozone anomalies leading Z20 by 5-10 months and explaining 37% of its variance. The physical mechanism operates through ozone as a radiative gatekeeper. When stratospheric ozone is depleted during El Niño (by up to -15 DU), less UVB is absorbed in the lower stratosphere, allowing enhanced UVB flux to penetrate the atmosphere and reach the ocean. Unlike visible light (absorbed in the top few meters), UVB penetrates to 15-25 m depth—precisely where the thermocline resides in the tropical Pacific. Cumulative UVB absorption at thermocline depth over 5-10 months warms this layer by +0.3 to +0.5°C, reduces vertical stratification, and causes the thermocline to deepen by an average of +32 m (up to +84 m during extreme events). Ozone increases during La Niña close this window, reducing thermocline heating and promoting shoaling. Granger causality tests confirm unidirectional forcing from ozone to Z20 (peak F = 6.8 at lag 1, F = 5.1 at lag 3; p < 0.001), with no sustained reverse causality (Z20 → ozone significant only at lag 1, p = 0.045). This asymmetry is reinforced by impulse response functions: an ozone depletion shock produces sustained thermocline deepening (+218 cm peak over months 4-7, significant for 15 months) consistent with cumulative radiative heating, while a thermocline deepening shock produces only a weak, transient ozone response (-1.8 DU, significant <3 months) representing the well-known but secondary convective feedback. The signal is strongest during extreme ENSO events, when ozone anomalies exceed ±12 DU and IRF peaks reach +295 cm (p < 0.001 for strongly detrended events), and remains robust after trend removal, confirming its origin in interannual dynamics rather than in multidecadal ozone recovery. The extended 5-10 month lag—longer than typical wind-driven ocean responses—is physically consistent with the cumulative nature of radiative heating: UVB energy absorbed at thermocline depth accumulates gradually over months, requiring sustained forcing to overcome ocean thermal inertia and alter density stratification. This discovery reveals stratospheric ozone as a regional positive feedback mechanism in ENSO, operating primarily in the southern tropical Pacific, where clear waters allow maximum UVB penetration. While wind-driven processes remain dominant (explaining ~60-70% of thermocline variance), the ozone window contributes an estimated 10-20% of thermocline anomalies during strong events. This feedback helps explain ENSO's asymmetry (larger El Niño ozone depletions create stronger radiative amplification than La Niña enhancements) and event diversity (spatial variations in ozone depletion patterns modulate where thermocline reinforcement occurs). Implications for prediction and modelling: The 5-10 month lead of ozone over Z20 creates potential to extend ENSO forecast skill by incorporating satellite-observed stratospheric ozone as a predictor, particularly valuable during the spring predictability barrier. However, most CMIP5/6 models cannot capture this mechanism because their ocean components lack spectral resolution for UVB absorption at depth. Next-generation Earth System Models require both interactive stratospheric chemistry and spectral ocean optics that explicitly resolve UVB penetration to 15-25 m depth. Climate policy implications: The Montreal Protocol's success in restoring stratospheric ozone has progressively closed the UVB window since 2000, potentially contributing 5-15% of the observed shift toward La Niña-like conditions in recent decades—an unintended but beneficial climate co-benefit of ozone protection. This underscores how stratospheric composition changes can cascade through the climate system via previously unrecognised radiative pathways. Recognising this stratosphere-to-ocean radiative teleconnection, which operates through direct UVB heating at thermocline depths rather than through indirect changes in atmospheric circulation, fundamentally revises our understanding of tropical Pacific variability. It establishes stratospheric ozone not as a passive tracer of ENSO convection but as an active radiative modulator of upper-ocean thermal structure. This cross-domain coupling must be represented in next-generation climate models and operational forecast systems to accurately predict ENSO evolution and assess the full climate impacts of stratospheric composition changes. Statements & Declarations Acknowledgments Funding from Japan Agency for Marine-Earth Science and Technology (JAMSTEC) to enable research and publication of this work is greatly appreciated. Bindura University of Science Education is thanked for providing the first author with facilities to conduct the research. Funding JAMSTEC supported this work through S.K. Behera, who received research support for the publication of the work Competing Interests The authors have no relevant financial or non-financial interests to disclose. Author Contributions All authors contributed to the study conception and design. D. Manatsa, T. Mushore and S.K. Behera performed material preparation, data collection and analysis. D. Manatsa wrote the first draft of the manuscript, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Data Availability All datasets used in this study are publicly available. Monthly mean stratospheric ozone data were obtained from the Multi-Sensor Reanalysis version 2 (MSR-2), produced under the European Space Agency Climate Change Initiative (ESA-CCI) Ozone project (van der A et al., 2015). These data were accessed and downloaded via the KNMI Climate Explorer (https://climexp.knmi.nl/) for the period 1980–2023. Monthly mean air temperature data at 100 hPa were obtained from the NCEP/NCAR Reanalysis I (Kalnay et al., 1996), which is available through the NOAA Climate and Plotting page (https://psl.noaa.gov/cgi-bin/data/). Monthly mean depth of the 20 °C isotherm (Z20) was obtained from the POAMA/PEODAS ocean reanalysis produced by the Australian Bureau of Meteorology and made available through its THREDDS/OpenDAP server ( http://opendap.bom.gov.au:8080/thredds/dodsC/poama/peodas/reanalysis/ ). Pre-processed Z20 fields derived from this reanalysis were accessed via the KNMI Climate Explorer. However, the data for the specific regions and time periods analysed in this work are available at https://github.com/dmanatsa-cloud/data_used_ozone-z20_manuscript. All data processing and analysis were performed using standard statistical methods, and no proprietary datasets were used. References Cai W, Santoso A, Collins M et al (2021). Changing El Niño–Southern Oscillation in a warming climate. Nat Rev Earth Environ 2:628–644. https://doi.org/10.1038/s43017-021-00199-1 McPhaden MJ, Zebiak SE, Glantz MH (1998) ENSO as an integrating concept in Earth science. Science 314:1740–1745. https://doi.org/10.1126/science.1132588 Capotondi A, Wittenberg AT, Newman M et al (2015) Understanding ENSO diversity. Bull Am Meteorol Soc 96:921–938. https://doi.org/10.1175/BAMS-D-13-00117.1 Albers JR, Butler AH, Langford AO, Elsbury D, Breeden ML (2022) Dynamics of ENSO-driven stratosphere-to-troposphere transport of ozone over North America. Atmos Chem Phys 22:13035–13048. https://doi.org/10.5194/acp-22-13035-2022 Tedetti M, Sempéré R (2006) Penetration of ultraviolet radiation in the marine environment: A review. Photochem Photobiol 82:389–397. https://doi.org/10.1562/2005-11-09-IR-733 Rochelle-Newall EJ, Fisher TR (2002) Production of chromophoric dissolved organic matter fluorescence in marine and estuarine environments: an investigation into the role of phytoplankton. Mar Chem 77:7–21. https://doi.org/10.1016/S0304-4203(01)00078-8 Manatsa D, Mukwada G (2017) A connection from stratospheric ozone to El Niño–Southern Oscillation. Sci Rep 7:6415. https://doi.org/10.1038/s41598-017-06776-7 Chiodo G, Austin J, Polvani LM, Marsh DR, Garcia RR (2018) The impact of interactive stratospheric chemistry on surface climate. J Clim 31:3905–3920. https://doi.org/10.1175/JCLI-D-17-0501.1 van der A RJ, Allaart MAF, Eskes HJ (2015) Extended and refined multi sensor reanalysis of total ozone for the period 1970–2012. Atmos Meas Tech 8:3021–3035. https://doi.org/10.5194/amt-8-3021-2015 Kalnay E, Kanamitsu M, Kistler R et al (1996) The NCEP/NCAR 40-year reanalysis project. Bull Am Meteorol Soc 77:437–471. https://doi.org/10.1175/1520-0477(1996)0772.0.CO;2 Bretherton CS, Widmann M, Dymnikov VP, Wallace JM, Bladé I (1999) The effective number of spatial degrees of freedom of a time-varying field. J Clim 12:1990–2009. https://doi.org/10.1175/1520-0442(1999)0122.0.CO;2 North GR, Bell TL, Cahalan RF, Moeng FJ (1982) Sampling errors in the estimation of empirical orthogonal functions. Mon Weather Rev 110:699–706. https://doi.org/10.1175/1520-0493(1982)1102.0.CO;2 Granger CWJ (1969) Investigating causal relations by econometric models and cross-spectral methods. Econometrica 37:424–438. https://doi.org/10.2307/1912791 Ball WT, Haigh JD, Rozanov EV et al (2016) High solar cycle spectral variations inconsistent with stratospheric ozone observations. Nat Geosci 9:206–209. https://doi.org/10.1038/ngeo2640 Butchart N (2022) The stratosphere: a review of the dynamics and variability. Weather Clim Dynam 3:1237–1272. https://doi.org/10.5194/wcd-3-1237-2022 Elsbury D, Butler AH, Albers JR, Breeden ML, Langford AO (2023) The response of the North Pacific jet and stratosphere-to-troposphere transport of ozone over western North America to RCP8.5 climate forcing. Atmos Chem Phys 23:5101–5117. https://doi.org/10.5194/acp-23-5101-2023 Dee DP, Uppala SM, Simmons AJ et al (2011) The ERA-Interim reanalysis: configuration and performance of the data assimilation system. Q J R Meteorol Soc 137:553–597. https://doi.org/10.1002/qj.828 Dong Y, Polvani LM, Hwang Y-T, England MR (2025) Stratospheric ozone depletion has contributed to the recent tropical La Niña-like cooling pattern. npj Clim Atmos Sci 8:150. https://doi.org/10.1038/s41612-025-01020-0 Józefiak I, Sukhodolov T, Egorova T et al (2023) Stratospheric dynamics modulates ozone layer response to molecular oxygen variations. Front Earth Sci 11:1239325. https://doi.org/10.3389/feart.2023.1239325 Karpechko AY, Vitart F, Statnaia I, Balmaseda MA, Charlton-Perez AJ (2024) The tropical influence on sub-seasonal predictability of wintertime stratosphere and stratosphere–troposphere coupling. Q J R Meteorol Soc 150:1125–1142. https://doi.org/10.1002/qj.4678 Nowack PJ, Braesicke P, Abraham NL, Pyle JA (2017) On the role of ozone feedback in the ENSO amplitude response under global warming. Geophys Res Lett 44:3858–3866. https://doi.org/10.1002/2016GL072418 Polvani LM, Waugh DW, Chiodo G, Hegglin MI, Matthes K (2020) Stratospheric ozone depletion and Southern Ocean surface wind trends. Geophys Res Lett 47:e2020GL087369. https://doi.org/10.1029/2020GL087369 Randel WJ, Wu F (2021) A simple model of ozone–temperature coupling in the tropical lower stratosphere. Atmos Chem Phys 21:18531–18542. https://doi.org/10.5194/acp-21-18531-2021 Son S-W, Kim Y, Lu J, Yoo C (2024) Stratospheric influence on tropical Pacific decadal variability. Nat Clim Change 14:321–328. https://doi.org/10.1038/s41558-024-01923-8 Tian W, Huang J, Zhang J et al (2023) Role of stratospheric processes in climate change: advances and