Earth's orbital changes are considered the primary drivers of Quaternary ice age dynamics. However, orbital theory alone cannot fully explain the shift from 41-kyr glacial cycles in the early Quaternary to 100-kyr cycles during the last 1 million years, a phenomenon known as the Mid-Pleistocene Transition (MPT). Understanding the MPT requires considering climate system feedbacks, such as land ice dynamics, possibly influenced by regolith removal and declining atmospheric CO₂ concentrations during glacial minima. Numerous hypotheses have been proposed over the past four decades to explain the MPT. A recent approach (T17) identifies the onset of interglacial periods based on caloric summer half-year insolation at 65°N, suggesting that deglaciations are primarily driven by local summer insolation and that the energy required to trigger deglaciation decreases over time due to increasing ice-sheet instability. In T17, interglacials are defined by thresholds in the detrended benthic LR04 δ¹⁸O stack, showing irregular interglacial appearances after the MPT, a pattern consistent with an increasing number of skipped obliquity-driven terminations. Other studies emphasize glaciological perspectives, suggesting that post-MPT glacial cycles are controlled by North American ice sheet dynamics, where the merging of ice sheets leads to thicker ice, pressure-melting point conditions, ice sliding, and facilitated deglaciation. While orbital theories successfully define deglaciation triggers, they often lack detail on glaciation processes, reducing land ice complexity and focusing on a single latitude (65°N). This study investigates land ice dynamics from an inverse deconvolution of the LR04 δ¹⁸O stack, defining interglacials by the absence of substantial northern hemispheric land ice outside Greenland, a criterion shown to be robust for the last 800 kyr. This approach avoids reliance on specific insolation metrics and considers the latitudinal distribution of land ice, aiming to address the limitations of previous studies and provide a more comprehensive understanding of the MPT.

Existing literature offers varied perspectives on the Mid-Pleistocene Transition (MPT). Some studies emphasize the role of orbital forcing, particularly obliquity, in driving glacial cycles. These studies often focus on insolation thresholds at specific latitudes to explain the timing of glacial terminations and the onset of interglacials. However, these models often simplify the complex interactions within the climate system, overlooking the crucial role of land ice dynamics and feedback mechanisms. Other research highlights the importance of land ice dynamics, particularly in North America, as a key driver of the MPT. The development and behavior of large ice sheets, their interactions with underlying geology, and their influence on ocean circulation and atmospheric CO2 are all implicated. Discrepancies exist in how the transition from shorter to longer glacial cycles occurred. Some suggest a gradual change in sensitivity to orbital forcing, others point to changes in ice sheet behavior and stability. This study aims to integrate these different perspectives by analyzing land ice volume dynamics in a physically consistent framework, resolving discrepancies in current explanations.

This study utilizes a model-based deconvolution of the LR04 benthic δ¹⁸O stack to reconstruct land ice dynamics. This deconvolution, performed using the 3-D ice-sheet model ANICE, separates the contribution of deep ocean temperature and land ice volume changes to the overall δ¹⁸O signal. The spatial resolution of the land ice data is 40x40 km² (20x20 km² for Greenland), and the temporal resolution is 2 kyr (interpolated to 1 kyr for analysis). The research defines interglacials as periods with minimal northern hemispheric land ice outside Greenland. Two thresholds are applied to the integrated land ice volume in North America and Eurasia to identify individual interglacials: a lower threshold to mark the onset of a new interglacial and an upper threshold to separate interglacials by intervening glacial periods. These thresholds are chosen to replicate findings for the last 800 kyr and to maximize the number of obliquity cycles containing interglacial onsets. The study minimizes reliance on insolation metrics, focusing instead on the presence or absence of interglacials within obliquity cycles. Alternative threshold values are also tested to assess the robustness of the results. The study further analyzes the latitudinal distribution of land ice volume and explores the relationship between land ice albedo feedback and orbital forcing. Comparisons are made with previous definitions of interglacials and alternative analyses based on different benthic δ¹⁸O records and age models, accounting for uncertainties in ice-sheet modeling and proxy data. The radiative forcing of CO2 is calculated using a established formula and is compared to existing reconstructions from ice cores, Allan Hills blue ice, marine proxies, and models. Frequency analysis was conducted using Analyseries software.

