An Innovation Exploration: Using ECMWF's Polytope to Accelerate and Simplify Scientific Workflows

ECMWF, the European Centre for Medium-Range Weather Forecasts, produces huge amounts of weather forecast data that can be regarded as high-dimensional data cubes. Traditional data extraction techniques based on bounding boxes are restrictive and do not scale well. To address these issues, ECMWF developed Polytope feature extraction: an efficient way of accessing petabyte-scale data cubes by allowing users to extract high-dimensional polytopes. This reduces I/O usage and post-processing needs after extraction. The datasets currently available via Polytope feature extraction are the operational ECMWF forecast, as well as the data produced by the Destination Earth Extremes and Climate digital twins. We illustrate the usage and benefits of Polytope through selected use cases.