A Footprint-Aware, High-Resolution Approach for Carbon Flux Prediction Across Diverse Ecosystems

2025-12-01

AI Summary

To address the low spatial extrapolation accuracy and poor interpretability of carbon flux estimates arising from the mismatch between satellite remote sensing pixel scales and eddy covariance tower footprints, this study proposes FAR (Footprint-Aware Regression), the first deep learning framework integrating footprint-aware modeling with pixel-level prediction. FAR jointly leverages Landsat imagery and multi-site eddy covariance observations to achieve high-accuracy, monthly net ecosystem exchange (NEE) estimation at 30 m resolution. Across diverse ecosystems and multiple validation sites, FAR achieves an R² of 0.78—significantly outperforming conventional approaches. Its footprint-weighted regression mechanism enhances physical consistency and spatial interpretability by explicitly incorporating spatially varying flux contributions. This work establishes a scalable, high-resolution, and verifiable paradigm for monitoring carbon flux dynamics, directly supporting natural climate solutions (NCS).

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Machine Learning: Deep Neural Architectures and Foundation Models
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Sustainability and carbon-aware systems for Web, mobile, and WoT
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems

Abstract

Natural climate solutions (NCS) offer an approach to mitigating carbon dioxide (CO2) emissions. However, monitoring the carbon drawdown of ecosystems over large geographic areas remains challenging. Eddy-flux covariance towers provide ground truth for predictive 'upscaling' models derived from satellite products, but many satellites now produce measurements on spatial scales smaller than a flux tower's footprint. We introduce Footprint-Aware Regression (FAR), a first-of-its-kind, deep-learning framework that simultaneously predicts spatial footprints and pixel-level (30 m scale) estimates of carbon flux. FAR is trained on our AMERI-FAR25 dataset which combines 439 site years of tower data with corresponding Landsat scenes. Our model produces high-resolution predictions and achieves R2 = 0.78 when predicting monthly net ecosystem exchange on test sites from a variety of ecosystems.

Problem

• Predicts high-resolution carbon flux at 30m scale
• Integrates satellite data with flux tower measurements
• Addresses spatial scale mismatch in ecosystem monitoring

Innovation

• Deep learning predicts carbon flux at 30m resolution
• Simultaneously models spatial footprints and pixel-level estimates
• Uses tower data and Landsat scenes for training