Surrogate cloud fields with measured cloud properties
Victor Venema, Susanne Crewell, Clemens Simmer
Abstract
This paper describes two new methods to generate 2D and 3D cloud fields based on 1D and 2D ground based profiler meas-urements. These cloud fields share desired statistical properties with real cloud fields. As they, however, are similar but not the same as real clouds, we call them surrogate clouds. One important advantage of the new methods is that the amplitude distribution of cloud liquid water is also exactly determined by the measurement: The surrogate clouds made with the classi-cal methods such as the Fourier method and the Bounded Cascade method are Gaussian and 'log-normal-like', respectively. Our first new method iteratively creates a time series with a measured amplitude distribution and power spectrum. Our sec-ond method uses an evolutionary search algorithm to generate cloud fields with practically arbitrary constraints. These clouds will be used to study the relation between radiation and cloud structure.
Create a lesson
Related papers
From Sea Surface Height to Velocity - A Machine Learning Approach
Yaswanth Reddy Yaradoddi, Jinbo Wang, Elyes Serghine
Tropical Cyclone-Following Dataset for Limited-Area Machine Learning Applications
Minh-Khanh Luong, Chanh Kieu
Can LLM Agents Improve Imperfect Earth System Model Parameterizations?
Kara D. Lamb
Six-hourly Probabilistic Precipitation Forecasts Created From Hourly Probabilities
Thomas M Hamill
New Wine in an Old Bottle -- NeuralPOM: A Hybrid Ocean Forecast Model with Differentiable Dynamics and Staged Learning
Ruiqi Shu, Guangliang Liu, Ruijian Gou et al.
Jovian Polar Vortex Generation in a Shallow-Water Model with Quasi-Stochastic Convection
Yuhang Ruan, Hao Fu, Cheng Li et al.