Monitoring harmful algae blooms in Darlings Lake, New Brunswick, using K-means clustering of multi-spectral 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
Darlings Lake, located in the Saint John River watershed, Canada, experienced lake-wide cyanobacteria blooms in the summers of 2021 and 2022. This study uses high spatial and temporal resolution satellite imagery from Planet Labs (Planet Labs, Inc., San Francisco, CA, United States of America) to understand the extent and severity of the blooms with a time series analysis of the normalized difference vegetation index (NDVI) and the normalized difference chlorophyll index (NDCI) over the lake using k-means clustering. We distinguish algae blooms from preexisting aquatic vegetation by creating a baseline map of mean aquatic vegetation extent, and subtracting this from each image in the time series. Additionally, results from a principal component analysis conducted on each year’s imagery corroborate the k-means finding, and align with spatial trends of bloom events observed in the lake. In this study, normalized difference chlorophyll index values are observed to be more reliable for estimating the severity of algal blooms, while NDVI is more sensitive to glare, haze, thin clouds, and signal over-saturation caused by blooms, aligning with preexisting research findings. We successfully fit a linear regression between NDCI values and in situ measurements of phycocyanin concentrations surrounding AlgaeTracker™ buoys ( R 2 :0.893). Furthermore we highlight bloom extent and severity for 2021 and 2022, revealing potential bloom hotspots in the lake. The methodology in this project can be extended to systematically analyze high-resolution satellite imagery in freshwater ecosystems to detect harmful algae blooms.
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 | 0.000 |
| 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 it