Tracking savanna vegetation structure in South Africa by extension of GEDI canopy metrics with Landsat, Sentinel-2, and PALSAR
Bibliographic record
Abstract
1 Abstract Adaptive management of savanna ecosystems requires frequent monitoring of woody vegetation structure, and although vegetation structure and changes may be captured with repeat airborne lidar, it is spatially and temporally limited across African savannas. As an alternative, this study evaluates the extension of spaceborne waveform lidar canopy metrics (RH98, Cover, Foliage Height Diversity) from the Global Ecosystem Dynamics Investigation (GEDI) across the Greater Kruger region in South Africa using moderate resolution optical sensors (Landsat and Harmonized Landsat Sentinel-2 [HLS]), L-band Synthetic Aperture Radar (PALSAR-1 and -2), and topographic and soil covariates. We compared the performance of 14 predictor sets incorporating different sensor combinations and temporal processing methods (LandTrendr and CCDC) in random forest models using temporal cross-validation to assess extrapolation accuracy. The most parsimonious fusion model (LandTrendr + SAR + topography/soils) achieved RMSEs of 3.04 m for RH98, 13.38% for Cover and 0.34 for FHD, which was comparable to more complex models using HLS and CCDC. All models demonstrated good temporal transferability with minimal bias but tended to overestimate low values and underestimate high values, which muted the estimated magnitude of change. Annual canopy structure maps derived from the best model captured expected spatial patterns and were used in model-based estimators to quantify changes in areas impacted by elephants, timber harvesting, fuelwood extraction, and woody encroachment. Extending GEDI metrics with moderate-resolution sensors thus offers a viable approach for large-scale savanna monitoring and detecting change in high impact areas.
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 | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 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".