Effects of paleogeographic changes and CO₂ variability on northern mid-latitudinal temperature gradients in the Cretaceous

Proxy records (δ18O in foraminifera and TEX86) indicate warm Cretaceous-to-Eocene surface oceans with relatively low pole-to-tropics temperature gradients, attributed largely to elevated atmospheric CO₂. Beyond CO₂, tectonically driven paleogeographic changes (topography, bathymetry, continental arrangement) can modulate SSTs by altering ocean circulation, deep-water formation sites, and seasonal contrasts. Previous work links Cretaceous paleogeography to damped seasonality, altered overturning/ventilation, and potential glacial events, as well as Paleogene changes such as Antarctic orography and gateway openings. Nonetheless, the separate roles and mechanisms by which CO₂ and paleogeography controlled regional SST gradients—especially northern mid-latitudes—remain uncertain. This study asks: how much of the observed increase in northern mid-latitudinal SST gradient from Early to Late Cretaceous is explained by CO₂ versus paleogeographically driven changes in wind-driven surface gyres? The authors use ensembles of paleoclimate simulations and a Stommel-like gyre framework to isolate the effect of gyral circulation changes on the evolution of the mid-latitudinal SST gradient.

- Studies have examined paleogeography impacts on pole-to-tropics gradients and ocean circulation during the Cretaceous–Paleogene (e.g., Poulsen et al., Donnadieu et al., Ladant et al.), but not specifically on surface gyres bounded by meridional continental margins. - Stommel (1948) introduced a simple gyre model where volumetric transport ψ scales linearly with wind stress amplitude; dependence on basin dimensions (Lx, Ly) is more complex. Real gyres are bounded by zero wind-stress curl latitudes rather than coastlines alone. - Prior theory and idealized studies suggest ψ and poleward heat transport H increase with gyral basin aspect ratio (Lx/Ly). Smaller aspect ratios can yield smaller H and larger meridional SST gradients. - Proxy compilations show nonmonotonic changes in meridional SST gradients through the Cretaceous with minima in the Aptian and maxima in the Maastrichtian. However, proxy coverage is sparse in Northern Hemisphere mid-latitudes, limiting direct constraints on ΔSSTy there. - CO₂ reconstructions for the CPE show long-term trends with uncertainties; CO₂ is known to drive polar amplification and thereby influence meridional SST gradients.

- Climate model ensembles: Two ensembles using the UK Met Office coupled model HadCM3L (HadCM3LB-M2.1aD), 3.75°×2.5° resolution, 19 atmospheric and 20 ocean vertical levels. CO₂ fixed at 560 ppmv (2×PI) and 1120 ppmv (4×PI) per ensemble; paleogeography varied across 19 geologic ages from Berriasian (~143 Ma) to Priabonian (~36 Ma). Solar constant increased monotonically by ~0.9% across ages; its forcing is small relative to CO₂ and paleogeography. - Paleogeographies: Reconstructions from geological constraints (lithology, tectonics, fossils, deep-sea data) at 0.5° resolution, mapped to model grid with smoothing for stability (Getech/BRIDGE datasets). - Diagnostic focus region: Mid-latitudinal North Pacific gyral basin (approx. rectangular, Stommel-like) where large ocean area concentrates surface mass transport through CPE. - Basin geometry (Methods): Identify latitudes φ1 and φ2 of zonal-mean wind extrema (zero wind-stress curl bounds). Define a trapezoidal gyral basin with typical longitudes λ1–λ4 along continental margins between φ1 and φ2. Meridional extent Ly = latitudinal distance between PQ and RS; zonal extent Lx = cosine-weighted mean of PQ and RS. Compute for each age; note grid-resolution dependence. - Ocean circulation metrics: Compute maximal volumetric (mass) transport Ψmax for the surface gyre and associated poleward heat transport H. Analyze relationship between Ly, Lx, Ψmax, H, and meridional SST gradient. - Temperature gradients: ΔSSTy defined as the zonal-mean SST gradient between 20°N and 50°N. Also compute global mean SST for context. - CO₂ effect estimation: Using ensemble pairs (560 vs 1120 ppmv) at fixed paleogeography to quantify average CO₂-induced change in ΔSSTy. Estimate expected ΔSSTy changes for observed long-term CO₂ trends via logarithmic scaling (per doubling) and typical climate sensitivity assumptions for global mean SST. - Simple theoretical model: Based on three assumptions: (1) ΔSSTy decreases linearly with oceanic H and atmospheric MOC heat transport; (2) atmospheric MOC transport proportional to ocean H; (3) Stommel-like gyre geometry yields large east-west SST contrast tied to ΔSSTy, with rapid western boundary advection and slower northern/eastern boundaries. Derive functional relationship: ΔSSTy = ΔSSTRad / [1 + (κ × Ψmax × ΔSSTRad)], where ΔSSTRad is radiative-equilibrium gradient (Ψ=0) and κ is empirical. - Regression: Fit the model to HadCM3L-derived ΔSSTy and Ψmax across 19 ages for each CO₂ ensemble to estimate ΔSSTRad and κ, and to quantify fraction of variance in ΔSSTy explained by Ψmax changes. Assess robustness via 1σ uncertainties and best-fit curves.

