Stereo Observations of Clouds for LES Validation and Sub-scale Cloud Parameterizations (SOCLES)
Final Report Abstract
This project combines high-frequency visual images of convective boundary layer clouds by a network of multiple hemispheric cameras at a meteorological site with high-resolution Large-Eddy Simulation (LES) output. The main objective is to better understand and quantify fine-scale spatial and temporal structures of such shallow cumulus cloud populations. These cloud fields are highly heterogeneous and quickly evolve on a broad range of scales, spanning from turbulence via mesoscale organization to diurnal evolution and beyond. This complicates their representation in numerical models for weather prediction and climate simulation. We approach this problem by exploiting the synergy between LES and stereo camera observations, thus filling a key scientific data gap on this topic. The high-resolution sampling capability of stereo reconstruction in four dimensions, and its considerable spatial coverage, allows us to capture cumulus cloud populations in unprecedented detail, which cannot be achieved with vertically pointing instrumentation or satellite sensors. The fine-scale model simulations supplement the observational data with four-dimensional data. This allows evaluation of the model against the stereo camera images, but also the virtual testing of measurement strategies, for example concerning the density of a stereo camera network. A key step towards these goals that was successfully realized in this project is the development of a dedicated instrument simulator algorithm with which the model fields can be fairly compared to the hemispheric cloud images. Ray tracing techniques well known from realistic rendering of landscapes and scenes in gaming technology is applied to the model cloud fields, yielding virtual hemispheric images that can directly compared to the real images from the camera network. The algorithm was tested for real cloud fields as observed during a dedicated campaign at the JOYCE meteorological site at Jülich, Germany, where a camera network was installed and operated. Analysis of the results showed that i) the ray tracing algorithm is capable of reproducing virtual images of model clouds that are good enough for threedimensional cloud reconstruction, and ii) the algorithm is a viable tool for evaluating fine scale geometry of cloud fields in high resolution models against camera network data. The main findings obtained in this project have been presented at workshops and major conferences, and have or are being been published in peer-reviewed scientific journals.
Publications
-
Comparison of Cloud Motion Vector Profiles Derived from Ground-Based Hemispheric Stereocameras to Wind-Lidar Observations. Meteorologische Zeitschrift, 30(6), 503-513.
Beekmans, Christoph; Schween, Jan; Lennefer, Martin & Simmer, Clemens
-
A Stereo Camera Simulator for Large‐Eddy Simulations of Continental Shallow Cumulus Clouds Based on Three‐Dimensional Path‐Tracing. Journal of Advances in Modeling Earth Systems, 16(3).
Burchart, Yannick; Beekmans, Christoph & Neggers, Roel
