Scalable Feature Extraction and Tracking (SCAFET): A general framework for feature extraction from large climate datasets

Abstract. This study describes a generalized framework, Scalable Feature Extraction and Tracking (SCAFET) to extract and track features from large climate datasets. SCAFET utilizes novel shape-based metrics that can efficiently identify and compare features from different mean states, datasets, and between distinct regions. Features of interest are extracted by segmenting the data based on a scale-independent bounded variable called shape index (SI). SI gives a quantitative measurement of the local geometric shape of the field with respect to its surroundings. To demonstrate the capabilities of the method, we illustrate the detection of atmospheric rivers, tropical and extratropical cyclones, sea surface temperature fronts, and jet streams. Cyclones and atmospheric rivers are extracted from the ERA5 reanalysis dataset to show how the algorithm extracts both locations and areas from climate datasets. The extraction of sea surface temperature fronts exemplifies how SCAFET effectively handles curvilinear grids. Lastly, jet streams are extracted to demonstrate how the algorithm can also detect 3D features. SCAFET can be implemented to extract and track most weather and climate features.