Bram Vaes

Warm, high-CO ₂ climates of Earth's past provide an opportunity to evaluate climate models under extreme forcing, and to explore mechanisms that lead to such warmth. One such time period is the early Eocene (∼56–41 million years ago), when global mean surface temperatures were ∼15 °C higher than preindustrial, and CO ₂ concentrations were ∼1500 ppmv. In this paper we present the experimental design for Phase 2 of the Eocene component of the Deep-time Model Intercomparison project (DeepMIP-Eocene-p2). The aim is to provide a framework for modelling groups to carry out a common set of simulations, thereby facilitating exploration of inter-model dependencies. The focus is on the early Eocene Climatic Optimum (EECO, ∼53.3–49.1 million years ago). Relative to Phase 1 of DeepMIP-Eocene, we provide a new paleogeography (topography, bathymetry) derived from several recent independent reconstructions that focused on different regions, a new vegetation distribution derived by merging paleobotanical data with vegetation model simulations, and a new CO ₂ specification derived from recent re-evaluations of proxy data. The core set of simulations consists of a preindustrial control, an abrupt increase to 4× preindustrial CO ₂ concentration from this preindustrial control, a standard control EECO simulation at 5× preindustrial CO ₂ concentration, and an EECO simulation with preindustrial CO ₂ concentration. In addition to these core simulations, we suggest a suite of optional sensitivity studies, which allow the impact of various factors to be explored, such as topography/bathymetry, greenhouse gases, land-surface parameters, astronomical and solar forcings, and internal model parameters. The updated boundary conditions and guidance on initialisation and spinup in Phase 2 will allow more robust model-data comparisons, more accurate insights into mechanisms influencing early Eocene climate, and increased relevance for informing future climate change projections.

Paleogeography, and particularly the paleolatitude, provides key context in the interpretation of paleoclimatic and paleobiological data but these fields are typically studied by scientists in different disciplines. To facilitate communication between these disciplines, a decade ago the online Paleolatitude.org calculator was developed. This provided for any coordinate on stable tectonic plates a paleolatitude estimate for any chosen Phanerozoic time interval, including an uncertainty that includes paleogeographic uncertainty and age uncertainty of a sample/fossil. Here, we provide a major update to this tool. First, we include in the calculator the first global paleogeographic model, including GPlates reconstruction files, back to 320 Ma that also restores paleogeographic units that are now thrusted over each other in orogenic (mountain) belts. Second, we include a recent, more precise paleomagnetic reference frame with updated statistical procedures, and provide the first update of its underlying database. Third, we introduce a new online interface with an easy-to-use tool with a batch option, and data and graph export functions. Finally, we illustrate differences with previous reconstructions and show an application by calculating a paleolatitudinal biodiversity gradient for the late Jurassic in which we use a bootstrap approach to propagate paleolatitude and age uncertainty into the result.

We welcome the critique of Hetényi and Cattin (2026, https://doi.org/10.1029/2025tc009214 ) on Sternai et al. (2025, https://doi.org/10.1029/2025tc009057 ), as it brings to light evaluative issues that require clarification and offers an opportunity to articulate more clearly how our work advances understanding of Himalayan‐Tibetan geodynamics. In this Reply, we clarify that most of the criticism arises from evaluating our process‐oriented geodynamic models as if they were intended to reproduce present‐day lithospheric structures, rather than to test the mechanical plausibility of Paleogene crustal‐doubling mechanisms. Moreover, non‐unique interpretations of present‐day geophysical data are treated in the Comment as definitive constraints and used as benchmarks to dismiss alternative physically viable scenarios, leading to circular reasoning. When the content and objectives of our study are correctly considered, the critiques do not invalidate our results or conclusions. More broadly, this Reply underscores the importance of reassessing established models when fundamental limitations are identified, as a constructive pathway toward progress in Earth sciences.

