The PalMod II‑3‑3‑1 project, carried out at the MARUM of the University of Bremen from 1 November 2019 to 31 January 2023, aimed to generate and evaluate simulations of the last glacial period using the coupled Earth system model CESM in its isotope‑enabled version. The main scientific objective was to compare model output with palaeodata, focusing on stable water isotopes, and to assess the consistency of the simulations with a palaeodata synthesis. To achieve this, the team performed a series of forward‑proxy runs, beginning with a control simulation under pre‑industrial boundary conditions, followed by a last glacial maximum (LGM) run that used oceanic initial conditions derived from a previous CESM simulation without isotopes. A subsequent equilibrium run for Marine Isotope Stage 3 (MIS 3) and a transient simulation of an abrupt climate change during the same period were also planned, although the latter could not be fully completed due to staffing delays and the time required to correct the isotope code.

Technical work involved debugging the isotope module of CESM 1.2, adjusting model parameters to ensure numerical stability, and migrating the model to the HLRN‑IV supercomputer in Berlin. Various processor layouts for the atmospheric, oceanic, land, and sea‑ice components were tested to optimise performance. The LGM simulation produced a global mean cooling of 3.6 °C relative to the pre‑industrial control, which is 1.9 °C greater than the cooling inferred from the GLOMAP reconstruction based on microfossils. The isotope results showed a pronounced latitudinal contrast in δ¹⁸O of precipitation, with stronger evaporation in the subtropics and northward‑transported precipitation leading to lower δ¹⁸O values over continental ice sheets, mirroring patterns seen in ice‑core and cave records. When the model’s δ¹⁸O precipitation field was plotted against palaeodata, the regression slope was 0.81 and the root‑mean‑squared error 3.44 ‰, indicating good agreement on a global scale. The model outputs were also compared with the ECHAM6‑wiso model under identical boundary conditions, providing a benchmark for inter‑model consistency.

Collaboration was organised within the Data‑Synthesis working group of the larger PalMod consortium. The project was funded entirely by the German Federal Ministry of Education and Research (BMBF) and initially lacked personnel and resources at the University of Bremen, leading to the creation of a full‑time scientist position that was filled by Dr. Tamás Kovács on 1 February 2020. Dr. André Paul served as project contact. The project relied on the long‑standing expertise of the MARUM in Earth system modelling and the Geosciences department’s geosystem‑modelling group. Travel and computational resources were covered by the BMBF grant, and the high‑performance computing needs were met by the HLRN. The project’s timeline was extended to 31 January 2023 to accommodate the delayed staffing and the additional time required for model setup, ensuring that the core objectives—generation of isotope‑enabled CESM simulations and their comparison with palaeodata—were achieved within the funded period.