arXiv Science⌕ Search

arXiv subjects

Rudolf Saathof

Publications and source records attributed to Rudolf Saathof.

5 recordsLinked to original sources

Operator-Aware Posterior Sampling for Multivariate Weather Downscaling

Statistical downscaling is the task of reconstructing high-resolution meteorological fields from coarse inputs. A principled approach formulates downscaling as a Bayesian inverse problem, using a diffusion prior over high-resolution fields and incorporating the coarse field only at inference through a likelihood. This decouples the prior from the coarse source, allowing one pretrained model to adapt to different sources. However, two challenges limit this flexibility: (1) when coarse and target fields come from different reanalysis systems, no fixed coarsening operator can capture their relationship; and (2) guidance gradients can differ by orders of magnitude across variables, making a single guidance strength ineffective. This paper addresses both challenges within the guidance loop, leaving the prior unchanged. We learn the forward operator from paired data and normalize the gradient per channel. We show that this normalization corresponds to steepest descent under a block-max norm, allowing a single guidance strength across variables. Our code is publicly available at https://github.com/CarloSaccardi/operator-aware-weather-downscaling.

cs.CV↗

OTProf: estimating high-resolution profiles of optical turbulence ($C_n^2$) from reanalysis using deep learning

Accurate, high-resolution vertical profiles of optical turbulence ($C_n^2$), which reflect local meteorology and topography, are crucial for ground-based optical astronomy and free-space optical communication. However, measuring these profiles or generating them with numerical weather models requires substantial operational or computational effort. In this work, we present OTProf, a deep-learning method that estimates high-resolution $C_n^2$ profiles from widely available coarse-resolution ERA5 reanalysis data. OTProf is based on the Squeezeformer architecture, modified for super-resolution by introducing grid-agnostic interpolation in the latent space. We train and evaluate OTProf in the Netherlands using high-resolution $C_n^2$ data obtained from the Weather Research and Forecasting (WRF) model. The commonly used Hufnagel-Valley model serves as the baseline. Overall, OTProf reproduces the vertical structure of WRF-based $C_n^2$ more accurately than Hufnagel-Valley and yields more accurate estimates of the Fried parameter $r_0$ and the scintillation index $σ_I^2$. Out-of-domain performance is assessed using a smaller WRF-based $C_n^2$ dataset covering the Canary Islands. Initial results indicate a performance drop because the orography differs markedly from the training domain. We show that output bias correction based on a few local simulations can help OTProf to adapt, but more work for domain adaptation is needed. OTProf is therefore primarily recommended for in-domain applications, e.g., generating $C_n^2$ climatologies from long ERA5 records.

physics.ao-ph↗

Assessing the Geographic Generalization and Physical Consistency of Generative Models for Climate Downscaling

Kilometer-scale weather data is crucial for real-world applications but remains computationally intensive to produce using traditional weather simulations. An emerging solution is to use deep learning models, which offer a faster alternative for climate downscaling. However, their reliability is still in question, as they are often evaluated using standard machine learning metrics rather than insights from atmospheric and weather physics. This paper benchmarks recent state-of-the-art deep learning models and introduces physics-inspired diagnostics to evaluate their performance and reliability, with a particular focus on geographic generalization and physical consistency. Our experiments show that, despite the seemingly strong performance of models such as CorrDiff, when trained on a limited set of European geographies (e.g., central Europe), they struggle to generalize to other regions such as Iberia, Morocco in the south, or Scandinavia in the north. They also fail to accurately capture second-order variables such as divergence and vorticity derived from predicted velocity fields. These deficiencies appear even in in-distribution geographies, indicating challenges in producing physically consistent predictions. We propose a simple initial solution: introducing a power spectral density loss function that empirically improves geographic generalization by encouraging the reconstruction of small-scale physical structures. The code for reproducing the experimental results can be found at https://github.com/CarloSaccardi/PSD-Downscaling

cs.LG↗

OTCliM: generating a near-surface climatology of optical turbulence strength ($C_n^2$) using gradient boosting

This study introduces OTCliM (Optical Turbulence Climatology using Machine learning), a novel approach for deriving comprehensive climatologies of atmospheric optical turbulence strength ($C_n^2$) using gradient boosting machines. OTCliM addresses the challenge of efficiently obtaining reliable site-specific $C_n^2$ climatologies near the surface, crucial for ground-based astronomy and free-space optical communication. Using gradient boosting machines and global reanalysis data, OTCliM extrapolates one year of measured $C_n^2$ into a multi-year time series. We assess OTCliM's performance using $C_n^2$ data from 17 diverse stations in New York State, evaluating temporal extrapolation capabilities and geographical generalization. Our results demonstrate accurate predictions of four held-out years of $C_n^2$ across various sites, including complex urban environments, outperforming traditional analytical models. Non-urban models also show good geographical generalization compared to urban models, which capture non-general site-specific dependencies. A feature importance analysis confirms the physical consistency of the trained models. It also indicates the potential to uncover new insights into the physical processes governing $C_n^2$ from data. OTCliM's ability to derive reliable $C_n^2$ climatologies from just one year of observations can potentially reduce resources required for future site surveys or enable studies for additional sites with the same resources.

physics.ao-ph↗

Π-ML: A dimensional analysis-based machine learning parameterization of optical turbulence in the atmospheric surface layer

Turbulent fluctuations of the atmospheric refraction index, so-called optical turbulence, can significantly distort propagating laser beams. Therefore, modeling the strength of these fluctuations ($C_n^2$) is highly relevant for the successful development and deployment of future free-space optical communication links. In this letter, we propose a physics-informed machine learning (ML) methodology, $Π$-ML, based on dimensional analysis and gradient boosting to estimate $C_n^2$. Through a systematic feature importance analysis, we identify the normalized variance of potential temperature as the dominating feature for predicting $C_n^2$. For statistical robustness, we train an ensemble of models which yields high performance on the out-of-sample data of $R^2=0.958\pm0.001$.

physics.ao-ph↗