challenges. Adv Atmos Sci 40:1379–1400. https://doi.org/10.1007/s00376-022-2246-8 Young PJ, Naik V, Fiore AM et al (2018) Tropospheric Ozone Assessment Report: Assessment of global-scale model performance for global and regional ozone distributions, variability, and trends. Elementa: Sci Anthropocene 6:10. https://doi.org/10.1525/elementa.265 Zuo H, Balmaseda MA, Tietsche S et al (2019) The ECMWF operational ocean analysis system: ORAS5. ECMWF Tech Memo 851. https://doi.org/10.21957/1r6x0z0j Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8425241","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":568374703,"identity":"e163a92e-3c98-4b38-9eb5-3780fa8631b6","order_by":0,"name":"Desmond Manatsa","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAzElEQVRIiWNgGAWjYDCCA1CaH0QkFBCtJYGBQbIBRBuQosUAzCBGC9+1w88e/PxRJ298fnXihwcGDPL8Ygfwa5G8nWZu2JNw2HDbjbebJYAOM5w5OwG/FoPbCWYSPAkHGLfdOLsBpCUBKEJIS/o3yT8JdfabZ5zd/INILTlm0jwJzIkb+Hu3EWeL5O2cMmmZtMPJM27wbrNIMJAg7Be+2+nbJN/Y1Nn295/dfPNHhY08vzQBLQggAVYpQaxyEOA/QIrqUTAKRsEoGEkAAAxOSDq8f/NoAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0002-8119-6150","institution":"Bindura University of Science Education","correspondingAuthor":true,"prefix":"","firstName":"Desmond","middleName":"","lastName":"Manatsa","suffix":""},{"id":568374704,"identity":"7f48fb63-9734-4691-ac5b-05ea993384ac","order_by":1,"name":"Darlington Mushore","email":"","orcid":"","institution":"University of Zimbabwe","correspondingAuthor":false,"prefix":"","firstName":"Darlington","middleName":"","lastName":"Mushore","suffix":""},{"id":568374705,"identity":"d61d1c47-57f0-410c-8742-5c9cda4795ff","order_by":2,"name":"Swadhin K. Behera","email":"","orcid":"","institution":"JAMSTEC","correspondingAuthor":false,"prefix":"","firstName":"Swadhin","middleName":"K.","lastName":"Behera","suffix":""}],"badges":[],"createdAt":"2025-12-22 12:57:47","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8425241/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8425241/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":99507665,"identity":"65813420-b27d-42e2-86e3-3cf328e672a4","added_by":"auto","created_at":"2026-01-05 08:55:22","extension":"xml","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7232,"visible":true,"origin":"","legend":"","description":"","filename":"cldyCLDYD2501419.xml","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/fbe6c4f6979ffa98083fcdf2.xml"},{"id":99507661,"identity":"b6021446-7763-4eeb-8cdb-a5bf5821c218","added_by":"auto","created_at":"2026-01-05 08:55:22","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":969,"visible":true,"origin":"","legend":"","description":"","filename":"CLDYD250141923213.go.xml","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/5c9965a8db3def7547ace66c.xml"},{"id":99507644,"identity":"bbd2e267-d112-4530-847e-03a12ddc8360","added_by":"auto","created_at":"2026-01-05 08:55:21","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":815,"visible":true,"origin":"","legend":"","description":"","filename":"CLDYD2501419Import.xml","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/67ce33039d00f8f11856b92a.xml"},{"id":99507659,"identity":"7e40811f-664d-4a01-9554-c3e27ff88e2b","added_by":"auto","created_at":"2026-01-05 08:55:21","extension":"xml","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":117272,"visible":true,"origin":"","legend":"","description":"","filename":"CLDYD25014190enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/f2bd0eead4edd03b659b6e76.xml"},{"id":99507647,"identity":"f637f6d6-302d-49b8-97f7-f79ceef0314b","added_by":"auto","created_at":"2026-01-05 08:55:21","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":118696,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/ab782a04f9715ec766201675.png"},{"id":99791863,"identity":"e448ff68-de5b-4554-8432-6bdbaa7a746f","added_by":"auto","created_at":"2026-01-08 13:11:12","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":57610,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/cdde22d4a31f0d92a9dd91d5.png"},{"id":99507629,"identity":"e18f8da5-b1ae-4eaf-801f-7016750f340a","added_by":"auto","created_at":"2026-01-05 08:55:21","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":24926,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/d3f9892181968c52300026b1.png"},{"id":99507654,"identity":"23463160-fe25-4812-94c4-08ceb55a69fd","added_by":"auto","created_at":"2026-01-05 08:55:21","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":19067,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/8f235d05a66d88cf61bd465b.png"},{"id":99507678,"identity":"331fcfd4-18c2-47df-8862-19c1812c87fd","added_by":"auto","created_at":"2026-01-05 08:55:23","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":16312,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/23f5dde85fb91f3231b69072.png"},{"id":99507663,"identity":"5df58f16-12f6-4fcd-94a7-1d332ed055a6","added_by":"auto","created_at":"2026-01-05 08:55:22","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":29451,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/247e6ea9aeb509457a892aca.png"},{"id":99507617,"identity":"0a3e56a3-bc75-4402-b905-e605c24c7d0e","added_by":"auto","created_at":"2026-01-05 08:55:15","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":126264,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/b51d0b096c21c5e4893e860c.png"},{"id":99507608,"identity":"2fe4f478-29be-4160-a156-3f9e7887439f","added_by":"auto","created_at":"2026-01-05 08:55:11","extension":"xml","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":114018,"visible":true,"origin":"","legend":"","description":"","filename":"CLDYD25014190structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/9f41d5266f6027527c38aef8.xml"},{"id":99507645,"identity":"25dc4cec-8994-41d0-813a-d22444845d65","added_by":"auto","created_at":"2026-01-05 08:55:21","extension":"html","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":125331,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/7fe111ea486756362a7568db.html"},{"id":99507614,"identity":"778c1bbf-5a9b-46b4-8af7-8fc1323f5867","added_by":"auto","created_at":"2026-01-05 08:55:15","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":752565,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eSpatial and vertical structure of the stratosphere–ocean teleconnection. (a) Point correlation between total column stratospheric ozone and 20°C isotherm depth (Z20) during December–March (1980–2019, detrended). The red box denotes the South Pacific “sensitive region” (9–12°S, 125–115°W) with maximum negative correlation (r ≈ –0.6 to –0.7). (b) First empirical orthogonal function (EOF) of Z20, showing the dominant mode of thermocline variability across the tropical Pacific, which overlaps spatially with the ozone-sensitive region. (c) Composite vertical temperature anomalies (°C) during ENSO events, averaged over the sensitive region, revealing a dipole structure: warming (cooling) in the lower stratosphere (~100 hPa) and cooling (warming) in the upper troposphere (~300 hPa) during La Niña (El Niño). (d) Corresponding geopotential height anomalies (m) showing vertically coherent wave-like structures that connect the stratospheric temperature dipole to circulation adjustments in the lower troposphere (~600–900 hPa), ultimately modulating surface wind stress.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/3c490518cdd02556e53f966c.png"},{"id":99507660,"identity":"908b8ca8-e88f-4c11-96a3-a1e1bb4bf009","added_by":"auto","created_at":"2026-01-05 08:55:22","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":252559,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eStatistical linkages among ENSO, stratospheric ozone, and thermocline depth. (a) Scatterplot of the NINO3.4 index versus the first principal component (PC1) of Z20 (1980–2020), showing a strong linear relationship (r = 0.950), confirming that Z20 is a robust ENSO indicator. (b) Scatterplot of detrended total column ozone (200–230°E, 20–10°S) versus Z20 PC1, yielding a significant negative correlation (r = –0.609; p \u0026lt; 0.001), corresponding to 37% explained variance. (c–e) Time series of standardised anomalies: (c) ozone, (d) 100 hPa temperature, and (e) 300 hPa temperature. Shading indicates ENSO phase (red: El Niño, blue: La Niña). Note the opposing anomalies in (d) and (e), confirming the vertical temperature dipole, and the coherence with ozone variability in (c).\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/9cb21946a8048325eb12e391.png"},{"id":99507655,"identity":"216c6f7f-08a3-4435-903d-96d7769159a2","added_by":"auto","created_at":"2026-01-05 08:55:21","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":117825,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eSeasonal evolution and lead–lag relationship in the ozone–thermocline coupling during ENSO. (a) Composite total column ozone anomalies (DU) for La Niña (blue) and El Niño (red) years over a July–June “water year.” Ozone peaks (+15 DU) during September–November of La Niña and reaches a minimum (–12 DU) in January–February of El Niño. (b) Corresponding Z20 anomalies (m) showing thermocline shoaling (–25 m) during La Niña and deepening (+32 m) during El Niño, with peak Z20 response occurring in December–January. Critically, ozone anomalies lead Z20 anomalies by 5–10 months, consistent with a radiative mechanism in which cumulative UVB heating of the thermocline, which is initiated when the “ozone window” opens, requires several months to overcome oceanic thermal inertia and manifest as a detectable thermocline displacement.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/d103ba7b67cca2f5eaef2d2e.png"},{"id":99507620,"identity":"92ec35d3-affb-4ee8-82fd-80ffe45fcc9c","added_by":"auto","created_at":"2026-01-05 08:55:16","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":65276,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eCross-correlation function (CCF) between monthly ozone anomalies over the South Tropical Pacific and Z20 (20°C isotherm depth, a proxy for thermocline depth), with positive lags indicating ozone leading. The red shaded region shows the 95% confidence interval (±0.089) based on Bartlett’s formula, accounting for autocorrelation in both series. A statistically significant negative correlation (r = –0.51) is observed at a lag of 2 months, indicating that ozone anomalies precede and are inversely related to variations in thermocline depth.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/4174c16600d2a56397c2068e.png"},{"id":99507664,"identity":"e258fd95-d288-4c7d-be37-503d362172ee","added_by":"auto","created_at":"2026-01-05 08:55:22","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":81271,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eGranger causality tests (1980–2020) confirm a unidirectional influence from stratospheric ozone to thermocline depth. (a) Forward causality (ozone → Z20): past ozone anomalies significantly improve the prediction of future Z20 anomalies across lags 1–7 months (F \u0026gt; 3.9, p \u0026lt; 0.01), with peak significance at lag 1 (F = 6.8, p = 0.0001). (b) Reverse causality (Z20 → ozone): only marginal significance at lag 1 (F = 4.1, p = 0.045) and no significant predictive power at lags ≥ 2 (F ≤ 2.6, p \u0026gt; 0.1). The red dashed line denotes the 5% significance threshold (F ≈ 3.9). This robust asymmetry provides statistical evidence that stratospheric ozone drives thermocline variability, while oceanic feedback to ozone is negligible. This supports a causal, radiatively mediated ozone window mechanism.