Forest Height Estimation in Jiangsu: Integrating Dual-Polarimetric SAR, InSAR, and Optical Remote Sensing Features
Bibliographic record
Abstract
Forest height is a key structural parameter for evaluating ecological functions, biodiversity, and carbon dynamics. While LiDAR and Synthetic Aperture Radar (SAR) provide vertical structure information, their large-scale use is restricted by sparse sampling (LiDAR) and temporal decorrelation (SAR). Optical remote sensing offers complementary spectral information but lacks direct height retrieval. To address these limitations, we developed a multi-modal framework integrating GEDI waveform LiDAR, Sentinel-1 SAR (InSAR and PolSAR), and Sentinel-2 multispectral data, combined with machine learning, to estimate forest canopy height across Jiangsu Province, China. GEDI L2A footprints were used as training labels, and a suite of structural and spectral features was extracted from SAR, GEDI, and Sentinel-2 data as input variables for canopy height estimation. The performance of two ensemble algorithms, Random Forest (RF) and Gradient Tree Boosting (GTB) for canopy height estimation, was evaluated through stratified five-fold cross-validation. RF consistently outperformed GTB, with the integration of SAR, GEDI, and optical features achieving the best accuracy (R2 = 0.708, RMSE = 2.564 m). The results demonstrate that InSAR features substantially enhance sensitivity to vertical heterogeneity, improving forest height estimation accuracy. These findings highlight the advantage of incorporating SAR, particularly InSAR with optical data, in enhancing sensitivity to vertical heterogeneity and improving the performance of RF and GTB in estimating forest height. The framework we proposed is scalable to other regions and has the potential to contribute to global sustainable forest monitoring initiatives.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".