Earth Embeddings as Products: Taxonomy, Ecosystem, and Standardized Access

TL;DR

This paper introduces a standardized framework and API for geospatial embedding data products, addressing interoperability issues and enabling more accessible Earth observation workflows.

Contribution

It formalizes a taxonomy for geospatial embeddings, surveys existing products, and extends TorchGeo with a unified API for standardized access.

Findings

Identified interoperability barriers in existing geospatial embedding products.

Developed a unified API extending TorchGeo for standardized data access.

Provided a roadmap for more transparent and reproducible Earth observation workflows.

Abstract

Geospatial Foundation Models (GFMs) provide powerful representations, but high compute costs hinder their widespread use. Pre-computed embedding data products offer a practical "frozen" alternative, yet they currently exist in a fragmented ecosystem of incompatible formats and resolutions. This lack of standardization creates an engineering bottleneck that prevents meaningful model comparison and reproducibility. We formalize this landscape through a three-layer taxonomy: Data, Tools, and Value. We survey existing products to identify interoperability barriers. To bridge this gap, we extend TorchGeo with a unified API that standardizes the loading and querying of diverse embedding products. By treating embeddings as first-class geospatial datasets, we decouple downstream analysis from model-specific engineering, providing a roadmap for more transparent and accessible Earth observation workflows.