Holocene thermal maximum mode versus the continuous warming mode: Problems of data-model comparisons and future research prospects

The Holocene, the most recent interglacial, provides an important time window for evaluating current global warming and predicting future temperature changes. With the development of new temperature proxies and improvements in climate models, significant progress has been made in understanding Holocene temperature changes. However, a major debate persists about whether global temperatures during the Holocene followed a pattern of gradual warming from the end of the Last Glacial Maximum, which culminated in a temperature maximum in the early to middle Holocene, followed by gradual cooling in the late Holocene (the thermal maximum mode); or whether there was a continuous warming trend that continued to the present day (the continuous warming mode). Significant discrepancies exist between different proxy records as well as between proxy records and models, which have resulted in the "Holocene temperature conundrum" that has challenged paleoclimatologists for the past decade. Here, we summarize the progress made to date in the study of Holocene temperature change via proxy reconstructions, climate model simulations, and paleoclimate data assimilation. We emphasize that the current research has limitations in terms of the multiplicity and seasonality of proxy records, the spatial heterogeneity of temperature records, and the incorporation of feedback processes (e.g., vegetation, cloud-radiation feedback) in climate models. These limitations have hindered a comprehensive understanding of the processes and mechanisms of Holocene temperature changes. To solve the "Holocene temperature conundrum", it is necessary to strengthen theoretical research on climate proxies from the perspective of the underlying processes and mechanisms, elucidate the seasonal response of various temperature proxies, emphasize regional differences in temperature changes, and expand quantitative temperature reconstructions to areas with limited records. However, it is also necessary to improve the simulation performance of complex feedback processes in climate models, reduce simulation errors, and advance the research on data assimilation of Holocene continental temperature records, which may ultimately lead to the optimal integration of paleoclimate records and simulations.