Cenozoic greenhouse climates offer important insights into Earth’s climate system and carbon cycle under elevated CO2 conditions. A major challenge in simulating these warm intervals lies in the accurate reconstruction of the paleogeography, yet its impact on modeled climates and their agreement with proxy data remains poorly quantified. In this study, we systematically assess the sensitivity of fully coupled climate simulations to alternative paleogeographic reconstructions for the Paleocene, early Eocene, and middle-late Eocene. Using the IPSL-CM5A2 Earth System Model, we find that regional climates are particularly sensitive to the paleolatitudinal position of landmasses and ocean basins. Latitudinal shifts of more than 5°, arising from the choice of mantle versus paleomagnetic reference frame, significantly alter modeled regional temperature and precipitation patterns, as well as ocean circulation patterns. Moreover, we demonstrate that reconciling simulated climates with temperature proxy data depends strongly on the reconstructed paleolatitude of the proxy sites. In regions such as the southwest Pacific, correcting for paleolatitude bias induced by a mantle frame reduces model-data temperature misfits by up to 5°C. Our results further show that the regional climatic impact of paleogeography can equal or even exceed that of a doubling of atmospheric CO2, particularly at mid-latitudes. These findings highlight the importance of using accurate paleogeographic reconstructions and an appropriate reference frame for improving paleoclimate simulations and their integration with proxy data.

Palaeomagnetic measurements provide the most robust, quantitative constraints on vertical axis rotations from outcrop scale blocks to entire tectonic plates. Vertical axis rotations are typically shown on maps as vectors representing local palaeodeclinations or individual rotations to a reference at the site level. When dealing with large and noisy datasets, however, this approach struggles to pick out underlying spatial patterns or finer scale regional displacements, especially when the data has variable density across regions.These limitations are compounded in areas where several rotations events have been recorded across multiple scales or are poorly recorded in the palaeomagnetic record. As a result, both the visualisation and interpretation of complex rotation fields are hindered.Here, we introduce PyRotate: a robust and scalable, end-to-end workflow for integrating and analyzing locality-level palaeomagnetic data from multiple sources. This framework allows users to curate and clean data through custom parameters, and to apply bootstrap techniques to determine the rotation of selected blocks and plates. Novel tools include kriging of rotations by age to simplify areas of highly variable rotation amounts and estimate interpolations between sites, giving an overlain interpolated rotation field as a visualisation tool, as well as automated outlier detection and labelling through cluster analysis. In addition, outputs are compatible with both PmagPy (Tauxe et al., 2016) and PyGPlates (Mather et al., 2024) as the basis of further analysis.To illustrate the capabilities of PyRotate, we analyse a curated paleomagnetic dataset from Japan, and multiple-source datasets directly obtained from the Magic database. This analysis tests the ability of the workflow to deal with noisy and unevenly distributed datapoints. The tools introduced in PyRotate enable easier identification of areas needing additional sampling, synthesis of disparate datasets, and interpretation of the rotation of blocks through time.

