New classes of climate model emulators to improve paleoclimate reconstructions

Abstract. Reconstructing spatial climate variability from proxy records requires forward models “emulators” that capture the dynamical structure of the climate system while remaining computationally efficient. Traditional emulators based on Empirical Orthogonal Functions and Linear Inverse Models (LIMs) face inherent limitations due to linearity and variance-based dimensionality reduction. Here we develop and evaluate a hierarchy of CMIP-class climate model emulators that integrate autoencoder-based dimensionality reduction with nonlinear prediction architectures, including Reservoir Computing (RC) and Recurrent Neural Networks (RNNs). Using a comprehensive experimental protocol applied to the IPSL-CM6A-LR model and 52 CMIP6s models, we show that combining autoencoder and RC (AERCn) provides the most robust performance across time scales and dynamical regimes when data are plentiful. The AERCn configuration captures nonlinear features of ENSO and Atlantic Multidecadal Variability, maintains high spatial reconstruction skill, and generalizes across distinct climate model structures. When training data are scarce, a multimodel pre-trained AERNN provides a data-efficient and competitive alternative. These properties make the proposed architectures particularly well suited for integration into Particle Filters and Ensemble Kalman Filter PDA frameworks. Our results highlight the importance of predictability-oriented dimensionality reduction and nonlinear dynamical memory for emulator design, and they provide a scalable path toward improved reconstructions of climate variability over the Common Era.