Testing methods for using high‐resolution satellite imagery to monitor polar bear abundance and distribution
Bibliographic record
Abstract
ABSTRACT High‐resolution satellite imagery is a promising tool for providing coarse information about polar species abundance and distribution, but current applications are limited. With polar bears ( Ursus maritimus ), the technique has only proven effective on landscapes with little topographic relief that are devoid of snow and ice, and time‐consuming manual review of imagery is required to identify bears. Here, we evaluated mechanisms to further develop methods for satellite imagery by examining data from Rowley Island, Canada. We attempted to automate and expedite detection via a supervised spectral classification and image differencing to expedite image review. We also assessed what proportion of a region should be sampled to obtain reliable estimates of density and abundance. Although the spectral signature of polar bears differed from nontarget objects, these differences were insufficient to yield useful results via a supervised classification process. Conversely, automated image differencing—or subtracting one image from another—correctly identified nearly 90% of polar bear locations. This technique, however, also yielded false positives, suggesting that manual review will still be required to confirm polar bear locations. On Rowley Island, bear distribution approximated a Poisson distribution across a range of plot sizes, and resampling suggests that sampling >50% of the site facilitates reliable estimation of density (CV <15%). Satellite imagery may be an effective monitoring tool in certain areas, but large‐scale applications remain limited because of the challenges in automation and the limited environments in which the method can be effectively applied. Improvements in resolution may expand opportunities for its future uses. © 2015 The Wildlife Society.
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 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 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| 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 teacher head, 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".