Latent-Constrained Conditional VAEs for Augmenting Large-Scale Climate Ensembles

TL;DR

Large climate ensembles are computationally expensive, motivating a generative approach to produce additional realizations that preserve statistical structure. The authors diagnose latent fragmentation when a CVAE is trained across multiple realizations and remedy it with a latent-constrained CVAE that aligns latent structure at a small set of geographic anchors. To further extend coverage to unsampled locations, they employ a multi-output Gaussian process to predict dense latent fields from sparse anchors and decode to full time series. Across ERA5 data, the method shows stability gains, demonstrates that benefits saturate after about five realizations, and yields plausible new realizations suitable for augmenting climate ensembles while preserving key statistics.

Abstract

Large climate-model ensembles are computationally expensive; yet many downstream analyses would benefit from additional, statistically consistent realizations of spatiotemporal climate variables. We study a generative modeling approach for producing new realizations from a limited set of available runs by transferring structure learned across an ensemble. Using monthly near-surface temperature time series from ten independent reanalysis realizations (ERA5), we find that a vanilla conditional variational autoencoder (CVAE) trained jointly across realizations yields a fragmented latent space that fails to generalize to unseen ensemble members. To address this, we introduce a latent-constrained CVAE (LC-CVAE) that enforces cross-realization homogeneity of latent embeddings at a small set of shared geographic 'anchor' locations. We then use multi-output Gaussian process regression in the latent space to predict latent coordinates at unsampled locations in a new realization, followed by decoding to generate full time series fields. Experiments and ablations demonstrate (i) instability when training on a single realization, (ii) diminishing returns after incorporating roughly five realizations, and (iii) a trade-off between spatial coverage and reconstruction quality that is closely linked to the average neighbor distance in latent space.

Figures (5)

• Figure 1: Latent-space fragmentation in a standard CVAE trained jointly across multiple realizations. Latent encodings tend to cluster by realization identity rather than by shared structure, degrading generalization to unseen realizations.
• Figure 2: Effect of the latent homogeneity constraint (LC-CVAE): latents corresponding to the same conditioning context become locally aligned across realizations, mitigating the realization-driven fragmentation shown in \ref{['fig:latent-fragmentation']}.
• Figure 3: Schematic of latent completion for an unseen realization. Sparse latent codes inferred at observed locations are used as training targets for (multi-output) GP regression, producing dense latent codes that are decoded into a completed realization.
• Figure 4: Ablation results from the poster study. Trends suggest instability in the low-realization regime and diminishing returns beyond a modest ensemble size.
• Figure 5: Qualitative examples of generation and/or completion from the poster-era pipeline. The model can generate spatially coherent fields and preserve basic temporal characteristics of the original series at selected locations.