Search papers, labs, and topics across Lattice.
This study explores three conditioning strategies for a denoising diffusion probabilistic model aimed at downscaling daily precipitation data, crucial for hydrological assessments. By comparing channel concatenation, learned convolutional encoder cross-attention, and frozen encoder cross-attention using the Prithvi WxC model, the authors find that cross-attention conditioning significantly enhances distributional realism and retains extreme precipitation events better than concatenation. Notably, the Prithvi-WxC model achieves strong performance with limited training data, suggesting its utility in data-scarce environments.
Cross-attention conditioning dramatically improves the realism of downscaled precipitation forecasts, especially for extreme events, outperforming traditional concatenation methods.
High-resolution precipitation fields are essential for hydrological impact assessment, yet global climate model outputs are too coarse and biased for direct use. AI-based statistical downscaling with diffusion models offers a promising approach, but the mechanism by which large-scale atmospheric predictors condition generation remains largely unexplored. We investigate three conditioning strategies for a denoising diffusion probabilistic model applied to daily precipitation downscaling: channel concatenation of upsampled coarse predictors, cross-attention conditioning with a learned convolutional encoder, and cross-attention conditioning with the frozen encoder of the pretrained Prithvi WxC weather foundation model. All strategies are evaluated against an unconditioned baseline under identical conditions using probabilistic, distributional, spectral, and extreme-event metrics for the Colorado River Basin. Concatenation conditioning achieves the lowest point-wise CRPS and MSE, but tends to produce over-smoothed fields that suppress high-intensity events. In contrast, cross-attention conditioning provides substantially better distributional realism and modest improvements in spectral fidelity. Improvements are greatest for extremes: the Prithvi-WxC conditioned model retains over half of>100mm/day events, although estimates are uncertain due to limited samples. When trained on the full dataset, the learned convolutional model performs similarly to the foundation model-conditioned approach while requiring lower computational resources. However, the Prithvi-WxC-conditioned model achieves comparable performance with only five years of training data. These results indicate that cross-attention conditioning offers advantages over simple concatenation for probabilistic precipitation downscaling, and that pre-trained foundation model representations may offer benefits in data-limited settings.
