Shallow cumulus clouds (ShCu) measurements are crucially important in evaluating Large-Eddy Simulations (LES) and ShCu-parameterizations in numerical weather and climate models. However, these data still mainly consist of one-dimensional profile data, often sampled by lidars or radars. A new method for adding multi-dimensional information is to use networks of multiple hemispheric cameras, which remotely observe ShCu in unprecedented spatial details constantly at high temporal frequency. These cameras provide a large field of view, enabling us to observe whole ShCu-life cycles. Thus, these networks strongly complement existing ground-based instruments. To objectively estimate camera networks' accuracy, we have to test them against virtual LES-cloud fields, that act as ground truth. However, for this purpose virtual camera projections of these cloud fields are needed.Our study aims to generate such projections by combining radiative transfer theory with open-source path-tracing. With these projections, we emulate our camera network, currently installed at the Jülich Observatory for Cloud Evolution (JOYCE), Germany as part of the ongoing SOCLES project. As input, we use LES-cloud fields. Via the emulated camera images, we reconstruct the cloud fields back in the same way the camera network does it from real-world images. However, by using artificial images over real-world images, we have the advantage of already knowing the whole cloud field. This knowledge enables us to statistically analyze and optimize our network. Concretizing this, here are our research objectives:Objectively estimate the efficiency of our camera networkAnalyze the capability of our camera network by investing how much of a cloud shell is on average visibleOptimize the camera network, using our new insightsOur camera network emulation works well in this workflow. For the selected days, about 70% of the mutually visible cloud grid boxes were rightly reconstructed by our artificial camera network. About 53% of a ShCu-cloud shell is averagely visible by a single stereo camera pair of our network at a single time point. With increasing distance between the two cameras of such stereo camera pairs, fewer cloud shell areas are detected. In fact, for every extra kilometer, about 3.3% of a cloud shell is lost on average.

The complex spatial and temporal structure of cumulus clouds complicates their representation in weather and climate models. Classic meteorological instrumentation struggles to fully capture these features. Networks of multiple high-resolution hemispheric cameras are increasingly used to fill this data gap, and provide information on this missing multi-dimensional spatial information. In this study, a path-tracing algorithm is used to generate virtual camera images of resolved clouds in Large-Eddy Simulations (LES). These images are then used as a camera network simulator, allowing reconstructions of three-dimensional cloud edges from the model output. Because the actual LES cloud field is fully-known, the combined path-tracing and reconstruction method can be statistically analyzed. The method is applied to LES realizations of summertime shallow cumulus at the Jülich Observatory for Cloud Evolution (JOYCE), Germany, which also routinely operates a camera network. We find that the Blender path-tracing method allows accurate reconstruction of up to 70% of the visible cloud edges, depending on camera distance and accuracy thresholds. Additionally, we conducted sensitivity tests and find that our method remains consistent and independent of changes in its hyperparameters. The sensitivity of the stereo reconstruction algorithm to cloud optical thickness is investigated, finding a cloud boundary placement error of approximately 182 m. This error can be considered typical for cloud boundary reconstruction using stereo camera imagery in general. The results provide proof of principle for future use of the method for evaluating LES clouds against real camera imagery, and for further optimizing the configuration of such camera networks.