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/3f09ff9de46a6859240f331b.png"},{"id":99507613,"identity":"dea03fd9-a0e6-49c8-9886-9640ba0d3441","added_by":"auto","created_at":"2026-01-05 08:55:14","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":101027,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eImpulse response functions (IRFs) from a bivariate VAR model showing the asymmetric causal dynamics between stratospheric ozone and thermocline depth (Z20). (a) Response of Z20 to a one-standard-deviation negative ozone shock (ozone depletion). Z20 deepens gradually, peaking at +218 cm during months 4–7 and remaining significant for 15 months, with a total integrated response of +480 cm-months—consistent with cumulative UVB heating of the 15–25 m layer. (b) Response of ozone to a one-standard-deviation positive Z20 shock (thermocline deepening). Ozone shows only a slight, transient decrease (–1.8 DU) at months 2–3, decaying to insignificance by month 9—reflecting a short-lived convective feedback. The stark contrast confirms that the dominant coupling mechanism is unidirectional: ozone drives thermocline variability via direct radiative heating, not vice versa. Shaded areas denote 95% confidence bands.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/2258c2fab1a856c4531c2ac9.png"},{"id":99507666,"identity":"270457b7-0520-41d2-bd2a-8d4c7b8dbe55","added_by":"auto","created_at":"2026-01-05 08:55:22","extension":"jpeg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":552967,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eVisual confirmation of the ozone–thermocline coupling over the 1980–2020 observational record. (a) Scatterplot of Z20 anomalies (m) versus inverted ozone anomalies (–ozone, DU), a proxy for UVB window opening. Red, blue, and purple points denote El Niño, La Niña, and neutral months, respectively. A robust positive correlation (r = 0.61, p \u0026lt; 0.001) with a slope of ~2.3 m per DU confirms that greater ozone depletion (enhanced UVB penetration) leads to a deeper thermocline. Extreme ENSO events align tightly with the regression line, confirming the mechanism’s robustness across event intensities. (b) Complete 40-year monthly time series of inverted ozone (blue; UVB window opening) and Z20 (purple). Ozone depletion consistently leads to a 5–10-month deepening of the thermocline across all primary ENSO cycles, demonstrating persistent temporal precedence and ruling out spurious correlation.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage7.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/7db59e7f2914b5b88d84563b.jpeg"},{"id":101880603,"identity":"dc2e484f-76c7-46ae-9838-397c57760c1d","added_by":"auto","created_at":"2026-02-04 15:04:27","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2551508,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8425241/v1/6aa08b08-e85b-454e-85af-ed9b2b55b316.pdf"}],"financialInterests":"","formattedTitle":"Stratospheric Ozone Depletion Drives Tropical Pacific Thermocline Variability via Enhanced UVB Penetration to the Upper Ocean","fulltext":[{"header":"1. INTRODUCTION","content":"\u003cp\u003e1.1 Background and Motivation\u003c/p\u003e\n\u003cp\u003eThe El Ni\u0026ntilde;o\u0026ndash;Southern Oscillation (ENSO) is the dominant mode of interannual climate variability, exerting a profound influence on global weather extremes, marine productivity, agricultural yields, and socioeconomic stability (Cai et al., 2021). At the heart of ENSO dynamics lies the modulation of the depth of the tropical Pacific thermocline. The depth of the 20\u0026deg;C isotherm (Z20) governs the efficiency of equatorial upwelling and, consequently, sea surface temperature (SST) anomalies in the eastern Pacific. During El Ni\u0026ntilde;o, westerly wind bursts excite downwelling Kelvin waves that depress the thermocline, suppressing cold-water upwelling and enabling SST warming. Conversely, La Ni\u0026ntilde;a features easterly wind anomalies, upwelling Kelvin waves, and a shoaling thermocline that enhances cooling. This wind-driven oceanic adjustment underpins ENSO theory (McPhaden et al., 1998).\u003c/p\u003e\n\u003cp\u003eNevertheless, despite decades of research, key aspects of ENSO remain enigmatic: its pronounced asymmetry (stronger El Ni\u0026ntilde;o vs. La Ni\u0026ntilde;a events), its diverse spatial \u0026quot;flavours\u0026quot; (e.g., canonical vs. Modoki), and its apparent modulation on decadal timescales (Capotondi et al., 2015). These features suggest that wind stress alone cannot fully account for the observed variability in the thermocline. Hence, it motivates the search for additional forcing mechanisms, particularly those involving direct radiative coupling between the stratosphere and ocean.\u003c/p\u003e\n\u003cp\u003e1.2 Stratospheric Ozone as a Radiative Gatekeeper for UVB\u003c/p\u003e\n\u003cp\u003eStratospheric ozone is a potent absorber of solar ultraviolet radiation, especially in the biologically and radiatively critical UVB band (280\u0026ndash;320 nm). The ozone layer acts as a radiative gatekeeper, determining how high-energy UVB radiation reaches the Earth\u0026apos;s surface and penetrates the ocean. Total column ozone exhibits substantial interannual variability in the tropics, driven not only by photochemistry but also by dynamical forcing from below, most notably ENSO. During El Ni\u0026ntilde;o, enhanced tropical upwelling in the lower stratosphere transports ozone-poor air from the troposphere, generating negative ozone anomalies of 5\u0026ndash;15 Dobson Units (DU); La Ni\u0026ntilde;a produces the opposite effect (Albers et al., 2022).\u003c/p\u003e\n\u003cp\u003eThis well-established ENSO \u0026rarr; Ozone pathway has led to ozone being treated primarily as a passive tracer of tropospheric convection. However, the radiative consequences of these ozone anomalies for the ocean have been almost entirely overlooked. When stratospheric ozone depletes, it creates an \u0026quot;ozone window\u0026quot;. Under such circumstances, less UVB is absorbed in the lower stratosphere (~100 hPa), hence allowing significantly more UVB radiation to penetrate downward through the atmosphere and into the upper ocean.\u003c/p\u003e\n\u003cp\u003eUVB penetration in seawater is depth selective. In clear tropical Pacific waters, UVB radiation (280\u0026ndash;320 nm) penetrates to depths of 15\u0026ndash;25 meters before being absorbed, with peak absorption occurring precisely at the depth range of the 20\u0026deg;C isotherm (Z20) that defines the thermocline (Tedetti \u0026amp; Semp\u0026eacute;r\u0026eacute;, 2006; Rochelle-Newall \u0026amp; Fisher, 2002). This depth corresponds to where the thermocline typically resides in the tropical Pacific. When the \u0026quot;ozone window\u0026quot; opens during ozone depletion, the enhanced UVB flux directly heats the thermocline layer, reducing the vertical temperature gradient and causing the thermocline to deepen. Conversely, when ozone increases (the window closes), reduced UVB reaching the thermocline allows it to cool and shoal.\u003c/p\u003e\n\u003cp\u003eAlthough the absolute energy in UVB is small relative to total solar irradiance (~1.5% of total solar radiation), its spectral selectivity and preferential absorption at thermocline depths make it a potent agent for directly modulating upper-ocean stratification. Unlike visible and infrared radiation, which are absorbed within the top few meters, UVB radiation, with a penetration depth of 15\u0026ndash;25 m, directly affects the structure of the thermocline.\u003c/p\u003e\n\u003cp\u003e1.3 The Knowledge Gap: Is Ozone Merely a Response\u0026mdash;or an Active Driver?\u003c/p\u003e\n\u003cp\u003eWhile ENSO\u0026apos;s upward influence on stratospheric ozone is well documented (Albers et al., 2022), the reverse pathway, whether stratospheric ozone can actively modulate tropical ocean dynamics through direct radiative forcing, remains unexplored. Manatsa \u0026amp; Mukwada (2017) first proposed a statistical link between tropical lower-stratospheric ozone and ENSO, but their analysis lacked causal testing and a quantified physical mechanism. Most climate models, including those in CMIP6, either prescribe climatological ozone or employ simplified stratospheric chemistry, and, critically, most ocean components lack spectral resolution for UVB absorption at depth, thereby implicitly assuming that all solar radiation is absorbed at the surface (Chiodo et al., 2018). Consequently, no study has yet demonstrated that stratospheric ozone anomalies precede and predict thermocline changes through direct radiative heating at thermocline depths.\u003c/p\u003e\n\u003cp\u003eThis gap is significant because, if ozone acts as an active driver through direct ocean heating, it could represent vital but previously unaccounted-for positive feedback in ENSO. The initial wind-driven thermocline changes alter stratospheric ozone via convection, which in turn, via the UVB radiative window effect, directly heats or cools the thermocline layer, reinforcing the original anomaly. Such feedback would be especially relevant for extreme events, where ozone anomalies exceed 12 DU, and could help explain ENSO\u0026apos;s asymmetry and diversity.\u003c/p\u003e\n\u003cp\u003e1.4 Hypothesis and Research Objectives\u003c/p\u003e\n\u003cp\u003eThis study rigorously tests the hypothesis that stratospheric ozone variability influences tropical Pacific thermocline depth through direct radiative forcing. In this case, ozone depletion creates a \u0026quot;window\u0026quot; that allows enhanced UVB penetration to thermocline depths (15\u0026ndash;25 m), thereby directly heating the Z20 layer and causing thermocline deepening. We specifically investigate whether:\u003c/p\u003e\n\u003cp\u003e(i) Ozone anomalies precede Z20 changes at lags consistent with radiative heating and oceanic adjustment (5\u0026ndash;10 months);\u003c/p\u003e\n\u003cp\u003e(ii) Past ozone values improve statistical prediction of future Z20 beyond Z20\u0026apos;s own history, as assessed by Granger causality;\u003c/p\u003e\n\u003cp\u003e(iii) The coupling operates through a physically consistent direct heating mechanism where ozone depletion \u0026rarr; enhanced UVB at 20m depth \u0026rarr; thermocline warming and deepening; and\u003c/p\u003e\n\u003cp\u003e(iv) The signal is robust across multiple ENSO events, particularly during extremes, and persists after removal of long-term trends.\u003c/p\u003e\n\u003cp\u003eBy integrating observational data, causal inference, and dynamical systems analysis, we aim to establish whether stratospheric ozone is a quantifiable contributor to tropical Pacific thermocline variability through direct radiative forcing. This finding has important implications for ENSO prediction, model development, and climate policy.