Extratropical cyclones are the primary source of precipitation in the mid-latitudes. The physical mechanisms that drive cyclones are well understood, and a variety of studies have demonstrated strong relationships between cloud structures and cyclone dynamics. However, past research has focused on simplistic cloud categorizations, which lack spatial and textural information that is included in more sophisticated classification schemes. Unsupervised deep learning approaches may have significant advantages over these past methods, allowing them to discover previously unidentified cloud information in large datasets. One such approach is the rotation-invariant cloud clustering (RICC), which combines a dimensionality reduction deep learning technique with rotation-invariant clustering of input cloud images. We employ the RICC, along with two established cloud clusters, to investigate the relationship between extratropical cyclones and horizontal cloud distributions. We focus on comparing these different sets of clusters to each other. First the spatial distributions and the physical properties of the identified cloud types are examined around cyclones and the results corresponding to each classification are compared in detail. Then the similarities of these distributions are quantified using structural similarity. Additionally, the evolution of spatial distributions of cloud over the lifetime of cyclones is compared between the different classifications. Interestingly, we identify the same broad physical developments in all sets of clusters. Notably, identified differences are likely due to differences in measurement processes and resolutions of the corresponding datasets.

Applications of X-ray computed tomography (CT) for porosity characterization of engineering materials often involve an extended data analysis workflow that includes CT reconstruction of raw projection data, binarization, labeling and mesh extraction. It is often desirable to map the porosity in larger samples but the computational challenge of reducing gigabytes of raw data to porosity information poses a critical bottleneck. In this work, we describe algorithms and implementation of an end-to-end porosity mapping code that processes raw projection data from a synchrotron CT instrument into a porosity map and visualization in the form of triangular face mesh. Towards this objective, we report the development of a novel subset reconstruction scheme for X-ray CT using filtered backprojection and a convolutional neural network that allows us to reconstruct arbitrarily-shaped subsets of a tomography object. We build upon this scheme to implement the complete code for porosity mapping. The code first detects possible voids by performing a coarse reconstruction on down-sampled projections and then improves the shape of those voids with higher detail offered by reconstructing selected subsets from the original raw data. We report measurements of the time taken by this code to perform complete processing from raw data to a triangular face mesh for several visualization scenarios on a single highperformance workstation equipped with GPU. We show that we can now visualize local porosity within a 8 gigavoxel CT volume (12 gigabytes raw data) within 1 to 2 minutes and a 64 gigavoxel CT volume (100 gigabytes of raw data) within 3 to 7 minutes.