The meridional overturning circulation (MOC), crucial for global heat, freshwater, and carbon transport, is highly sensitive to deep-ocean mixing. Mixing is geographically inhomogeneous, often enhanced over rough topography due to breaking internal tides. The global energy conversion from barotropic to baroclinic tides is substantial (~1 TW), a significant portion of the power required for maintaining the MOC. Internal tides exhibit vertical structures described by eigenmodes related to ocean stratification. Low-mode internal tides (e.g., modes 1–3) radiate long distances, contributing to background mixing, while high-mode internal tides (≥4) dissipate locally, creating mixing hotspots. Current coupled global climate models (CGCMs) typically assume constant tidal energy conversion and modal partition. However, global warming intensifies stratification, particularly in the upper ocean, leading to a prevailing, yet potentially incomplete, view that this enhanced stratification weakens ocean mixing. Regional studies have shown conflicting results on the effects of enhanced stratification on tidal energy conversion. A comprehensive understanding of the global-scale response of tidal energy conversion and its modal partition to global warming is critical for accurate climate change projections, given the role of tidal mixing in the MOC.

Several studies have explored the connection between internal tides, ocean mixing, and the MOC. Some research focuses on the spatial variability of turbulent mixing in the abyssal ocean, highlighting the importance of topographic features. Others have investigated the role of internal tides in mixing the deep ocean and the geographical distribution of mixing hotspots. The influence of stratification on tidal energy conversion and its modal partition has also been investigated, with regional studies showing contrasting results depending on the specific location and characteristics of the topography. However, a comprehensive, global-scale assessment of these effects under global warming remains lacking, which this study aims to address.

This study uses a linear model of internal tide generation applied to simulations from an ensemble of 25 CMIP6 CGCMs under a high carbon emission scenario (SSP585). The model estimates tidal energy conversion and its modal partition based on near-bottom (Nb) and depth-averaged (N) buoyancy frequencies. The dependence of the nth mode energy conversion (En) on Nb is nearly linear when Nb is much larger than tidal frequencies. The effect of N on En is more complex, depending on the spectral density of tidal energy conversion in horizontal wavenumber space. A larger N reduces the modal horizontal wavenumber (Kn), impacting En differently depending on the slope of the spectral density. The climatological mean (1995–2004) Nb and N from observations (World Ocean Atlas 2018) and CMIP6 CGCMs are compared to assess model fidelity. Projected changes (2091–2100 minus 1995–2004) in Nb and N are analyzed. Tidal energy conversion (E1-50) and its partition into low (E1-3) and high (E4-50) modes are estimated using St. Laurent and Garrett’s formulation based on Bell’s linear theory. Calculations are limited to the first 50 modes due to bathymetric data resolution. The effects of Nb and N changes on tidal energy conversion are separately analyzed by fixing one while varying the other. The M2 internal tide is used as a representative example to demonstrate the effects of Kn changes on energy conversion. The spectral density of M2 tidal energy conversion in horizontal wavenumber space is analyzed for regions with enhanced and reduced E1-3 under global warming. Finally, a tidal mixing parameterization for high-mode internal tides is used to explore the effects on deep-ocean mixing (diapycnal diffusivity, κ).

Analysis of CMIP6 CGCM simulations reveals a spatially heterogeneous response of low-mode tidal energy conversion to global warming, resulting in a negligible global increase (0.9 ± 0.2%). Conversely, high-mode tidal energy conversion exhibits a universal increase, projected to rise by 7.8 ± 0.7% by the end of the century. This increase in high-mode energy is robust across individual CGCM members and different time periods and emission scenarios. Both Nb and N changes influence tidal energy conversion. Increased Nb increases both low and high-mode energy, while increased N increases high-mode but reduces low-mode energy. The reduction of Kn due to increased N consistently increases high-mode energy, but its effect on low-mode energy is regionally dependent, influenced by the shape of the topographic spectrum. Parameterization of deep-ocean diapycnal diffusivity (κ) shows that mixing hotspots over rough topography become “hotter” under global warming, with a global average increase of 6% and local increases exceeding 20%. Preliminary analysis suggests a potential 10% acceleration of the Atlantic MOC's lower limb and a 10% increase in the peak of globally integrated diautral upwelling due to enhanced high-mode internal tide energy.

The findings demonstrate that global warming significantly enhances high-mode internal tide energy conversion, despite near-unchanged low-mode conversion. This is primarily driven by the interplay between intensified near-bottom and depth-averaged stratification, with the latter's effect on the modal horizontal wavenumber being crucial. The enhanced high-mode energy leads to stronger deep-ocean mixing over rough topography. While enhanced stratification can suppress mixing, the increased energy conversion into high-mode tides dominates, resulting in 'hotter' mixing hotspots. The potential implications for the MOC and diautral upwelling are significant, although further research is needed to refine the parameterizations and reduce uncertainties. The results highlight the need to integrate stratification changes into tidal mixing parameterizations in CGCMs for more accurate climate change projections.

This study provides strong evidence that global warming will intensify the generation of high-mode internal tides, leading to enhanced deep-ocean mixing over rough topography. The nearly unchanged low-mode tidal energy conversion contrasts sharply with the substantial increase in high-mode energy, which has significant implications for the MOC and diautral upwelling. This underscores the necessity for incorporating the dynamic interaction between stratification and tidal mixing into future climate models to improve their accuracy and reliability.

The study is limited by the resolution of the bathymetric data, restricting the analysis to the first 50 vertical modes of internal tides. The tidal mixing parameterization used also introduces uncertainties, particularly concerning the mixing efficiency. Future work should aim to address these limitations through improved data resolution and refined parameterization schemes. The preliminary analyses on the MOC and diautral upwelling need further validation and investigation.