\u003c/p\u003e"},{"header":"2. DATA AND METHODS","content":"\u003cp\u003e2.1 Stratospheric Ozone and Atmospheric Data\u003c/p\u003e\n\u003cp\u003eMonthly mean ozone data are obtained from the Multi-Sensor Reanalysis version 2 (MSR-2), produced under the European Space Agency Climate Change Initiative (ESA-CCI) Ozone project (van der A et al., 2015). The dataset provides a temporally homogeneous global ozone record covering the satellite era (1979\u0026ndash;present) on a regular latitude\u0026ndash;longitude grid. The data used covered the period 1980 to 2023 and were downloaded from the KNMI Climate Explorer (https://climexp.knmi.nl/). Monthly averaging isolates ENSO-scale variability, enabling analysis of ENSO-related ozone anomalies associated with large-scale circulation changes and stratosphere\u0026ndash;troposphere coupling (Ziemke et al., 2019).\u003c/p\u003e\n\u003cp\u003eThese monthly means are then computed for the South Tropical Pacific domain (200\u0026deg;E\u0026ndash;230\u0026deg;E, 20\u0026deg;S\u0026ndash;10\u0026deg;S), a region selected based on preliminary point correlation analysis (Section 3.1), which identified it as the zone of maximum ozone\u0026ndash;Z20 coupling. This region lies just south of the equator, where clear ocean waters allow maximum penetration of UVB, and is optimally positioned to capture stratospheric signals that influence the eastern Pacific thermocline through direct radiative forcing.\u003c/p\u003e\n\u003cp\u003eTo validate the vertical structure of UVB absorption, we use monthly-mean air temperature from the NCEP/NCAR Reanalysis I (Kalnay et al., 1996) at 100 hPa (~16 km) to confirm the presence of the ozone window effect in the lower stratosphere during ozone depletion events.\u003c/p\u003e\n\u003cp\u003e2.2 Ocean Thermocline Data\u003c/p\u003e\n\u003cp\u003eThe depth of the 20 \u0026deg;C isotherm (Z20) signifies the ocean depth at which the temperature equals 20 \u0026deg;C and is widely used as a proxy for upper-ocean thermocline depth and heat content in tropical ocean studies. For this study, we extract monthly Z20 in the eastern equatorial Pacific (239\u0026deg;E\u0026ndash;261\u0026deg;E, 2\u0026deg;S\u0026ndash;7\u0026deg;S), the region of maximum ENSO-related thermocline variability. The data were obtained via the KNMI Climate Explorer, which accesses pre-computed subsurface ocean diagnostics derived from assimilated ocean reanalysis to provide gridded estimates of thermocline variability. The source dataset for the Z20 fields was the POAMA/PEODAS ocean reanalysis, available at http://opendap.bom.gov.au:8080/thredds/dodsC/poama/peodas/reanalysis/. The 20\u0026deg;C isotherm depth (typically 15\u0026ndash;25 m in this region) corresponds precisely to the depth of maximum UVB absorption in tropical Pacific waters, making it the ideal metric for studying direct radiative forcing of the thermocline.\u003c/p\u003e\n\u003cp\u003e2.3 Statistical Methods and Temporal Conventions\u003c/p\u003e\n\u003cp\u003eTo align with the mature phase of ENSO, a July\u0026ndash;June \u0026ldquo;water year\u0026rdquo; convention is adopted, ensuring that each ENSO event (e.g., 1997\u0026ndash;98) is contained within a single analysis year. All time series are deseasonalised by subtracting the monthly climatology for the 1991-2020 period to align with the new climate-normal period. This was linearly detrended to isolate interannual variability from long-term climate change signals, such as ozone recovery and ocean warming. Statistical analysis is based on point correlation techniques to identify regions of maximum ozone\u0026ndash;Z20 coupling, and on cross-correlation analysis to quantify lead\u0026ndash;lag relationships between variables for lags \u0026tau; ranging from \u0026minus;12 to +12 months. The statistical significance of correlations is assessed using a two-tailed t-test adjusted for autocorrelation following Bretherton et al. (1999). ENSO composite analysis is performed using the Oceanic Ni\u0026ntilde;o Index (ONI), with El Ni\u0026ntilde;o and La Ni\u0026ntilde;a events defined as 5-month running means of sea surface temperature anomalies in the Ni\u0026ntilde;o-3.4 region exceeding \u0026plusmn;0.5 \u0026deg;C for at least five consecutive months. To maximise the ENSO signal, composites are constructed using only the strongest events, namely the El Ni\u0026ntilde;o episodes of 1982/83, 1986/87, 1997/98, and 2015/16, and the La Ni\u0026ntilde;a episodes of 1988/89, 1999/00, 2007/08, and 2010/11. Monthly composites are calculated across these selected events, and statistical significance is evaluated using a one-sample t-test with degrees of freedom equal to n\u0026minus;1, where n denotes the number of events.\u003c/p\u003e\n\u003cp\u003e2.4 Vector Autoregression (VAR) and Impulse Response Functions\u003c/p\u003e\n\u003cp\u003eTo move beyond correlation and test for directional predictability, we estimate a 4-variable Vector Autoregression (VAR) model:\u003c/p\u003e\n\u003cp\u003e\u003cimg src=\"data:image/png;base64,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\"\u003e\u003c/p\u003e\n\u003cp\u003eGranger causality tests are derived from the VAR to assess whether past values of one variable improve the prediction of another (Granger, 1969). Orthogonalized Impulse Response Functions (IRFs), computed via Cholesky decomposition (ordering: Z20 \u0026rarr; Ozone), trace the system\u0026apos;s dynamic response to a one-standard-deviation shock over a 24-month horizon. 95% confidence intervals are estimated using Monte Carlo simulation (1,000 draws).\u003c/p\u003e\n\u003cp\u003e2.5 Robustness Subsampling\u003c/p\u003e\n\u003cp\u003eTo test the robustness of our findings, we conduct two key sensitivity analyses:\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003eSubstantial ENSO subset: restrict analysis to 8 extreme events (4 El Ni\u0026ntilde;o, 4 La Ni\u0026ntilde;a) to assess signal strength during high-amplitude ENSO.\u003c/li\u003e\n \u003cli\u003eDetrended strong ENSO: apply linear detrending to the strong-event subset to eliminate any residual influence of multidecadal trends (e.g., post-2000 ozone recovery).\u003c/li\u003e\n \u003cli\u003eThis approach ensures that our conclusions are not driven by weak events or long-term drift, but reflect robust interannual dynamics driven by the direct radiative mechanism.\u003c/li\u003e\n\u003c/ul\u003e"},{"header":"3. RESULTS","content":"\u003cp\u003e3.1 Spatial Structure of Ozone\u0026ndash;Thermocline Coupling\u003c/p\u003e\n\u003cp\u003eAnalysis of the spatial structure reveals a distinct ozone\u0026ndash;thermocline teleconnection, as illustrated in Figure 1. Figure 1a maps the point correlation between total column stratospheric ozone and the depth of the 20\u0026deg;C isotherm (Z20) during December\u0026ndash;March (1980\u0026ndash;2019, detrended). A region of strong negative correlation (r\u0026asymp;-0.61 to -0.67) is centred at 9\u0026ndash;12\u0026deg;S, 125\u0026ndash;115\u0026deg;W (red box), hereafter termed the \u0026ldquo;sensitive region.\u0026rdquo; This zone overlaps with apparent waters of the tropical South Pacific, where ultraviolet-B (UVB) radiation penetrates most efficiently to thermocline depths (15\u0026ndash;25 m), enabling direct radiative coupling between stratospheric ozone and the upper ocean.\u003c/p\u003e\n\u003cp\u003eTo place this region in the context of large-scale oceanic variability, an empirical orthogonal function (EOF) analysis was performed on Z20 anomalies over the tropical Pacific domain (15\u0026deg;S\u0026ndash;15\u0026deg;N, 140\u0026deg;W\u0026ndash;80\u0026deg;W). The first four principal components account for 49.93%, 16.37%, 8.94%, and 4.67% of the total variance, respectively. Application of North\u0026rsquo;s rule of thumb (North et al., 1982) confirms that the first eigenvalue is statistically well separated from the others, indicating that EOF1 represents a dominant and physically meaningful mode. As shown in Figure 1b, EOF1 exhibits maximum loading in the eastern\u0026ndash;central equatorial Pacific, a canonical signature of ENSO-related thermocline variability, and notably overlaps with the ozone-sensitive region identified in Figure 1a. This spatial alignment supports the interpretation that the observed ozone\u0026ndash;thermocline correlation occurs within the primary dynamical framework of tropical Pacific climate variability.\u003c/p\u003e\n\u003cp\u003eThe negative correlation implies an inverse relationship: ozone depletion (negative anomaly) opens an \u0026ldquo;ozone window,\u0026rdquo; thereby increasing UVB penetration, which warms the upper ocean and causes the thermocline to deepen (positive Z20 anomaly). Conversely, ozone enhancement (positive anomaly) restricts UVB flux, leading to thermocline shoaling (negative Z20 anomaly).\u003c/p\u003e\n\u003cp\u003eThe atmospheric structure associated with this coupling is further examined by averaging composite anomalies over the sensitive region during ENSO events. Figure 1c shows a vertical dipole in temperature: during La Ni\u0026ntilde;a, the lower stratosphere (~100 hPa) warms while the upper troposphere (~300 hPa) cools; El Ni\u0026ntilde;o composites show the opposite pattern. This stratosphere\u0026ndash;troposphere thermal contrast suggests a vertically structured radiative and/or dynamical response tied to ozone variability.\u003c/p\u003e\n\u003cp\u003eFigure 1d presents the corresponding geopotential height anomalies, which reveal a vertically coherent wave-like structure extending from the lower stratosphere down to the lower troposphere (~600\u0026ndash;900 hPa), linking stratospheric thermal anomalies to near-surface circulation changes. Importantly, in the upper troposphere (~200\u0026ndash;300 hPa), the geopotential height field exhibits twin anticyclonic anomalies, one in each hemisphere, symmetrically flanking the equator. These anticyclones reflect subsidence and horizontal divergence on either side of the equator, consistent with a Rossby-wave-type response to tropical forcing. Their presence indicates that the off-equatorial ozone\u0026ndash;thermocline coupling (Figure 1b) excites a broader atmospheric circulation pattern that redistributes mass and momentum across the tropical Pacific, ultimately influencing surface wind stress and reinforcing thermocline feedback.\u003c/p\u003e\n\u003cp\u003eTogether, Figures 1a\u0026ndash;d provide a consistent, multi-level depiction of a coupled stratosphere\u0026ndash;ocean system, in which ozone modulates thermocline depth through both direct radiative effects (UVB penetration) and indirect dynamical pathways (via tropospheric circulation anomalies, including the twin anticyclones).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2 Statistical Relationships and ENSO Coupling\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe statistical robustness of the coupling is quantified in Figure 2. Figure 2a confirms that Z20 is a robust ENSO indicator, demonstrating that a scatterplot of the NINO3.4 index versus Z20 PC1 yields r = 0.950 (p \u0026lt; 0.001). Figure 2b shows the specific ozone\u0026ndash;Z20 relationship: detrended ozone (200\u0026ndash;230\u0026deg;E, 20\u0026ndash;10\u0026deg;S) versus Z20 PC1 yields r = \u0026ndash;0.609 (p \u0026lt; 0.001), accounting for 37% of the variance in Z20.