Linking Structural Forest Heterogeneity and Ecological Processes Using Sentinel-2 and FAD-Based Zoning
Bibliographic record
Abstract
Structural heterogeneity strongly influences forest ecological function, yet zone-specific spectral diagnostics remain limited. This study integrated Sentinel-2 imagery (2016, 2020, 2024) with field-observed ecological attributes across Foreground Area Density (FAD)-based structural zones in the Tuchola Forest Biosphere Reserve, Poland. The aim was to evaluate whether open Sentinel-2 vegetation indices can capture ecological variation across structurally distinct forest zones using interpretable machine-learning models. Correlation and cluster analyses of 17 vegetation indices revealed substantial multicollinearity, supporting the selection of a reduced set of spectrally distinct indices for modelling. Extra Trees (ET) and LightGBM (LGBM) produced comparable predictive performance, although ET achieved equal or lower RMSE values in most zone × year combinations and was retained for interpretation. Test-set RMSE remained below one degradation class (0.70-0.96), 2.04 moisture units, 2.35 site-type categories, and 34.4 years for stand age. Permutation importance and partial dependence analyses revealed clear zone-specific spectral-ecological relationships. Rare zones exhibited stronger stress-related spectral responses and greater variability in moisture and stand-age-related patterns, whereas Core zones displayed more stable response surfaces across years. NDRE emerged as the most consistent predictor across ecological attributes, while MCARI, NDMI, and CVI provided complementary information depending on the response variable. By combining FAD-based structural stratification, cluster-driven multicollinearity reduction, and interpretable ensemble learning, this framework provides a reproducible approach for linking spectral traits to ecological gradients across fragmentation contexts and supports open-data monitoring of fragmented forest landscapes.
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) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 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".