FORMSpoT: A Decade of Tree-Level, Country-Scale Forest Monitoring

2025-12-18

AI Summary

Current satellite-based forest disturbance products suffer from insufficient spatial resolution (>100 m²), hindering tree-level (<100 m²) monitoring. To address this, we developed a 1.5-m-resolution time series of forest canopy height and annual disturbance vector maps for mainland France (2014–2024). Our method uniquely integrates high spatiotemporal-resolution SPOT-6/7 imagery with a hierarchical Transformer architecture (PVTv2) and introduces a co-registration–spatiotemporal total variation (ST-TV) denoising framework tailored for heterogeneous remote sensing data. This significantly enhances detection sensitivity for small-scale disturbances. In mountainous, fragmented forest landscapes, our approach achieves an F1-score of 0.44—representing a tenfold improvement over state-of-the-art products. Validation against 19 airborne LiDAR (ALS) re-measurement sites and 5,087 national forest inventory plots confirms superior accuracy and robustness in disturbance detection. This work establishes the first operational, national-scale, tree-level, high-frequency forest dynamic monitoring system.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Machine Learning: Ensemble Methods
• Search and Optimization: Distributed Search

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Web Mining and Content Analysis: Normalization, clustering, classification, and summarization of Web text

Abstract

The recent decline of the European forest carbon sink highlights the need for spatially explicit and frequently updated forest monitoring tools. Yet, existing satellite-based disturbance products remain too coarse to detect changes at the scale of individual trees, typically below 100 m$^{2}$. Here, we introduce FORMSpoT (Forest Mapping with SPOT Time series), a decade-long (2014-2024) nationwide mapping of forest canopy height at 1.5 m resolution, together with annual disturbance polygons (FORMSpoT-$Δ$) covering mainland France. Canopy heights were derived from annual SPOT-6/7 composites using a hierarchical transformer model (PVTv2) trained on high-resolution airborne laser scanning (ALS) data. To enable robust change detection across heterogeneous acquisitions, we developed a dedicated post-processing pipeline combining co-registration and spatio-temporal total variation denoising. Validation against ALS revisits across 19 sites and 5,087 National Forest Inventory plots shows that FORMSpoT-$Δ$ substantially outperforms existing disturbance products. In mountainous forests, where disturbances are small and spatially fragmented, FORMSpoT-$Δ$ achieves an F1-score of 0.44, representing an order of magnitude higher than existing benchmarks. By enabling tree-level monitoring of forest dynamics at national scale, FORMSpoT-$Δ$ provides a unique tool to analyze management practices, detect early signals of forest decline, and better quantify carbon losses from subtle disturbances such as thinning or selective logging. These results underscore the critical importance of sustaining very high-resolution satellite missions like SPOT and open-data initiatives such as DINAMIS for monitoring forests under climate change.

Problem

• Detects individual tree-level forest disturbances at national scale
• Monitors forest canopy height changes with high-resolution satellite data
• Improves carbon loss quantification from subtle forest management practices

Innovation

• Hierarchical transformer model processes SPOT imagery for canopy height
• Post-processing pipeline uses co-registration and denoising for change detection
• Tree-level monitoring at national scale with 1.5 m resolution mapping