Physically Interpretable AlphaEarth Foundation Model Embeddings Enable LLM-Based Land Surface Intelligence

TL;DR

This paper analyzes Google AlphaEarth's dense satellite embeddings, revealing their physical interpretability and demonstrating how they can be used to build a land surface intelligence system that answers environmental queries with high fidelity.

Contribution

The study provides a comprehensive interpretability analysis of satellite embeddings and develops a retrieval-based system for environmental assessment using LLMs.

Findings

Embedding dimensions map onto specific land surface properties

High fidelity reconstruction of environmental variables (R^2 > 0.90)

Robust and stable interpretability across space and time

Abstract

Satellite foundation models produce dense embeddings whose physical interpretability remains poorly understood, limiting their integration into environmental decision systems. Using 12.1 million samples across the Continental United States (2017--2023), we first present a comprehensive interpretability analysis of Google AlphaEarth's 64-dimensional embeddings against 26 environmental variables spanning climate, vegetation, hydrology, temperature, and terrain. Combining linear, nonlinear, and attention-based methods, we show that individual embedding dimensions map onto specific land surface properties, while the full embedding space reconstructs most environmental variables with high fidelity (12 of 26 variables exceed $R^2 > 0.90$; temperature and elevation approach $R^2 = 0.97$). The strongest dimension-variable relationships converge across all three analytical methods and remain robust under spatial block cross-validation (mean $\Delta R^2 = 0.017$) and temporally stable across all seven study years (mean inter-year correlation $r = 0.963$). Building on these validated interpretations, we then developed a Land Surface Intelligence system that implements retrieval-augmented generation over a FAISS-indexed embedding database of 12.1 million vectors, translating natural language environmental queries into satellite-grounded assessments. An LLM-as-Judge evaluation across 360 query--response cycles, using four LLMs in rotating generator, system, and judge roles, achieved weighted scores of $\mu = 3.74 \pm 0.77$ (scale 1--5), with grounding ($\mu = 3.93$) and coherence ($\mu = 4.25$) as the strongest criteria. Our results demonstrate that satellite foundation model embeddings are physically structured representations that can be operationalized for environmental and geospatial intelligence.