- Geometry and circulation: - North Pacific mid-latitudinal gyral basin Lx remained ~constant (~13,000 km) through the CPE, while Ly varied non-monotonically: decreased from ~3900 km to ~2800 km, then increased to ~3600 km. - Smaller Ly (and correlated smaller aspect ratio Lx/Ly) led to weaker Ψ and reduced poleward H. The Valanginian (~135 Ma) gyre transport was about twice that of the ~30% narrower Maastrichtian (~68 Ma) basin. - Meridional SST gradients (ΔSSTy between 20°N–50°N): - At 560 ppmv: ΔSSTy increased from 12.5 °C (Valanginian) to 15.7 °C (Maastrichtian); decreased from 16.5 °C (Danian, ~63 Ma) to 15.5 °C (Priabonian, ~36 Ma). - At 1120 ppmv: ΔSSTy increased from 11.1 °C to 13.9 °C (Cretaceous) and decreased from 14.3 °C (Danian) to 13.7 °C (Priabonian). - Across ages, doubling CO₂ (560→1120 ppmv) decreased ΔSSTy by 0.7–2.6 °C (average −1.8 °C), consistent with polar amplification. - Temporal ΔSSTy change from Valanginian to Maastrichtian is ~3 °C in both ensembles despite <1 °C SD in global mean SST; PE variation is ~0.8 °C, comparable to global SST SD. - CO₂-trend implications: - Estimated long-term CO₂ decrease from ~700 to ~570 ppmv (Valanginian→Maastrichtian) implies ~−1.5 °C global mean SST change and ~+0.6 °C increase in ΔSSTy; Danian→Priabonian (~590→~530 ppmv) implies ~−0.7 °C global mean SST change and ~+0.3 °C ΔSSTy increase. - Regression/model parameters: - For 560 ppmv: ΔSSTRad = 22.6 ± 1.4 °C; κ = (3.9 ± 0.5) × 10⁻⁴ Sv⁻¹ °C⁻¹. Reduction in Ψmax explains ~80% of ΔSSTy variance from Berriasian to Priabonian. - For 1120 ppmv: ΔSSTRad = 18.0 ± 0.9 °C; κ = (3.7 ± 0.6) × 10⁻⁴ Sv⁻¹ °C⁻¹. Reduction in Ψmax explains ~75% of ΔSSTy variance. - An abrupt ΔSSTy increase (~130–100 Ma) at 1120 ppmv arises from SST cooling at 50°N between Barremian and Albian. - Comparison with proxies: - Combined effect of reduced CO₂ and paleogeography from Valanginian to Maastrichtian yields an estimated −3.6 °C change in ΔSSTy, within ~55% of proxy-inferred variation (−6.5 °C) for broader Equator-to-pole gradients, acknowledging sparse NH mid-latitude proxy coverage.

Results demonstrate that tectonically driven Northern Hemisphere paleogeographic changes reduced the horizontal aspect ratio of the North Pacific mid-latitude basin, weakening wind-driven gyre transport (Ψmax) and poleward ocean heat transport (H). This weakening increased the northern mid-latitudinal SST gradient through the Cretaceous. The theoretical relationship ΔSSTy = ΔSSTRad / [1 + κ Ψmax ΔSSTRad] captures most of the variance in simulated ΔSSTy, highlighting the dominant role of surface gyre dynamics in setting regional gradients under varying paleogeographies. CO₂ primarily modulates gradients via polar amplification, with doubling CO₂ reducing ΔSSTy by ~1.8 °C on average. During the Paleogene, the anti-correlation between Ψmax and ΔSSTy persists, but the net ΔSSTy change is small relative to ensemble global SST variability. Lack of a clear Southern Hemisphere trend likely reflects concurrent tectonic reorganizations (e.g., Antarctica–Australia breakup) that violate the simple gyral framework. Additional climate-system processes (deep-sea temperatures, gateway changes, marine ice extent, atmospheric circulation changes) may modulate Ly/Lx and H, contributing to deviations (e.g., Barremian–Albian cooling) not fully captured by the simple model.

The study quantifies the relative roles of CO₂ and paleogeography on northern mid-latitude SST gradients during the Cretaceous–Paleogene. A simple Stommel-based framework, validated against HadCM3L ensembles, shows that paleogeographically driven reductions in the North Pacific gyral aspect ratio weakened surface gyre transport and poleward heat transport, explaining ~75–80% of ΔSSTy variability from Early to Late Cretaceous and into the Paleogene. CO₂ variations exert a secondary but significant control via polar amplification, decreasing ΔSSTy with higher CO₂. These findings underscore the importance of basin geometry in regulating surface ocean heat transport and regional temperature gradients in greenhouse climates. Future work should co-vary CO₂ and paleogeography across models, use model hierarchies with idealized continents, and expand proxy coverage in NH mid-latitudes to refine constraints and test model generality across geologic periods.

- Atmospheric CO₂ and paleogeography do not coevolve in the simulations, limiting direct comparison with time-evolving proxy records. - Significant uncertainties in reconstructed CO₂ levels propagate into estimated CO₂-driven ΔSSTy changes. - Diagnostic focus on Northern Hemisphere mid-latitudinal North Pacific; Southern Hemisphere trends are not captured due to complex tectonic evolution (e.g., Antarctica–Australia breakup). - The identification of Ly and Lx and derived metrics depend on model grid resolution and a specific trapezoidal basin definition. - The simple model relies on first-order assumptions (proportionality between oceanic and atmospheric heat transports, Stommel-like gyre behavior) and may miss processes such as gateway dynamics, deep-ocean changes, or atmospheric circulation shifts. - An abrupt Barremian–Albian SST change at 1120 ppmv is not investigated mechanistically. - Proxy spatial sparsity, especially between 30°N–80°N, limits validation of mid-latitudinal gradients.