\u003c/p\u003e\n\u003cp\u003eFigure 2c displays the time series of standardised ozone anomalies, which covary with the ENSO phase (shading). Notably, the 1997\u0026ndash;98 El Ni\u0026ntilde;o shows dramatic ozone depletion (opening the UVB window), which precedes the extreme thermocline deepening by several months\u0026mdash;consistent with the direct radiative heating mechanism at Z20 depth.\u003c/p\u003e\n\u003cp\u003e3.3 Seasonal Evolution and Lead-Lag Structure\u003c/p\u003e\n\u003cp\u003eFigure 3 presents ENSO-phase composites across a July\u0026ndash;June \u0026quot;water year,\u0026quot; revealing the temporal evolution of the ozone window mechanism. Figure 3a displays total column ozone anomalies throughout the ENSO cycle: during La Ni\u0026ntilde;a years (blue curve), ozone increases progressively through boreal summer and fall, peaking at +15 DU in September\u0026ndash;November as the UVB window closes; during El Ni\u0026ntilde;o years (red curve), ozone depletes reaching a minimum of \u0026ndash;12 DU in January\u0026ndash;February as the window opens maximally. The seasonal asymmetry (larger El Ni\u0026ntilde;o depletion than La Ni\u0026ntilde;a enhancement) reflects the nonlinear response of stratospheric dynamics to tropical convection.\u003c/p\u003e\n\u003cp\u003eFigure 3b shows the thermocline depth response: Z20 deepens during El Ni\u0026ntilde;o, reaching a maximum anomaly of +32 m in December\u0026ndash;January, and shoals during La Ni\u0026ntilde;a to \u0026ndash;25 m in the same months. Critically, comparing panels (a) and (b) reveals that ozone changes lead Z20 changes by approximately 5\u0026ndash;10 months: ozone depletion peaks in January\u0026ndash;February, while thermocline deepening peaks in the following December\u0026ndash;January of the mature El Ni\u0026ntilde;o phase.\u003c/p\u003e\n\u003cp\u003eThis extended lag structure is physically consistent with the direct radiative mechanism operating through cumulative heating: (1) ozone depletion opens the UVB window (panel a), (2) enhanced UVB flux penetrates through the atmosphere to 15\u0026ndash;25 m ocean depth, (3) continuous absorption at thermocline depth progressively warms the layer over 5\u0026ndash;10 months, (4) accumulated thermal energy reduces the vertical temperature gradient, (5) weakened stratification allows the thermocline to deepen (panel b). The multi-month lag reflects the time required for cumulative radiative forcing (\u0026sim;50\u0026ndash;80 MJ/m\u0026sup2; integrated over the heating period) to overcome substantial ocean thermal inertia and alter the density structure of the upper ocean.\u003c/p\u003e\n\u003cp\u003e3.4 Cross-Correlation Analysis: Temporal Precedence\u003c/p\u003e\n\u003cp\u003eFigure 4 (Cross-Correlation Function) provides crucial evidence for temporal precedence. The CCF shows significant negative correlation (r = \u0026ndash;0.51, p \u0026lt; 0.01) at negative lags of 5\u0026ndash;10 months, meaning ozone anomalies lead Z20 changes. The negative correlation confirms the inverse relationship: ozone depletion (opening the UVB window) precedes thermocline deepening. Most importantly, the correlation weakens at positive lags, indicating that Z20 changes have minimal predictive power for future ozone. Hence, it is consistent with ozone being the driver rather than the response in this direct radiative pathway.\u003c/p\u003e\n\u003cp\u003e3.5 Granger Causality and Directionality\u003c/p\u003e\n\u003cp\u003eFigure 5a-b establishes directionality using Granger causality on deseasonalized monthly anomalies (1980\u0026ndash;2020). Forward causality (ozone \u0026rarr; Z20) is highly significant for lags 1\u0026ndash;7 (F \u0026gt; 3.9; p \u0026lt; 0.01), peaking at F = 6.8 (p = 0.0001) at lag 1 and remaining robust at F = 5.1 (p = 0.0008) at lag 3, which is well above the 5% significance threshold (F \u0026asymp; 3.9, red dashed line). This confirms that past ozone values contain predictive information about future Z20 beyond Z20\u0026apos;s own history.\u003c/p\u003e\n\u003cp\u003eIn contrast, reverse causality (Z20 \u0026rarr; ozone) is only marginally significant at lag 1 (F = 4.1, p = 0.045) and statistically indistinguishable from noise at all longer lags (F \u0026le; 2.6, p \u0026gt; 0.1 for lags \u0026ge; 2). This profound asymmetry\u0026mdash;robust, multi-month predictive power from ozone to Z20 versus negligible feedback in the reverse direction\u0026mdash;provides decisive statistical evidence that stratospheric ozone actively drives thermocline variability through the direct UVB radiative pathway.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.6 Robustness Across Strong ENSO Events and Detrending\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo test whether the ozone\u0026ndash;thermocline coupling is robust to event strength and long-term trends, we restrict the analysis to 8 strong ENSO events (4 El Ni\u0026ntilde;o: 1982\u0026ndash;83, 1986\u0026ndash;87, 1997\u0026ndash;98, 2015\u0026ndash;16; 4 La Ni\u0026ntilde;a: 1988\u0026ndash;89, 1998\u0026ndash;99, 2007\u0026ndash;08, 2010\u0026ndash;11) and apply linear detrending. Granger causality strengthens markedly in this subset: the p-value for ozone \u0026rarr; Z20 improves to \u0026lt; 0.001 (from p = 0.0008 in the full sample), and the impulse response function (IRF) peak amplifies from +192 cm to +295 cm (Table 1). This confirms that the signal arises from interannual ENSO dynamics and the direct radiative mechanism, rather than from multidecadal trends.\u003c/p\u003e\n\u003cp\u003eDuring these strong events, ozone anomalies exceed \u0026plusmn;12 DU, creating large perturbations to the UVB window. The amplified thermocline response (+295 cm) during extreme ozone depletion demonstrates the \u003cstrong\u003enonlinear sensitivity\u003c/strong\u003e of the radiative heating mechanism: larger ozone depletions allow proportionally more UVB to penetrate to thermocline depths, producing disproportionately large heating and deepening effects.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 1.\u003c/strong\u003e \u003cem\u003eGranger causality and impulse response function (IRF) peak magnitude for the causal link from stratospheric ozone to thermocline depth (Z20) across three data subsets. Results confirm that the ozone \u0026rarr; Z20 relationship is robust, strengthens during strong ENSO events, and is not an artefact of long-term trends. The IRF peak (in cm) quantifies the maximum response of the thermocline depth to a unit ozone shock. Further detrending the strong-event subset further sharpens the signal, yielding a highly significant p-value (\u0026lt; 0.001) and an amplified IRF peak of +295 cm.\u003c/em\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 17.3077%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eDataset\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20.1923%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSample Period\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 15.3846%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eN (months)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 27.8846%;\"\u003e\n \u003cp\u003e\u003cstrong\u003ep-value (Ozone \u0026rarr; Z20)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.2308%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eIRF Peak (cm)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 17.3077%;\"\u003e\n \u003cp\u003eFull period\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20.1923%;\"\u003e\n \u003cp\u003e1980\u0026ndash;2020\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 15.3846%;\"\u003e\n \u003cp\u003e~492\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 27.8846%;\"\u003e\n \u003cp\u003e0.0008\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.2308%;\"\u003e\n \u003cp\u003e+192\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 17.3077%;\"\u003e\n \u003cp\u003eStrong ENSO years only\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20.1923%;\"\u003e\n \u003cp\u003e8 events (Jul\u0026ndash;Jun)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 15.3846%;\"\u003e\n \u003cp\u003e~96\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 27.8846%;\"\u003e\n \u003cp\u003e\u0026lt; 0.005\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.2308%;\"\u003e\n \u003cp\u003e+265\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 17.3077%;\"\u003e\n \u003cp\u003eStrong ENSO + detrended\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 20.1923%;\"\u003e\n \u003cp\u003e8 events (Jul\u0026ndash;Jun)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 15.3846%;\"\u003e\n \u003cp\u003e~96\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 27.8846%;\"\u003e\n \u003cp\u003e\u0026lt; 0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.2308%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e+295\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2.\u0026nbsp;\u003c/strong\u003e\u003cem\u003eStatistical significance of monthly composite anomalies during the four strongest El Ni\u0026ntilde;o events (1982\u0026ndash;83, 1986\u0026ndash;87, 1997\u0026ndash;98, 2015\u0026ndash;16). For each month from July to June, mean anomalies and two-sided p-values are shown for (left) Z20 (cm) and (right) total column ozone (DU). All ozone anomalies from July to June are statistically significant at p \u0026lt; 0.05, while Z20 anomalies are significant throughout most months, with marginal significance (p \u0026lt; 0.05 but \u0026gt; 0.01) in April and May. Bold significance (\u0026ldquo;Yes\u0026rdquo;) is assigned for p \u0026le; 0.01; \u0026ldquo;p \u0026lt; 0.05\u0026rdquo; denotes marginal significance.\u003c/em\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 275px;\"\u003e\n \u003cp\u003eZ20 Anomalies (cm) \u0026ndash; El Ni\u0026ntilde;o Phase\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 288px;\"\u003e\n \u003cp\u003eOzone Anomalies (DU) \u0026ndash; El Ni\u0026ntilde;o Phase\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003e\u003cem\u003eMonth\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e\u003cem\u003eMean Anom\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e\u003cem\u003ep-value\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e\u003cem\u003eSignificant?\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u003cem\u003eMean Anom\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e\u003cem\u003ep-value\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003e\u003cem\u003eSignificant?\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003eJul\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e+445\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.008\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u0026ndash;9.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.009\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003eAug\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e+498\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.006\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u0026ndash;8.7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e0.032\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003e\u003cem\u003ep \u0026lt; 0.05\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003eSep\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e+612\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.002\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u0026ndash;10.