Changes in the geological carbon cycle and associated surface input and output CO2 fluxes drive long-term Cenozoic climate trends mainly through magmatic emissions and weathering of silicate minerals. Proxy records, which indirectly reconstruct past climate conditions, demonstrate a steady decline in both surface CO2 and temperature since ˜50 million years ago (Ma), punctuated by shorter periods of climatic optima and hyperthermals such as the PETM, EECO, MECO and MMCO. However, lack of constraints in terms of input and output CO2 fluxes prevents the assessment of responsible processes for these trends. Here, we use a newly developed technique based on a reversible-jump Markov chain Monte Carlo algorithm (rj-McMC) to invert the temporal CO2 changes from the Proxy Integration Project (CENCO2PIP) (Hönisch et al., 2023) and obtain estimates of the surface input and output CO2 fluxes throughout the Cenozoic. We base the inversion on a general formulation of the geological carbon cycle that includes a degassing source and a temperature-dependent sink term, with the temperature time history (Hansen et al., 2023) used as an additional constraint. Reconstructed fluxes reveal that perturbations of the carbon cycle are stronger during the early Cenozoic (i.e., ˜66 – 34 Ma), while these reduce since˜34 Ma. We hypothesise that stronger degassing from the solid-Earth during the EECO and MECO prevent an earlier onset of the Antarctic ice cap during the Eocene. We discuss that the higher carbon emissions during these periods can partially link to the evolution of the Neo-Tethyan magmatic margin, which extinction occurs ˜34 Ma. Results show that carbon flux stabilization since the Oligocene could be due to temperature dependent processes like albedo increase and enhanced silicate weathering in the context of Tibetan Plateau uplift. Finally, we estimate that the net amount of CO2 removed since ˜34 Ma is four times greater than that of the first half of the Cenozoic. References Hansen, J. E., Sato, M., Simons, L., Nazarenko, L. S., Sangha, I., Kharecha, P., Zachos, J. C., von Schuckmann, K., Loeb, N. G., Osman, M. B., Jin, Q., Tselioudis, G., Jeong, E., Lacis, A., Ruedy, R., Russell, G., Cao, J., & Li, J. (2023). Global warming in the pipeline. Oxford Open Climate Change, 3(1). https://doi.org/10.1093/oxfclm/kgad008.Hönisch, B., Royer, D. L., Breecker, D. O., Polissar, P. J., Bowen, G. J., Henehan, M. J., Cui, Y., Steinthorsdottir, M., McElwain, J. C., Kohn, M. J., Pearson, A., Phelps, S. R., Uno, K. T., Ridgwell, A., Anagnostou, E., Austermann, J., Badger, M. P. S., Barclay, R. S., Bijl, P. K., … Zhang, L. (2023). Toward a Cenozoic history of atmospheric CO2. Science, 382(6675). DOI: 10.1126/science.adi517.

Warm, high-CO₂ climates of Earth's past provide an opportunity to evaluate climate models under extreme forcing, and to explore mechanisms that lead to such warmth. One such time period is the early Eocene, when global mean surface temperatures were 10–17 °C higher than preindustrial, and CO₂ concentrations were ~1500 ppmv. In this paper we present the experimental design for Phase 2 of the Eocene component of the Deep-time Model Intercomparison project (DeepMIP-Eocene-p2). The aim is to provide a framework for modelling groups to carry out a common set of simulations, thereby facilitating exploration of inter-model dependencies. The focus is on the early Eocene Climatic Optimum (EECO, ~53.3–49.1 million years ago). Relative to Phase 1 of DeepMIP-Eocene, we provide a new paleogeography (topography, bathymetry) derived from several recent independent reconstructions that focused on different regions, a new vegetation derived by merging paleobotanical data with vegetation model simulations, and a new CO₂ specification derived from recent reevaluations of proxy data. The core set of simulations consists of a preindustrial control, an abrupt increase to 4x preindustrial CO₂ concentrations from this preindustrial control, a standard control EECO simulation at 5x preindustrial CO₂ concentrations, and an EECO simulation with preindustrial CO₂ concentrations. In addition to these core simulations, we suggest a suite of optional sensitivity studies, which allow the impact of various factors to be explored, such as topography/bathymetry, greenhouse gases, land-surface parameters, astronomical and solar forcings, and internal model parameters. The updated boundary conditions and guidance on initialisation and spinup in Phase 2 will allow more robust model-data comparisons, more accurate insights into mechanisms influencing early Eocene climate, and increased relevance for informing future climate change projections.

