Branching Algorithm to Identify Bottom Habitat in the Optically Complex Coastal Waters of Atlantic Canada Using Sentinel-2 Satellite Imagery
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Sentinel-2 satellite imagery has been successfully used to map submerged seagrasses in clear waters, and surface-canopy forming seaweed habitats in a range of water types. We examined the ability to use Sentinel-2 remote sensing reflectance to classify fully submerged seagrass and seaweed habitats in optically complex, temperate waters within a high priority management region in Atlantic Canada. To do so, we determined the “best” Sentinel-2 image available between 2015 and 2019 based on tidal height, absence of sun glint and clouds, and water transparency. Using the full Sentinel-2 tile, we atmospherically corrected the image using ACOLITE’s dark spectrum fitting method. Our classification goal was a two-class prediction of vegetation presence and absence. Using information obtained from drop-camera surveys, the image was first partially classified using simple band thresholds based on the normalized difference vegetation index (NDVI), red/green ratio and the blue band. A random forest model was built to classify the remaining areas to a maximum depth of 10 m, the maximum depth at which field surveys were performed. The resulting habitat map had an overall accuracy of 79% and ~231 km2 of vegetated habitat were predicted to occur (total area 345.15 km2). As expected, the classification performed best in regions dominated by bright sandy bare substrate, and dense dark vegetated beds. The classification performed less well in regions dominated by dark bare muddy substrate, whose spectra were similar to vegetated habitat, in pixels where vegetation density was low and mixed with other substrates, and in regions impacted by freshwater input. The maximum depth that bottom habitat was detectable also varied across the image. Leveraging the full capacity of the freely available Sentinel-2 satellite series with its high spatial resolution and resampling provides a new opportunity to generate large scale vegetation habitat maps, and examine how vegetation extent changes over time in Atlantic Canada, providing essential data layers to inform monitoring and management of macrophyte dominated habitats and the resulting ecosystem functions and services.
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.001 |
| 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 it