Expanding a data fusion approach for 4-D cloud nowcasting

We present an update on our complete 4-D cloud solution: Deep Generative Modelling of Satellite (DGMS) alongside the Vertical Cloud Profiler (VCP). We now explore new design decisions designed to facilitate a better model, including explicitly penalizing dissipation of clouds in the generative algorithm in DGMS, an expanded dataset for DGMS for downstream use in the Algorithm Working Group's (AWG) Cloud Height Algorithm (ACHA), and new data sources and expanded functionality for VCP. We provide more thorough verification by adding additional verification metrics against new baselines and persistence forecasts. We demonstrate the ability to generate spatio-temporally consistent predictions across channels and across a variety of meteorological conditions.