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.004\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003eOct\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e+789\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u0026ndash;12.6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003eNov\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e+845\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u0026ndash;13.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003eDec\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e+756\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u0026ndash;14.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003eJan\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e+548\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.003\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u0026ndash;14.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003eFeb\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e+512\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.004\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u0026ndash;14.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003eMar\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e+421\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.009\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u0026ndash;12.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.002\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003eApr\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e+312\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e0.042\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e\u003cem\u003eMarginal\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u0026ndash;9.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e0.028\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003e\u003cem\u003ep \u0026lt; 0.05\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003eMay\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e+298\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e0.048\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e\u003cem\u003eMarginal\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u0026ndash;10.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e0.035\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003e\u003cem\u003ep \u0026lt; 0.05\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003eJun\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 101px;\"\u003e\n \u003cp\u003e+367\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 72px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.007\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 114px;\"\u003e\n \u003cp\u003e\u0026ndash;11.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.003\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eYes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e3.7 Impulse Response Quantification of the Direct Radiative Pathway\u003c/p\u003e\n\u003cp\u003eImpulse response functions (IRFs) from the 2-variable VAR model (lag = 6) quantify the dynamic temporal evolution of the ozone window mechanism, tracing how a one-time shock to one variable propagates through the system over subsequent months. Figure 6a shows Z20\u0026rsquo;s response to a one-standard-deviation ozone depletion (a negative ozone shock that opens the UVB window). The thermocline deepens gradually, as expected from cumulative radiative heating. It decreases by 92 cm in month 1 as UVB begins to warm the 15\u0026ndash;25 m layer. Deepening continues, peaking at +218 cm during months 4\u0026ndash;7 as accumulated heat overcomes stratification. The response remains statistically significant within the 95% confidence bands for 15 months. The total integrated deepening reaches +480 cm-months.\u003c/p\u003e\n\u003cp\u003eThis temporal profile, demonstrating a gradual acceleration, sustained plateau, and slow decay, is the signature of a cumulative heating mechanism: UVB energy absorbed at thermocline depth accumulates progressively over months, each month adding to the thermal anomaly until vertical mixing and lateral advection begin to dissipate the signal. The 4\u0026ndash;7 month lag to peak response represents the time required for \u0026sim;50\u0026ndash;80 MJ/m\u0026sup2; of cumulative UVB heating to sufficiently warm the thermocline layer (by \u0026sim;0.3\u0026ndash;0.5\u0026deg;C), thereby weakening vertical density stratification and deepening the thermocline.\u003c/p\u003e\n\u003cp\u003eFigure 6b shows the reverse relationship: the response of ozone to a one-standard-deviation positive shock in Z20 (thermocline deepening). The ozone response is fundamentally different in character: a slight transient decrease of \u0026ndash;1.8 DU peaking at months 2\u0026ndash;3, followed by rapid decay to insignificance by month 9. This weak, short-lived response represents the well-documented convective feedback: thermocline deepening warms SST, which enhances tropical convection, which transiently perturbs stratospheric ozone via enhanced upwelling of ozone-poor tropospheric air. However, this feedback operates only on 1\u0026ndash;3-month timescales and cannot account for the sustained 5\u0026ndash;10 month lead of ozone over Z20 observed in the cross-correlation analysis (Figure 4).\u003c/p\u003e\n\u003cp\u003eThe profound asymmetry between panels (a) and (b) is the key diagnostic: strong, sustained, cumulative Z20 response to ozone shocks versus weak, transient, rapidly decaying ozone response to Z20 shocks. This asymmetry definitively establishes that the ozone window \u0026rarr; direct UVB heating at 15\u0026ndash;25 m depth \u0026rarr; progressive thermocline warming \u0026rarr; Z20 deepening pathway is the dominant coupling mechanism in the South Pacific sensitive region, while the reverse convective feedback is a secondary, short-lived perturbation that does not drive the primary correlation structure.\u003c/p\u003e\n\u003cp\u003e3.8 Visual Confirmation: Scatterplot and Time Series\u003c/p\u003e\n\u003cp\u003eFigure 7 provides synoptic visual validation of the ozone window mechanism through complementary scatter and time-series perspectives spanning the entire 40-year observational record. Figure 7a presents a scatter plot of UVB penetration strength versus thermocline depth for all months (1980\u0026ndash;2020). The x-axis shows inverted ozone anomalies (\u0026ndash;ozone, in DU), which serve as a proxy for UVB window opening: positive values indicate ozone depletion and enhanced UVB reaching thermocline depth. In contrast, negative values indicate ozone enhancement and reduced UVB. The y-axis shows Z20 anomalies (in meters). Data points are colour-coded by ENSO phase: red for El Ni\u0026ntilde;o months (when ozone typically depletes), blue for La Ni\u0026ntilde;a months (when ozone typically increases), and purple for neutral conditions.\u003c/p\u003e\n\u003cp\u003eThe scatterplot reveals a robust positive linear relationship (r = 0.61, p \u0026lt; 0.001) with a slope of approximately 2.3 m of Z20 change per 1 DU change in ozone. This positive relationship confirms the direct mechanism: greater UVB penetration to thermocline depth (occurring when ozone is depleted and the window opens) \u0026rarr; greater cumulative heating at 15\u0026ndash;25 m \u0026rarr; more thermocline deepening. Critically, extreme events cluster tightly along the regression line rather than scattering randomly: the most significant El Ni\u0026ntilde;o events (e.g., 1997\u0026ndash;98 with \u0026ndash;15 DU ozone depletion) produce the largest Z20 deepening (+84 m), while the strongest La Ni\u0026ntilde;a events (e.g., 2010\u0026ndash;11 with +12 DU ozone increase) produce the largest Z20 shoaling (\u0026ndash;45 m). This tight clustering across a wide dynamic range confirms the mechanism operates robustly from weak to extreme conditions.\u003c/p\u003e\n\u003cp\u003eFigure 7b displays the complete 40-year monthly time series, allowing assessment of temporal precedence and phase relationships. The blue curve shows inverted ozone (\u0026ndash;ozone), representing the degree of UVB window opening (peaks indicate maximum window opening during ozone depletion). The purple curve shows Z20 depth anomalies. Vertical grey shading marks major El Ni\u0026ntilde;o events for reference.\u003c/p\u003e\n\u003cp\u003eClose inspection of the time series reveals that ozone depletion (blue peaks) consistently precedes thermocline deepening (purple peaks) by approximately 5\u0026ndash;10 months at every central ENSO turning point across four decades:\u003c/p\u003e\n\u003cul\u003e\n \u003cli\u003e1982\u0026ndash;83 El Ni\u0026ntilde;o: Ozone depletes (blue peak) in early 1982, Z20 deepens (purple peak) in late 1982\u003c/li\u003e\n \u003cli\u003e1997\u0026ndash;98 El Ni\u0026ntilde;o: Ozone depletes dramatically (most prominent blue peak in record) in early-mid 1997, Z20 reaches extreme depth (most prominent purple peak) in late 1997\u0026ndash;early 1998\u003c/li\u003e\n \u003cli\u003e2015\u0026ndash;16 El Ni\u0026ntilde;o: Ozone depletion in mid-2015, Z20 deepens in early 2016\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eSimilarly, during La Ni\u0026ntilde;a events, ozone increases (blue troughs) precede thermocline shoaling (purple troughs) by 5\u0026ndash;10 months. This consistent temporal precedence across multiple independent ENSO cycles\u0026mdash;spanning different climate background states, volcanic eruptions, and solar cycle phases\u0026mdash;demonstrates that the ozone window\u0026ndash;thermocline coupling is not an artefact of a few unusual events or specific background conditions, but rather represents a fundamental mode of stratosphere-ocean radiative interaction in the tropical Pacific.\u003c/p\u003e\n\u003cp\u003eThe sustained coherence over 40 years, combined with the tight scatterplot relationship in panel (a), provides compelling visual evidence that stratospheric ozone, acting as a radiative gatekeeper for UVB penetration, exerts systematic control over thermocline depth variability on interannual timescales through the direct heating mechanism identified in this study.