Menno Veerman | Searcharxiv

arXiv2023

Observed Patterns of Surface Solar Irradiance under Cloudy and Clear-sky Conditions

Surface solar irradiance varies on scales as small as seconds or meters due to scattering and absorption by the atmosphere. Clouds are the main driver of this variability, but moisture structures in the atmospheric boundary layer and aerosols have an influence too, and depend on wavelength. The highly variable nature of solar irradiance is not resolved by most atmospheric models, yet it affects most notably the land-atmosphere coupling, which in turn can change the cloud field, and the quality of solar energy forecasting. Spatially and spectrally resolved observational datasets of solar irradiance at such high resolution are rare, but they are required for characterising observed variability, understanding the mechanisms, and developing fast models capable of accurately resolving this variability. In 2021, we deployed a spatial network of low-cost radiometers at the FESSTVaL (Germany) and LIAISE (Spain) field campaigns, specifically to gather data on cloud-driven surface patterns of irradiance, including spectral effects, with the aim to address this gap in observations and understanding. We find in case studies of cumulus, altocumulus, and cirrus clouds that these clouds generate large spatiotemporal variability in irradiance, but through different mechanisms and at difference spatial scales, ranging from 50 m to 30 km. Spectral irradiance in the visible range varies at similar spatial scales, with significant blue enrichment in cloud shadows, most strongly for cumulus, and red enrichment in irradiance peaks, particularly in the case of semi-transparent clouds or near cumulus cloud edges. Under clear-sky conditions, solar irradiance varies significantly in water vapour absorption bands at the minute scale, due to local and regional variability in atmospheric moisture.

physics.ao-ph↗

arXiv2020

Deep-learning enhancement of large scale numerical simulations

Traditional simulations on High-Performance Computing (HPC) systems typically involve modeling very large domains and/or very complex equations. HPC systems allow running large models, but limits in performance increase that have become more prominent in the last 5-10 years will likely be experienced. Therefore new approaches are needed to increase application performance. Deep learning appears to be a promising way to achieve this. Recently deep learning has been employed to enhance solving problems that traditionally are solved with large-scale numerical simulations using HPC. This type of application, deep learning for high-performance computing, is the theme of this whitepaper. Our goal is to provide concrete guidelines to scientists and others that would like to explore opportunities for applying deep learning approaches in their own large-scale numerical simulations. These guidelines have been extracted from a number of experiments that have been undertaken in various scientific domains over the last two years, and which are described in more detail in the Appendix. Additionally, we share the most important lessons that we have learned.

cs.CE↗