| 30 Jan 2026
Tropical Instability Vortices reduce Pacific Ocean ENSO-Driven CO2 outgassing
Abstract. The relationship between the intensity of Pacific Ocean Tropical Instability Vortices (TIVs), ENSO variability, and dissolved inorganic carbon (DIC) remains poorly constrained. Here, we use a 30-year-long eddy-resolving ocean biogeochemistry simulation to quantify the effects of TIVs on DIC budget components at both synoptic and interannual timescales. At synoptic scales, TIVs primarily influence DIC through advection, especially along the leading edge of the wave fronts, while vertical diffusion and biological processes play secondary roles. To investigate interannual variability, we develop a TIV index to classify strong and weak TIV phases within each ENSO state. In the upper 50 meters, TIV-driven advection shapes large-scale DIC transport pathways while enhancing, yet spatially confining, primary production. Consequently, during El Niño, TIVs tend to amplify oceanic CO2 uptake, associated with a 57 % decrease in CO2 partial pressure (pCO2). During La Niña, they suppress CO2 outgassing, even reversing the ocean's role from a source to a sink. TIVs also affect the upper thermocline carbon inventory by modulating both biological activity and lateral transport. Strong TIVs during El Niño reduce DIC inventories in the upper thermocline by 8.5 GtC due to increased vertical mixing and enhanced transport, while during La Niña, strong TIVs lead to a 77 % higher DIC accumulation compared to weak TIVs. These findings underscore the critical role of TIVs in regulating the equatorial Pacific carbon budget and highlight the need to accurately represent them in Earth system models.
Review of “Tropical instability vortices reduce Pacific Ocean ENSO-driven CO2 outgassing” by Casaroli et al.
This study investigates the DIC variability in the eastern equatorial Pacific as a function of the activity of tropical instability vortices and ENSO phase, using simulation results from a state-of-the-art ocean physical-biogeochemical model. Their approach which extracts intrinsic TIV variability independent of ENSO and examines composites of strong and weak TIV regimes separately for El Niño and La Niña phases, is unique. Their findings, showing that the contributions from TIV regimes differ substantially between ENSO phases, are important and appropriate for publication in this journal. I have one major concern, as described below, which I hope can be addressed before publication.
To my understanding, the mean states differ among the composites: upper-layer stratification and thermocline depth as well as EUC depth vary between ENSO phases; differences in TIV activity imply that the SEC, NECC, and/or other currents as well as meridional density gradients, associated with the TIV generation, are remarkable. These differences in the mean states are not explicitly described in the present manuscript but could affect their analysis, particularly the estimates of the advection term. Thus, I suggest a more careful interpretation of the advection term. Note that the drivers of TIVs would be the mean states; hence, descriptions such as “TIV-driven” might be misleading. In addition, TIVs sometimes enhance northward transport, sometimes enhance southward transport, and sometimes trap tracers locally. The physical interpretation should account for both the mean state differences and the TIV activity differences. I suggest revising the descriptions regarding these points.
Specific comments
L27-28: Most of these references investigated tropical instability waves (TIWs). I suggest separating the references between TIWs and TIVs and adding a description of the relationship between these phenomena. For this purpose, the description in Zheng et al. (2016), already cited, and a recent analysis by Toyoda et al. (2023, doi:10.1038/s41598-023-41159-5) may be useful. In addition, many other studies on TIWs exist, and the references in the manuscript do not fully cover the literature. Thus, I suggest using “e.g.,” before the references, here and in L25. Please also check similar reference descriptions (e.g., L150-151 and L181).
L44: CO2 -> CO2
Section 2.2: To my understanding, the authors aim to analyze the variabilities of SST anomaly and TIV index on a similar (interannual) time scale, although the treatment of these variables differs. I suggest adding an explanation of the influence of this difference.
L97: I consider that “minimum” might more appropriately be “mean”.
L104: I understand that Figure 2 (upper panel) shows the “intrinsic TIV variability independent of ENSO”. If so, I think that the figure caption and title need to be improved.
L126: What is the baseline of the sea surface temperature anomaly? Is seasonality included?
L159: Although the distribution of mixed layer depth is partly shown in Fig. 6, a brief explanation of the meridional distribution of mixed layer depth and its influence on the analysis would help in understanding the subsequent interpretation of the results.
Figure 8c, d: Please add an explanation for why the mean El Niño lines are not located between the strong and weak TIV composite lines.
L231: “(( ))” -> “( )”
Figure 10: Using latitude on the y-axis, as in other figures, would improve readability.
L243: What is the intrinsic dynamical difference between strong and weak TIVs? Why does one enhance the mean advection term while another traps carbon locally?
L247 “strong TIVs disrupt near-equatorial northward flow”: Please add an explanation of how the flow field changes and why.
L258 “strong TIVs cause shoaling of the equatorial thermocline during El Niño”: How can this causal relationship be understood dynamically?
L308-309: Why does “the remineralization of POM is also shifted south of this point” suggest “a larger production of particulate organic matter in strong TIVs north of 3°N”? Please add an explanation.
L323 “silicate is advected zonally”: The difference in the mean current as well as TIV intensity can affect this.
L328 “south of the equator”, L332 “across the equator”, L341 “cross-equatorial exchange”: To my understanding, the authors do not explicitly conduct analyses supporting these statements.
Figure 16: I wonder how the integrated DIC concentration (or anomaly) over this vertical-meridional cross section differs between weak and strong TIV composites. The analyzed terms are conservative when integrated over the domain, except for the SoSi term. The discussion of dominant terms might be meaningful when the overall (integrated) DIC concentration does not differ significantly between the TIV regimes.