Warm, high-CO₂ climates of Earth's past provide an opportunity to evaluate climate models under extreme forcing, and to explore mechanisms that lead to such warmth. One such time period is the early Eocene, when global mean surface temperatures were 10–17 °C higher than preindustrial, and CO₂ concentrations were ~1500 ppmv. In this paper we present the experimental design for Phase 2 of the Eocene component of the Deep-time Model Intercomparison project (DeepMIP-Eocene-p2). The aim is to provide a framework for modelling groups to carry out a common set of simulations, thereby facilitating exploration of inter-model dependencies. The focus is on the early Eocene Climatic Optimum (EECO, ~53.3–49.1 million years ago). Relative to Phase 1 of DeepMIP-Eocene, we provide a new paleogeography (topography, bathymetry) derived from several recent independent reconstructions that focused on different regions, a new vegetation derived by merging paleobotanical data with vegetation model simulations, and a new CO₂ specification derived from recent reevaluations of proxy data. The core set of simulations consists of a preindustrial control, an abrupt increase to 4x preindustrial CO₂ concentrations from this preindustrial control, a standard control EECO simulation at 5x preindustrial CO₂ concentrations, and an EECO simulation with preindustrial CO₂ concentrations. In addition to these core simulations, we suggest a suite of optional sensitivity studies, which allow the impact of various factors to be explored, such as topography/bathymetry, greenhouse gases, land-surface parameters, astronomical and solar forcings, and internal model parameters. The updated boundary conditions and guidance on initialisation and spinup in Phase 2 will allow more robust model-data comparisons, more accurate insights into mechanisms influencing early Eocene climate, and increased relevance for informing future climate change projections.

Warm, high-CO₂ climates of Earth's past provide an opportunity to evaluate climate models under extreme forcing, and to explore mechanisms that lead to such warmth. One such time period is the early Eocene, when global mean surface temperatures were 10–17 °C higher than preindustrial, and CO₂ concentrations were ~1500 ppmv. In this paper we present the experimental design for Phase 2 of the Eocene component of the Deep-time Model Intercomparison project (DeepMIP-Eocene-p2). The aim is to provide a framework for modelling groups to carry out a common set of simulations, thereby facilitating exploration of inter-model dependencies. The focus is on the early Eocene Climatic Optimum (EECO, ~53.3–49.1 million years ago). Relative to Phase 1 of DeepMIP-Eocene, we provide a new paleogeography (topography, bathymetry) derived from several recent independent reconstructions that focused on different regions, a new vegetation derived by merging paleobotanical data with vegetation model simulations, and a new CO₂ specification derived from recent reevaluations of proxy data. The core set of simulations consists of a preindustrial control, an abrupt increase to 4x preindustrial CO₂ concentrations from this preindustrial control, a standard control EECO simulation at 5x preindustrial CO₂ concentrations, and an EECO simulation with preindustrial CO₂ concentrations. In addition to these core simulations, we suggest a suite of optional sensitivity studies, which allow the impact of various factors to be explored, such as topography/bathymetry, greenhouse gases, land-surface parameters, astronomical and solar forcings, and internal model parameters. The updated boundary conditions and guidance on initialisation and spinup in Phase 2 will allow more robust model-data comparisons, more accurate insights into mechanisms influencing early Eocene climate, and increased relevance for informing future climate change projections.

Studying geodynamic processes at the plate‐mantle interface requires knowledge of how plates move over and subduct into the mantle. The absolute motion of plates relative to the mantle since ∼120 Ma is classically estimated using hotspot tracks. However, for older times only hybrid frames are available that integrate multiple types of constraints, such as trench kinematics, plate velocity, net lithosphere rotation, true polar wander, and eruptions of kimberlites and large igneous provinces. This diminishes the number of independent constraints that we can use to learn about geodynamic processes. Here, we compute absolute plate motions by minimizing absolute continent velocity in a global plate model, in 10 Ma timesteps back to 1 Ga, assuming that continental keels resist relative continent‐mantle motion. We estimate uncertainty by adding a ±5 Ma temporal uncertainty to each step and find absolute plate motion is reliable for the last 350 Ma, but not before. The “continent frame” predicts a Cretaceous episode of high net lithosphere rotation due to fast Izanagi plate motion that may signal an anomalously weak NW Pacific upper mantle. Furthermore, the continent frame predicts that Indo‐Atlantic hotspots moved parallel to the edges of large low shear velocity provinces and kimberlite eruption sites moved parallel to plate motion. We show that by constructing a global absolute plate motion model based on one assumption, minimal continent motion, allows placing the mantle in a plate tectonic reference frame and unlocks all other constraints of plate‐mantle interaction to kinematically reconstruct mantle convection.