\u003c/p\u003e"},{"header":"4. DISCUSSION","content":"\u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003e4.1 The Ozone Window Mechanism and Its Physical Basis\u003c/h2\u003e \u003cp\u003eOur analysis reveals a robust, physically coherent link between stratospheric ozone and tropical Pacific thermocline depth, mediated by a direct radiative pathway we term the \u0026ldquo;ozone window\u0026rdquo; mechanism. When total column ozone decreases by up to \u0026minus;\u0026thinsp;15 DU during strong El Ni\u0026ntilde;o events, enhanced UVB radiation penetrates the atmosphere and reaches ocean depths of 15\u0026ndash;25 m, precisely where the 20\u0026deg;C isotherm (Z20) resides in the clear waters of the South Pacific (9\u0026ndash;12\u0026deg;S, 125\u0026ndash;115\u0026deg;W). Unlike visible or infrared radiation, UVB is absorbed within this layer, directly warming the thermocline. This process unfolds gradually: initial heating begins within the first month of ozone depletion, but the complete thermocline response\u0026mdash;peaking at a depth of +\u0026thinsp;218 cm\u0026mdash;requires 4\u0026ndash;7 months to develop, with statistical significance persisting for 15 months and an integrated response of +\u0026thinsp;480 cm-months. The 5\u0026ndash;10 month lag between peak ozone depletion (typically January\u0026ndash;February) and peak Z20 deepening (December\u0026ndash;January during the mature El Ni\u0026ntilde;o phase) is consistent with the time required for cumulative UVB energy (~\u0026thinsp;50\u0026ndash;80 MJ/m\u0026sup2;) to overcome ocean thermal inertia and weaken vertical stratification. The reverse occurs during La Ni\u0026ntilde;a: ozone enhancement (+\u0026thinsp;12 to +\u0026thinsp;15 DU) closes the UVB window, reduces deep heating, and shoals the thermocline by ~\u0026ndash;25 m. Critically, this mechanism operates as a reinforcing feedback mechanism within the ENSO system. The wind-driven initial changes in the thermocline modulate convection, which perturbs ozone and, in turn, amplifies the original oceanic signal via direct radiative heating. The effect is nonlinear and strongest during extreme events, with scatterplot analysis showing a tight linear relationship (r\u0026thinsp;=\u0026thinsp;0.61) between inverted ozone anomalies and Z20 across four decades, and the most significant ENSO events aligning precisely with the regression slope of ~\u0026thinsp;2.3 m per DU.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec23\" class=\"Section2\"\u003e \u003ch2\u003e4.2 Causal Directionality and Statistical Robustness\u003c/h2\u003e \u003cp\u003eThe ozone\u0026ndash;thermocline relationship is not merely correlative but exhibits clear predictive causality in one direction. Granger causality tests confirm that past ozone anomalies significantly improve forecasts of future Z20 across lags 1\u0026ndash;7 months (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01), with peak statistical strength at lag 1 (F\u0026thinsp;=\u0026thinsp;6.8, p\u0026thinsp;=\u0026thinsp;0.0001). In stark contrast, the reverse\u0026mdash;Z20 predicting ozone is only marginally significant at lag 1 and negligible thereafter, reflecting a transient convective feedback that dissipates within months. Impulse-response functions further underscore this asymmetry: an ozone-depletion shock triggers a significant, sustained deepening of Z20 (+\u0026thinsp;218 cm), whereas a Z20 shock produces only a weak, short-lived ozone decrease (\u0026ndash;1.8 DU). This directional imbalance persists and even strengthens when analysis is restricted to the eight strongest ENSO events: Granger p-values improve to \u0026lt;\u0026thinsp;0.001, and the IRF peak amplifies to +\u0026thinsp;295 cm after detrending, confirming that the signal originates from interannual dynamics rather than long-term trends. Monthly composites during major El Ni\u0026ntilde;o events show statistically significant ozone and Z20 anomalies across nearly all months (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e), with ozone consistently leading Z20 by 5\u0026ndash;10 months in every major cycle since 1980. This multi-decadal coherence rules out spuriousness and affirms the mechanism\u0026rsquo;s robustness across varying background states.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003e4.3 Broader Implications for Climate Science\u003c/h2\u003e \u003cp\u003eThese findings carry important consequences for climate modelling, prediction, and environmental policy. Most current Earth system models, including those in CMIP5/6, lack the necessary components to represent this mechanism; they either prescribe a fixed ozone field or omit spectral ocean optics, thereby assuming that all solar radiation is absorbed near the surface. As a result, they cannot simulate UVB\u0026rsquo;s unique 15\u0026ndash;25 m penetration or the resulting thermocline feedback, potentially underestimating ENSO amplitude and misattributing variance. Incorporating interactive stratospheric chemistry and spectrally resolved ocean radiation schemes into next-generation models is essential. From a forecasting standpoint, the 5\u0026ndash;10-month lead of ozone over Z20 offers a tangible opportunity to improve ENSO predictions, particularly during the \u0026ldquo;spring predictability barrier,\u0026rdquo; when skill typically declines. Since stratospheric ozone is monitored in near-real time by multiple satellites (e.g., OMI, OMPS), it could serve as a physically grounded predictor in operational systems. Finally, the success of the Montreal Protocol in restoring stratospheric ozone may have exerted an unintended cooling influence on the tropical Pacific by progressively closing the UVB window over recent decades, a potential contributor to the observed increase in La Ni\u0026ntilde;a-like conditions. While rigorous attribution is needed, this represents a plausible climate co-benefit of ozone protection policy.\u003c/p\u003e \u003c/div\u003e"},{"header":"5. CONCLUSIONS","content":"\u003cp\u003eThis study demonstrates that stratospheric ozone variability actively influences tropical Pacific thermocline depth through a direct radiative pathway: ozone depletion opens a \u0026quot;window\u0026quot; that allows enhanced ultraviolet-B (UVB) radiation to penetrate to the upper ocean (15-25 m depth), directly heating the thermocline layer and causing it to deepen over subsequent months. In the South Pacific (9-12\u0026deg;S, 130-110\u0026deg;W), stratospheric ozone and the depth to the 20\u0026deg;C isotherm (Z20) are strongly anticorrelated (r = -0.61), with ozone anomalies leading Z20 by 5-10 months and explaining 37% of its variance.\u003c/p\u003e\n\u003cp\u003eThe physical mechanism operates through ozone as a radiative gatekeeper. When stratospheric ozone is depleted during El Ni\u0026ntilde;o (by up to -15 DU), less UVB is absorbed in the lower stratosphere, allowing enhanced UVB flux to penetrate the atmosphere and reach the ocean. Unlike visible light (absorbed in the top few meters), UVB penetrates to 15-25 m depth\u0026mdash;precisely where the thermocline resides in the tropical Pacific. Cumulative UVB absorption at thermocline depth over 5-10 months warms this layer by +0.3 to +0.5\u0026deg;C, reduces vertical stratification, and causes the thermocline to deepen by an average of +32 m (up to +84 m during extreme events). Ozone increases during La Ni\u0026ntilde;a close this window, reducing thermocline heating and promoting shoaling.\u003c/p\u003e\n\u003cp\u003eGranger causality tests confirm unidirectional forcing from ozone to Z20 (peak F = 6.8 at lag 1, F = 5.1 at lag 3; p \u0026lt; 0.001), with no sustained reverse causality (Z20 \u0026rarr; ozone significant only at lag 1, p = 0.045). This asymmetry is reinforced by impulse response functions: an ozone depletion shock produces sustained thermocline deepening (+218 cm peak over months 4-7, significant for 15 months) consistent with cumulative radiative heating, while a thermocline deepening shock produces only a weak, transient ozone response (-1.8 DU, significant \u0026lt;3 months) representing the well-known but secondary convective feedback.\u003c/p\u003e\n\u003cp\u003eThe signal is strongest during extreme ENSO events, when ozone anomalies exceed \u0026plusmn;12 DU and IRF peaks reach +295 cm (p \u0026lt; 0.001 for strongly detrended events), and remains robust after trend removal, confirming its origin in interannual dynamics rather than in multidecadal ozone recovery. The extended 5-10 month lag\u0026mdash;longer than typical wind-driven ocean responses\u0026mdash;is physically consistent with the cumulative nature of radiative heating: UVB energy absorbed at thermocline depth accumulates gradually over months, requiring sustained forcing to overcome ocean thermal inertia and alter density stratification.\u003c/p\u003e\n\u003cp\u003eThis discovery reveals stratospheric ozone as a regional positive feedback mechanism in ENSO, operating primarily in the southern tropical Pacific, where clear waters allow maximum UVB penetration. While wind-driven processes remain dominant (explaining ~60-70% of thermocline variance), the ozone window contributes an estimated 10-20% of thermocline anomalies during strong events. This feedback helps explain ENSO\u0026apos;s asymmetry (larger El Ni\u0026ntilde;o ozone depletions create stronger radiative amplification than La Ni\u0026ntilde;a enhancements) and event diversity (spatial variations in ozone depletion patterns modulate where thermocline reinforcement occurs).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eImplications for prediction and modelling:\u003c/em\u003e The 5-10 month lead of ozone over Z20 creates potential to extend ENSO forecast skill by incorporating satellite-observed stratospheric ozone as a predictor, particularly valuable during the spring predictability barrier. However, most CMIP5/6 models cannot capture this mechanism because their ocean components lack spectral resolution for UVB absorption at depth. Next-generation Earth System Models require both interactive stratospheric chemistry and spectral ocean optics that explicitly resolve UVB penetration to 15-25 m depth.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eClimate policy implications:\u003c/em\u003e The Montreal Protocol\u0026apos;s success in restoring stratospheric ozone has progressively closed the UVB window since 2000, potentially contributing 5-15% of the observed shift toward La Ni\u0026ntilde;a-like conditions in recent decades\u0026mdash;an unintended but beneficial climate co-benefit of ozone protection. This underscores how stratospheric composition changes can cascade through the climate system via previously unrecognised radiative pathways.\u003c/p\u003e\n\u003cp\u003eRecognising this stratosphere-to-ocean radiative teleconnection, which operates through direct UVB heating at thermocline depths rather than through indirect changes in atmospheric circulation, fundamentally revises our understanding of tropical Pacific variability. It establishes stratospheric ozone not as a passive tracer of ENSO convection but as an active radiative modulator of upper-ocean thermal structure. This cross-domain coupling must be represented in next-generation climate models and operational forecast systems to accurately predict ENSO evolution and assess the full climate impacts of stratospheric composition changes.\u003c/p\u003e"},{"header":"Statements \u0026 Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFunding from Japan Agency for Marine-Earth Science and Technology (JAMSTEC) to enable research and publication of this work is greatly appreciated. Bindura University of Science Education is thanked for providing the first author with facilities to conduct the research.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFunding\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eJAMSTEC supported this work through S.K. Behera, who received research support for the publication of the work\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eCompeting Interests\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eThe authors have no relevant financial or non-financial interests to disclose.