Spatiotemporal Interpolation of GEDI Biomass with Calibrated Uncertainty

2026-04-04

AI Summary

This study addresses the spatiotemporal coverage gaps in GEDI LiDAR data caused by irregular orbital sampling and operational interruptions, which particularly hinder continuous and uncertainty-quantified estimation of aboveground biomass density (AGBD) during disturbance events. To overcome this limitation, we propose a novel approach based on Attentive Neural Processes (ANPs), extending ANPs for the first time to sparse spatiotemporal settings. Our method integrates geospatial foundation model embeddings with symmetric spatiotemporal modeling, enabling cross-temporal spatial substitution and joint interpolation. Through ensemble learning, the model produces well-calibrated predictive uncertainty intervals and demonstrates robust performance across diverse disturbance scenarios, thereby effectively supporting forest carbon monitoring, reporting, and verification (MRV) applications.

Abstract

Monitoring deforestation-driven carbon emissions requires both spatially explicit and temporally continuous estimates of aboveground biomass density (AGBD) with calibrated uncertainty. NASA's Global Ecosystem Dynamics Investigation (GEDI) provides reliable LIDAR-derived AGBD, but its orbital sampling causes irregular spatiotemporal coverage, and occasional operational interruptions, including a 13-month hibernation from March 2023 to April 2024, leave extended gaps in the observational record. Prior work has used machine learning approaches to fill GEDI's spatial gaps using satellite-derived features, but temporal interpolation of biomass through unobserved periods, particularly across active disturbance events, remains largely unaddressed. Moreover, standard ensemble methods for biomass mapping have been shown to produce systematically miscalibrated prediction intervals. To address these gaps, we extend the Attentive Neural Process (ANP) framework, previously applied to spatial biomass interpolation, to jointly sparse spatiotemporal settings using geospatial foundation model embeddings. We treat space and time symmetrically, empirically validating a form of space-for-time substitution in which observations from nearby locations at other times inform predictions at held-out periods. Our results demonstrate that the ANP produces well-calibrated uncertainty estimates across disturbance regimes, supporting its use in Measurement, Reporting, and Verification (MRV) applications that require reliable uncertainty quantification for forest carbon accounting.

Problem

• spatiotemporal interpolation
• biomass
• uncertainty calibration
• GEDI
• deforestation

Innovation

• Spatiotemporal Interpolation
• Attentive Neural Process
• Calibrated Uncertainty
• Geospatial Foundation Model
• Biomass Mapping