The study's key findings challenge the simplistic view of obliquity-driven interglacial terminations throughout the Quaternary. By defining interglacials based on the absence of significant northern hemispheric land ice outside Greenland, the researchers find irregular appearances of interglacials both before and after the Mid-Pleistocene Transition (MPT). Before the MPT (2.6 to 1.6 Myr BP), only 67% of obliquity cycles contained a new interglacial, compared to 88% during the MPT and 52% after the MPT. This indicates a more complex mechanism driving glacial terminations than previously assumed. The study reveals that the glacial/interglacial amplitudes of northern hemisphere land ice variability outside Greenland are considerably smaller before the MPT than after. The detrending of the LR04 δ¹⁸O used in previous studies proves to be imprecise for correcting for the effects of deep-ocean temperature and ice volume in Antarctica and Greenland prior to the MPT. The researchers find that the deconvolution reveals a relatively large contribution of ice sheets in Antarctica and Greenland to δ¹⁸Osw in the early Quaternary compared to the late Quaternary. The latitudinal integration of land ice volume distribution shows a southward shift of northern hemisphere land ice over time, mainly located north of 65°N before the MPT. The 50% line (latitude where half of northern hemisphere ice volume change is located north) shifts from ~70°N before 1.8 Myr BP to 62°N at the Last Glacial Maximum (LGM). The study also finds a large overlap between their skipped termination identifications and those identified previously in the literature (for the last 2 million years), supporting the robustness of their findings. The study suggests that a simple rule to define interglacials might not be applicable across the entire Quaternary period and alternative definitions might be necessary for earlier periods.

The findings challenge the prevailing notion that obliquity cycles consistently trigger interglacials throughout the Quaternary. The irregular appearance of interglacials both before and after the MPT, particularly the lower realization fraction before the MPT, suggests a more complex interplay of factors determining glacial terminations. The smaller amplitudes of northern hemispheric land ice outside Greenland prior to the MPT, as revealed by the deconvolution, highlight the limitations of previous methods that relied on detrending the LR04 δ¹⁸O record. The southward shift of land ice suggests a changing energy budget, impacting the importance of insolation changes at specific latitudes and questioning the validity of simple rules based on insolation at 65°N. The robustness of the findings is discussed in the context of uncertainties in the chosen thresholds, the LR04 δ¹⁸O stack, and ice-sheet models. While some inconsistencies exist with previous classifications, particularly for the early Quaternary, the study emphasizes the overall shift in the pattern of climate change across the Quaternary, from an interglacial-dominated period to a glacial-dominated one. Further research is needed to refine the understanding of the complex interplay between orbital forcing, land ice dynamics, carbon cycle feedbacks, and other factors driving the long-term evolution of climate during the Quaternary.

This study demonstrates that the definition and interpretation of interglacial periods require careful consideration of land ice dynamics. The irregular appearance of interglacials, both before and after the MPT, suggests a more complex interplay of factors than previously thought. The study highlights the limitations of simpler models focusing solely on orbital forcing and underscores the need for more sophisticated models integrating land ice dynamics and feedback mechanisms. Future research should focus on refining models of land ice behavior, improving proxy data resolution, and further investigating the interplay of orbital forcing, carbon cycle feedbacks, and land ice dynamics to explain the observed climate patterns throughout the Quaternary. A less binary approach to classifying climate stages might be more appropriate than strict interglacial/glacial classifications, emphasizing gradual climate shifts over time.

The study relies on a model-based deconvolution of the LR04 benthic δ¹⁸O stack, introducing uncertainties inherent in model parameters and data interpretation. The coarse temporal resolution of the ice-sheet model and the benthic δ¹⁸O stack also limits the ability to analyze millennial-scale variability. The chosen thresholds for defining interglacials are subject to uncertainty, potentially influencing the classification of some periods. The availability of independent data on northern hemispheric ice volume outside Greenland is limited, hindering the verification of the results.