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eAuthor Contributions\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAll authors contributed to the study conception and design. D. Manatsa, T. Mushore and S.K. Behera performed material preparation, data collection and analysis. D. Manatsa wrote the first draft of the manuscript, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eData Availability\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAll datasets used in this study are publicly available. Monthly mean stratospheric ozone data were obtained from the Multi-Sensor Reanalysis version 2 (MSR-2), produced under the European Space Agency Climate Change Initiative (ESA-CCI) Ozone project (van der A et al., 2015). These data were accessed and downloaded via the KNMI Climate Explorer (https://climexp.knmi.nl/) for the period 1980\u0026ndash;2023. Monthly mean air temperature data at 100 hPa were obtained from the NCEP/NCAR Reanalysis I (Kalnay et al., 1996), which is available through the NOAA Climate and Plotting page (https://psl.noaa.gov/cgi-bin/data/).\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eMonthly mean depth of the 20 \u0026deg;C isotherm (Z20) was obtained from the POAMA/PEODAS ocean reanalysis produced by the Australian Bureau of Meteorology and made available through its THREDDS/OpenDAP server (\u003c/em\u003e\u003cem\u003ehttp://opendap.bom.gov.au:8080/thredds/dodsC/poama/peodas/reanalysis/\u003c/em\u003e\u003cem\u003e). Pre-processed Z20 fields derived from this reanalysis were accessed via the KNMI Climate Explorer.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eHowever, the data for the specific regions and time periods analysed in this work are available at https://github.com/dmanatsa-cloud/data_used_ozone-z20_manuscript. All data processing and analysis were performed using standard statistical methods, and no proprietary datasets were used.\u0026nbsp;\u003c/em\u003e\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eCai W, Santoso A, Collins M et al (2021). Changing El Ni\u0026ntilde;o\u0026ndash;Southern Oscillation in a warming climate. Nat Rev Earth Environ 2:628\u0026ndash;644. https://doi.org/10.1038/s43017-021-00199-1\u003c/li\u003e\n \u003cli\u003eMcPhaden MJ, Zebiak SE, Glantz MH (1998) ENSO as an integrating concept in Earth science. Science 314:1740\u0026ndash;1745. https://doi.org/10.1126/science.1132588\u003c/li\u003e\n \u003cli\u003eCapotondi A, Wittenberg AT, Newman M et al (2015) Understanding ENSO diversity. Bull Am Meteorol Soc 96:921\u0026ndash;938. https://doi.org/10.1175/BAMS-D-13-00117.1\u003c/li\u003e\n \u003cli\u003eAlbers JR, Butler AH, Langford AO, Elsbury D, Breeden ML (2022) Dynamics of ENSO-driven stratosphere-to-troposphere transport of ozone over North America. Atmos Chem Phys 22:13035\u0026ndash;13048. https://doi.org/10.5194/acp-22-13035-2022\u003c/li\u003e\n \u003cli\u003eTedetti M, Semp\u0026eacute;r\u0026eacute; R (2006) Penetration of ultraviolet radiation in the marine environment: A review. Photochem Photobiol 82:389\u0026ndash;397. https://doi.org/10.1562/2005-11-09-IR-733\u003c/li\u003e\n \u003cli\u003eRochelle-Newall EJ, Fisher TR (2002) Production of chromophoric dissolved organic matter fluorescence in marine and estuarine environments: an investigation into the role of phytoplankton. Mar Chem 77:7\u0026ndash;21. https://doi.org/10.1016/S0304-4203(01)00078-8\u003c/li\u003e\n \u003cli\u003eManatsa D, Mukwada G (2017) A connection from stratospheric ozone to El Ni\u0026ntilde;o\u0026ndash;Southern Oscillation. Sci Rep 7:6415. https://doi.org/10.1038/s41598-017-06776-7\u003c/li\u003e\n \u003cli\u003eChiodo G, Austin J, Polvani LM, Marsh DR, Garcia RR (2018) The impact of interactive stratospheric chemistry on surface climate. J Clim 31:3905\u0026ndash;3920. https://doi.org/10.1175/JCLI-D-17-0501.1\u003c/li\u003e\n \u003cli\u003evan der A RJ, Allaart MAF, Eskes HJ (2015) Extended and refined multi sensor reanalysis of total ozone for the period 1970\u0026ndash;2012. Atmos Meas Tech 8:3021\u0026ndash;3035. https://doi.org/10.5194/amt-8-3021-2015\u003c/li\u003e\n \u003cli\u003eKalnay E, Kanamitsu M, Kistler R et al (1996) The NCEP/NCAR 40-year reanalysis project. Bull Am Meteorol Soc 77:437\u0026ndash;471. https://doi.org/10.1175/1520-0477(1996)077\u0026lt;0437:TNYRP\u0026gt;2.0.CO;2\u003c/li\u003e\n \u003cli\u003eBretherton CS, Widmann M, Dymnikov VP, Wallace JM, Blad\u0026eacute; I (1999) The effective number of spatial degrees of freedom of a time-varying field. J Clim 12:1990\u0026ndash;2009. https://doi.org/10.1175/1520-0442(1999)012\u0026lt;1990:TENOSD\u0026gt;2.0.CO;2\u003c/li\u003e\n \u003cli\u003eNorth GR, Bell TL, Cahalan RF, Moeng FJ (1982) Sampling errors in the estimation of empirical orthogonal functions. Mon Weather Rev 110:699\u0026ndash;706. https://doi.org/10.1175/1520-0493(1982)110\u0026lt;0699:SEITEO\u0026gt;2.0.CO;2\u003c/li\u003e\n \u003cli\u003eGranger CWJ (1969) Investigating causal relations by econometric models and cross-spectral methods. Econometrica 37:424\u0026ndash;438. https://doi.org/10.2307/1912791\u003c/li\u003e\n \u003cli\u003eBall WT, Haigh JD, Rozanov EV et al (2016) High solar cycle spectral variations inconsistent with stratospheric ozone observations. Nat Geosci 9:206\u0026ndash;209. https://doi.org/10.1038/ngeo2640\u003c/li\u003e\n \u003cli\u003eButchart N (2022) The stratosphere: a review of the dynamics and variability. Weather Clim Dynam 3:1237\u0026ndash;1272. https://doi.org/10.5194/wcd-3-1237-2022\u003c/li\u003e\n \u003cli\u003eElsbury D, Butler AH, Albers JR, Breeden ML, Langford AO (2023) The response of the North Pacific jet and stratosphere-to-troposphere transport of ozone over western North America to RCP8.5 climate forcing. Atmos Chem Phys 23:5101\u0026ndash;5117. https://doi.org/10.5194/acp-23-5101-2023\u003c/li\u003e\n \u003cli\u003eDee DP, Uppala SM, Simmons AJ et al (2011) The ERA-Interim reanalysis: configuration and performance of the data assimilation system. Q J R Meteorol Soc 137:553\u0026ndash;597. https://doi.org/10.1002/qj.828\u003c/li\u003e\n \u003cli\u003eDong Y, Polvani LM, Hwang Y-T, England MR (2025) Stratospheric ozone depletion has contributed to the recent tropical La Ni\u0026ntilde;a-like cooling pattern. npj Clim Atmos Sci 8:150. https://doi.org/10.1038/s41612-025-01020-0\u003c/li\u003e\n \u003cli\u003eJ\u0026oacute;zefiak I, Sukhodolov T, Egorova T et al (2023) Stratospheric dynamics modulates ozone layer response to molecular oxygen variations. Front Earth Sci 11:1239325. https://doi.org/10.3389/feart.2023.1239325\u003c/li\u003e\n \u003cli\u003eKarpechko AY, Vitart F, Statnaia I, Balmaseda MA, Charlton-Perez AJ (2024) The tropical influence on sub-seasonal predictability of wintertime stratosphere and stratosphere\u0026ndash;troposphere coupling. Q J R Meteorol Soc 150:1125\u0026ndash;1142. https://doi.org/10.1002/qj.4678\u003c/li\u003e\n \u003cli\u003eNowack PJ, Braesicke P, Abraham NL, Pyle JA (2017) On the role of ozone feedback in the ENSO amplitude response under global warming. Geophys Res Lett 44:3858\u0026ndash;3866. https://doi.org/10.1002/2016GL072418\u003c/li\u003e\n \u003cli\u003ePolvani LM, Waugh DW, Chiodo G, Hegglin MI, Matthes K (2020) Stratospheric ozone depletion and Southern Ocean surface wind trends. Geophys Res Lett 47:e2020GL087369. https://doi.org/10.1029/2020GL087369\u003c/li\u003e\n \u003cli\u003eRandel WJ, Wu F (2021) A simple model of ozone\u0026ndash;temperature coupling in the tropical lower stratosphere. Atmos Chem Phys 21:18531\u0026ndash;18542. https://doi.org/10.5194/acp-21-18531-2021\u003c/li\u003e\n \u003cli\u003eSon S-W, Kim Y, Lu J, Yoo C (2024) Stratospheric influence on tropical Pacific decadal variability. Nat Clim Change 14:321\u0026ndash;328. https://doi.org/10.1038/s41558-024-01923-8\u003c/li\u003e\n \u003cli\u003eTian W, Huang J, Zhang J et al (2023) Role of stratospheric processes in climate change: advances and challenges. Adv Atmos Sci 40:1379\u0026ndash;1400. https://doi.org/10.1007/s00376-022-2246-8\u003c/li\u003e\n \u003cli\u003eYoung PJ, Naik V, Fiore AM et al (2018) Tropospheric Ozone Assessment Report: Assessment of global-scale model performance for global and regional ozone distributions, variability, and trends. Elementa: Sci Anthropocene 6:10. https://doi.org/10.1525/elementa.265\u003c/li\u003e\n \u003cli\u003eZuo H, Balmaseda MA, Tietsche S et al (2019) The ECMWF operational ocean analysis system: ORAS5. ECMWF Tech Memo 851. https://doi.org/10.21957/1r6x0z0j\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Stratospheric ozone, Thermocline variability, UVB radiative heating, ENSO feedback, Granger causality, Ozone window mechanism","lastPublishedDoi":"10.21203/rs.3.rs-8425241/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8425241/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eStratospheric ozone is a key modulator of Earth's radiative balance, yet its role in driving tropical ocean variability remains uncertain. Here, using 40 years (1980\u0026ndash;2020) of satellite and reanalysis data, we demonstrate that stratospheric ozone anomalies precede and predict changes in thermocline depth in the tropical Pacific through a direct radiative pathway. In this case, ozone depletion opens a \"window\" that allows enhanced ultraviolet-B (UVB) radiation to penetrate to the upper ocean (15\u0026ndash;25 m depth), directly heating the thermocline layer. In the South Pacific (9\u0026ndash;12\u0026deg;S, 130\u0026ndash;110\u0026deg;W), stratospheric ozone and the depth to the 20\u0026deg;C isotherm (Z20) are strongly anticorrelated (r = \u0026minus;\u0026thinsp;0.61), with ozone leading Z20 by 5\u0026ndash;10 months and accounting for 37% of its variance. Granger causality tests confirm unidirectional forcing from ozone to Z20 (peak F\u0026thinsp;=\u0026thinsp;5.1 at 3-month lag; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), with no significant reverse causality. This establishes ozone as an active driver rather than a passive response. Impulse response functions quantify a robust pathway in which a one-standard-deviation ozone-depletion shock deepens the thermocline by 218 cm within 4\u0026ndash;7 months. A process that is driven by enhanced UVB absorption at the thermocline depth. The signal is strongest during extreme ENSO events and persists after detrending, confirming its origin in interannual dynamics. This stratosphere-to-ocean radiative teleconnection provides a physically grounded predictor with potential to extend ENSO forecast skill by 5\u0026ndash;10 months.\u003c/p\u003e","manuscriptTitle":"Stratospheric Ozone Depletion Drives Tropical Pacific Thermocline Variability via Enhanced UVB Penetration to the Upper Ocean","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-01-05 08:54:14","doi":"10.21203/rs.3.rs-8425241/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"e37a8e21-a388-4557-957d-6411808e1eee","owner":[],"postedDate":"January 5th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-02-02T11:26:59+00:00","versionOfRecord":[],"versionCreatedAt":"2026-01-05 08:54:14","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8425241","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8425241","identity":"rs-8425241","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}